University Microfilms, a Xerpkcompany, Ann Arbor

72-15,188

CAMPBELL, Gerald Allan, 1946- THE CHEMISTRY OF ARYL NITRENIUM IONS.

The Ohio State University, Ph.D., 1971 Chemistry, organic

University Microfilms, A XERPKCompany, Ann Arbor. Michigan

THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED THE CHEMISTRY OF ARYL HITREPTIUM IONS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By Gerald Allan Camphell, B.S. * * * * #

The Ohio State University 1971

Approved by

Advisor Department of Chemistry ACKNOWLEDGMENTS

The author wishes to thank Dr. Paul G. Gassman for suggesting this

problem and for his many ideas throughout the completion of this work.

The author is also deeply grateful to his wife, Jo Ann, for her patience and encouragement during the course of this work. In addition, the author wishes to thank his parents for their help and financial

support during his college education.

11 VITA

Gerald Allan Campbell, son of Robert R. and Eubatinia Campbell, was born on May $0, 19^6 in Cincinnati, Ohio. He obtained his primary and secondary education in Deer Park, Ohio. In September, 1964 he

entered The University of Cincinnati where he received his B.S. in

Chemistry in August, 1 9 67. In September, 196% he entered the Graduate

School of The Ohio State University. He married Jo Ann Broxterman of

Deer Park, Ohio in December, 1967. While at The Ohio State University, he held the positions of a National Defense Education Act Fellow and a Stauffer Fellow. In December, 1971 he received his Ph.D. in organic

chemistry from The Ohio State University.

Ill TABLE OF COIITEOTS

Page

ACKNOWLEDGafflJTS ...... 11

VITA ...... ill

TABLES ...... vlil

ILLUSTEATIONS ...... ix

INTRODUCTION ...... 1

PART I. Historical ...... 1

PART II. The Problem ...... 20

RESULTS AND DISCUSSION ...... 51

PART I. Synthesis and Solvolysls of N-Chloro-N- alkylanlllnes ...... 51

PART II. Synthesis and Solvolysls of N-Alkyl-N- arylhydroxylamine p-Nitrobenzoates ...... 58

PART III. Nucleophilic Aromatic Substitution of Indole Derivatives ...... ÔJ

EXPERIMENTAL ...... 87

N-t-Butylanlllne 87

N-t-Butyl-p-tolul dine 87

N-t-Butyl-p-anls idine (Ipla) ...... 88

N-t-Butyl-o-anisi dine (102a) ...... 88

N -t-Butyl -o-t oluldine ...... 89

N-t-Butyl-p-fluoroanillne ...... 89

IV Page

U_t-Butyl-£-chloroaniline (l04a) ...... 89

N -t-Batylamino ) b ipheny 1 ...... 90

Ethyl p-(N-t-Butylamino)benzoate ...... 90 p-(N-t-Butylamino)nitrobenzene ...... 91 p-(N-t-Butylamino)benzonitrile (9%) ...... 91

Chlorination and Solvolysis of N-t-Butylaniline (9^ ..... 92

Chlorination and Solvolysis of N-t-Butyl-p-tolni dine( 9 8 ) .. 95

4-methyl-4 -methoxy-2,5 -cyclohexadienone N-t-Butyl- imine ( 10^ ) ...... 94

Chlorination and Solvolysis of N-t-Butyl-p-fluor o- aniline ...... 7 ...... 95

Chlorination and Solvolysis of N-t-Butyl-p-chloroaniline ( 104a) ...... 96

Chlorination and Solvolysis of p-( N-t-Butylamino) biphenyl .. 97

Chlorination and Solvolysis of Ethyl p^-( N-t-Butylamino) - benzoate ...... 7...... 98

Chlorination and Solvolysis of p-( N-t-Butylamino) nitro benzene ...... 100

Chlorination and Solvolysis of p-( N-t-Butylamino)- benzonitrile ( 9%) ...... 101

Chlorination and Solvolysis of N-t-Butyl-o-toluidine ...... 102

Chlorination and Solvolysis of N-Methylaniline ...... 105

2,6,N-Trimethylaniline ...... 104

Chlorination and Solvolysis of 2,6,N-Trimethylaniline ..... 104

Chlorination and Solvolysis of o-( N-Methylamino) - biphenyl ...... I06

N-t-Butyl-N-phenylhydroxylamine p-Nitrobenzoate (l42) lOT Page

Solvolysis of H-t-Butyl-H-phenylhydroxylamine £-Nitro- benzoate ( .7...... 108

N-t-Butyl-N-(£-tolyl)hydroxylamine ...... IO9

N-t-Butyl-N-(p-tolyl) hydroxy lamine p_-Nitrobenzoate ( 151) ••• 110

Solvolysis of N-t-Butyl-N-(p-tolyl)hydroxylamine p-nitrobenzoate fl^l) ...... Ill

Transestérification of o-( N-t-Butylamino) phenol p-Nitrobenzoate ( l ^ ...... 112

3 -Chloro-2,3 -d-imethylindolenine ( 168) ...... 112

5-Methoxy-2,5-âimethylindolenine ( l6^ ) ...... II5

2 -Carboethoxy-3 -methoxy-3 -methylindolenine ...... 113

2-Oxobutyric Acid PhenyIhydrazone (17^) ...... Il4

2 -Carboethoxy-3 -methylindole ( lj6) ...... 115

2 -%rdroxymethyl-3 -methylindole (]J%) ...... 115

2 -%rdroxymethyl-3 -methyl Indole Acetate ( 174) ...... 115

Chlorination and Acetolysis of 2,3 -Dimethylindole ...... II6

Nydride Reaction of 2-Hydroxymethyl-3-methylindole Acetate (ijCit) ...... II6

2-Methoxymethyl-3-methylindole ( IT^) ...... 117

Chlorination and Methanolysis of 2,3 -Dimethylindole ...... II8

N-%rdroxy-2 -phenylindole ( 178] ...... II8

N-Hydroxy-2-phenylindole p-Nitrobenzoate ( IJ^) ...... II8

2-Phenyl-3-indolol p-Nitrobenzoate (180) ...... II9

2-Nitrobenzylidinemalonate ( l82) ...... 120

Diethyl o-Cyano-Q'-nitrobenzylmalonate ( 183^ ...... 120

N-%rdroxy-3 -cyanoindole-2-carboxylic Acid ...... 120

VI Page

ïï-Iîydroxy-2 -carlDoethoxy-5 -cyanoindole ( l84) ...... 120

N-Hydroxyoxindole ( l88 ) ...... 120

N-îîydroxyoxindole g-Nitrobenzoate ( 189) ...... 120

Transestérification of N-Hydroxyoxindole p-Nitro- benzoate (l8p) ...... 121

Methanolysis of N-Hydroxyoxlndole p-Toluenesvilfonate ..... 121

Methanolysis of N-H&rdroxy ox indole p-Nitrobenzene- sulfonate ...... 122

hydrolysis of N-I^rdroxyoxindole p^-Toluenesulfonate ...... 123

Methanolysis of N-hydroxy-2-carboethoxy-3-cyanoindole £-Toluene sulfonate ...... 123

KINETICS ...... 124

Reagents ...... 124

Procedure ...... 125

Product Studies ...... 125

Product Yields ...... 125

REFERENCES ...... 126

VI1 TABLES

Table Page

1. Migration Aptitudes in Ketone Oxime Rearrangements ...... 6

2. Rates and Products of N-Chloroaziridine Solvolysis ...... 13 3* N-Chloro-N-alkylanilines ...... 55

4. Percentage of Methoxy Products from the Silver Ion- Assisted Methanolysis of K-Chloro-N-t-butylanilines ...... 46

5. Rates of Rearrangement of R-Chloroanilines in Ethanol Buffered with 0.1 R Acetic Acid - 0.1 R Sodium Acetate ... 49

6. Products from the Solvolysis of pa£a-Substituted R-Chloro-R-t-butylanilines in Buffered Ethanol ...... 52

7 . Products from the Solvolysis of para-Substituted R - Chlor o -R -t-butylaniline s in Pure Ethanol ...... 55

Vlll IIIiUSTRATIOITS

Figure Page •4* 1. a 2 Plot for the Solvolysis of U-Chloro-W-t-butylanilines in Buffered Ethanol...... 51

2. Kinetic Behavior of Ethyl p-(N-Chloro-N-t-butylamino)- benzoate in Buffered and Unbuffered Ethanol ...... 56

XX IliffiRODUCTION

Part I. Historical

The chemistry of carhonium ions has received a vast amount of

attention from organic chemists during the past $0 years. In view of

the extensive information available on the trivalent, electron-

deficient carbon species (l), it is rather surprising that until the 1 recent work of Gassman and co-workers very little was known about the

(l) P. G. Gassman, Accounts Chem. Res., ^ 26 (l9T0),

nitrenium ion, the nitrogen analogue of the carbonium ion. This

divalent, electron-deficient nitrogen species (2_) has six electrons in

its valence shell and is isoelectronic with the carbonium ion. On

El El 1+ u Eg— C ;N I I E g E g

1 2

comparing the relative reactivities of 1 and 2_, the electronegativities

of carbon and nitrogen indicate that of the two isoelectronic species, the nitrenium ion should be the more reactive. However, an important factor which requires consideration is that the nitrenium ion differs 2

from most positive ions of interest to the organic chemist in that it possesses a nonhonding pair of electrons. This unique characteristic

enables the nitrenium ion to exist in either a singlet spin state or a 2 triplet spin state. Sclvolytically generated nitrenium ions would

(2 ) P. G. Gassman and R. L. Cryberg, J. Amer. Chem. Soc., 91, 517^ (1969).

initially be, of necessity, in the singlet state with the electron

spins paired (J^). The singlet species could, however, undergo a spin inversion to yield the triplet nitrenium ion (5 ). The singlet would

Cl I + R — N-R --- > R-R-Rê ^ R-R-R-K-6

thus be expected to exhibit carbonium ion-like behavior while the triplet would be expected to resemble a nitrogen cation radical in its chemical reactivity.

The first example of nitrenium ion chemistry to appear in the literature, althou^ unrecognized as such, was reported by Julius 3 Stieglitz and co-workers during the period 1913-1916. He reported that trityliiydroxylamines and trltyl-R-chloramines rearranged readily to give, after hydrolysis, benzophenone and aniline, but that disub- (3) J. Stieglitz and P. N. Leech, Ber., k6^ 2lhj (l915); J. Stieglitz and P. 1Î. Leech, J. Amer. Chem. Soc., ^6^ 2T2 (l9l4); J. Stieglitz and B. A. Stagner, ibid., ^8 , 2046 (1916).

stituted hydroxyl amines and disubstituted U-chloramines either did not rearrange or that they rearranged only under extremely forcing conditions. From this data, Stieglitz proposed that the rearrangement was occurring via the intermediacy of a mono-valent nitrogen (nitrene) species. More recent investigators, however, have been able to smoothly rearrange the disubstituted coinpounds. Thus it now appears likely that the rearrangements observed by Stieglitz actually involved

(4) R. T. Conley and H. Brandman, unpublished results.

a divalent electron-deficient nitrogen intermediate (nitrenium ion) rather than a nitre ne intermediate.

0 R H PCI5 ^ I I . 0 — c— n— OH C=H— C=0 + R-N-P 1

Support for this nitrenium ion mechanism was provided by the 5 study of migratory aptitudes of aryl groups in the rearrangement of

(5 ) M. S. Newman and P. M. Hay, J. Amer. Chem. Soc., 2522 (1955)- 4 trityUiydroxylamines. It was found that as the para- subst ituent of the aryl groups in the tritylhydroxylamine "became more electron- donating, the migratory aptitude of the aryl group increased. This behavior is consistent with the generation of positive charge on nitrogen in the transition state of the rearrangement reaction.

Another exançle of a classicial reaction which may in some cases involve a nitrenium ion intermediate is the Beckmann rearrangement.

In this reaction ketoximes or their esters are converted to substituted amides. Usually it is the group trans to the hydroxyl which migrates and this fact has lead many researches to propose a concerted mechanism 6 for the transformation. Subsequently, however, it has been shown that

(6 ) For a review of the Beckmann rearrangement, see P. Smith in “Molecular Rearrangements, ” P. Mayo, ed., Vol. 1, p. 48^-507, Interscience Publishers, Inc., New York, W.Y., 19^5•

II — > h-m--Rz

in some cases the group cis to the hydroxyl migrates and that when both Ri and Rg are alkyl, mixtures of the two possible amides are pro duced. Thus it seems that a concerted, backside attack by the migrating trans group does not always occur, and that a discrete unsaturated nitrenium ion may be involved in the process. 0 11 Eg. El Eg El Ea —C— NH-Ei C ^ Il ^ Il 0 Il (V ^ Ê7 \i) Bi—C^Œ^Rs

Support for this concept of unsaturated nitrenium ions (iminium T cation) was provided hy Lanshury and co-workers who studied the

(t) p. Lanshury and N. Mancusc, Tetrahedron Ijetters, 2kk'^ (1965).

Beckmann rearrangement of indanone oximes. These workers reported that when steric considerations precluded the possibility of prior isomerization of the oxime and made aryl migration unlikely due to the steric compression in the transition state, the alkyl group cis to the hydroxyl group migrated preferentially. Iminium cations were considered as possible intermediates in this reaction and support for this proposal was obtained in a novel intramolecular insertion reaction. a In this reaction, Lansbury and co-workers reported that 4-bromo-

(8 ) P. Lansbury, J. Colson, and N. Mancuso, J. Amer. Chem. Soc., 86, 5225 {196k).

T-t-butyl-1-indanone oxime (6) was converted to imine (T_) in T5^ yield by the treatment with polyphosphoric acid. In the insertion reaction leading to 7 , it is possible that the species attacking the proximal Table 1

Migration Aptitudes in Ketone Oxime Rearrangement

Oxime i> Aryl Migration $ Alkyl Migration

,CH

37 63

0 (013)3 i®

15 85 O

,0H

Br Br

C-H bond could be either the cationic nitrogen intermediate or the vinyl nitrene (§_). In order to distinguish between these intermediates,

+ ■H I

Br Br 8 T the reaction was carried out in polyphosphoric acid. If the reaction was proceeding through the iminium cation 8 no deuterium would he found in the product, whereas, if the nitrene were the intermediate the product should have one deuterium atom at the carbon adjacent to the imine carbon. Since no deuterium was found in the product, the authors proposed that the nitrenium ion was the intermediate.

Gassman and co-workers' work in the chemistry of nitrenium ions began with establishing that alkyl migration could occur to a diimlent electron-deficient nitrogen species where both substituents on nitrogen were alkyl groups. This approach was taken because of the tendency of alkyl groups to migrate to cationic centers and not to radical centers, in contrast to the behavior of aryl groups and hydrogen which are known to migrate to both cationic and radical centers. Thus, the migration of an alkyl group from carbon to nitrogen would indicate that the nitrogen was significantly electron deficient. This concept was substantiated by the conversion of N-chloroisoquinuclidine (lO) into

2-methoxy-l-azabicyclo[5.2.l]octane (l5) in yield by treatment 9 with a methanolic solution of silver nitrate. The reaction could

(9 ) P. G. Gassman and B. L. Fox, J. Amer. Chem. Soc., ^ 338 (1967).

involve loss of chloride ion to leave the nitrenium ion as an inter mediate which could undergo an alkyl migration to give ^ followed by solvent capture to give ]2. An alternate route would involve concerted loss of chlorine and migration of the alkyl group with its pair of 8

10 11 12

iDonding electrons to give ^ directly. Regardless of which mechanism was followed an electron-deficient nitrogen species must have been involved since the alkyl group must have migrated with its electron pair. This rearrangement was the first example of alkyl group migration to a nitrenium ion center.

The generality of this reaction was demonstrated by the solvolysis 10 of 4,T,T-trimethyl-2-chloro-2-azabicyclo[$.2.l]heptane (l^,). In

(lO) P. G. Gassman and R. L. Cryberg, J. Amer. Chem. Soc., gO, 1555 (1968); P.G. Gassman and R. L. Cryberg, ibid., 91. 204T (1969).

refluxing methanol ^ gave ^ ^ and ^ in 59%, 20%, and 7% yields, respectively. In a much faster reaction, when ^ was treated with a

CH CH. tl;

Cl H 13 14 15 16 methanolic silver perchlorate solution the same products were produced in 78%, 8%, and 4% yields, respectively. The increase in chlorine containing product in the presence of silver ion was explained by either a very tight ion-pair or a hi^ly concerted transition state as illustrated by

CH; CH;

CH. 6+ :C1 A g

Alkyl migration to nitrenium ion centers in a non-bicyclic system was exemplified in a novel series of ring expansion and contraction 11 reactions. Treatment of N-chloro- 1-phenyl-H-methylcyclobutylamine

(ll) P. G. Gassman and A. Carrasquillo. Tetrahedron Letters, 109 (1971).

(3^ with silver ion in methanol resulted in ring expansion to give the iminium cation ^ which was reduced ^ situ with borohydride to give 20 as the major product.

Î V -if-GHs •3 4- N-CHc Cl ■ à Ô

18 20 10

When the nitrogen atom "was in the four-membered ring, the genera

tion of the nitrenium ion resvilted in ring contraction. Thus W-chloro-

2-phenylazetidine (2^) reacted with silver ion in methanol to give the

aziridinium salt (22^) which suffered hydrolysis to give henzaldehyde

and a mixture of 2^ and 2^.

: w - c «— > \. -Nv H H Cl 22 21

(J 0-C-H + H2ll(CH2)20CH3 + HsNfCHgïsCH2 /S'"

2g_ 24

The rearrangements discussed above firmly established the occur rence of alkyl group migration from carbon to nitrogen. Since the end products involved the addition of a nucleophile to a carbonium ion, the alkyl groups must have migrated with their electron pairs. Thus, these reactions must have involved heterolytic cleavage of the N-Cl bond to yield a nitrenium ion and chloride anion.

After studying alkyl migration to cationic nitrogen, Gassman next provided kinetic evidence for the heterolytic cleavage of the W-Cl bond from the study of the kinetics of N-chloroaziridine solvolysis. 12 The molecular orbital symmetry rules of Woodward and Hoffman have 11

(l2) R. B. ¥ood'ward and R. Hoffman, J. Amer. Chem. Soc., 87, 595 (1 9 65 ). See also H. C. Longuet-Higgins and F. W. Abrahamson, ibid., 87 . 2045 (1965).

predicted that in the solvolysis of cyclopropyl tosylates and chlor ides, the C2 -C3 bond of 2^ should cleave in a disrotatory process in which the groups trans to the leaving group rotate outwards. This 13, 14 prediction was confirmed experimentally when a kinetic study of

(1 3 ) C. H. LePuy, L. G. Schnack, J. ¥. Hausser, and ¥. ¥iedemann, ibid., 8%^ 4006 (1965).

(14) p. von R. Schleyer, G. ¥. VahDine, U. Schollkopf, and J. Paust, ibid., 8^ 2868 (1966); U. Schollkopf, K. Fellenberger, M. Patsch, P. von R. Schleyer, T. Su, and G. ¥. VahDine, Tetrahedron Letters, 3^39 (1967)*

the concerted ring opening of cyclopropyl cations to allylic cations indicated that a consideration of electronic and steric effects coupled with molecular orbital symmetry arguments permitted predictions of the observed rates. N-chloroaziridines, which are stereochemically stable

R i R:

25 26 12 15 to inversion, should show behavior similar to cyclopropyl tosylates

(15 ) S. J. Brois, J. Amer. Chem. Soc., 22/ 508 (1968).

and chlorides if the W-Cl bond is heterolytically cleaved to give a is cationic nitrogen intermediate. Gassman, Bygos, and Trent reported

(16) P. G. Gassman and D. K. Dygos, J. Amer. Chem. Soc., 9I; 15^3 (1 9 69 ); P. G. Gassman, D. K. Dygos, and J. E. Trent, ibid., 92; 208U (1970).

that concerted solvolytic ring opening of 2^ gave an intermediate 2^

Cl Ri + /R4 HgO 0 0 C=N=C n H Rs E1CR2 + R3CR4 + EH4C1

28 29

which hydrolyzed under aqueous conditions to give two moles of carbonyl compound and ammonium chloride. As indicated in Table 2 , the rates of this reaction are in complete accord with the consideration of heterolytic N-Cl bond cleavage in a concerted disrotatory ring opening reaction coupled with the electronic and steric effects of the substi tuents on the ring. Thus, when no carbonium ion stabilizing groups 13

Table 2

Rates and Products of H-Chloroaziridine Solvolysis

N-Chloroaziridine Mode of Ring Opening rel Products

Cl 0 II 2 HCH + m^Gl

H H H H

36 57

Ô+ 15 CH3CH + HCH + ÏÏH4CI

38 39

0- C1 ^ 1 M (I Ô+ H 210 CH3ŒI + HCH + MH4CI n IZV » I > ; H CH3 H

ho 4l 14

N -Chlor oaz ir idine Mode of Ring Opening P r o d u c t s

C l C l

1490 2 C H 3C H + M Î 4 CI

42 45

C l ^ f i Ü i860 CH 3CCH 3 + HCH + HH 4 CI

CH 3 H

44 45

6- C l 0 -X' 11 H 1 5 5 , 0 0 0 2 C H 3 C H + ÎJH4C1

46 47 15 are present, as in ^ the reaction is slow. However, when methyl groups are available to stabilize the incipient carbonium ion, the effect must be balanced with steric strain generated during the disro tatory ring opening. Thus ^8 reacts slower than because of the methy1-hydrogen steric interaction in the transition state of The combined effects of carbonium ion stabilization and relief of steric interaction is seen in the solvolysis of ^ which reacts 155,000 times faster than The solvolysis of M-chloroaziridines, therefore, has conclusively indicated that the H-Cl was cleaved to give an electron- deficient divalent nitrogen intermediate and chloride anion.

Although the examples above have indicated that the concept of an electron-deficient divalent nitrogen species is a sound one, the dis tinction could not be made between a nitrenium ion with a unit positive charge or a slightly electron-deficient nitrogen species which reacts further before a unit positive charge on nitrogen can develop. The first evidence for a discrete nitrenium ion intermediate wrLth a full XT positive charge was provided by a study of the spin inversion of

(i t ) P. G. Gassman and R. L. Cryberg, J. Amer. Chem. Soc., 91» 5176 (1969).

singlet nitrenium ions to triplets. As noted above, the solvolytically generated singlet nitrenium ion can undergo spin inversion to the triplet which can abstract two hydrogen atoms from solution to give the protonated form of the secondary starting amine. These researchers found that when nitrenium ions were generated in solvents containing l6

H H % CH 3 CH 1 + C H 3 O H 1 + R— H— R ---- > R-R-R > R—N-RR— N------> R— N - R I (t) é H

heavy atoms, such as bromoform and chloroform, the percentage of the secondary starting amine in the product mixture increased. Since heavy 18 atoms solvents are known to catalyze spin inversion, the authors

(18) A. G. Anastassiou, J. Amer. Chem. Soc., 8 8 , 2^22 (1966); C. D. Dijkgraaf and G. J. Hotjtink. Tetrahedron Supplement No. 2, 179 (1965).

proposed that the occurrence of the secondary starting amine in the product mixture was due to hydrogen atom abstraction by the triplet nitrenium ion followed by the neutralization of the protonated amine in work-up. In order for spin inversion to occur, the nitrenium ion must have existed as a discrete entity with unit positive charge on nitrogen.

Recently, ab initio SCFCI molecular orbital calculations have 19 been reported which are in agreement with and lend interpretative

(19) S. T. Lee and K. Morokuma, private communication to P. G. Gassman. 17 support to the above experimental work on singlet and triplet nitre nium ions. These calculations have indicated that for the nitrenium + ion, the ground state is a linear or almost linear triplet with a very flat potential energy curve for H-IT-H bending. About 4$ kcal/ mole above it is the lowest singlet state with H-ÎÎ-H bond angle of about

115°. A population analysis suggests that the singlet EHg will react 4- like a carbonium ion and that the triplet EfHa will react like triplet methylene. The net charge distributions for the singlet state were calculated as indicated. This charge distribution suggests that

+ 0.87

0 H ”

-0.51

electrophilic attack by these reagents may take place with the 2p ^ orbital pointing toward the substrate, although this is not a necessary conclusion.

The calculations also indicated that the energy separation between the triplet and the singlet states decreases as the H-N-H angle de creases. This suggests that if the nitrenium ion is formed in a ring structure where it is clamped at both ends near its singlet equilibrium angle, the singlet state gains an extra stability and hence the singlet reaction could predominate in such cases. 18

Once a firm basic for the nitrenium ion theory had been estab

lished, several examples of the synthetic utility of electron-deficient

divalent nitrogen began to appear in the literature. One of the more

interesting of these reports concerned the generation of azabicyclics

via intramolecular addition of a nitrenium ion intermediate to a double

bond in a “jt route" synthesis. ’ Thus the silver ion assisted

(20) p. G. Gassman, F. Hoy da, and J. lygos, J. Amer. Chem. Soc., 9Q_, 2Ti6 (1968).

(21) p. G. Gassman and J. Dygos, Tetrahedron Letters, 47^5 (1970).

solvolysis of cyclopentenyl chloramine ^ in a variety of nucleophilic

solvents gave the azabicyclic ^ in synthetically useful yields. Also,

cyclization of ^ gave ^ and ^ as the major bicyclic products. It was established that the nucleophile added trans to the attacking

nitrenium ion.

(CH2)2N-CHs RO H

RQH

CH2-|-CH3 Cl +

53 54 19 The solvolysis of N-chloroaziridines T^as used hy Horwell and 22 Rees in a synthesis of isoquinoline. Thus the aziridine (56) was

(22) D. c. Horwell and C. W. Rees, Chem. Conmiun., 1428 (1969).

converted to its U-chloro analogue which was solvolyzed with heterolytic cleavage of the H-Cl bond to give the intermediate 5^ which lost a proton to give isoquinoline (5â)* This reaction not only illustrates the synthetic utility of nitrenium ion chemistry but also adds support ing evidence for the electron-deficient nitrogen intermediate proposed IS in the solvolysis of H-chloroaziridine.

H Cl o

56 57 58 I

A novel transannular insertion reaction of a nitrenium ion into a

C-H bond was accomplished in synthetically useful yield by Edwards and 23 co-workers. N-chloroazacyclononane (60), on treatment with silver ion, gave a 68% yield of indolizidine (6]J . The reaction was proposed 20

(23) G. E. Edwards, D. Vocelle, J. W. ApSitnon, and F. Haq.ue, J. Amer. Chem. Soc., 8 ^ 678 (1965).

to proceed through a mechanism involving hydride abstraction followed by ring closure to give the protonated tertiary amine.

60 61

N ^ Hie- — > ''N-H + C— — » N -C • ^ R / ^ r /

Part II. The Problem

The research project with which this thesis is involved is the

development of a new method for the nucleophilic aromatic substitution of aniline derivative via the phenyl nitrenium ion. Nucleophilic aromatic substitution normally occurs in aromatic systems abundantly

substituted with electron-withdrawing groups and in the special case 24 of diazonium salts. In the case of anilines the possibility of nucleophilic substitution seemed remote due to the electron-donating 21

(24) For reviews of nucleophilic aromatic substitution see: J. F. Bunnett and R. E. Zohler, Chem. Rev., 275 (l95l); E. D. Hughes and C. K. Ingold, Quart. Rev., ^ 5^ (1952); J. F. Bunnett, ibid.. 12. 1 (1958); R. Sauer and R. Huisgen, Angew. Chem., 72 , 294 (i960); S. D. Ross, Progr. Phys. Org. Chem., 1, 51 ( 1905)/

power of the nitrogen atom. It appeared likely, however, that if a

suitable leaving group could be situated on the nitrogen atom of an aniline derivative, the solvolysis of such a compound should produce an electron-deficient phenyl nitrenium ion (anilenium ion) which should undergo facile nucleophilic substitution. On considering the relative electronegativities of nitrogen and carbon, it was felt that this posi-

< —

anilenium ion tive charge would be predominantly delocalized from the nitrogen atom to the ortho and para-posit ions in the aromatic ring and that attack by a nucleophile in solution would occur mainly at these two positions.

This process seemed very reasonable in view of the acid-catalyzed rearrangement of arylhydroxylamines which was elegantly studied by 25-32 Bamburger at the turn of the century. It was found that p-phenyl- hydroxy lamine (6^ rearranged to p-aminophenol ( ^ in good yield when 22

(25 H. E. Heller, E. D. Hughes, and C. K. Ingold, Hature, l68, 909 (1951). (26 R. A. Abramovitch and B. A. Davis, Chem. Rev., 64, 176 (1964).

(27 E. Bamburger, Ber., 2J, 154-7 (1894).

(28 E. Bamburger, ibid., 27, 1548 (1894).

(29 E. Bamburger, ibid., 2 8 , 245 (1895).

(50 E. Bamburger and F. Brody, ibid., 53.. 3642 (19OO).

(31 E. Bamburger, ibid., 4o, 1906, 1918 (1907).

(32 E. Bamburger, Ber., 5^. 360O (19OO).

27 treated with aq.ueous sulfuric acid.

HHOH

& &

It soon became evident from this work that the rearrangement was intermolecular. Thus treatment of ^ with ethanolic sulfuric acid 28 gave mainly the aminoethers ^ and M s

65 23 Also, if ethanolic hydrogen chloride was used instead of ethanolic 29 sulfuric acid, the products were not only ^ and ^ but also the

ring-chlorinated compounds 6^ and Since the rearranged amino-

NHa

phenols were stable in these reaction médias, it became apparent to

Bamburger that the products must have come from an intermole cular

reaction.

When the para position in a methylphenylhydroxylamine was vacant rearrangement occurred in the expected way, but that when a methyl group

occupied the para position, the product was a p-hydroquinone in which the para methyl group had migrated to an adjacent carbon atom. Thus 29 68 gave ^ as the major product.

CH. OH

In order to explain this methyl migration, Bamburger proposed the formation of the iminoquinol X2. which was hydrolyzed to give quinol %1^. 2k The quinol could then undergo the dienone-phenol rearrangement to give the final product %2 . In support of this mechanism, the hydroxy lamine

I I

E OH R CH

70 71 72

52, was rearranged to the dienone Jk which was isolated and further con- 30 verted to Also, the iminoquinol was isolated and then hydro- 31 lyzed to %8.

NHOH OH H2SO4

cold warm

73 74 II

MHOH 0 CH; CH; H2SO4 CH;■3 H2O

Eton OEt OEt

75 77 78 25 In terms of a general reaction mechanism to explain all of these reactions, Bamburger proposed that the hydroxylamine suffered dehydra 32 tion to yield a phenyl nitrene intermediate. This nitrene could

79 80 81 then add the molecule HX in two ways, either to the nitrogen to give

or partly to the ring and partly to nitrogen to give 8o_ and 8^. 25 In more modern mechanistic terns, Ingold proposed that the reaction was actually proceeding through the phenyl nitrenium ion.

Protonation of the hydroxylamine could give 8g_ which would lose water

H m

+ 82

to give the de localized nitrenium ion intermediate. Nucleophilic attack by solvent at the para and ortho-positions could then explain all of the products found by Bamburger. It is not known if the reac- 26 tion occurs via an Sjj2 ’ reaction on the protonated hydroxylamine 82_ or ty an Sj^l’ reaction on the anilenium ion itself. However, the reaction is first order in acid and the nucleophile is not involved in the rate determining step. Thus it appears that the acid catalyzed rearrange ment of phenylhydroxylamines is indeed occurring through a phenyl nitrenium ion intermediate, and tliat this reaction represents a method for the nucleophilic substitution of aniline derivatives.

Nucleophilic aromatic substitution of phenylhydroxylamine deriva- 33 tive8 has been proposed to be the process responsible for the carcino-

(55) For a comprehensive review of this subject see: James A. Miller, Cancer Research, 5P- 559 (19T0), and references therein.

genic behavior of some aromatic amines and amides. Most carcinogens possess groups with an electrophilic site which are attacked in a Spj2 manner by a nucleophilic site in the nucleic acids of the mutant cell.

Carcinogenic aromatic amines and amides such as 2 -acetylaminofluorene

( ) , however, seemed to lack such an electrophilic site. It was felt,

§2. % 27 therefore, that these compounds must be metabolized in the body to some

ultimate carcinogen, the compound which actually causes the mutation.

Accordingly, Miller and co-workers discovered that 8^ was being

converted vivo to the N-hydroxy compound 8^. Synthetic 8^ proved

CCHcf ■> ÏÏADPH \OH

83 85

to be a much stronger carcinogen than 8^. However, since 8^ itself did

not react with proteins and nucleic acids, some further activation to the ultimate carcinogen must be occurring. This was subsequently

shown to be conversion of the H-hydroxy compound 8^ to its sulfuric

acid ester 8 ^ which provides the good leaving group needed for nucleo

philic aromatic substitution.

IcHs 85 ‘OSOc

86 28

The sulfuric acid esters "were then reacted with protein and

nucleic acid residues to produce compounds which were also isolated

from test animals. For example, 86 reacts with methionine to produce

8j in 655^ yield, which is also isolated from the livers of rats fed

8 6 . As indicated in Chart I, Miller proposed a reaction involving a

Cj 86 + CH3S(CH2)2CHC0H NH;:

ICH;

87 nucleophilic attack by sulfur on the nitrenium ion precursor in either a Sjj2 or Sj^l manner. Support for the reaction going through the 34 sulfonium salt ^ was provided by Gassman and co-workers, who

(34) p. G. Gassman, R. Smith, and G. Greutzmacher, unpublished results.

pyrolyzed salt § 2 to yield ^ and methyl chloride as the major products. 29

RS-CHs / 86

CCH. ESCH3

CHs I

‘S CH3 4"

CCH;

.CH3

8 1 CHs— HHs 30

+ yCHs s: Cl H (H3C)3C-W'' GHs (HsC)3C-N/

IMF > + CH3CI

The concept of nucleophilic aromatic substitution via the phenyl nitrenium ion, therefore, seemed quite sound. This work, thus, involves developing the concept into a practical synthetic method for making a wide variety of organic compounds and provides physical evi dence for the existence of the phenyl nitrenium ion. Finally, the method will be extended to a variety of indole derivatives. RESULTS AED DISCUSSION

Part I. Synthesis and Solvolysis of N-chloro-N-alkylanilines

In order to generate the phenyl nitrenium ions needed for this study, a suitable leaving group had to be situated on the nitrogen atom of the aniline derivative. Chloride ion proved to be a particularly

suitable leaving group because of the ease in which the W-chloro compound could be made from the parent secondary N-alkylaniline deri vative and because silver ion could be used to accelerate the solvolysis of these N - chloramine s. The N-alkyl groups on the secondary aniline derivative were either N-methyl or N-t-butyl.

The N-methylanilines were commercially available or were obtained from the primary aniline by treatment with ethyl formate followed by lithium aluminum hydride reduction of the resulting formamide. For example, 2,6-dimethylaniline (^ij was converted to the formamide (^2) in 0 ^ yield which was reduced to 2,6-N-trimethylaniline in $2^

0 I! .-CSV ■TO '

21 22 21

31 32 yield. The N-t-hutylanilines were synthesized by condensing the 35 corresponding primary aniline hydrochloride with t-butyl alcohol or

(35) A. Bell and M. B. Knowles, U- S. ^Patent 2,692, 287 ( 1936)j Chem. Abstr., 50, 2666e (1956).

36 by nucleophilic displacement of the aryl fluoride by t-butyl amine.

(36) H. Suhr, Justus Liebig’s Ann. Chem., 175 (I961).

Thus, heating aniline hydrochloride (gjtj with t-butyl alcohol at 150° in a steel bomb gave 63^ of N-t-butylaniline (9^ after workup. Reac tion of p-fluorobenzonitrile (90 with jb-butylamine in EMSO gave a yield of £-(lI-t-butylamino)benzonitrile ( 9jj .

+ Rife Cl" H-N-C( C%)s

21 31

P h - r - c (CH3)3

O) ------

C®tl

3 â 3 L 55 Once the parent N-alkylanillne derivative was on hand, the IT- chloro analog was easily prepared by dissolving the N-alkyl compound in a non-polar solvent followed by treatment with solid calcium hypo chlorite, aqueous sodium hypochlorite solution, or with t-butylhypochlorite to give an essentially quantitative yield of the H-chloro-N- 37,38,39 alky Ian iline derivative. These U-chloramine s exhibit a wide

(57) R. S. Neale, R. G. Schepers, and M. R. Walsh, J. Qrg. Chem., 2ÿ. 3390 (1964).

(38) p. Haberfield and D. Paul, J. Amer. Chem. Soc., 8j, 5502 (1965)» (59) The yields of the N-chloro compounds were determined by either isolation or by treatment of an acidified solution of the chlor amine with potassium iodide followed by titration of the liberated iodine with thiosulfate.

range of stabilities. Some may be isolated as neat oils or crystalline solids, whereas others rapidly decompose at room temperature. For example, N-t-butyl-£-toluidine (gS) was dissolved in pentane and stirred with solid calcium hypochlorite at 0 ° for 1 hour. The salts were then removed by filtration and the pentane evaporated to give the N-chloro compound (99) 9^^ yield as an oil.

H-N-C( 053)3 C l - N -0(053)3

28 22. 34 When very unstable N-chloroaniline compounds were needed, the

method of choice was treatment of the parent E-alkylaniline in a non

polar solvent with t-butyIhypochlorite at -78°. Table 3 lists the

N-alkylaniline derivatives and their N-chloro analogs synthesized in

this study.

Two methods were used for the solvolysis of the N-chloro-N-alkyl

aniline derivatives: silver ion-assisted methanolysis and uncatalyzed

solvolysis in pure or buffered alcohol. The first compounds studied were simply N-chloro-N-t-butylaniline ( 100a) and N-chloro-N-methyl- 40 41 aniline ( lOObj . Silver ion-assisted methanolysis of lOOa and ip_Ob_

(40) For a preliminary report of this work see; p. G. Gassman, G- Campbell, and R. Frederick, J. Amer. Chem. Soc., 9Q, 7377 (1968).

(41) Dr. Ronald Frederick carried out theinitial work on the chemistry of N-chloro-N-t-butylaniline ( 100a). This author thanks Dr. Frederick for his assistance in this work.

R -N-Cl R-N-H R-N-H R-N-H R-N-H (J. O J OCHb + [ X o ,C1 OCHe lOQa R=^-butyl 101^.3# 102a W 104a lOpb R^ethyl 101b 102b 1^ . 104b Table 3

N-chloro-N-alkylaniline s

Compound Source Chlorination Reagent Temp. Yield

N-t-Butylî-£-aminob iphenyl a Ca( 001)2 0° N-t-Buty l-£-t oludine a Ca( 001)2 0° N -t-Butyl-g^-fluoroaniline a Oa(001)2 25 0 N-t-Butylaniline a Oa( 001)2 - 8° N -t-Butyl-p^-chloroaniline a Oa( 001)2 25O Ethyl p-(N-t-Butylamino)benzoate a 0a(00l)2 23° I £-( N -t-Butylamino) benzonitr lie b 0a(00l)2 250 25 0 £-( N-t-Butylamino) nitrobenzene b Ca(0Cl)2 W: N-t-Butyl-o-tolui dine a t-Butylhypochlorite N -methylaniline c NaOOl - 8° 2 ,6-N-Trimethylaniline d t-Butylhypochlorite N -Met hyl-o -aminob iphenyl d t-Butylhypochlorite -78°

^Arylamine hydrochloride and t-butyl alcohol, ^luorobenzene derivative and t-butyl amine. ^Commercial sample, deduction of the formamide. ^Isolated yield. ^Titration yield. ^Yield not determined due to instability. 56

gave mixtures of anisidines as major products and rearranged C-chloro-

aniline derivatives as minor products. Silver chloride also precipi

tated from the reaction mixture. The structures of the anisidines and

ring chlorinated products "were verified hy comparison with authentic

samples prepared from the corresponding aniline hydrochlorides and t- 37 hutyl alcohol or with published spectra.

The overall yield (5i^) from lOOb was considerably lower from that

of 100a due to the slow oxidation of the reaction products by silver

ion. The W-t-butylaniline products were shown to be stable to silver

ion under the reaction conditions.

When IpOa and lOgtL were solvolyzed in methanol without the assis

tance of silver-ion, however, no anisidines were formed and the major products were simply the rearranged ortho and para chi oroanilines.

For example, ipOa gave 52^ of 10^a^. l4^ of ip4a, and 8% of N-t-butyl

aniline. Qualitatively, it was also noted that the silver-ion assisted

reaction proceeded at a faster rate than the solvolysis in methanol without silver ion.

Since the nucleophilic character of the silver ion-assisted solvoly

sis of lOpa and ipob was clearly indicated by the formation of anisi

dines as the major products, it seemed reasonable to propose that the

reaction involved nucleophilic attack by the solvent methanol on some

electron-deficient intermediate. This reaction pathway could involve

initial heterolysis of the N-Cl bond to give the delocalized phenyl

nitrenium ion, followed by attack of methanol predominantly at the para position in a Sjjl' manner. A second mechanistic possibility in volves a silver ion-assisted displacement by methanol in a 8 ^ ' manner 57

Rv +

<— »

R - N - Cl

R —N—E

%C 0 H

on the N-chloro compound. Both of these mechanisms would produce a dienimlne intermediate which would rearomatize to give the observed major product.

The precipitation of silver chloride and the fact that the silver- ion assisted reaction is faster than the uncatalyzed reaction also supports either of the above mechanisms. It was noted, however, that the presence of sodium methoxide in the reaction medium did not increase the percentage of anisidine products formed. This tends to indicate that the cleavage of the N-Cl bond occurs before the attack of solvent, that is, that the nucleophilic attack occurs on the phenyl nitrenium ion in a 8^' manner. Additional support for this contention will be presented later. 38 If this méthoxylation reaction indeed proceeds through the dienimine intermediate, then replacement of the para hydrogen hy a group less prone to migrate should enable one to trap the dienimine as a stable compound. This was accomplished in the solvolysis of N-chloro-

IT-t-butyl-£-toluidine (22) • When 22,’was treated with a methanolic solution of silver trifluoroacetate, a 7C^ yield of 10^ and 17^ of 106 was produced.

^ . 0 ( C H 3 ) 3 h -N-C(CH s )3

HgC OCHs CHs

22. 101 106

An authentic sample of 10^, which is a stable crystalline compound, was synthesized in 52^ yield from 4 -methoxy-4-methylcyclohexa-2,5- 42 dienone ( 10%) by the initial addition of lithio-t-butylamine in ether

(42) E. Hecker and R. Lattrell, Annalen, 662 , 48 (1965).

43 followed by the addition of 1 equivalent of methyllithium. 39

(45) We thank Dr.. R. Steppel for suggesting this synthetic approach.

CCCHaîb 0 N

IfeC^^OCHs IfeC HbC^^OCHb

101

Likewise, reaction of 4 -(N-chloro-N-t-tutylamino)biphenyl ( l ^ with silver ion in methanol, followed hy hasic aqueous workup and chromatography gave 62^ of and 5^ of 110.

Cl-N-C(CHb)3 0 H-F-C(CH3)3

0 OCHa 1>

108 102, 110 4o When ^ was solvolyzed in methanol in the absence of silver ion, however, the major product was the o-chloro compound ip6. This non silver ion catalyzed reaction gave bl^ of 106 and 32^ of 103.

The isolation of these stable dienone derivatives adds considerable supporting evidence to the proposed nitrenium ion mechanism given above for the production of anisidines in the solvolysis of N - chloro -N-alkyl - anilines. The presence of a carbonium ion-stabilizing group in the para position of the N-chloroaniline derivative increases the amount of solvent-capture products. Thus lOOa^gives 43% of solvent incorpora tion while 9P and IO8 give 7C% and 6^ of methoxyl products, respec tively. This behavior exemplifies the electron-deficient nature of the aryl nitrenium ion intermediate, which is expected to be stabilized by electron-donating substituents.

N-chloroanilines substituted in the para position with a carbonium ion-stabilizing group can thus be converted into derivatives of cyclohexa-2 ,5-dienone, which are valuable intermediates for the synthesis 44 of a variety of organic compounds.

(44) E. J. Corey, S. Borca, and G. Klotmann, J. Amer. Chem. Soc., 91. 4782 (1969).

The reaction is also convenient in that one can isolate the ketone or stop at the imine depending on the method of workup. Thus ketone

109 was isolated following an aqueous sodium hydroxide workup, while the imine 103 was obtained after a weakly basic ammonium hydroxide workup. k-1 Since the solvolysis of N-chloroaniline derivatives substituted

in the para-posit ion with carbonium ion-stabilizing groups give syn

thetically useful yields of cyclohexa-2,5 -dienones, it was hoped that

the solvolysis of ortho-substituted compounds would yield derivatives

of cyclohexa-2 ,4 -dienone. However, N-chloroanilines substituted in the

R — N -Cl

OCIfe

ortho-position with carbonium ion stabilizing groups gave para-anisidines

as the major product. Thus on silver ion-assisted solvolysis of 111,

we obtained of 112_ and 17% of 11^ . Likewise, Ilk gave 51% of lig

and k% of U .;6 while 117 yielded kC% of ll8 . Solvolysis of 111_ in

methanol without silver ion gave of 112 and 37% of II3. while

uncatalyzed methanolysis of Ilk gave lk% of II5 and 35% of II6. As in

the case of the p-methyl compound 99, these compounds gave methoaylated

products even in the absence of silver ion. This indicates the ease

of generation of the phenyl nitrenium ion intermediate when electron-

donating groups are present on the aryl nucleus of the H-chloroaniline

derivative. lj-2

C% HbC o pife

OCIfe Cl

111 112 m .

HaCx ^C1 %T

ll4 m . 116

C1 ^^_C(C% ) 3 H . ^ ^ C ( C % ) 3

ÜI

These reactions are in agreement with Bamburger*s work, who reacted 29 lj-9 to produce ]2 0 . The failure of these ortho-substituted compounds to give derivatives of cyclohexa-2 ,4 -dienone is probably a reflection of the already crowded steric situation at the ortho position. Even though the positive charge is expected to reside significantly on the carbon atom bearing the carbonium ion-stabilizing group, therefore, the nucleophilic attack by solvent occurs at the para position. 43

An Increase in the yield of the rearranged C-chloroaniline deriva tive was noted when the reaction was carried out with H-chloro-H-alkylanilines substituted in the para position with electron-withdrawing

groups. In these reactions only small amounts of products resulting

from nucleophilic attack by solvent were observed. Instead, minor but significant amounts of the starting secondary anilines were pro duced in the silver ion-assisted reaction along with the corresponding amounts of silver chloride precipitation. In this regard, silver ion- assisted methanolysis of p-(N-chloro-H-t-butylamino)nitrobenzene (l2 ^

gave 65^ of o-chloro compound ( 1 2 2 ) and 2 ^% of starting aniline ( 123).

In the same manner, the p-carboethoxy chloramine ( 124 j yielded 38^ of

2 2 ^ and 2 ^ of 126 .

(45) ¥. Duerckheimer and L. Cohen, Biochemistry, ^ 1948 (1964 ). kh

0 1 ^ ^ ^ 0 (053)3 3 h^ ^ ^ c(c% ) 3

122

Cl^^^C(C%)a H^^^C(CHa)3 ^^^^C{cSq )q

OOgEt

126

A complex reaction mixture resulted from the silver ion-assisted solvolysis of îî,p-dichloro-îî-t-butylanilin.e {12J ), the £-chloro group being only weakly electron-withdrawing. Thus 1 2 J gave 25% of the o-chloro confound 128, 2 ^ of 12^, IC^ of 130, and 12^ of starting aniline 1^ 1 . When the solvolysis of 1 2 1, 1 2 4 , and were carried out in pure methanol without silver ion, only trace amounts of starting aniline resulted in the product mixture. For example, 121 gave 99 ^ of

122 and of I2g , yielded 93^ of 1 2 j _ and h$> of 1 2 6 while the solvoly sis of 12%_ in pure methanol resulted in 74 % of 128 and iSh of starting aniline. 45

I I

H3CO' 'OCH3 Cl Cl

iSL lêâ. 3^ 1 ^ ]Ji

The electron-deficient nature of the aryl nitrenium ion interme diate is again substantiated by the occurrence of these o-chloroaniline derivatives, which are the major products from the silver ion-assisted

solvolysis of the chloroamine compounds with electron-withdrawing para-

substituents. It was noted above that with the N-chloroanilines with electron-donating substituents, nucleophilic attack by solvent was the major reaction pathway. Conversely, as the substituents became strongly electron-withdrawing, nucleophilic solvent incorporation did not occur.

Table 4 lists the percentage of solvent incorporation in the products from the silver ion-assisted methanolysis of the 2-substituted H-chloro-

N-t-butylanilines. A good correlation is noted between the electronic nature of the para -sub st ituent and. the percentage of methoxy products. 46

Table 4

percentage of Methoxy Products from the Silver Ion-assisted Methanolysis of N-Chloro-KT-^b-butylanilines

para-Substituent $ Methoxy Products

Phenyl 62 ? H 4 ^ Cl 36 ^, COsEt NOa - ^

^No methoxy products were observed within the limits of detection.

This behavior can be explained in terms of a tight anilenium ion- chloride ion-pair which should become tighter and resist nucleophilic displacement by solvent as the substituents on the aryl nucleus became electron-withdrawing. It becomes attractive to suggest that not o n l y the anisidines and cyclohexa-2, $ - dienone derivatives, but also the o-

chloroanilines are formed via the intermediacy of the phenyl nitrenium

ion.

The actual mechanism of these ring-chlorination reactions, however, remained a mystery. Several mechanistic schemes could be suggested for the conversion of 1^2 into and 1^ 4 . of the various possibilities. k'l the most likely choices are heterolytic cleamge of the chloroamine to 87 produce amide anion and positive chlorine, or heterolytic cleavage of

the N-Cl bond to yield divalent electron-deficient nitrogen (nitrenium

ion) and chloride anion. Either of these ion-pairs could then recombine

B x .01

* ( ê ( ' “ •

to give the dienimine intermediate, -which is expected to rearomatize

Cl l o r ' ,4- 01- 48 under the reaction conditions. In order to distinguish between these divergent concepts, we have studied the kinetics of the facile rear rangement of para-substituted N-chloroanilines in ethanol buffered with 46 0.1 N acetic acid - 0.1 N sodium acetate.

(46) For a preliminary account of this work see: P. G. Gassman and G. A. Canÿbell, J. Amer. Chem. Soc., 9$. 2^ 6j (1971).

Table 5 lists the rates of solvolysis of a series of para-substituted N-chloro-N-t-butylanilines in buffered ethanol. Inspection of

Table 5 shows that the presence of electron-withdrawing substituents results in a marked decrease in the rate of rearrangement of the six anilines studied. The overall rate decrease in going from the para- methyl derivative to the para-cyano case 1^6 was 1.8 x 10®. This decrease in rate indicated that the transition state for the rearrange ment of 1^2 involved the generation of a positive charge adjacent to the aromatic ring. Confirmation of this concept was provided by Figure 47 1 which shows a plot of log k vrs. the Brown substituent constants.

(47) H. C. Brown and Y. Okamoto, J. Amer. Chem. Soc., 80, 4979 (1958).

For the six compounds studied, we obtained an excellent correlation at

25° with The P for the reaction is -6.55 and the correlation co efficient is 0.996• This linear correlation plus the fact that all of the compounds listed showed good first order kinetics indicates that the ring chlorination of these aniline derivatives occurs via phenyl Table 5

Rates of Rearrangement of N-Chloroanilines in Ethanol Buffered with 0.1 N Acetic Acid - 0.1 N Sodium Acetate

k Compound Temp (± 0 .02°C) (sec~^) rel (kcal/mole) As^ (e.u.)

Cl - 10.0 ( + . ) X 10 5.27 0 05 -1 CHb-^^N-C(CHa)3 0.0 (2.57 + O.Ol) X 10 20.A -0.5 10.0 a (8.79 + 0.01) X 10 22. 25.00' 5.55 X 10“® 1,800,000

Cl I 20.0 (7.^7 + 0.01) X 10“ F -^O^N-C( 0113)3 30.0 (2.60 + 0.01) X 10-4 21.2 -5.0 -4 40.0 r (8.18 + 0 .0%) X 10 121 25.oof l.îb X 10“4 %%,000

Cl 50.0 (5.83 + 0.03) X 10 H - ( ^ W - C ( C H 3)3 %0.0 (1.27 + 0.02) X 10-4 21.9 - 6.6 50.0. (3.91 + 0 .01) X 10-4 100a 25.00° 2.06 X 10-5 6,500 Table 5 (Continued)

k Compound Temp ( ± 0 .02°C) rel Ah ^ (kcal/mole) A s ’^ (e.n.)

Cl hO.O (1.12 +0.02) X 10 Cl-@-M-C(C%); 50.0 (3.60 + 0 .02) X 10-4 23.0 -3.1 60.0 (1.09 +0.01) X 10-3 .a m . 25.00 1.55 X 10^ 5,200

01 I 70.0 (l.OT + 0 .02) X 10-3 EtOC-^-N-C( C%)s 80.0 (3.97 + 0 .02) X 10-5 26.3 -4.9 90.0 ( + 0.32) X 10-3 a 9.35 124 25.00 2.96 X 1 0 - ®

-5 Cl 90.0 (2.04 + 0.08) X 10

100.0 (6.10 + . ) X 1Q-® - m = 0 - @ - N - C ( 003)3 0 04 _4 28.4 2.2 110.0 (1.68 + 0 .04) X 10 321 3.19 X 1Q-®

^Extrapolated from other temperatures.

^Rates were determined by iodometric titration of active chlorine. Unreacted N-chloroanilines were first reacted with acidic potassium iodide followed by titration with thiosulfate using a dead- stop titration method. VJlO 51

-2

-3

-4 log k -5

-6

p = — 6.35 -7 l/bl = ,996

-S

-9

-,4 0

Figure 1 . a p plot for the solvolysis of N-chlor o -N -{k-butyl-

anilines in buffered ethanol. 52 nitrenium ion intermediates.

The products of this buffered reaction are also in agreement with a nitrenium ion mechanism and these are given in Table 6. There are

Table 6

Products from the Solvolysis of ^-Substituted N-Chloro-N-t-butylanilines in Ethanol Buffered with 0.1 IT Sodium Acetate and 0.1 N Acetic Acid

p-Substituent ^ o-Chloroaniline % Starting Aniline a Clfe 85% F 64% 8%

h" 48% l6%

Cl 72% 17%

COgEt 78% 15% ON 64% 29%

^IC^ of 4 -methoxy-4 -metbyl-2 ^ ^ -cyclohexadienone -E-t-butylimine (10^) was also produced in this reaction.

^11^ of N-t-butyl-p-chloroaniline and 5^ of IT-t-o,p-dichloroaniline were also formed in this reaction.

three major points to be brought out in this product study. The first

is that all six of the anilines gave the o-chloro derivative as the major product. Since there was no silver ion in the kinetic runs, the 55 nitreniuni ion-chloride ion-pair simply recombined instead of undergoing nucleophilic displacement by solvent. Secondly, as noted before, the unsubstituted compound (lOpa) gave a mixture of the ortho and para- chloroaniline derivatives. But the o to p-chloro-N-t-butylaniline ratio was 4.6 in the silver ion-assisted reaction of 100a and 3*7 in the uncatalyzed reaction. Thus it appears that the formation of p- chloro-N-t-butylaniline was also the result of the nitrenium ion- chloride ion-pair which could be partially intercepted by silver ion.

Thirdly, solvolysis of 22. buffered ethanol gave IC^ of 10^. indi cating the presence of positive charge on the ring.

Another interesting feature of the product study is the occurrence of the starting aniline derivative in the product mixture. Table 6 shows that as the p-substituent becomes more electron-withdrawing the

amount of this product gradually increases. An attractive explanation is that as the substituent on the aryl nucleus becomes more electron- withdrawing, the charge density increases on the nitrogen atom of the delocalized anilenium ion allowing a spin inversion to the triplet nitrenium ion followed by two hydrogen atom abstractions to give proto- 54 nated form of the starting aniline derivative. An alternate mechanism involves a simple hydride abstraction by the nitrenium ion to go directly to the starting aniline derivative. Regardless of which of these two processes represents the actual mechanism, both are still derived from an electron-deficient nitrogen intermediate. Therefore, all of the products from the solvolysis of E-chloro-R-t-butylanilines in buffered media can be adequately explained by the intermediacy of the aryl nitrenium ion: ring chlorinated products, anisidines, cyclo hexa-2 ,5-dienone derivatives, and starting aniline compounds.

On comparing the solvolysis of N-chloro-l-alkylanilines with that of substituted 1 -chloro-1 -arylethanes, it appears that due to the relative electronegativities of nitrogen and carbon, the nitrogen cation is much more dependent on charge delocalization into the aromatic ring than was the analogous carbonium ion. Thus the p for the solvolysis of 1 -chloro-1 -arylethanes in 20:80 water-dioxane has been reported to 48 be -4 .50. This is also apparent from a comparison of products from

(48) C. Mechelynek-David and p. J. C. Fierens, Tetrahedron, 6, 232 ( 1959). the two reactions. While the E-chloroanilines give mostly products incorporating substitution on the aryl nucleus, the 1-chloro-l-aryl- ethane compounds give only derivatives of benzyl alcohol when solvolyzed 49 5 0 under Sjjl conditions.

(49) J. Steigman and L. P. Hammett, J. Amer. Chem. Soc., 59j 2536 (1937). 55 (50) E. D. Hughes, C. K. Ingold, and A. D. Scott, J. Chem. Soc. ,1201 (1937).

An investigation was then made into the rates and products of the

solvolysis of the same series of para-substituted N-chloro-N-t-butyl

aniline s in pure, unbuffered ethanol. The kinetic data demonstrated

that the buffer was unimportant for compounds 22^ lj^_, Ippa, and 12p

since these four chloramines showed essentially the same rates and pro

ducts, irrespective of the presence of buffer. Tne solvolysis of 12h

and 1^6 in unbuffered ethanol, however, gave kinetic data indicative

of a rapidly increasing rate of reaction. Figure 2 contrasts the dif

ference in the behavior of compound 22h_ in the two solvent systems.

The product ratios for the solvolysis of 124 and 1^6 in pure ethanol

were also quite different from the ratios obtained in buffered ethanol

The products from this unbuffered solvolysis are given in Table J.

Table 7

Products from the Solvolysis of p-Substituted H-chloro-N-t-butylanilines in Pure Ethanol p-Substituents % o-Chloroaniline Starting Aniline

GHs^ 7^ 6 # ll3^ 51^ # Cl 9^ 5^ COgEt m 9K1>

^12^ of 4 -methoxy-4 -methyl-2,5-cyclohexadienone N-t-butylimine ( 10^ ) was also formed in the reaction. ~

^24^ of N-t-butyl-£-chloroaniline and of H-t-butyl-o,p-dichloroaniline were also formed in this reaction. r— I O I z L_J

Pure Ethonol Ethonol .lj\| NoOAc —.I HOAc

0 1000 2000 3000 Tl ME(sec.)

Figure 2. Kinetic Behavior of Ethyl £_-(N-chloro-N-t-'butylamino)benzoate in

Buffered and Unbuffered Ethanol. VJl ON 5T Comparison of Table 6 and Table 7 indicates that vhen the solvolyses are done in pure ethanol, the amount of starting amine produced is

significantly reduced for compounds 12J. 124. and 1^6 .

In view of the excellent first-order kinetic data obtained for

124 and Ip6 in buffered ethanol, the observations for the unbuffered

solvolysis indicated that a change in mechanism was occurring. The

change in product ratios for the two solvent systems is also in accord with this conclusion. Clearly the solvolysis of 124 and lp6 could be acid catalyzed by small amounts of hydrogen chloride produced in the

initial stages of the reaction, while the solvolysis of 100^ and 127 was not sensitive to small amounts of acid. The unbuffered reaction for 124 and lp6_ could then proceed via the classical Orton 51 rearrangement, which is known to be specifically catalyzed by strong mineral acids. The product differences in buffered and unbuffered ethanol can also be accounted for by this change in mechanism.

(51) For a discussion of the Orton rearrangement see: E. S. Gould, “Mechanism and Structure in Organic Chemistry," Holt, Rinehart, and Winston, Hew York, H.Y., 1959? P« 65O, and references therein.

HCl ^ + HCl 58 Part II. Synthesis and Solvolysis of N-Alîsyl-H-arylhydroxylamine p-Eitr ohenzoat e s

In order to investigate the effect of changing the leaving group on the nitrogen atom in these solvolysis reactions, we synthesized and studied the methanolysis of two arylhydroxylamine p-nitrobenzoates.

Several examples of the chemistry of esters of arylhydroxylamines have 52 appeared in the literature. In 1919 it was reported that 0 ,N-diben-

(52) E. Bamburger, K. Blaskopf, and A. Landau, Ber., 52, III6 (1919) *

zoyl-N-phenylhydroxylamine ( 1^7) violently explodes on extreme heating. 53 On investigating this reaction further, Homer and Steppan found that

(55) L- Homer and H- Steppan, Ann. Chem., S06, 24 (1957)-

l^J rearranges smoothly in a melt at 150° to give a $<3 yield of the o-aminophenol derivative 1^8 . The reaction was also carried out in

0 0 0 ^ If II . II 0 0 -C-N-OC 0 0 - C - W - H II 59 a variety of non-nucleophilic solvents, in an attempt to determine the mechanism of this reaction, a study was made into the effect of

changing the acyl group. These workers found that acyl derivatives

of stronger acids rearrange more easily than the corresponding acyl

derivative of a weaker acid. In this regard 1^^_ did not react under the conditions where lAO was converted in 98^ yield to l4l. Also,

0 0 II II / — \ R-C-IT-OC-/Q)-0CH8

■ #

0 0 II II / — \ R-C-N-OC-/Q VNOg R-C- -H || —

l4o i4l

the p-methoxy compound l42_ could not he isolated hut rearranged spon taneously at room temperature ..to l4^ . 6o

0 0 0 Il II II 0 0 -C-K-OC - 0 0C-ÎI-H II 25 O & 0 0 % 0 0 %

üi2 ü a

The compound 1 ^ prepared from N-acetylphenylhydroxylamine and dichloroacetyl chloride and labeled at the dichloro acetyl carbonyl 54 ■with 0^ was studied by Cox. He obtained both the ortho and para-

(54) J. Cox and M. Dunn, Abstracts, 152nd meeting of the AOS, Abstrt. SI62, Ne-w York, September ( 1966J.

dichloroacetyloxyacetanilides, and the phenol oxygen after hydrolysis contained half of the 0^ label. Thus the two oxygens of the migrating

0 0 “ 0 0I8 %OC“N-000%012 %CC-H-H jj ,0^®C0H01g 6l group became equivalent in the process. Together with Homer's data, this points to an ion-pair, or to an even more dissociated interme diate in the rearrangement of O-acylphenylhydroaylamine derivatives.

0 + . II 0 R-W-OCR' R-N OCR’ R-W-H

In contrast to these results, Lwowski reported data which indi cates that sulfonate esters of phenylhydroxylamines rearranged via a concerted mechanism. The nosy late lk6^ from H-benzoylphenylhydroxyl- amine and p-nitrobenzenesulfonyl chloride-0^® rearranged to give the nosylate of o-aminophenol which was hydrolyzed to give the aminophenol

]Aj_wlth 5C^ of the 0^ in the phenolic oxygen. Thus the 0^® was

(55) G. Tisue, M. Grassmann, and W. Lwowski, Tetrahedron, 2^, 999 (1968).

not scrambled in the rearrangement. Both an ion pair and a radical- pair mechanism are ruled out. The authors, therefore, proposed a concerted mechanism. This result was une3^ected since sulfonate esters are considered to be better leaving groups than the benzoate and acetate 62 esters, which seem to rearrange via an ion-pair mechanism.

0^ ( 50^) 0

- N — 0 - (50^) Q I 0i8 (5C^)

146

The purpose of this investigation is to effect nucleophilic sub

stitution of N-alkylphenylhydroxylamines by solvolysis of their p-

nitrobenzoate esters. So that a valid comparison could be made with the ïï-chloro-ïï-t-butylaniline derivatives, the K-alkyl group was chosen

to be t-butyl. N-t-butyl-lî-phenylhydroxylamine (l4^ was synthesized by the addition of phenyllithium to t-nitrobutane. The reaction has been

OLi (lfeC)3 C - i r - 0 H

Q 1 + (ïfec)3C-ro2 O J + (HbC)3C-N=0

148 63 56 proposed to go through t-nitrosobutane as an intermediate. The p-

(56) K. Hoffman, A. Feldman, and E. Gelblum, J. Amer. Chem. Soc., 8 6 .

6k6 (1964).

methyl compound was made in 48^ yield in the same manner. The p-nitro-

benzoate esters were made in 79 and 88^ yield, respectively, by treat

ment of the hydroxylamlnes in ether at low temperature with p-nitrobenzoyl chloride.

Solvolysis of N-t-butyl-H-phenylhydroxylamine-^-nitrobenzoate ( lk-9 )

in methanol buffered with sodium acetate, followed by column chromato

graphic separation of the products, gave 52?^ of H-t-butyl-o-aminophenol

( 1^ ^ , 18% of ipia, 5^ of 102a, and 3^ of N-t-butylaniline. In addi

tion, 63^ of methyl p-nitrobenzoate and 3C^ of p-nitrobenzoic acid were

isolated. Compound IjO was identified by comparison to an authentic

(H3C)3C-N-0FKB H - H - C ( C H s ) 3 H-H-0(053)3 5-5-0(053)3 5-5-0(053);

0 1 • (if* (o) 00H3

149 1^0 101a 102 a 64 sample made from o-aminophenol hydrochloride and t-butylalcohol. Like

wise, hnffered methanolysis of £-nitrohenzoate Ijl and separation of

the product mixture by chromatography on basic alumina yielded 4C^ of

the dienone 10^, 4^ of 1^2_, of N-t-butyl-p-toluidine. She struc

ture of the o-aminophenol 1^2 was assigned on the basis of its ir and

nmr spectra, elemental analysis, and by analogy with compound 1^.

/C(C%)s (HsCJsC-N-OBTB -N-C(0% ) s H-N-C

o T ' HsC 0 0 % T 1 0 % c%

m . 101 28

It was proven by a control experiment that the o-aminophenol 1^0

was a result of a transestérification reaction of the rearranged p-

nitrobenzoate Thus treatment of with p-nitrobenzoyl chloride

gave 1^3 in 82^ yield. Solution of 1^ in methanol instantaneously gave nearly quantitative yields of 1^0 and methyl p-nitrobenzoate.

0 (%C)3C-N-H 0-00% JvOHIB — > no 65 From, this data it is apparent that solvolysis of phenylhydrozyl-

amine p-nitrotenzoates proceed, at least in part, via heterolytic

cleavage of the N-0 bond to generate a phenyl nitrenium ion inteime-

diate and p-nitrohenzoate anion. This is shovm by the occurrence of

25^ of anisidine products from 1^ and of the cyclohexa-2 ,5-dienone

derivative from 1^1 . One question which cannot be answered on the

R-N-OHIB R-R’*’ "o ENB

on the basis of the data presently available is whether the formation

of the rearranged p-nitrobenzoates occurs in a separate concerted reac

tion or whether they are also arising by means of a tight ion-pair.

The second possibility has the advantage that a single type of hetero

lytic cleavage can account for all of the products. In view of the 53 54 work of Horner and that of Cox, it is reasonable to favor the ion- pair mechanism.

A ccmparison of the behavior of the N-chloroanilines and hydroxylamine £-nitrobenzoates shows that in nitrenium ion chemistry the £- nitrobenzoate anion is a slightly better leaving group than the chloride

ion. This conclusion can be made on the grounds that slightly more nucleophilic solvent capture occurs in the solvolysis of hydroxylamine 66 p-nitrolienzoates than in the non-silver ion-assisted N-chloroaniline solvolysis. Thus l4g gave 25^ of anisidine products while the uncata lyzed reaction of 100^ produced no anisidines. Also, 1^1 yields 40^ of the dienimine 10^ while the solvolysis of 22. S&ve 525^ of 10^. Quali tatively, the rates of solvolysis of the p-nitrobenzoates and the ET- chloroanilines are also similar.

These relative leaving group abilities are in marked contrast to the analogous carbonium ion chemistry. In carbonium ion chemistry, chlorides can solvolyze as much as 10^ times faster than the correspon- 57 ding £-nitrobenzoates. An adequate explanation of this phenomena.

(57) For comparisons of carbon £-nitrobenzoates and chlorides see: (a) A. Fainberg and S. Winstein, J. Amer. Ghem. See., jQ, 2770 (1956); (b) C. Wilcox, Jr. and M. Mesirov, ibid., 84, 2757 (I962); (c) 0 - Benfey, E. Hughes, and C. K. Ingold, J. Chem. Soc., 2488 (1952); (d) M. Silver, J. Amer. Chem. Soc., 8^, 4o4 ( 1961) ; (e) R. Buckson and S. Smith, J. Org. Chem., 32. 634 (1967).

which appears to be general, must await further experimentation. This 58 behavior has also been found by other workers in comparing nitrogen p-nitrobenzoates and chlorides.

(58) p. G. Gassman, G. Hartman, R. Cryberg, and J. Trent, unpublished results. 67 Part III. Bucleophillc Aromatic Substitution of Indole Derivatives.

Since N-chloroaniline derivatives cotild be methoxylated in the

para-position under extremely mild conditions, an attempt was made to

extend the reaction to indoles. It was felt that if H-chloroindoles

such as 1^4 could be made, then methanolysis of these compounds would

yield derivatives of 5 -methoxyindole ( Ij^). This would be an extremely

useful synthetic method because 5-oxygenated indole derivatives are found in a wide variety of naturally occurring alkaloids. Also, many

of these derivatives play an important physiological role and are mar keted as therapeutic agents. Thus many members of the iboga family

of alkaloids such as ibogaine ( lj6) possess a methoxy group at a position para to the indole nitrogen atom. 5 -Hydroxy groups are also

common in alkaloid systems. 5 -Hydroxy functions occur in hunterbumine

(l^X) and sarpagine (l^â)* 5-%rdroxytryptamine has been detected in many animal sources as well as in the stinging nettle and the banana.

It is a powerful smooth muscle stimulator and the closely related N-

acetyl-5-methoxytryptamine, a consitituent of the pineal gland, has 59 potent stimulating properties. 68

i5i HO 25L

(59) W. I. Taylor, “indole Alkaloids," Pergamon Press, Ne-w York, N.y., 1966, p. 30.

It became apparent, hcwever, from literature reports and from our o"wn work that N-chlorination of indole derivatives could not be achieved.

The reactions of indoles with various chlorinating agents have been 60-67 investigated by several groups of workers. In general it has been

(60) W. Godtfredsen and S. Vangedal, Acta. Chem. Scand., lOj lln4 (1956).

(61) n. Pinch and W. Taylor, J. Amer. Chem. Soc., 8 ^ 1318 (I962).

(62) J. Shave 1 , Jr. and H. Zinnes, ibid., 8 4 , 1321 (1962).

(63) N. Finch and W. Taylor, ibid., 84_, 387I ( 19^2).

(64) G. Buchi and R. Manning, ibid., 88^ 2532 (1966).

(65) H. Zinnes and J. Shovel, Jr., J. Org. Chem., 1765 (1966). 69

(66) L. Dolby and G. Gribtle, ibid., ^2 , 1591 ( 19^7) «

(67) M. Ohno, T. Spande, and B. Witkop, J. Amer. Chem. Soc., 9 0 . 6521(1968).

proposed that indole derivatives of the geneiul type 159 react with halogenating agents such as t-butyl hypochlorite to yield 5 -haloin- dolenines ( 160). Also, it has been suggested that ij.6p^ equilibrated

Cl » jj CHgR CHR

H H

159 160 1 ^

with the corresponding 2-aIkylidene-5-chloroindolines (161) with ease under the reaction conditions. Starting with 160 and 1^ a variety of reaction paths have been described to explain a variety of products. 61-66 These reactions are summarized below. Several workers have evoked

160 as the first intermediate in the oxidative conversion of indoles into oxindoles (16^ . They propose initial addition of the solvent to

160 to yield l62 followed by concerted loss of hydrogen chloride and rearrangement to yield 165. which on hydrolysis would give the oxindole l64. The formation of the oxygenated indoles represented by l66^ has 70

ClfeR ClfeR R'OH

H 160 162

CHaR CHsR R'OH» CHR CHR 0

OR'

R'OH

'CHR

H 161

"been discussed in terms of both the l60-l6^-166 and the 16O-16I-166 64-66 routes.

With these reactions in mind, we set out to investigate the nucleophilic substitution of indoles, especially the concept of oxygenating the 5-position. 2 ,3 -Dimethylindole ( 167). a commercially available material, was chosen to be the model compound. Reaction of l6j with aqueous sodium hypochlorite solution at -5 to -10° or with t-butyl

hypochlorite at -78° gave the very unstable 3 -chloro-2,3 -dimethylindolenine (168) . Under no conditions could any N-chloro compound be observed or isolated. One question was whether initial chlorination occurred at the 1-position or the 3 -position of the indole nucleus. By 71

Cl. CHb

N ' t:h3

i 6t 168

analogy with the chlorination of anilines it might have "been anticipated

that the initial attack would be on the more nncleophilic nitrogen atom,

followed by rapid migration of the chlorine from the nitrogen to the 68 5 -position. Even though the N-chloro compound was not obtained, the

(68) Support for this process can be found in the rearrangement of 17^ to give 180.

widespread reports of the instability of $ -chloroindolenines prompted

us to investigate the chemistry of 168.

The structure of 168 was established on the basis of several pieces

of evidence. The nmr spectrum of 168 taken at -25° in methylene chloride

showed two sharp three proton singlets at t 7-T^ and 8.55, assigned to

the C-2 and C-5 methyl groups, respectively. This is to be contrasted

with the methyl groups of 2,5-dimethylindole ( l6j) which appear as

sharp singlets at t 7*85 and Additional nmr evidence was provided by the nmr spectrum of 2,5 -dimethyl-5 -methoxyindolenine (l6^) which

showed the C-2 and C-5 methyl groups at t J .82 and 8 .64, respectively.

As expected chlorine was slightly more de shielding than methoxy. 72

The -ultraviolet spectrum of 1_68 in methylene chloride showed a maximum at 266 mp with a shoulder at 292 mp,. 2,3 ,3 -Trimetbylindole- 69 ■ • nine and 169 show maxima in alcoholic solvents at 255 and 256 np,

(69) The uv data was obtained by Dr. Goverdhan Mehta.

respectively. These values are to be contrasted with the characteristic ultraviolet absorption of 2,5-dimethylindole which appears at 280 mii>.

When a cold methylene chloride solution of 168 was added to a solu tion of silver trifluoroacetate in methanol at -10°, the precipitation of silver chloride occurred and work-up gave a 9^^ yield of 2,5-dimethyl- 70 3 -methoxyindolenine ( 169) • The structure of 169 was based on a com-

(70) Dr. Goverdhan Mehta carried out the conversion of l68_ to 169. This author thanks Dr. Mehta for his help in this area.

CHa

168

bination of its ir, nmr, and uv spectra and mass spectral molecular weight determination. It is unknown whether this reaction occurs in a

% 2 manner or proceeds via a reaction pathway involving the discrete 75 delocalized, ion 170. Nucleophilic substitution of this ion would be

expected to occur at the stabilized tertiary, benzylic carbonium ion.

( o n " +

In an attempt to deactivate the five-member ed ring and force the méthoxylation of the benzene nucleus, the reaction was carried out with

indoles having electron-withdrawing groups in the 2 or 5 -positions. In

this regard, the chlorination and silver ion-assisted methanolysis of

2-carboethoxy-^ -methylindole still gave a yield of 2-carboethoxy-5- methoxy-3 -methylindolenine. Chlorination and solvolysis of 3 -carbo- methoxyindole gave intractible tars. Silver ion-assisted methanolysis

of 5-chloroindolenines, therefore, did not give the desired 3-methoxy

derivatives.

OCHb

COgEt Tk- Although methylene chloride solutions of l68 were stable for

several hours when kept at -10 both spectroscopic and iodometric anal- 71 ysis showed that 1 @ underwent a dramatic change upon warming the

(71) It is interesting to note that 1^ . being a vinylogous N-chloramine, retained its oxidizing properties. Thus, could be reacted with acidic potassium iodine solution and the liberated iodine titrated with thiosulfate.

methylene chloride solution to room tenperature. At this temperature all

active chlorine was rapidly lost. On standing, this solution darkened

and intractible material was produced. However, if the methylene

chloride solution of 1 ^ was rapidly warmed to room temperature followed by immediate cooling to -25° a solution of a new compound was obtained.

The nmr of this solution showed that only trace amounts of 168 remained.

In place of the two methyl absorptions of 1^ , there appeared two new

absorptions as singlets at t 5 «26 and 7.68. These absorptions had

relative intensities of two to three, respectively. This spectrum

provided convincing evidence that 3,^ had undergone a dramatic change

at room temperature to give a new compound IJj^. On considering the nmr

data, four structures are possible for the product from 168.

Cl C% o CHs ^ 0 % -N' 'CH2CI H H Cl H Cl"

m a m b ms. m i 15 The structure of 171 can be further defined by its ultraviolet

spectrum in methylene chloride at -25°. The uv absorption due to 168

completely disappeared and a new absorption maximum approximately five

times as intense as that of the solution of 1 ^ appeared at 285 mji.

Although the extinction coefficient of this absorption could not be

deteimined accurately it is estimated to be around 20,000. Since 2,3-

dialkylated indole derivatives have extinction coefficients around 72 6,500, the proposed structure 171c can be ruled out. The uv spectrum

(72) Compounds 17^ and IJk exhibit absorption maxima at 283 mp, (e = 6,400) and 279 mp. ( s = 6 ,300), respectively.

73 of 171 can be conpared with that of Fischer's base This com-

(73) 0. Riester, Chimia, 3 ^ 75 (l9^l)

pound in benzene solution exhibits an absorption at 283 mp with an

extinction coefficient of 22,400. This spectrum agrees very well with

the uv data of IJl. The uv data, therefore, bests fits structure 171a-

CHb 0%

I 0 %

m . 76 for the thermal product from 1^. although structures l%lb and IJld

cannot be ruled out on this data. Structure l^ld can be eliminated

by the chemical evidence presented below.

In view of the unstable nature of IJl we decided to obtain chemical

evidence for its presence through trapping experiments. If an anionic

species were added to a solution of IJl. Michael type addition could

occur to yield an indole derivative. When a methylene chloride solution

of 168 was quickly warmed to room temperature and immediately added to

a methanolic solution of sodium methoxide, an 8c6^ yield (based on 2,3-

dimethylindole) of 2-methoxymethyl-3-methylindole ( 173) was obtained.

The use of thallium acetate in acetic acid in place of the sodium 74 methoxide solution gave the corresponding acetate, 174, in 88^ yield.

(74) Dr. Goverdhan Mehta initially carried out the conversion of 1 @ to 1J3. and Ijhj This author, however, significantly m,odified the procedures and only these will be reported.

1 7 1 N-" XlfeOCIfe

168

H 77 The structwes of and IJh were confirmed by independent syn thesis. 2-Oxobutyric acid phenylhydrazone ( IJ^) underwent the Fisher TS indole cyclization to give the indole lj6 in 765^ yield. The ester

(75) W. Wislicenus and E. Arnold, Ber., 20, 3595 (188%).

was reduced with lithium aluminum hydride to produce an 88^ yield of

2-hydroxymethyl-5-methylindole ( 177) • Treatment of this alcohol with

\ N 0 II / / » q n CHs CH3CH2C - COH ^ > Eton

H i

- Ijjt. 4-

CH2ÏÏ2

BFs -EtgO H m .

acetic anhydride in pyridine gave ±jk in 78^ yield. Although reaction

(76) W. Taylor, Helv. Chem. Acta., l6k (l950) ., 78

of IJJ^-with sodium hydride and methyl iodide gave intract ible tars,

1JJ_ was smoothly converted in 72^ yield to 17^ by treatment with diazomethane and boron trifluoride etherate con^lex.

The formation of compounds XD l 1?4 effectively eliminated

as a possible structure for the intermediate formed on warming

3^ . From all of the above data, therefore, it appears that only

171a and 171b are possible structures for this intermediate. Unfor tunately, attempts to trap 171 by a Diels-Alder reaction with tetra-

cyanoethylene, dicarbomethoxyacetylene, or 4-phenyl-1 ,2 ,4-triazoline-

3,5-dione resulted in the production of resinous materials. The ultimate structure determination of IJX must await additional evidence.

The final series of compounds investigated in this work are a group of U-hydroxyindoles. In continuing the search for a new synthe

sis of 5 -oxygenated indole derivatives, it was hoped that the preference of nucleophilic attack at the 3 -position could be avoided by solvolyzing

U-hydroxyindole derivatives instead of attempting to make N-chloroindole derivatives. Since N-hydroxyindoles are stable compounds, it was felt that they would not rearrange to the 3 -position prior to solvolysis.

The desired process, therefore, involves synthesizing esters of the

E-hydroxy compounds and solvolyzing them to the 5-oxygenated derivatives.

The hydroxy group had to be converted to a better leaving group because,

in contrast to Bamburger's phenylhydroxylamine compounds, these indole 77 78 derivatives were stable to acids. 79

(77) E. Fisher, Ber., 28^ 1258 (1895).

(78) W. Wright, Jr. and K. Collins, J. Amer. Chem. Soc., l8j 221 (195-6)......

P.OH ^ Es -OX I H OX

The first compound to be studied was N-hydro3qy-2 -phenylindole ( 178).

This cozopound was isolated in 75^ yield by treatment of benzoin oxime T7 with concentrated sulfuric acid, according to the method of Fisher.

The nitrogen £-nitrobenzoate (IJS) "^sis obtained as a stable crystalline solid from in 75^ yield by reaction with £-nitrobenzoyl chloride in ether solution. Solvolysis of 17ÿ_ in refluxing methanol produced a 65^ isolated yield of a rearranged p-nitrobenzoate 180. The structure of l80_

0 II C-Cl

o m B CH3OH

OH opm H 8 0

■was verified as 2 -pheriyl-5 -irLdolol-£-nitroTDenzoate by a combination of

ir and nmr spectra and elemental analysis. This reaction is in agree- 79 ment with some results obtained by Sandberg, who reacted lj8^ with

(79) R" J. Sandberg, J. Qrg. Chem., ^0. $6o4 (1965).

tosyl chloride to obtain the rearranged tosylate ]3l in 22^ yield. In

this case, the jN-tosylate was not isolated.

pyridine

It appears, then, that esters of indole hydroxylamines undergo the same rearrangement to the 5 -position as was postalated for the chlori nation of indole derivatives. Again it is not known whether this reac tion is concerted or proceeds via heterolytic clea'vage of the N-0 bond to give a delocalized nitrenium ion intermediate. In order to provide more mechanistic evidence and to try to force nucleophilic substitution at the 5 -position, a N-hydroxyindole with electron-withdrawing groups at the 2 and 5 -positions was synthesized. Diethyl 2-nitrobenzylidene- 80 malonate (182) was prepared by the method of Loudon and Wellings and

(8 0) J. Loudon and I. Wellings, J. Chem. Soc., ^h62 (1960). 8i

was reacted with hydrogen cyanide to give l8l . The cyano compound was SI cyclized by treatment with base to give IT-hydroxy-cyanoindole-2 -

(81) R. Ache son, C. Brookes, D. Deamaly, and B. Quest, ibid., ^Ok (1968).

carboxylic acid, which was esterified with ethanolic sulfuric acid

to yield the ethyl ester 62^ yield.

C=C( C0sEt)2 CHCHCCOsEt) HCTT ^ R ^ COgEt

182

The IT-tosylate of iffl was made at -î8 ° in tetrahydrofuran solution by treatm^^nt of l84^ with tosyl chloride in the presence of triethylamine. The tosylate proved too reactive to isolate and was solvolyzed

in situ in methanol at -78° to give the 3 -tosyloxyindolenine l8j in

6i(-^ yield along with 2C^ of 3 -cyano-2-carboethoxyindole (l8^ . The

structures of 3 ^ and 186 were verified by a combination of ir and nmr

spectra and elemental analysis. Again the major reaction pathway

involved migration of the N-tosylate to the 5 -position. The indole

186 could have been formed via hydrogen atom abstraction by the triplet nitrenium ion or by a simple hydride abstraction by the nitrogen cation. 82

.CN Œ THF ClfeOH 'COsEt COsEt COgEt H

The ease with which this rearrangement to the 3 -position occurs

indicates that the nature of substituents at the 5 -position have little

effect on the course of the reaction. On the basis of this data, it

cannot he determined whether the reaction is concerted or proceeds through an ion-pair. Regardless of which mechanism is correct, an explanation for the preferential rearrangement to the 3-position rather than substitution of the benzene nucleus may be found in comparing the 82 resonance energies of benzene and pyrrole. Benzene has been calculated

(82) A m o Liberies, “introduction to Molecular Orbital Theory,” Holt, Rinehart, and Winston, Inc., New York, N.Y., 1^66, p. Jh; 171-

to have a resonance energy of 36 kcal/mole whereas pyrrole has a resonance energy of 29 kcal/mole. Thus it costs less energy to destrqy the aromaticity of the five-membered ring, as in the transition state leading to 183. than it costs to destroy the aromaticity of the benzene nucleus. This concept offers a plausible explanation for the reactivity 83 of the five -membered ring. The question of the relative aromaticities of the two rings can be avoided in oxindole systems where the five- membered ring is no longer aromatic. N-Hydroxyoxindole (l88) was syn- 78 thesized by the reductive cyclization of o-nitrophenylacetic acid

( 18%) . The p-nitrobenzoate ( l8gij was made in 85^ yields but solvolysis of this conrpound led only to transestérification. Thus l8g_ gave l88^ in 995^ yield along with an 85?^ yield of methyl p-nitrobenzoate. It

CH2COH

H2SO4 (SO, 0 I OH igz. 188

appeared that a better leaving group would be needed for the solvolysis

188 > I n I I r ril + 188

OFEB

would be needed for the solvolysis of H-hydroxyoxindole derivatives.

Sulfonate esters proved to be suitable leaving groups. Treatment of l88_ with tosyl chloride in tetrahydrofuran at -78° followed by the 84 addition of methanol resulted in a 42$^ yield of 5 -metho%yoxindole (igo) and 3^^ of the T-tosylate ( l^l). The structure of IgO was confirmed

TSCI THF CH3OH I I OH H

188 12a m

83 by con^arison of its ir, nmr, and uv spectra with published data.

(83) T. Wieland and 0. Unger, Ber., $6^ 255 ( 190) .

The structure of 1 ^ was assigned on the basis of its ir and nmr spec- 55 tra and by analogy to the work of Lwowski and coworkers. Solvolysis of the tosylate from 188 in aqueous dioxane resulted in a 22^ yield ^84 . of 3 -hydroxyoxindole ( 192) and 29^ of Igj..

(84) A. Beckett, R. Daisley, and J. Walker, Tetrahedron, 2 ^ 6093 (1968).

The solvolysis of the nosylate ester from gave a better yield of the 3 -methoxy compound 1^0 than did the corresponding tosylate.

Thus the reaction of l88_ with p-nitrobenzenesulfonyl chloride in tetra hydrofuran solution at followed by solvolysis in methanol gave a 85 52^ yield of 190 and of the rearranged T-nosylate ip^ .

NsCl ^ THF 12Q 9 d . CHbOH I H ONs H

188 m

The solvolysis of sulfonate esters of N-hydroxyoxindole, therefore,

give synthetically useful yields of 5“ and T-oxygenated oxindole deri 85 vatives. Since oxindoles can be converted back to indole derivatives.

(85) R. Sundberg, “The Chemistry of Indoles,'' Academic Press, New York, N.Y., 1970, p. 556.

this reaction may have some synthetic potential in both the areas of

indole alkaloid and oxindole chemistry. The synthesis of N-hydroxy-

oxindole derivatives with electron-donating substituents at the 5 - or

7 -positions, followed by conversion to their sulfonate esters, should provide convenient routes to the unknown compounds l^|f and 19^.

■N^ 0 R / OR & 121 86 59 In view of the physiological activity of certain indole derivatives, these compounds may he worthy of some future research program. EXPERIMENTAL

Elemental analysis mere performed by the Scandinavian Microanaly- tical Laboratory, Herlev, Denmark. Melting points and boiling points are uncorrected. Infrared spectra were taken on a Perkin-Elmer Model

157 Infracord as neat liquids, solutions in carbon tetrachloride, or powered solids in potassium bromide disks. Nuclear magnetic resonance spectra were obtained on a Varian Associates A-60-A and HA-100 spectro meters and reported in tau (t) units relative to tetremethylsilane (t =

10.00) as in the internal standard.

N-t-Butylaniline • Compound was prepared by the method of Bell 3B and Knowles.

N-t-Butyl-g^-toluidine (g8). A mixture of 50 g ( .5^8 moles) of £- toluidine hydrochloride and 100 ml of t-butyl alcohol was sealed in a steel bomb and heated at 155° for 2k hr. The bomb was then cooled in a dry ice-isopropanol bath and opened. The contents were dissolved in 300 ml of water, made basic with aqueous ammonium hydroxide solution, and extracted with three 200-ml portions of ether. The combined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and concentrated on the rotary evaporator to leave a red oil. The oil was added to a mixture of 20 g of acetic anhydride in 300 ml of water and stirred for 3 hr. Excess sodium bicarbonate was slowly added and

87 88 the mixture steam distilled until the distillate was neutral to litmus paper. The distillate was extracted with three 100-ml portions of ether. The combined ethereal extracts were dried over anhydrous magne sium sulfate, filtered, and concentrated on the rotary evaporator to leave a dark oil. The oil was fractionally distilled to give 35*0 (6^ ) of pure bp 64° (0.35 mm) ; n^ = l.$l64; ir (neat) BlO, 1540, 1630,

5350 cm"^; nmr (CClJ 8.73 T (9H, s), 7-79 ^ (5H, s), 3 .35 t (4h, q).

Anal. Calcd for CxiHisNCl: C, 66.14; H, 9 .08; IT, 7 .01; Cl, 17.75-

Found: C, 66.47; H, 9-12; IT, 6.95; Cl, 17-5^•

IT-t-Butyl-g_-anisidine ( lOlaj . The procedure was identical to that of

98. In this manner lOla was synthesized from p-anisidine in 54/^ yield; bp 54-57° (0.04 mm); nj^ = 1.5225; ir (neat) 3300, 1580, 1290, IO80 cm“^; nmr (CCI4) 8.87 t (9H, s), 6.45 t (3H, s ), 6.99 ^ (IH, s), 3.33 t (4h, m).

Anal. Calcd for CnHiyNO: C, 75.70; H, 9-56; IT, 7-81.

Found: c, 73-79; H, 9-59; N, 7-79-

N-t-Butyl-o-anisidine (l02aj. The procedure was identical to that of g8_.

In this manner 102a was synthesized in 42^ yield from o-anisidine: bp

48-49° (0.06 mm); n^^ = 1.5273; ir (neat) 3450, 1610, 1230, 1120, 738 cm-^; nmr (CCI4) 8.67 T (9H, s), 6.22 T (3H, s), 6.00 T (ih, s) , 3-53 t

(4h, m) .

Anal. Calcd for C11H17NO: C, 73-70; H, 9-56; N, 7-81.

Found: c, 73-69; H, 9-63; N, 7-93- 89 N-t-Butyl-o-toluidine. A mixture of 5 5 g ( *25 moles) of o-toluidine hydrochloride and 2^0 ml of t-hutyl alcohol was sealed in a steel homb and heated to 150° for 8 hr . The bomb was cooled and opened and the reaction mixture was added to 500 ml of water, made basic with ammonium hydroxide solution, and extracted with three 200-ml portions of ether.

The combined ethereal extracts were dried over anhydrous magnesium sul fate, filtered, and the solvent removed in vacuo to leave a dark oil.

The oil was chromatographed on 250 g of Activity I basic alumina (hexane) to separate the product from starting material. Distillation of the purified product gave 5-2 g ( l4^) of N-t-butyl-o-toluidine : bp 50-51°

(0.57 mm); n^ = 1.5178; nmr (CCI4) 8.62 T (9H, s), 7-92 t (5H, s) ,

6.75 T (IH, s), 5.10 T (4H, m).

Anal. Calcd for C11H17N: C, 80.92; H, 10.50; N, 8 .58.

Found: G, 80.97; H, 10.4l; N, 8 .60.

H-t-Butyl-£-fluoroaniline. The procedure was identical to that for 98^.

In this manner the desired compound was synthesized from £-fluoroaniline in yield: bp 48.5 -(0.2 mm); n^ = 1.4992; ir (neat) 786, 850, 1250,

1510 cm"^; nmr (CCI4) 8.62 T (9H, s), 7-92 t (5H, s ) , 6.75 ^ (m, s),

5.10 T (4h, m ) .

Anal. Calcd for C10H14NF: C, 71-82; H, 8.44; IT, 8 .5 8 .

Found: C, 71-62; H, 8.54; N, 8.47-

3T TT-t-Butyl-g^-chloroaniline ( 104a) . The procedure was identical to that for In this manner the desired compound was synthesized from p- chloroaniline in 54^ yield: bp 67° (0-6 mm); n^ = 1.5420. 90 p-(N-t-Butylamino)biphenyl. A mixture of l6 g of p-aminobiphenyl hydrochloride and 8o ml of t-butyl alcohol was placed in a steel bomb and heated at 155° for 12 hr. The bomb was then cooled in a dry ice- isopropanol alcohol bath and opened. The contents were added to 250 ml of water, made basic with ammonium hydroxide solution, and extracted with three 100-ml portions of ether. The ethereal extracts were com bined, dried over anhydrous magnesium sulfate, filtered, and concen trated on the rotary evaporator to give a dark oil. The oil partially crystallized on standing but recrystallzing from hexane failed to purify the reaction mixture. Chromatography on Activity I basic alumina

(hexane-ether) gave 10.T g (6]^) of pure £-(N-t-butylamino)biphenyl: mp 63.0-6 4 .5°; ir (KBr) 696, 764, 828, 1580 cm“^; nmr (CCI4) 8.64 t

(9H, s), 6.68 T (iH, s), 3.32T (2H, d), 2.60 T (7H, m).

Anal. Calcd for CisHigïT: C, 8 5 .28; H, 8 .50; N, 6.22.

Found: C, 85.47; H, 8.48; N, 6 .2 5 .

Ethyl p-( N-br-butylamino)benzoate. A mixture of 24 g of benzocaine hydrochloride and 80 ml of t-butyl alcohol was sealed in a steel bomb and heated to 135-l40° for I6 hr. The bomb was cooled in a dry ice- isopropanol bath and opened. The contents were added to 200 ml of water, made basic with ammonium hydroxide solution, and extracted with three 200-ml portions of ether. The combined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and the solvents removed ^ vacuo to leave a viscous yellow oil. The oil was chromato graphed on Activity II basic alumina ( hexane-ether) to give the desired product as a white crystalline solid. Recrystallization from hexane 91 gave 9.02 g (5^^) of the aminoester: mp 70-72°; ir (KBr) 704, 773,

843, 1690, 1700 cm-i; imr (CCI4) 8.64- t (i2h, m), 5.70 t (2h, q),

2.78 T (4H, q).

A m i . Calcd for CisHisNOg: C, 70.55,* H, 8.65; K, 6.33.

Fo\md; C, 70.65; H, 8.79; N, 6.4l,

£-(N-t-Butylamlno)nitrobenzene. This compound was prepared by the 36 method of Suhr.

p-(N-t-Butylamino)benzonitrile (gZ)- To a solution of 5*00 g (4l.3

mmoles) of £-fluorobenzonitrile (2^ in 50 ml of dry dimethylsulfoxide

was added 9.05 g ( 123.9 nmoles) of t-butyl amine and the resulting

solution was sealed in a glass lined steel bomb. After heating at

150° for five days, the bomb was cooled to 0-5° and opened. The con tents were added to 250 ml of water and extracted with two 100-ml portions of chloroform. The combined chloroform extracts were washed with two 100-ml portions of saturated sodium bicarbonate solution and

100 ml of saturated sodium chloride solution, dried over anhydrous magnesium sulfate, filtered, and concentrated on the rotary evaporator

to give a pale yellow oil which partially crystallized. The semi

solid reaction mixture was washed with 100 ml of pentane to remove the unreacted starting material which gave the crystalline product. Re cry

stallization from hexane gave 3.50 g (4-9%) of 9%,: mp 103-105°; ir

(KBr) 834, 1625, 2280, 3550 cm“^; nmr (CCI4) 8.59 ^ (9H, s), 5.62 T

(IH, s), 3.00 T (4h, q). 92

Anal. Calcd for CnHi^g: C, 75-82; H, 8.10; N, 16.0 8 .

Found: C, 75-75; H, 8.15; W, 16.II.

Chlorination and Solvolysis of U-t-Butylaniline - A solution of

2.1 g of 25- ill 250 ml of carton tetrachloride m s vigorously stirred

with 250 ml of ^ sodium hypochlorite solution at -8° under nitrogen ty means of a Vibro-mixer mechanical stirrer for 1 hr. The organic

layer was then separated and dried over anhydrous magnesium sulfate in

the cold. The cold mixture was then filtered into a solution of 9-5 g

of silver trifluoroacetate in 5 00 ml of methanol at -8 °. Silver

chloride began to precipitate. After stirring in the cold for 18 hr,

the reaction mixture was allowed to come to room temperature, when

excess lithium chloride was added to precipitate the excess silver ion.

The silver chloride was removed by filtration and the solvents evaporated

in vacuo to leave a red oil. The residue was added to 100 ml of 5^

sodium hydroxide solution and the resulting mixture was extracted with

three 50-ml portions of pentane. The combined pentane extracts were

dried over anhydrous magnesium sulfate, filtered, and concentrated on

the rotary evaporator to leave a red oil. Distillation gave I.9I g of

a clear oil, bp 60-70° (0.75 mm). Gas chromatography on a 5^ Amine 220

on Diaport S column indicated the presence of five products in the

reaction mixture. Small amounts of the five products were isolated

in pure form by preparative gas chromatography on the same column. The

first component was N-t-butylaniline (25}* The second component was

N-t-butyl-o-chloroaniline (l^a), which was identified by a canparison 37 of published physical properties. The third component was N-t-butyl- 95 o-anisidine ( 102a,) ; bp 48-^9° (0.06 mm) ; = 1 .5273; Ir (neat)

738, 1330, 1530, l6lO cm“^; nmr (CClJ 8.67 t (9H, a), 6.23 t (3H, s ),

3.30 T (4-H, m). The fourth compound was N-t-butyl-2 -anisidine (lOla): bp 54-57° (0.04 mm); n ^ = 1.5225; ir (neat) 820, 1050, 1230, 1510 cm ; nmr (CCI4) 8.88 t (9H, s), 6.46 t (Jh, s ), 3*31 ^ (4h, s ) . The fifth compound was H-t-butyl-p-chloroaniline (l04a) which was identical to an authentic sample.

The reaction was repeated and analyzed on a 3^ Amine 220 on Diaport

S column using naphthalene as an internal standard giving ^ of

39^ of 101a, 28^ of 103a. of 102a, and ^ of 104a.

Solvolysis of N-t-butyl-IT-chloroaniline in pure methanol without silver trifluoroacetate gave 525^ of lOga, l4^ of 104a, and 8^ of N-t- butylaniline as determined by vapor phase chromatography in the same manner.

Chlorination and Solvolysis of N-t-butyl-p-toluidine (98) . A mixture of 2.0 g of 98, in 50 ml of carbon tetrachloride and 11.3 g of solid calcium hypochlorite was vigorously stirred at 0-5° for 1 hr . The salts were then removed by filtration to leave a clear solution of the

N-chloro compound 99,* This solution was then added to a solution of

8.15 g of silver trifluoroacetate in 100 ml of methanol at 0-5°. Silver chloride immediately precipitated and the mixture was allowed to stir for 3 hr at 0-5°. Silver chloride was removed by filtration and the solvents removed on the rotary evaporator to leave a viscous residue.

The residue was added to 50 ml of water, made basic with ammonium hydroxide solution, and extracted with three 50-ml portions of ether. 9^ The combined ethereal extracts were dried over anhydrous magnesium sul

fate, filtered, and concentrated vacuo to leave a viscous oil. The components of the oil were separated via preparative vapor phase chroma tography on a 20^ 4;1 Apiezon,L:KOH on ^°/so Firebrick column. The first component was 4-methyl-4-metho%y-2 ,5 -cyclohexadienone-IT-t-butyl-

imine (lO^): mp 56-56°; ir (neat) IO9O, 1220, I37O, 1600, 1650 cm“^; nmr (CCI4) 8.71 t (3H, s), 8.68 t (9h, s) , 6.96 t (jh, s), 4.01 t (4h, m).

Anal. Calcd for C12H19NO: C, 74.57; H, 9-91; W, 7-25.

Found: C, 74.44; H, 10.00; N, 7.25-

The second component was N-t-butyl-p-toluidine (98). The third component 25 was N-t-butyl-o-chloro-£-tolui dine (106): n^ = I.5278; ir (neat) 807,

1225, 1540, 1630 cm“^; nmr (CCI4) 8.64 t (9H, s), 7-77 T (3H, s), 5-l4 t

(3H, m) .

Anal. Calcd for CxiHieNCl: C, 66.82; H, 8 .16; N, 7-09; Cl, 17-95-

Found: C, 67.06; H, 8.10; N, 6 .98; Cl, 18.02.

Vpc analysis of a repeat reaction ( versus j-04a^ as an internal standard) yielded 70^ of 10^ 17^ of 10^ and 2^ of 98.

Solvolysis of 99. pure methanol without silver ion gave 6lj^ of ip6 , 32^ of 8,nd 3^ of 9^ as deteimined by vpc analysis in the same manner.

4 -methyl-4-methoxy-2,5 -cyclohexadienone-N-t-butylimine ( 1Q5) . A solution of lithio-t-butylaminewas prepared by the addition of 0.264 g of t- butylamine in 10 ml of dry ether to 2.24 ml of a 5 -02^ solution of methyllithium in ether at room temperature. To this solution was then 42 added O.5OO g of 4-methyl-4-methoxy-2,5-cyclohexadienone (10%) in 25 95 ml of dry ether. After stirring at room temperature for 5 hr, one equivalent of methyllithium was added followed by 12 hr at reflux. The reaction was quenched by the addition of 10 ml of water, followed by extraction with three 40-ml portions of ether. The combined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and the solvent evaporated to leave an oil. The desired product was iso lated by preparative vapor phase chromatography on a 2d^ 4:1 Apiezon

L:KOH on ^°/so Firebrick column to give 350 mg (5^) of Ifg. It was shown to be identical to the product from the solvolysis of nmr, and a mixture melting point (56-58°).

Chlorination and Solvolysis of N-t-Butyl-p-fluoroaniline. A mixture of

2.56 g of N-t-butyl-p-fluoroaniline in 50 ml of pentane and 15 g of finely powdered calcium hypochlorite was stirred at room temperature for 2 hr. The salts were removed by filtration and the solvent was evaporated in vacuo at 0-5° to leave the chloramine as a neat oil. The oil was dissolved in 100 ml of dry ethanol and the resulting solution was stirred at room temperature for 2 hr. Removal of the solvent on the rotary evaporator left a brown oil which was separated into two compo nents by preparative gas chromatography on 20^ 4:1 Apiezon L:KOH on

°°/so Firebrick. The first component was N-t-butyl-£-fluoroaniline. The second component was N-t-butyl-o-chloro-£-fluoroaniline: bp 70°

(.25 mm) ; n^^ = 1.5104; ir (neat) 807, 859, 900, 1220, 1505 cm“^; nmr (CCI4) 8.65 t (9H, s ) , 3-06 t (^h, m) .

Anal. Calcd for C10H13NCIF: C, 59*55; H, 6.50; N, 6.95 ; Cl, 17*58.

Found: c, 59*^5; H, 6.40; N, 6.88; Cl, 17*7^* 96

A repeat procedure was analyzed via vapor phase chromatography on the same column using m-chloroaniline as an internal standard to give of the o-chloro derivative and 11^ of IT-t-hutyl-p^-fluoroani line.

Chlorination and Solvolysis of N-t-Butyl-£-chloroaniline ( . A mixture of 2.0 g of lC4a in 50 ml of carbon tetrachloride and l4.$ g of finely powdered hypochlorite was stirred at room temperature for two hours. The mixture was then filtered and the chloramine solution was added to a solution of 5.02 g of silver trifluoroacetate in 100 ml of methanol at 0-5°. After stirring for 20 min, excess lithium chloride was added. The silver chloride was removed by filtration and the sol vents evaporated m vacuo to leave a red oil. The oil was taken up in

50 ml of water, made basic with ammonium hydroxide solution, and ex tracted with three 50-ml portions of ether. The combined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and the solvent removed on the rotary evaporator to leave a red oil. The oil was separated into four components via preparative gas chromatography on a 2C6^ 4:1 Apiezon L:KOH on °°/so Firebrick column. The first com pound was 4,4-dimethoxy-2,5-cyclohexadienone (IJO), which was identified 45 by comparison ir, uv, and nmr spectra with published data. The second , . 25 component was IT-t-butyl-p-chloro-o-anisidine (129): n^ = 1.5425; ir

(neat) 869, 1050, 1230, 1510, 16OI cm"^; nmr (CCI4) 8.69 T (9H, s),

6.28 T (5H, s), 7.10 T (5H, m).

Anal. Calcd for CnHisNOCl: C, 6l.82; H. 7-55; N, 6.55; Cl, 16.59•

Found: C, 61.95; H, 7-49; N, 6.47; Cl, 17-03. 97 The third product was N-t-butyl-o,£-dichloroaniline ( 128) ; =

1.5^77; ir (neat) 769, 805 , 868, 1220, 15IO, 161O cm"^; nmr (CCI4)

8.63 T (9H, s), 7.15 T (5H, m).

Anal. Calcd for C10H13NGI2 ; C, 55-06; H, 6.01; N, 6.42; Cl, 52.51.

Found; C, 55-29; H, 6.00; N, 6.42; Cl, 32.43.

The fourth component was 104a.

The reaction was repeated and analyzed on a 10^ Carbowax 2CM:K0H on ®°/so Chrom N column (versus naphthalene as an internal standard) to yield 25^ of 2^ of 122, 1C0 of 0^ and 12^ of 104a.

Solvolysis of the chloramine 127 in pure methanol without silver ion gave ?4^ of ^ 8 and iSf of 104a as indicated by vpc analysis as above.

Chlorination and Solvolysis of p-( N-1-Butylamino) biphenyl. A mixture of 1.5 g of p-(N-t-butylamino)biphenyl in 50 ml of pentane and 10.00 g of powdered calcium hypochlorite was stirred at 0-5° for 2 hr. The salts were then removed by filtration and the solution concentrated at

0-5° on the rotary evaporator to 10 ml at which time the chloramine precipitated. The solid was collected by filtration and air dried to give 1.505 g of 1^^ mp 86-88° (dec.). Pure lp8 was not analyzed since it rearranged to 3 -chloro-4-( N-t-butylamino) biphenyl ( l io) on storage in the refrigerator.

To a solution of 2 .94 g of silver trifluoroacetate in 50 ml of methanol at room temperature was added 1.00 g of 1C8. Silver chloride immediately precipitated. After stirring the reaction mixture over night, 2.0 g of sodium chloride was added. The silver chloride was 98

removed by filtration and the solvents evaporated in vacuo. The residue

■was taken up in 50 ml of 5^ sodium hydroxide solution and the mixture was extracted "with three 100-ml portions of ether. The ethereal ex tracts were combined, dried over anhydrous magnesium sulfate, filtered,

and the solvent e-vaporated to leave a red oil. Chrcmatography on 60 g of Activity I basic alumina ( hexane-ether) gave two conopounds. The

first eluted was 50 mg (5^) of 110; n^ = 1 .6152,* ir (neat) 695,

T6i, 1220, 1605, 5500 cm"^; nmr(CClJ 8.57 t (9H, s), 2.6o t (8h, m).

Anal. Calcd for GisHisIîCl: C, 73-97; H, 6 .98; E, 5*59; Cl, 15 .65.

Found; C, 74.05; H, 7*08; W, 5.4l; Cl, I5 .6O.

The second product eluted was h95 mg (62^^) of 4-methoxy-4-phenyl-2,5 -

cyclohexadienone (IO9): mp 91 “93°; ir ( KBr) 704, 760,865 , 1200, 167O,

1690 cm"^; nmr (CCI4) 6.56 t (3h, s) , 5.48 t (4n, q), 2.6l t (5h, m) .

Anal. Calcd for C13H12O2 : C, 77-98; H, 6.o4.

Found: c, 78.07; H, 6 .05.

Chlorination and Solvolysis of Ethyl £-(N-t-Butylamino)benzoate. To a

solution of 1.0 g of the amine in 50 ml of pentane was added 10.0 g of finely powdered calcium hypochlorite. Vigorous stirring was initiated and continued at room temperature for 4 hr. The salts were removed by filtration and the solvent evaporated in •yacuo while keeping the flask

cold to give the neat chloramine 124 as a pale yellow oil. Infrared analysis showed the complete absence of the E-H band of the starting amine. The chloramine was then added to a solution of 5 .02 g of silver trifluoroacetate in 50 ml of methanol. Silver chloride slowly precipi 99 tated. Stirring was continued for 2 hr at which time an iodometric

test for active chlorine was negative. Excess silver ion was then pre

cipitated with lithim chloride and the silver chloride was removed

"by filtration. The solvents were removed on the rotary evaporator

and the residue was taken up in 50 ml of 5^ sodium bicarbonate solu

tion and extracted with three 50-ml portions of ether. The combined

ethereal extracts were dried over anhydrous magnesinm sulfate, filtered,

and the solvent evaporated to leave a yellow residue. The residue was

separated by preparative vapor phase chromatography on a 2C^ 4:1 60 Apiezon LrKOH on /so Firebrick column to give two components. The

first component was recrystallized from hexane to give ethyl 3-chloro-

4-(N-t-butylamino)benzoate ( I25) : mp 61-65°; ir (KBr) 725, 7^9, 1290,

1710,. 5550 cm“^; nmr (CCI4) 8.55 t (9H, s), 8.65 t (3h, t), 5-71 t

(2H, q), 3.15 T (IH, d), 2.10 T (2H, m) .

Anal. Calcd for CisHisNOgCl: C, 6l.05; H, 7.09; N, 5.48; Cl, 15.87.

Found: C, 61.19; H, 7.12; ÏÏ, 5.48; Cl, 15.47.

The second component was the starting amine, 126.

The reaction was repeated and analysis via gas chrcmatography on

the same type of column (versus ethyl g^-aminobenzoate as the internal

standard) yielded 58^ of 12^ and 2 ^ of the starting amine 1^6.

Solvolysis of the chloramine 124 in pure methanol without silver

ion gave 95^ of 123_ and 4^ of the starting amine 12^ as indicated by analytical gas chromatography as above. 100

Chlorination and Solvolysis of p-(E-t-Bntylamino)nitrobenzene. To a

solution of 2.0 g of p -( N-t -butylamino ) nitrobenzene in 100 ml of 1:1

carbon tetrachloride rpentane was added 20 g of powdered calcium hypo

chlorite. After stirring for 4 hr at room temperature, the salts were

removed by filtration and the solution was concentrated to 50 ml on the

rotary evaporator. The solution was then diluted to 100 ml with pen

tane in a volumetric flask and separated into two 50 ml portions. The

solvents were removed iu vacuo to give two samples of the neat chlor

amine 121. Infrared analysis of 121 showed the complete absence of

the N-H band of the starting amine. One sample was added to a solution

of 5.42 g of silver trifluoroacetate in 50 ml of methanol. Silver

chloride began to slowly precipitate. After stirring overnight, an

iodometric test for active chlorine was negative. Excess lithium

chloride was added and the silver chloride was removed by filtration.

The solvent was removed on the rotary evaporator and the residue was

taken up in 100 ml of ether and stirred over sodium hydroxide pellets

for 1 hr. The ethereal solution was decanted and the solvent evaporated

to give yellow crystals. Chromatography on Activity I basic alumina

separated two compounds. The first compound eluted (hexane-ether) was

761 mg (65^) of 5 -chloro-4-( N-t-butyl) nitrobenzene (l22) ; mp 115-115°;

ir (KBr) 1130, 1220, 1300, 1350, 1600 cm“^j nmr (CDCI3) 8.49 ^ (9H, s),

7.00 T (IH, d), 1.72 T (5H, m).

Anal. Calcd for C10H13N2O2CI; C, 52.52; H, 5-73; N, 12.25; Cl, 15-51.

Found: C, 52-58; H, 5-75; E, 12.12; Cl, 15-50.

The second product to be eluted was starting p-( N-t-butylamino) nitro benzene . 101

The second sample of the chloramine 121 v?as dissolr-.d in 50 ml of

methanol and stirred overnight at room temperature. The methanol was

removed iu vacuo to leave a yellow crystalline residue which was dis

solved in 50 ml of ether and stirred over sodium hydroxide pellets for

1 hr. The ethereal solution was filtered and the solvent evaporated

to leave a yellow crystalline product which was chromatographed on

Activity I basic alumina to give I.160 g (99^) of 122 and of p- ( U-t-butylamino) nitrobenzene.

Chlorination and Solvolysis of £-( N-t-Buty lamino)benzohitrile (g%) . To

a solution of 0.452 g of in 25 ml of 1:1 ether rpentane was added

10 g of powdered calcium hypochlorite and the mixture was stirred at

room temperature for 3 hr. The salts were removed by filtration and

the solvents evaporated iu vacuo at 0-5° to give the neat chloramine.

Infrared analysis showed the complete absence of the N-H band of the

starting amine. The neat chloramine was then dissolved in 25 ml of

ethanol buffered 0.1 N in sodium acetate and 0.1 N in acetic acid and

the solution was sealed in a glass tube and heated at 100° for 24 hr.

The tube was cooled and opened and the ethanol was removed on the rotary

evaporator. The residue was taken up in 25 ml of water and excess

sodium bicarbonate was added. The mixture was then extracted with

three 25-ml portions of ether. The combined ethereal extracts were

dried over anhydrous magnesium sulfate, filtered, and the solvent

evaporated to leave a yellow oil. The oil was triturated with 25 ml of pentane which brought about the precipitation of white crystals. Cooling 102 to 0-5° and filtration yielded. 135 mg (2^ ) of 2L’ The filtrate was concentrated to leave another crystalline solid which was re cry stallized from hexane to yield 264 mg { & ^ ) of 3-chloro-4-(N-t-butyl amino) henzonitrile : mp 81-8 3 .5°; ir (KBr) 830, 882, 1525, 1601,

2200 cm'^; mnr (CCI4) 8.53 T ( 9H, s), 5 -08 T (ih, d), 2.50 T (2H, m) .

Anal. Calcd for C11H13N2CI: C, 63.31; H, 6.28; N, 13-42; 01, 16.99 .

Found: G, 63.59,* H, 6.26; N, 13-22; 01, I6 .8 8 .

Chlorination and Solvolysis of N-t-Butyl-o-toluidine. To a solution of 0.5 g (3-07 mmoles) of the amine in 230 ml of pentane at -78° was added a solution of O .367 g (3 -38 mmoles) of t-butylhypochlorite in 20 ml of pentane. After stirring for 2 hr at -78 the reaction mixture was added to a solution of 2.04 g (9 .31 mmoles) of silver trifluoro acetate in 300 ml of methanol at -78°. The solution was then allowed to slowly warm to room tengerature. After stirring for 1 hr at room temperature, excess lithium chloride was added and the silver chloride was removed by filtration to leave a yellow solution. The solvents were evaporated to give a yellow oil which was taken up in 50 ml of water, made basic with sodium hydroxide pellets, and extracted with three 75-nil portions of ether. The combined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and the solvents were evaporated to yield an oil. The oil was chromatographed on

Activity I basic alumina (skelly B-ether) to give 50 mg (IC^) of N-t- butyl-o-tolui dine and 224 mg (38^) of N-t-butyl-4 -methoxy-2-methylaniline (II8}: n^^‘® = 1 .5192; ir (neat) 3300, 1500, 1230, 1055,

810 cm"^; nmr (CClJ 8.75 ^ (9H, s), 7-90 t (3H, s ) , 7.20 t (i h , s ) , 105 6 .52t. T (5H, s), 5.45 T (3H, m).

Molecular v?eight. Calcd for C12H19NO: 195 .14665 .

round: 195.i46t8.

Chlorination and Solvolysis of H-methylanillne. A solution of I .5 g of

N-methylaniline in 250 ml of carton tetrachloride and 250 ml of a 6^

sodium hypochlorite solution were vigorously stirred with a Vitro-

mixer mechanical stirrer for 45 min at -6 under nitrogen. The organic

layer was separated, dried over anhydrous magnesium sulfate, filtered,

and added to a solution of 2.0 g of silver trifluoroacetate in 500 ml

of methanol at -20°. After stirring for 50 min, lithium chloride was

added to precipitate the excess silver ion. The silver chloride was

removed hy filtration and the solvents evaporated ^ vacuo. The residue

was taken up in 100 ml of 5^ sodium hydroxide solution and extracted with three 50-ml portions of ether. The comhined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and concentrated

on the rotary evaporator to leave a dark oil. The oil was separated

into five components by preparative gas chromatography on a 5^ amine

220 on ®°/ioo Diaport S column: E-methylaniline, N-methyl-p-anisidine

( 101b). N-methyl-o-anisidine ( 102bJ . U-methyl-o-chloroaniline ( lO^b), and N-methyl-p-chloroaniline ( 104b) . All of the compounds were iden tified by comparison of their ir, nmr, and refractive indices with those of authentic samples. The reaction was repeated and analysis via gas chromatography on the same type of column (versus 2 ,6 -dichloroaniline as an internal standard) yielded of 101b, iS^ of 102b, ^ of ip4b. 9^ of Ifgb, and 5^ of N-methylaniline. 104

2,6 ,N-Trlmethylanllirie. A solution of 15*0 g ( .124 moles) of dimethylaniline in 50 ml of ethyl formate was refluxed for 4 days. The excess ethyl formate was then removed on the rotary evaporator to leave a damp solid. The wet solid was then washed with two 50 ml portions of ether to remove unreacted 2,6-dimetbylaniline. After drying ^ vacuo the crude formamide weighed 10.0 g {&é>). To a refluxing solution of

4.0 g of lithium aluminum hydride in 500 ml of dry THF was added 10.0 g of the crude formamide in small portions. After the addition was complete, the reaction was refluxed for 2 days. The excess hydride was destroyed by dropwise addition of l6 ml of water followed by a 2 hr reflux. The salts were removed by filtration and the solvents removed in vacuo to give the crude secondary amine. Short path distillation gave 8.3 g of the desired product, bp 45-44*^ (0.5 mm) for a $2^ yield from the formamide.

Chlorination and Solvolysis of 2,6 ,IT-Trimethylaniline. To a solution of 1.0 g of 2 ,6,n-trimethylaniline in 100 ml of methanol at -78 was added a solution of 0.884 g of t-butyl hypochlorite in 20 ml of methanol.

After stirring for 20 min at -78°, a solution of 1.64 g of silver tri fluoroacetate in 20 ml of methanol was added dropwise. Silver chloride immediately precipitated. The mixture was stirred for an additional

40 min at -78° and then allowed to come to room temperature. The silver chloride was removed by filtration and the solvents were evaporated in vacuo to leave a blue oil. The oil was taken up in 50 ml of water, made basic with sodium hydroxide pellets, and extracted with three 105 100-ml portions of ether. The combined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and the solvent evaporated to leave a red oil. The oil was separated into two components by chromatography on 150 g of Activity I basic alumina ( hexane-ether), although some overlap of the two bands occurred. The first component eluted was 0.100 g {&f°) of ^-chloro-2 ,6,IT-trimethylaniline (13^ ; n^®’^ = 1.5505; ir (neat) 865, 882, 1250, l48o cm“^; nmr (CCI4) 7.80 t

(6h, s ), 7*30 T (5H, s), 7.26 T (iH, s), 5 .13 T (2H, s ) . The compound was analyzed as the hydrochloride.

Anal. Calcd for CgHxsNCla: C, 52.67; H, 6.39; N, 6.83; Cl, 34.11.

Found: C, 52.73; H, 6.40; N, 6 .8 8 ; Cl, 33.84.

The second component eluted was 2,6 ,TT-trimethyl-4-anisidine (llSj : n^°‘® = 1.5342; ir (neat) IO71, 1155, l480, 1510 cm'^; nmr (CCI4) 7-79 ^

(6h, s ), 7.38 T (5H, s), 7.53 T (m, s), 6.55 T (3H, s), 5.58 T (2H, s ). The compound was analyzed as the hydrochloride.

Anal. Calcd for CioHieNOCl: C, 59-67; H, 8.02; N, 6 .96; Cl, 17-37-

Found: C, 59-39; H, 8.09; N, 6 .98; Cl, 17-88.

The reaction was repeated and analyzed on a ICÇ^ Carbowax 2CM:K0H on

®°/so Chrom W (versus o-tolui dine as an internal standard) to yield 77^ of 112, 17^ of 115, and 2^ of 2 ,6 ,N-trimethylaniline.

When the reaction was carried out in pure methanol without silver ion, vpc analysis gave 49^ of 112, 57^ of Ijj, and 4^ of 2 ,6 ,W-trimethylaniline. 1 0 6

Chlorination and Solvolysis of o-(N-methylamino)T3iphenyl. To a solu tion of 0.510 g of the amine in 40 ml of methylene chloride at -78°

■was added a solution of 0.525 g of t-hutyl hypochlorite in 10 ml of methylene chloride. After stirring at -78° for 15 min, a solution of 0.606 g of silver trifluoroacetate in 75 isl of methanol -was added dropwise over a 50 min period. The mixture -was allowed to stir at

-78° for 1 hr and then gradually allowed to "warm to room temperature to give a "blue solution with much silver chloride precipitation. The salts were removed "by filtration and the solvent evaporated ^ vacuo to give a dark oil. The oil was taken up in 50 ml of IC^ sodi-um hydroxide solution and extracted with three 75-nil portions of ether.

The combined ethereal extracts were dried over anhydrous magnesium sulfate, filtered, and the solvents removed on the rotary e-vaporator to leave a red oil. Chromatography on Activity I basic alumina ( et her- hexane) separated the oil into three components. The first component was o-(ÎÎ-methylamino)biphenyl. The second compound eluted was 5- chloro-2-(N-methylamino)biphenyl (ll6j ; n^^'^ = 1.6225 j ir (neat)

705, 764, i4i5, i48o, 1501 cm"^; nmr (CCI4) 7.58 t (3h, s ) , 6.10 T

(IH, s), 5.15 T (5H, m), 2.69 T (5H, m).

Anal. Calcd for C13H12HCI: C, 71-87; H, 5-57; H, 6.45; 01, 16.ll.

Found: C, 71.59; H, 5-75; H, 6.45; Cl, 16.5 4 . The third component eluted was 5 -methoxy-2-(H-methylamino)biphenyl

( 115) : ir (neat) 707, 1045, 1220, 1510, 5500 cm“^; nmr (CCI4) 7.28 T

(5H, s), 6.60 T (IH, s), 6.51 T (5H, s), 5.58 T (5H, m), 2.67 t (5H, m). lOT

Anal. Calcd for : C, 78.84; H, 7.07; N, 6 .57.

Pound; C, 78-50; H, 7-00; H, 6.43-

The reaction was repeated and analysis via vapor phase chromato graphy on a 2C0 4:1 Apiezon L:KOH on °/ioo Firebrick column (versus

£-( IT-t-butylamino)biphenyl as an internal standard) yielded 51^ of U.5 ,

4^ of 116, and 7^ of o-(lT-methylamino)biphenyl.

When the reaction was carried out in an analogous manner in pure methanol without silver ion, vpc analysis gave 35^ of II6 . l4^ of II3 , and 7^ of the starting material.

N-t-butyl-IT-phenylhydro3!ylamine p-nitrobenzoate (l4^) • To a solution of 1.0 g of N-t-butyl-n-phenylhydroxylamine (l48) in 10 ml of dry ether at -78° under nitrogen was added 5-0 g of powdered sodium hydroxide followed by dropwise addition of a solution of 1.24 g of p-nitrobenzoylchloride in 20 ml of dry ether. The mixture was stirred for 1 hr at -78° and allowed to warm to room temperature. Excess sodium hydroxide was removed by filtration and the solvents evaporated in vacuo to give a yellow solid. The solid was recrystallized from 50 ml of hexane to give 1.6 g (79^) of the desired p-nitrobenzoate (ijj^) : mp 101-103°; ir (KBr) 7-17, 1070, 1260, 1530, I760 cm'^; nmr (CCI4)

8.72 T (9H, s), 2.69 T (5H, s), 1.74 T (4h, s ).

Anal. Calcd for C17H18N2O4 : C, 64.95; H, 5-77; N, 8 .91 .

Found: C, 64.86; H, 5-86; N, 8 .8 1 . io8 Solvolysis of N-t-butyl-N-phenylhy dr oxylamine p-nitrobenzoate .

A solution of 500 mg of l4^ in 25 ml of dry methanol -which was 0.1 N

in anhydrous sodium acetate -was stirred at room temperature for 48 hr.

The solvent was removed on the rotary evaporator to leave a semi-solid

residue which was taken up in 50 ml of saturated aqueous sodium bicar

bonate solution and extracted with three 50-ml portions of ether.

The aqueous solution was set aside and the combined ethereal solution

was extracted with two 50-ml portions of dilute hydrochloric acid solu

tion. The ethereal solution was dried over anlnydrous magnesium sulfate, filtered, and the solvent evaporated to leave a crystallized solid.

The solid was re crystallized from 5 ml of hexane to give 172 mg ( 65^)

of methyl-£-nitrobenzoate, mp 95-97°* The aqueous sodium bicarbonate

solution was acidified with hydrochloric acid and extracted with two

25-ml portions of ether. The combined ethereal extracts were dried

over anhydrous magnesium sulfate, filtered, and solvent evaporated in

vacuo to yield 80 mg (5C0) of p-nitrobenzoic acid, mp 24o-24l°. The

initial hydrochloric acid solution was carefully neutralized with

sodium bicarbonate and extracted with three 50-ml portions of ether.

The combined ethereal e:ctracts were dried over anhydrous magnesium

sulfate, filtered, and the solvent evaporated to leave a semi-solid material. The material was separated into two bands by chromatography

on 50 g of Activity I basic alumina. The first band was a mixture of

compounds which were further separated by preparative -vapor phase

chromatography on a 3^ Amine 220 on ^^/loo Diaport 8 column to give

N-t-butyl-o-anisidine ( 102a), N-t-butylaniline (2^, and N-t-butyl 109 anlsidine ( 101a). 'which were identified by comparison with authentic samples. The second component to be eluted from the alumina column was a solid which was re crystallized from pentane to give 137 mg (5^ ) of N-t-butyl-o-aminophenol ( 1^ O) : mp 90-92°^ ir (KBr) 7^0, 7^7, 1110,

1250 cm"^; nmr (CGlJ 8.80 t (9H, s), 5.38 t (2H, s ) , 3 .I6 t (4n, m) .

Anal. Calcd for C10H15WO: C, 7 2 .69; H, 9-15,' N, 8.48.

Found: C, 72.50; H, 9-09; N, 6.45.

The reaction was repeated and the amine fraction was analyzed via e ! 80 a 3® Amine 220 on /loo Diaport S column (versus naphthalene as an internal standard) to yield 3^ of N-t-butylaniline, 5^ of ip2a, and

of 101a.

N-t-butyl-N-( 2 'tolyl) hydroxylamine. To a solution of g^-me thy Iphenyl- lithium prepared from 17 .I g of g-bromotoluene and 6.4 g of n-butyl- n 86 lithium at 0-5 by the method of Oilman, was added a solution of

(86) H. Oilman in Organic Reactions, Vol. VI, (A. C. Cope, ed.), Ch. 7, p. 3^9 ( 1951):

2.58 g of t-nitrobutane in 10 ml of ether. A vigorous reaction began and subsided after about 10 min. The reaction was then allowed to stir at room temperature overnight to give a pale yellow solution. The reaction was quenched with 100 ml of water and the ethereal layer was separated, dried over anhydrous magnesium sulfate, filtered, and the solvent removed me vacuo to leave a red oil. The oil was triturated with 20 ml of cold pentane to crystallize the product which was collected 1 1 0 by filtration and ^washed -with 10 ml of cold pentane. Recrystallization from 10 ml of hexane gave l.Jl g ($8^) of N-t-'bntyl-E-( g^-tolyl) hydroxylamine: mp 121-125°; ir (KBr) 828, II90, 14-90, 5150 cm"^; nmr ( CCI4) 8.89 T (9H, s), 7.70 T (5H, s), 2.98 T (IfH, m), 2.81 T (IH, s) .

Anal. Calcd for CnHi^KO: C, 73-70; H, 9-58; N, 7-8l.

Pound: C, 73-61; ÏÏ, 9-39; N, 7-76.

N-t-Butyl-N-(p-tolyl) hydroxy lamine p-nitrobenzoate (l^l) . To a solu tion of 500 mg of the hydroxy lamine and 580 mg of triethylamine in 50 ml of ether at -78° was added a solution of 520 mg of p-nitrobenzoyl chloride in 10 ml of ether. The reaction mixture was then allowed to warm slowly to room temperature. Triethylamine hydrochloride began to precipitate and the reaction appeared to be over after 4-5 min at room temperature. The salts were removed by filtration and the ethereal solution concentrated to yield a yellow oil which slowly crystallized.

The solid was re crystallized at -50° from 25 ml of 1:1 ether :pentane to give 810 mg (88^) of the p-nitrobenzoate 15I : mp 79-80°; ir (KBr)

718, 1081, 1550, 1760 cm-^.

Anal. Calcd for CisRsoNgO^: G, 65 .84-; H, 6.l4; îî, 8.53-

Pound; C, 65-8 7 ; H, 6.22; N, 8 .5 2 .

Solvolysis of N-t-Butyl-N-( p-tolyl) hydroxy lamine £-nitrobenzoate ( 151).

A solution of 724- mg of in 25 ml of methanol which was 0.1 E in anhydrous sodium acetate was stirred overnight at room temperature to give a light red solution. The methanol was removed ^ vacuo and the residue was taken up in 50 ml of ether and extracted with two 50-ml Ill portions of ether. The combined ethereal extracts ■were dried over anhydrous magnesium sulfate, filtered, and the solvent removed on the rotary evaporator to give a red oil. The oil was chromatographed on

Activity I basic alumina to give two components. The first compound was eluted with ether to give 136 mg (30^) of 4-methoxy-it--methyl-2,5- cyclohexadienone K-t-butylimine (10^ , which was identified by a canparison of ir spectra and a mixture melting point with an authentic sample. The second compound was eluted with 1&> methanol in ethyl acetate to yield 126 mg (3 lS^) of 2-(w-t-butylamino)-5 -methylphenol

( 132). The phenol was sublimed at 8o° (0.05 mm), and recrystallized from hexane to give an analytical sample; mp 95-95°; ir (KBr) 808,

1110, 1280, 3000 cm"^.

Anal. Calcd for CnHirNC: C, 75-70; H, 9-56,• N, 7-8l.

Found; c, 75-^8; H, 9-48; ÏÏ, 7-59- The reaction was repeated and analysis by vpc on a IC^ Carbowax

2QM:K0H on °°/8 o Chrom W (versus naphthalene as an internal standard) gave kdf of 103, 42^ of 1^2 , and 3^ of N-t -butyl-p-toluidine.

o-( N-t-Butylamino) phenol g_-nitrobenzoate ( 1^5 ) - To a solution of 300 mg of 1^ and 306 mg of triethylamine in 25 ml of ether at 0-5° was added a solution of 565 mg of p-nitrobenzoyl chloride in 10 ml of ether.

Triethylamine hydrochloride immediately precipitated. After stirring for 1 hr at 0-5°, the salts were removed by filtration and the filtrate was concentrated ^ vacuo to leave yellow crystals. Eecrystallization from hexane gave 78I mg of the p-nitrobenzoate 153 - 120-122°; ir

(KBr) 719, 1065, 1520, 1750 cm-^. 1 1 2

Anal. Calcd for Ci7HiaN204: C, 6k.9^; H, 5-77j N, 8 .91.

Found: C, 6k.9k’, H, 5-94; KT, 8 .88.

Transestérification of o-( N-t-Butylamino) phenol g_-nitrobenzoate ( 1^^) .

Asolution of 513 mg of in 25 ml of dry methanol ’was stirred at

room temperature for 2 hr. The methanol was removed on the rotary

evaporator and the residue was dissolved in 50 ml of ether. The

ethereal solution was extracted with two 50-ml portions of dilute

hydrochloric acid, dried over anhydrous magnesium sulfate, filtered,

and the solvent evaporated to yield 29k mg (99^) of crude methyl p-

nitrobenzoate, mp 95-97°. The acidic aqueous extracts were then neu

tralized with sodium bicarbonate and extracted with 50-ml portions

of ether. The combined ethereal extracts were dried over anhydrous

magnesium sulfate, filtered, and the solvent evaporated to yield 268

mg (99^) of crude 1^0, mp 90-92°. The compounds were identified by

comparison with authentic samples.

3 -Chloro-2,3-dimethylindolenine ( 1 ^ . A solution of 1.45 g of 2,3-

dimethylindole in 50 ml of methylene chloride was added under a

nitrogen atmosphere to a vigorously agitated mixture of 15 0 ml of

^ sodium hypochlorite solution and 100 ml of methylene chloride at

-8 °. After stirring for 30 min, the organic layer was separated,

washed with 30 ml of cold saturated sodium chloride solution, dried

over anhydrous magnesium sulfate in the cold, and filtered. This

solution showed the presence of 85^ active chlorine as determined by a 113

low temperataire iodometric titration. When the solvent was removed in

vacuo only a dark intractible material was obtained. However, iff was

stable in solution for several hours at -10°.

5-methoxy-2,3-dimethylindolenine (l6â) • A cold methylene chloride

solution of 4.33 g (0.03 moles) of iff was added to a stirred solution

of 6.6 g (0.03 moles) of silver trifluoroacetate in 130 ml of methanol at -10°. The reaction mijcbure was stirred for 4 hr at -10° and then

allowed to warm to room temperature. The precipitate of silver

chloride was removed by filtration and washed with 30 nil of methylene

chloride. The filtrate and washings were combined, washed with two

100-ml portions of IC^ sodium bicarbonate solution and two 100-ml

portions of saturated sodium chloride solution. The organic layer was

dried over anhydrous magnesium sulfate, filtered, and the solvent re moved on the rotary evaporator to give 3-1 g of a pale yellow oil.

Distillation gave 4.9 g (94^) of pure 1^; bp 71-73° (1.3 mm); n^° =

1 .3338; ir (neat) 160I, 1133, IO6O, 773, 737 cm“^; nmr (CCI4) 8.64 t

(3H, s), 7.82 T (5H, s), 7.21 T (3H, s), 2.64 T (4h, m).

Molecular weight. Calcd for C11H13WO: 175*09970.

Found: 173.O983.

2-Carboetho%y-3-metho%y-3-methylindolenine. A solution of 1.0 g (4.93

mmoles) of 2-carboethoxy-3 -met hylindole ( 1 7 in 100 ml of methylene

chloride was cooled to -78° and chlorinated by dropwise addition of a

solution of 0.535 g (4.93 mmoles) of t-butyl hypochlorite in 10 ml Ilk of methylene chloride. After stirring for i hr at -78°, the clear solution was added to a solution of 3.0 g of silver trifluoroacetate in 50 ml of methanol at -78°. The solution was then allowed to warm to room temperature to hring about the slow precipitation of silver chloride. After stirring for 2k hr, excess lithium chloride was added and the salts were removed by filtration. The filtrate was concen trated 3^ vacuo to leave a yellow residue which was taken up in 50 ml of 1(^ sodium bicarbonate solution and extracted with three 50-ml portions of methylene chloride. The combined extracts were washed with

50 ml of saturated sodium chloride solution and dried over anhydrous magnesium sulfate. The dess leant was removed by filtration and the solvent evaporated to leave 1.20 g of crude product. Recrystallization from hexane gave 0.85 g (7^^) of 2 -carboethoxy-5 -methoxy-5 -methyl indolenine: mp 81-83°; ir (KBr) 1710, 1500, 1120, 757 cm"^; nmr (CCI4)

8.55 T (3H, t), 8.38 T (3H, s), 7-13 T (3H, s), 5.60 T (2H, q),2.J2 T

{kE, m) .

Anal. Calcd for C13H15NO3 : C, 66.93; H, 6 M; N, 6.01.

Found: C, 66.9O; H, 6.46; N, 6.13.

2-Oxobutyric Acid Phenylhydrazone ( 175) « The hydrazone was prepared 75 by the general procedure of Wislicenus and Arnold. To a mixture of

10.6 g (98 mmoles) of phenyl hydrazine and 100 ml of water was added just enough acetic acid to idssolve the hydrazine. To this solution was added 10.0 g (98 mmoles) of 2-oxobutyric acid. The hydrazone immediately precipitated. After cooling the mixture in an ice bath. 115 the product was collected hy filtration and dried vacuo to constant

weight. Recrystallization from aqueous ethanol gave 1 6.O (85^) O 7 5 of 175 ; mp 148-150 (lit. mp l49°).

2-Carhoethoxy-3 -metl^rlindole ( 1X8). The ester was prepared hy the

75 . general procedure of Wislicenus and Arnold. To 15O ml of i w

ethanolic sulfuric acid was added 16.O g (83.4 mmoles) of After

refluxing for 10 hr, the solution was added to 300 ml of water to precipitate the product. The mixture was cooled in an ice bath and the product was collected hy filtration. Recrystallization from absolute ethanol gave 12.85 g (7^ ) of 176: mp 136-138° (lit. mp

155- 134° ) .

2-Rydroxymethyl-3 -methylindole ( IJJ). This amino alcohol was prepared

in 88^ yield hy the lithium aliminum hydride reduction of 176 according 76 o , 76 to the procedure of Taylor: mp 125-128 (lit. mp 123-124 ).

2-Hydrozymethyl-3-methylindole Acetate ( 1J4). To a solution of 500 mg

of 177 in 5 ml of dry pyridine was added 316 mg of acetic anhydride.

After stirring overnight at room temperature, the solution was poured

into 20 ml of water. An oil which slowly crystallized, came out of

solution. After the mixture was cooled in an ice hath, the solid was collected hy filtration. After drying, the product was recrystallized from Shelly B to give 492 mg (78^) of the acetate nip 91 «0-95 *5°*

Anal. Calcd for C12H13NO2 : C, 70.91; H, 6.45; W, 6 .8 9 .

Found: C, 70.89; H, 6.55; N, 6 .9 7 . n 6 Chlorination and Acetolysls of 2,3-Dimethylindole. The procedure was a modification of that of G. Mehta and P. G. Gassman. To a stirred solution of 1.^5 g (0.01 moles) of 2,3-dimethylindole in 150 ml of methylene chloride at -78° was added dropwise a solution of I .19 g

(0.11 moles) of t-hutyl hypochlorite in 10 ml of methylene chloride.

The solution was stirred at -78 for 15 min and then quickly warmed to

15° with an oil bath (5-4 min warming time) . A solution of 2.65 g

( .01 moles) of thallium acetate in 50 ml of glacial acetic acid was then immediately added. Thallium chloride precipitation began at once.

After stirring at room temperature for five hr, the salts were removed by filtration and the pale yellow filtrate was added to 500 ml of water. Excess sodium bicarbonate was slowly added to neutralize the acetic acid. The organic layer was separated, washed with 100 ml of saturated sodium chloride solution, and dried over anhydrous magnesium sulfate. The drying agent was removed by filtration and the solvents evaporated in vacuo to give I .80 g of a pale yellow oil. A 500 mg sample of the oil was sublimed to give 452 mg of 1J4. Re crystallization from Skelly B gave 450 mg (76.5% from 2,5-dimethylindole) of lj4_: mp

90-92°. A mixture melting point with authentic 1J4^ showed no depression.

The remainder of the crude reaction product was used in the next reac tion.

Hydride Reduction of 2-Hiydroxymethyl-5-methylindole Acetate (lj4). To a re fluxing suspension of 500 mg of lithium aluminum hydride in 75 ml of dry ether was added I .50 g of the crude 174 from the chlorination 117 and acetolysls of 2,3-dimethylindole. After refluxing for 6 hr, 2.0 ml

of water was added dropwise and refluxing was continued for an addi

tional hr. The salts were removed by filtration and the solvent eva

porated to give a pale yellow oil which slowly crystallized. Recry

stallization from ethanolic Skelly B gave 903 mg (78^ from 2,3-

dimethylindole) of IJX : mp 125-127°. The ir and nmr spectra were

identical to those of an authentic sample. A mixture melting point

with authentic was satisfactory: mp 124-127°.

2-Methoxymethyl-3-methylindole ( ■ To a solution of 1.00 g (6.20

mmoles) of IJJ^ in 30 ml of dry ether at room temperature was added a

solution of 87 mg of boron trifluoride ethereate in 10 ml of dry ether.

Immediately a solution of about 800 mg (19 mmoles) of diazomethane in

50 ml of dry ether was added dropwise. Nitrogen was vigorously evolved.

After 10 min, 50 ml of ICF sodium hydroxide solution was added and

stirred for i hr. The mixture was decanted from precipitated poly methylene and the organic layer was separated and dried over anhydrous magnesium sulfate. The dessicant was removed by filtration and the

solvent evaporated to yield 1.00 g of a pale yellow oil which partially

solidified on standing. The oil was distilled via a molecular still

to give 781 mg (72^) of the methyl ester 173 which was a semi-solid

colorless compound at room temperature : n ^ = 1.5910. The compound

slowly decomposed in the air and a satisfactory elemental analysis was not obtained.

Molecular weight. Calcd for C11H13NO : 175•0997•

Found: 175-0999- Il8

Chlorination and, Methanolysis of 2, g-Dimethylindole. The procedure is a modification of that of G. Mehta and P. G. Gassman. To a stirred solution of l.i|-5 g ( .01 moles) of 2 ,5 -dimethylindole in 150 ml of o methylene chloride at -To was added dropwise a solution of 1.19 g

(0.011 moles) of t-hutyl hypochlorite in 10 ml of methylene chloride.

The solution was stirred at -78 for 15 min and then quickly warmed to

15° with an oil hath (5-^ min warming time). A freshly prepared solu tion of methanolic sodium methoxide made from 0.25 g of sodium metal and 50 ml of absolute methanol was immediately added. Sodium chloride slowly hegan to precipitate. After stirring at room temperature for

5 hr, the sodium chloride was removed hy filtration and the filtrate was washed with 100 ml of water and then 100 ml of saturated sodium chloride solution and dried over anhydrous magnesium sulfate. The dessicant was removed hy filtration and the solvent evaporated to leave

2.105 g of a light red oil. The oil was distilled via a molecular still to give 1.4-55 g (82^) of IJp as a pale yellow semi-solid

26 material: n^ = 1.5933. The ir and nmr spectra were identical to those of an authentic sample.

N-Hydroxy-2-phenylindole ( 178). This compound was prepared in 75^ yield hy the treatment of henzoin oxime with concentrated sulfuric acid 7 7 according to the method of Fisher.

N-Hydroxy-2-phenylindole £-Nitrohenzoate ( 179). To a solution of 500 mg (2.59 mmoles) of 1J8 and 5 00 mg of triethylamine in 20 ml of dry 119 ether at 0-5° was added a solution of kh3 mg (2.39 mmoles) of g-nitrobenzoyl chloride in 10 ml of ether. A precipitate immediately came out of solution. After stirring for an additional & hr, the ether was removed vacuo to leave a damp solid. The solid was dissolved in

50 ml of chloroform and the resulting solution was washed with 50 ml of saturated sodium bicarbonate solution and 50 ml of saturated sodium chloride solution. The organic layer was dried over anhydrous magnesium sulfate, filtered, and the solvent evaporated to leave an oil which

slowly crystallized. Re crystallization from 5C^ benzene-hexane gave

6k0 mg (75^) of ]J9,: mp 128° dec.; ir (KBr) 1T6o, 1510, 1530, 1000,

750, 708 cm-^; nmr (CDCI3) 3-50 t (i h , s ) , 2.80 t (gy, m), I .80 t

(4h , s ).

Anal. Calcd for C21H14N2O4 : C, 70.38; H, 3-9^; N, 7-82.

Found: C, 70.57; H, 4.02; N, 7-77-

2-Phenyl-5-indolol p-TTitrobenzoate ( 180). A solution of 10.0 g of 179,

in 250 ml of 5 methanol-tetrahydrofuran was refluxed for two days to precipitate a yellow solid. The solid was collected by filtration and the mother liquor concentrated to 50 ml to precipitate a second

crop of yellow solid which was collected by filtration. The solids were combined and recrystallized from 5C^ tetrahydrofuran-ethyl acetate to yield 6.52 g (65^) of l8o_: mp 284-285°; ir (KBr) 5550, 1740, 1550,

1250, 1075 cm-^; nmr (DMSO-dg) 2.55 T (9H, m), 1.55 ^ (4h, s ) .

Anal. Calcd for C21H14N2O4 : C, 70.58; H, 5-94; R, 7.82.

Found: c, 70.15; H, 4.06; N, 7-71. 1 2 0

2-Nitrobenzylidinemalonate (l8g) . This catnpound was prepared in 75^ yield from o-nitrobenzaldehyde and ethyl malonate by the method of so Louden and Sellings.

Diethyl a-Cyano-a-2-nitrobenzylmalonate ( l8g). This compound was pre pared from 182 and hydrogen cyanide in 7w yield according to the 80 method of Louden and Sellings.

N-%rdroxy-$-cyanoindole-2-carboxylic Acid. This compound was prepared ^ 81 in 80> yield by the method of Ac he son and coworkers.

N-%rdroxy-2-carboethoxy-3-cyanoindole (l8 jjj . A solution of 3*0 g (15«1 mmoles) of the acid in 100 ml of IC^ ethanolic sulfuric acid was refluxed for 5 hr. After cooling, the solution was added to 400 ml of crushed ice. The product precipitated and was collected by filtration and dried in vacuo at room temperature. Recrystallization from 20 ml of 5C^ benzene-hexane gave 2.1 g (6^ ) of l84; mp 119-120° (lit. mp 116°).

N-%rdroxyoxindole (]^8) . This compound was made in 33% yield from the reduction of o-nitrophenylacetic acid ( l8j) by the method of Wright 78 and Collins.

N-%rdroxyoxindole p-Witrobenzoate ( iS^) ' To a solution of 1.00 g (6.71 mmoles) of l88_ in 10 ml of dry pyridine was added in portions 1 .2^3 g 1 2 1

(6.71 mmoles) of p-nitrobenzoyl chloride. The solution heated up to

45° and pyridine hydrochloride precipitated. After standing for i hr at room temperature, the mixture was added to ^0 ml of water. The pro duct precipitated and was collected by filtration and dried 3^ vacuo to yield a yellow powder. Recrystallization from benzene gave 1.695 g

(85^) of 082.: mp 159-161°; ir (KBr) 1T6o, 1730, 1500, 1230, 705 cm“^; nmr (CDCI3 ) 6.32 t (2H, s ) , 2.79 ^ (^H, m), 1.62 T (4E, s) .

Anal. Calcd for CasHioHsOs : C, 60.40; H, 3-58; 9-39"

Found; c, 60.53; H, 3-47; N, 9-^1-

Transestérification of N-Bydroxyoxindole p-Kitrobenzoate (189) . A solution of 1.00 g (3-36 mmoles) of iS^ in 60 ml of dry methanol was refluxed for 24 hr. The methanol was then removed on the rotary evaporator to yield a crude yellow solid. The solid partially dis solved in 50 ml of Skelly B and the insoluble material was collected by filtration to yield 500 mg (99^) of l^r mp 196-198° (lit. mp

199-201°). A mixture melting point with authentic 188 showed no depression. The filtrate was then concentrated to 15 ml and cooling precipitated 517 mg (85%) of methyl p-nitrobsnzoate: mp 93-97°. A mixture melting point with authentic material showed no depression.

Methanolysis of N-Rydroxyoxindole p-Toluenesulfonate . To a solution of

5.00 g (33-6 mmoles) of 188 and 5-0 g of triethylamine in 200 ml of dry tetrahydrofuran at -78° was added dropwise a solution of 6.36 g (33-6 mmoles) of p-toluenesulfonyl chloride in 50 ml of dry tetrahydrofuran. 1 2 2

Triethylamine hydrochloride "began to precipitate immediately. After stirring for 2 hr at -78°, $00 ml of methanol at -78° -was added.

After stirring for an additional hr at -78°, the solution -was allowed to warm to room temperature. The solvents were then removed vacuo to leave a red oil which was dissolved in 100 ml of methylene chloride.

The solution was washed with 100 ml of saturated sodium 'bicarbonate solution and then with 100 ml of saturated sodium chloride solution, dried over anhydrous magnesium sulfate, filtered, and the solvents evaporated to leave a red oil. The oil was then separated into two components by chromatography on 250 g of silica gel. Elution with 5C(^ benzene-ethyl acetate gave 2.50 g (42^) of 5-methoxyoxindole ( 3^o ) . o , 83 which was recrystallized from 25 ml of benzene: mp 155-154 (lit. mp 155-154°) . The ir, nmr, and uv spectra of IgO were identical to published data for this compound. Elution with pure ethyl acetate gave 5-01 g (54%) of 7-hydroxyoxindole p-toluenesulfonate (ig^ which was recrystallized from tetrahydrof-uran: mp 199"201°; ir (KBr) I700,

1475, 1185, 1170, 885, 710 cm“^; nmr (CF3CO2H)7-50 t (5H, s ), 6.18 t

(2H, s), 2.92 T (3H, m), 2.59 1 (4h, q).

Anal. Calcd for C15H13EO4S: C, 59-59; H, 4.52; N, 4.62; S, 10.57-

Found: C, 59-34; H, 4.55; N, 4.66; S, 10.28.

Methanolysis of E-%rdroxyoxindole p-Nitrobenzenesulfonate. The proce dure was the same as for the methanolysis of N-hydroxyoxindole p- tolue ne sulfonate. Thus, 500 mg (5-36 mmoles) of ^ 8 and 745 mg (5-36 mmoles) of p-nitrobenzenesulfonyl chloride gave 285 mg (52^) of 19O 123 and 536 mg (3C^) of 7 -hydroxyoxindole £-nitrobenzenesnlfonate (12^ which was recrystallized from tetrahydrofuran; mp 246-248 ; ir (KBr)

1710, 1480, 1175, 828 cm-^i nmr (iMSO-ds) 6.55 ^ (2H, s ) , 5.28 t

(3H, m), 1.78 T (4h, q), -.4o t (m, s).

Anal. Calcd for Ci^HiotTsOeS: C, 50.30; H, 3.02; N, 8 .3 8 ; S, 9-59.

Found: C, 50.40; H, 3.29; M, 8 .30; S, 9-6l.

Hydrolysis of N-iyrdroxyoxindole p-Toluenesulfonate. To a solution of

500 mg of ^ 8_ and 500 mg of triethylamine in 20 ml of dry tetrahydro

furan at -78 was added dropwise a solution of 636 mg of g^-t oluene-

sulf onyl chloride in 10 ml of tetrahydrofuran. After stirring for 1 hr at- 78°, the mixture was added directly to 100 ml of 5C^ aqueous-

dioxane to give a red solution. The solution was allowed to stir at room temperature for 1 hr and was extracted with 100 ml of methylene

chloride. The organic extract was washed with 50 ml of saturated

sodium chloride, dried over anhydrous magnesium sulfate, filtered, and the solvents were evaporated in vacuo to leave a red oil. Chromato graphy on 55 g of silica gel (ethyl acetate) yielded 256 mg (2q9^) of

191 _ and 105 mg (2^ ) of 5 -hydroxyoxindole (ig^ which was re cry

stallized from 5C^ aqueous ethanol: mp 162-163° (lit. mp 165-166°). 84 The ir and uv spectra of I92 were identical to published data.

Methanolysis of N-%rdroxy-2-carboethoxy-3-cyanoindole g-Toluenesulfonate,

To a solution of 500 mg (2.42 mmoles) of l8^ and 300 mg of triethylamine in 20 ml of dry tetrahydrofuran at -78° was added a solution of 462 mg 124

(2.42 mmoles) of tosyl chloride in 5 inl of tetrahydrofuran. Triethyl

amine hydrochloride immediately precipitated. Afber stirring for 1 hr

at -78°, 50 inl of methanol was added and the resulting solution was

allowed to warm to room temperature. The solvents were removed in

vacuo to leave a residue which was separated into two components by

chromatography on 25 g of silica gel. Elution with benzene gave 591'™g

(64^) of 2 -carboethoxy-5 -cyano-5 -tosylo%yindolenine ( l8^ ) which was

re crystallized from 5<^ benzene-hexane : mp 100-110°; ir (KBr) 2220,

1750, 1580, 1185, 1175, 748 cm-^; nmr (CDCI3) 8.55 ^ (5H, t), 7-50 t

(5H, s), 5.58 T (2H, 4), 2.70 T (4h, m), 2.58 T (4h, q).

Anal. Calcd for CigHigNsOsS: C, 59-57,• H, 4.20; E, 7-29; S, 8.54.

Pound: c, 59-25; H, 4.24; n, 7.54; s, 8.55-

Elution with ld° ethyl acetate in benzene gave 101 mg (2C^) of 2-carbo-

ethoxy-5 -cyanoindole (l8^ which was recrystallized from benzene: mp

l64-l66°; ir (KBr) 5500, 2220, 1695, 126O, 742 cm“^; nmr (CDCI3) 8.50 T

(5H, t), 5.42 T (2H, q), 2.50 T (4h, m).

Kinetics. Reagents. Commercial absolute ethanol was dried by the 87 method of Lund and Bjerrum. Standard sodium thiosulfate solution

(ca. 0 . 0 2 0 n) was used to titrate the liberated iodine.

(87) Lund and Bjerrum in A. Vogel, practical Organic Chemistry, 5rd . ed., J. Wiley and Sons, Inc. , New York, w'.Y.1956, p. I67. 125 Procedure. Stock solutions of the N-chloramines in a non-polar solvent

■were prepared as indicated above for each compound. For each kinetic run, an aliquot of the stock solution -was taken and the solvent was evaporated ^ vacuo to leave the neat chloramine. The chloramine was then dissolved in ethanol buffered 0.1 IT in anhydrous sodium acetate and 0.1 IT in acetic acid, or in pure ethanol without buffer. The

ethanolic solution was quickly thermostatted to the desired temperature and aliquots were withdrawn at the desired time intervals. The aliquots were quickly added to UT methanolic potassium iodide solution and acidified with a few drops of 0.1 IT hydrochloric acid solution to

liberate the iodine. The iodometric titrations were quickly done via

sodium thiosulfate solution using a dead stop technique employing two platinum electrodes and a Fisher Accumet Model 510 pH meter.

Product Studies. The N-chloroaniline derivatives were solvolyzed for

20 half-lives. The reaction mixtures were worked up and the products

separated as given above for each compound.

Product Yields. The yields were determined by analytical glpc or by

isolation of the products as given above for each compound. REFERENCES

1. p. G. Gassman, Accounts Chem. Res., ^ 26 (1970).

2. p. G. Gassman and R. L- Cryberg, J. Amer. Chem. Soc., 9.1- 5176 (1969)• 5- J. Stieglitz and P. N. Leech, Ber., 21^7 (1915) j J* Stieglitz and p. N. Leech, J. Amer. Chem. Soc., 56, 272 ( 191^) j J. Stieglitz and B. A. Stagner, ibid., 204-6 (191^.

4. R. T. Conley and H. Brandman, unpublished results.

5- M. S. Ne-wman and P. M- Hay, J. Amer. Chem. Soc., 75, 2522 (1955)-

6. For a review of the Beckmann rearrangement, see P. Smith in "Molecular Rearrangements," P. Mayo, ed., vol. 1, p. 485-507, Interscience Publishers, Inc., New York, I965.

7 . p. Lansbury and N. Mancuso, Tetrahedron Letters, 2#-5 (1965).

8 . p. Lansbury, J. Colson, and N. Mancuso, J. Amer. Chem. Soc., 5225 (1964).

9 . p. G. Gassman and B. L. Fox, J. Amer. Chem. Soc., 538 (1967).

10. p. G. Gassman and R. L. Cryberg, ibid., QO, 1555 (I968) ; P. G. Gassman and R. L. Cryberg, ibid., 2047 (I969) -

11. p. G. Gassman and A. Carrasquillo, Tetrahedron Letters, IO9 (l97l)-

12. R. B. Woodward and R. Hoffiaan, J. Amer. Chem. Soc., 8j_, 595 (1965)* See also H. C . Longuet-Higgins and F. W. Abrahams on, ibid., 8%^ 2045 (1965). 15. C. H. DePuy, L. G. Schnack, J. W. Hausser, and W. Wiedemann, ibid., ^ 4006 (1965). l4. p. von R. Schleyer, G. W. VanDine, U. Schollkopf, and J. Paust, ibid., 88 , 2868 (1966); U. Schollkopf, K. Fellenberger, M. Patsch, P. von R. Schleyer, T. Su, and G. W. VanDine, Tetrahedron Letters, 5659 (1967) .

1 2 6 127

15- s. J. BroiSj J. Amer. Chem. Soc., 90, 506, $08 (1968).

16. P. G. Gassman and D. K. Hjrgos, J. Aoaer. Chem. Soc., $1, 15^3 (1969); P. G. Gassman, D. K. Pygos, and J. E. Trent, ibid., 22^ 2084 (1970).

IT- p. G. Gassman and R. L- Cryberg, J. iüaer. Chem. Soc., 21, 51T6 (1969). 18. A. G. Anastassiou, ibid., 8 8 , 2J22 (I966) ; C. D. Dijkgraaf and G.J. Hotjtink, Tetrahedron Supplement Ho. 2, 179 ( 190) -

19* S. T. Lee and K. Morokuma, private communication to P. G. Gassman.

20. p. G. Gassman, E . Hoy da, and J. pygos, J. Amer. Chem. Soc., 22j 2 7 1 6 (1968). ' ------

21. p. G. Gassman and J. Pygos, Tetrahedron Letters, 4749 (I970).

22. P. C. Horwell and C. ¥. Rees, Chem. Commun., l428 (1969).

23. 0. E.Edwards, P. Vocelle, J. W. ApSimon, and F. Hague, J. Amer. Chem. Soc., 8 7 , 678 (1969).

24. For reviews of nucleophilic aromatic substitution see: j. F. Bunnett and R. E. Zohler, Chem. Rev., 49, 273 (l95l); E. P. Hughes and C. K. Ingold, Q,uart. Rev., 6, 3^ (1932); J. F. Bunnett, ibid., 12, 1 (1998); R. Sauer and R. Huisgen, Angew. Chem., 72, 294 (i960); S. P. Ross, Progr. Phys. Org. Chem., ^ il~(l9^).

29. H. E. Heller, E. P. Hughes, and C . K. Ingold, Nature, 168, 909 (1951).

26. R. A. Abramovitch and B. A. Pavies, Chem. Rev., 176 (1964).

27. E. Bamburger, Ber., 2J_, 1347 (I894) .

28. E. Bamburger, ibid., 2J, 1948 (1894) .

29. E. Bamburger, ibid., 2J, 249 (1895) ■

3 0 . E. Bamburger and F. Brody, ibid., 3$, 3642 (1900).

3 1 . E. Bamburger, ibid., 40, 1906, 1918 (I907)-

3 2. E. Bamburger, ibid., 33_« 3600 (19OO).

33 - For a comprehensive review of this subject see; James A. Miller % Cancer Research, 30, 959 (1970), and references therein. 128

$4. p. G. Gassman, R. Smith, and G. Greutzmacher, unpublished results.

55* A. Bell and M. B. Knowles, U. S. Patent 2,692, 287 (1956); Chem. Abstr., 50. 2666e (1956).

3 6 . H. Suhr, Justus Liebig's Ann. Chem., @J, 175 ( I961) •

57- R. S. Neale, R. G* Schepers, and M. R. Walsh, J. Org. Chem., 22, 3 3 9 0 (1964).

3 8 . P. Haberfield and D- Paul, J. Auer. Chem. Soc., 8j^ 5502 (1963).

39" The yields of the N-chloro compounds were determined by either isolation or by treatment of an acidified solution of the chlor amine with potassium iodide followed by titration of the liberated iodine with thiosulfate.

40. For a preliminary report of this work see: P. G. Gassman, G. Campbell, and R. Frederick, J. Amer. Chem. Soc., gO, 7377 (1968).

41. Dr. Ronald Frederick carried out the initial work on the chemistry of N -chloro-N-t-butylaniline ( 100a). This author thanks Dr. Frederick for his assistance in this work.

42. E. Hecker and R. Lattrell, Annalen, 66^ 48 ( I963).

4 3 . We thank Dr. R. Steppel for suggesting this synthetic approach.

44. E. J. Corey, S. Borcza, and G- KLotmann, J. Amer. Chem. Soc., 91, 4782 (1969). ------

4 5. W. Duerckheimer and L. Cohen, Biochemistry, ^ 1948 (1964).

46. For a preliminary account of this work see: p. G. Gassman and G. A. Campbell, J. Amer. Chem. Soc., 95, 2567 (19Tl)•

47. H. C. Brown and Y. Okamoto, ibid., 80, 4g79 (1958).

48. C. Mechelynek-David and P. J. C. Fierens, Tetrahedron, ^ 2^2 ( 1959) -

49. J. Steigman and L- P. Hammett, J. Amer. Chem. Soc., 59. 2536 (1937). 50. E. D. Hughes, C. K- Ingold, and A. D. Scott, J. Chem. Soc., 1201 (1937).

51. For a discussion of the Orton rearrangement see: E. S. Gould, “Mechanism and Structure in Organic Chemistry,” Holt, Rinehart, and Winston, New York, N.Y., 1959, p. 650, and references therein. 129

52. E. Bamburger, K. Blaskopf, and A. Landau, Ber., 5K» (1919).

53- L. Horner and H. Steppan, Ann. Chem., 606, 24 (1957) •

5 4. J. Cox and M. Dunn, Abstracts, 152nd meeting of the ACS, Abstrt. SI62, New York, September ( I90ÔJ.

55* G. Tisue, M. Grassmann, and W. Lwowski, Tetrahedron, 24, 999 (1968).

5 6. K. Hoffman, A. Feldman, and E. Gelblum, J. Amer. Chem. Soc., 8 6 . 646(1964). ------

57* For comparisons of carbon g^-nitrobenzoates and chlorides see; (a) A. Fainberg and S. Winstein, J. Amer. Chem. Soc., J8j 2JJ0 (1956); (b) C. Wilcox, Jr. and M. Mesirov, ibid.TW ^ 2757 (1962) ; (c) 0. Benfey, E. Hughes,, and G. K. Ingold,' J. Chem. Soc., 2488, (1952); (d) M. Silver, J. Amer. Chem. Soc., ^ , 404 (Ï96I) ; (e) R. Buckson and S. Smith, J. Org. Chem., 32, ^4- (1967)-

5 8 . P. C. Gassman, G. Hartman, R. Cryberg, J. Trent, unpublished results.

59' W. I. Taylor, “indole Alkaloids,'' Pergamon Press, New York, N.Y., 1966, p. 3 0 .

60. W. Godtfredsen and S. Vangedal, Acta. Chem. Scand., 10, l4l4 (1956).

61. N. Finch and W. Taylor, J. Amer. Chem. Soc., 84, 1318 (1962).

62. J. Shavel, Jr. and H. Zinnes, ibid., 84, 1321 (1962) .

63. N. Finch and W. Taylor, ibid., 84, 3871 (1962).

64. G. Buchi and R. Manning, ibid., 88 , 2532 (I966).

65. H. Zinnes and J. Shavel, Jr., J. Org. Chem., 3 1 , 1765 (1966).

66. L. Dolby and G. Gribble, ibid., 32, 1391 (19^7)«

67' M. Cimo, T. Spande, and B. Witkop, J. Amer. Chem. Soc., 90. 6521 (1968).

68. Support for this process can be found in thp rearrangement of 179_ to give 180.

69. The uv data was obtained by Dr. Goverdhan Mehta. 130

70. Dr. Goverdhan Mehta, a postdoctoral student of P. G. Gassman, carried out the conversion of l ^ t o This author thanks Dr. Mehta for his help in this area.

71. It is interesting to note that 1@, being avinylogous N-chloramine, retained its oxidizing properties. Thus, l68 could bereacted t ■with acidic potassium iodine solution and the liberated iodine titrated with thiosulfate.

72. Compounds and exhibit absorption maxima at 283 m|i ( e = 6;4oo) and 279 (s = 6,300), respectively.

73. 0. Riester, Chimia, ^ 75 (1961).

7^. Dr. Goverdhan Mehta initially carried out the conversion of 168 to 17^ and 174. This author, however, significantly modified the procedures and only these will be reported.

75. W. Wislicenus and E. Arnold, Ber., 20, 3395 (1887).

7 6 . ¥. Taylor, Helv. Chem. Acta., (1950).

7 7 . E. Fisher, Ber., 28^ 1238 (1895).

7 8 . ¥. Wright, Jr. and K. Collins, J. Amer. Chem. Soc., %8_, 221 (1956).

79* R. J. Sundberg, J. Org. Chem., 3P.. 3^04 (1965).

8 0 . J. Loudon and I. Wellings, J. Chem. Soc., 3462 (1960).

8 1 . R. Acheson, C. Brookes, D. Dearnaly, and B. Quest, ibid., 504 (1968).

8 2 . Arno Liberies, “introduction to Molecular Orbital Theory," Holt, Rinehart, and Winston, Inc., New York, W.Y., I966, p. 74; 171.

8 3 . T. Wieland and 0. Unger, Ber., 9 6 , 253 ( 19^3).

84. A. Beckett, R. Daisley, and J. Walker, Tetrahedron, 24 , 6093 (I988).

8 5 . R. Sundberg, “The Chemistry of Indoles," Academic Press, New York, N.Y., 1970, p. 356.

8 6 . H. Gilman in Organic Reactions, Vol VI, (A. C. Cope, ed.). Ch. 7, p. 5^9 (1951). 8 7 . A. Vogel, Practical Organic Chemistry, 3rd ed., J. Wiley and Sons, Inc., New York, N.Y., 1956, p. 167.