the gabriel synthesis of primary amines

12
polymer crystallized from melt or solution [1OJ. Occa- sionally a less stable or low temperature structure can be produced. An example is the polycaprolactam made by zone polymerization which has not the ideal orthorhombic crystal structure, but a paracrystalline structure which has a less perfect alignment of H- bonds[331. The metaphosphates can be produced in a low temperature crystal form first which converts into the high temperature modification at higher tem- perature 31. Polyoxymethylene may under certain conditions of crystallization during polymerization form orthorhombic instead of hexagonal crystals 1121. Poly-p-xylylene also crystallizes during polymeriza- tion first as a metastable orthorhombic structure, to change at about 240 "C into the stable, denser hexa- gonal crystal structure "91. Received: July 15, 1968 [A 672 IEI German version: Angew. Chem. 80, 1009 (1968) The Gabriel Synthesis of Primary Amines BY M. S. GIBSON AND R. W. BRADSHAWI*I Reaction of potassium phthalimide with halogenoalkanes and with a variety of other alkylating agents leads to the N-alkylphthalimide. N-Substituted phthalimides may be converted into the corresponding primary amine by hydrolysis or hydrazinolysis. The scope and limitations of these reactions, which together constitute the Gabriel Synthesis of primary amines, are reviewed. 1. Introduction The Gabriel Synthesis of primary amines and amino compounds may be summarized schematically as follows: Scheme 1. + K-NHP Potassium phthalimide is generally used for the first stage but (a) sodium phthalimide and (b) phthalimide and potassium carbonate or sodium carbonate have been used to some extent; the second stage is normally accomplished by hydrolysis or hydrazinolysis [I]. The importance of the Gabriel Synthesis lies in (i) the absence of secondary or tertiary amine contamination of the primary amine, (ii) the toleration of a very wide range of other functional groups in the molecule, and (iii) the mild conditions now available for accomplish- ing both stages. High yields of primary amines are generally obtainable, rendering the synthesis attractive, [*] Prof. M. S. Gibson Department of Chemistry, Brock University, St. Catharines, Ontario (Canada) Dr. R. W. Bradshaw Department of Chemistry, John Dalton College of Technology Manchester 1 (England) 11 I G. Spielberger in Horrben- Weyl-Miiller: Methoden der orga- nischen Chemie. 4th Edit., Thieme-Verlag, Stuttgart 1957, Vol. 11/1, p. 79. inter alia, for the preparation of isotopically labeled amines and amino acids (21. The successful conversion of methyl and ethyl iodides into the N-alkylphthalimides by reaction with potassium phthalimide at 150°C was reported by Craebe and Pictet in 1884r3J. It was already known that N-alkyl- phthalimides could be hydrolyzed to phthalic acid and the primary amine, but it was Gabriel who appreciated the significance of these reactions and formulated the Synthesis which bears his nameI41. The scope of the synthesis was explored by Gabriel and his colleagues in Berlin for over 30 years. In the early days, examples slowly accumulated of the preparation of primary alkylamines and aralkylamines from alkyl bromides (including ally1 bromide) 151 and aralkyl chlorides and bromides [4,61. However, from the beginning, greater interest attached to applications of the reaction to bi- and poly-functional halogeno compounds (with potassium phthalimide) and to transformations, including hydrolysis, of the products. In the aralkyl series, early examples included the chlo- romethylnitro- and chloromethylcyano-benzenes [4,71. Reaction occurred readily (with potassium phthal- [2] A. Murray and D. L. WiNiams: Organic Syntheses with Iso- topes. Part 1, Chap. 2 and 4; Part 2, Chap. 17, Interscience, New York 1958. 131 C. Graebe and A. Pictet, Ber. dtsch. chem. Ges. 17, 1173 (1884); Liebigs Ann. Chem. 247, 302 (1888). 141 S. Gabriel, Ber. dtsch. chem. Ges. 20, 2224 (1887). [5] a) A. Neumann, Ber. dtsch. chem. Ges. 23, 994 (1890); b) S. Gabriel, ibid. 24, 3104 (1891). 161 H. Strassmann, Ber. dtsch. chem. Ges. 21, 576 (1888). [7] a) S. Gabriel and H. Hendess, Ber. dtsch. chem. Ges. 20, 2869 (1887); b) H. K. Gunfher, ibid. 23, 1058 (1890). Angew. Chem. internat. Edit. Vol. 7 (1968) [No. I2 919

Upload: kybernetikum

Post on 10-Oct-2014

1.772 views

Category:

Documents


13 download

TRANSCRIPT

Page 1: The Gabriel Synthesis of Primary Amines

polymer crystallized from melt o r solution [1OJ. Occa- sionally a less stable or low temperature structure can be produced. An example is the polycaprolactam made by zone polymerization which has not the ideal orthorhombic crystal structure, but a paracrystalline structure which has a less perfect alignment of H- bonds[331. The metaphosphates can be produced in a low temperature crystal form first which converts into the high temperature modification at higher tem-

perature 31. Polyoxymethylene may under certain conditions of crystallization during polymerization form orthorhombic instead of hexagonal crystals 1121.

Poly-p-xylylene also crystallizes during polymeriza- tion first as a metastable orthorhombic structure, to change at about 240 "C into the stable, denser hexa- gonal crystal structure "91.

Received: July 15, 1968 [A 672 IEI German version: Angew. Chem. 80, 1009 (1968)

The Gabriel Synthesis of Primary Amines

BY M. S . GIBSON AND R. W. BRADSHAWI*I

Reaction of potassium phthalimide with halogenoalkanes and with a variety of other alkylating agents leads to the N-alkylphthalimide. N-Substituted phthalimides may be converted into the corresponding primary amine by hydrolysis or hydrazinolysis. The scope and limitations of these reactions, which together constitute the Gabriel Synthesis of primary amines, are reviewed.

1. Introduction

The Gabriel Synthesis of primary amines and amino compounds may be summarized schematically as follows:

Scheme 1. + K-NHP

Potassium phthalimide is generally used for the first stage but (a) sodium phthalimide and (b) phthalimide and potassium carbonate or sodium carbonate have been used to some extent; the second stage is normally accomplished by hydrolysis or hydrazinolysis [I]. The importance of the Gabriel Synthesis lies in (i) the absence of secondary or tertiary amine contamination of the primary amine, (ii) the toleration of a very wide range of other functional groups in the molecule, and (iii) the mild conditions now available for accomplish- ing both stages. High yields of primary amines are generally obtainable, rendering the synthesis attractive,

[*] Prof. M. S. Gibson Department of Chemistry, Brock University, St. Catharines, Ontario (Canada) Dr. R. W. Bradshaw Department of Chemistry, John Dalton College of Technology Manchester 1 (England)

11 I G. Spielberger in Horrben- Weyl-Miiller: Methoden der orga- nischen Chemie. 4th Edit., Thieme-Verlag, Stuttgart 1957, Vol. 11/1, p. 79.

inter alia, for the preparation of isotopically labeled amines and amino acids (21.

The successful conversion of methyl and ethyl iodides into the N-alkylphthalimides by reaction with potassium phthalimide at 150°C was reported by Craebe and Pictet in 1884r3J. It was already known that N-alkyl- phthalimides could be hydrolyzed to phthalic acid and the primary amine, but it was Gabriel who appreciated the significance of these reactions and formulated the Synthesis which bears his nameI41. The scope of the synthesis was explored by Gabriel and his colleagues in Berlin for over 30 years.

In the early days, examples slowly accumulated of the preparation of primary alkylamines and aralkylamines from alkyl bromides (including ally1 bromide) 151 and aralkyl chlorides and bromides [4,61. However, from the beginning, greater interest attached to applications of the reaction to bi- and poly-functional halogeno compounds (with potassium phthalimide) and to transformations, including hydrolysis, of the products. In the aralkyl series, early examples included the chlo- romethylnitro- and chloromethylcyano-benzenes [4,71.

Reaction occurred readily (with potassium phthal-

[2] A. Murray and D. L. WiNiams: Organic Syntheses with Iso- topes. Part 1, Chap. 2 and 4; Part 2, Chap. 17, Interscience, New York 1958. 131 C. Graebe and A. Pictet, Ber. dtsch. chem. Ges. 17, 1173 (1884); Liebigs Ann. Chem. 247, 302 (1888). 141 S. Gabriel, Ber. dtsch. chem. Ges. 20, 2224 (1887). [5] a) A . Neumann, Ber. dtsch. chem. Ges. 23, 994 (1890); b) S . Gabriel, ibid. 24, 3104 (1891). 161 H. Strassmann, Ber. dtsch. chem. Ges. 21, 576 (1888). [7] a) S. Gabriel and H. Hendess, Ber. dtsch. chem. Ges. 20, 2869 (1887); b) H. K. Gunfher, ibid. 23, 1058 (1890).

Angew. Chem. internat. Edit. Vol. 7 (1968) [ N o . I2 919

Page 2: The Gabriel Synthesis of Primary Amines

imide) and control of temperature was important; sometimes the reactions were necessarily moderated by diluting the reactants with sodium chloride181 or (in one case) benzyl cyanide191, but years were to elapse before the use of solvents became commonplace.

With the N-(cyanobenzyl)phthalimides, the vigorous treatment used for hydrolysis led to the corresponding amino acidtlol. In the aliphatic series, such bi- and poly-functional compounds as monochloroacetone, ethyl chloroacetate and 1,3-dichloro-2-propanoi [II] ,

1 -brorno-2-phenoxyethane [12J and 4-bromobutane- nitrile [I31 were examined; acidic hydrolysis gave the expected amino compounds, the ester and nitrile groups also undergoing hydrolysis.

Greatest interest attached to the particular cases of 1,2- dihalogenoethanes and 1,3-dihalogenopropanes in which, by choice of conditions, one or both halogen atoms could be displaced in Stage 1. If both halogen atoms were replaced, completion of the Gabriel Syn- thesis gave ethylenediamine and trimethylenediamine respectively [4,141. If one halogen atom were displaced, the resulting N-(halogenoalky1)phthalimide could be subjected to hydrolysis, giving halogenoalkylamine salts, or to metathetical reactions with various nucleo- philes. Thus, reaction of N-(2-bromoethyl)phthalimide with (a) aniline [151, (b) potassium hydrogen sul- fide 1161, and (c) sodium benzenesulfinate 1171 gave (a) N-(monoanilinoethy1)- and N-(dianilinoethy1)- phthalimides, (b) the N-(mercaptoethy1)phthalimide and N,N'-thiodiethylenediphthalimide, and (c) phenyl- sulfonylethylphthalimide respectively, which could be hydrolyzed to the corresponding amino compounds. Again, use of sodio derivatives of diethyl malonate and its monoalkylated derivatives provided syntheses of amino acids 1181.

Other useful developments hinged on the stability of the phthalimido group under a variety of conditions which permitted the modification of other functional groups. These may be illustrated by the use of 3- phthalimidopropionyl chloride in the Friedel-Crafts acylation of benzene 1191, the K-bromination of w- phthalimidoalkylmalonic esters in the preparation of a,w-diamino acids 1201, and the oxidation of N-(w-

[8] C. WOK Ber. dtsch. chem. Ges. 25, 3030 (1892); G. Eanse, ibid. 27, 2161 (1894). [9] a) S. Gabriel and R. Jansen, Ber. dtsch. chem. Ges. 24, 3091 (1891); b) see J . v. Eraun, ibid. 37, 3583 (1904).

[lo] S. Gabriel and W. Landsberger, Ber. dtsch. chem. Ges. 31, 2732 (1898); F. Ehrlich, ibid. 34, 3366 (1901). [ll] C. Goedeckemeyer, Ber. dtsch. chem. Ges. 21, 2684 (1888). [I21 C. Schmidt, Ber. dtsch. chem. Ges. 22, 3249 (1889). 1131 a) S. Gabriel, Ber. dtsch. chem. Ges. 22, 3335 (1889); b) 23, 1767 (1890). [14] S. Gabriel and J . Weiner, Ber. dtsch. chem. Ges. 21, 2669

I151 S. Gabriel, Ber. dtsch. chem. Ges. 22, 2223 (1889). [161 S. Gabriel, Ber. dtsch. chem. Ges. 24, 1110 (1891). I171 S. Gabriel and J . Colman, Ber. dtsch. chem. Ges. 44, 3628 (1911). 1181 W. Aschan, Ber. dtsch. chem. Ges. 23, 3692 (1980); 24, 2443 (1891). [19] S. Gabriel, Ber. dtsch. chem. Ges. 41, 242 (1908). [20] E. Fischer, Ber. dtsch. chem. Ges. 34, 454, 2900 (1901).

(1888).

hydroxyalky1)- and N-(w-mercaptoalky1)phthalimides to the corresponding carboxylic and sulfonic acids 116,211.

The general usefulness of the Gabriel Synthesis was thus established, though limitations were imposed by the reaction conditions in common use. Most of the alkylation reactions (Stage 1) were carried out by heating the reactants together at temperatures in the range 120-240 "C for periods of up to several hours, even though such conditions were recognized as un- desirable or even dangerous in some cases [221. In this context, it is perhaps remarkable that Beck's demon- stration in 1893 that 1-chloromethyl-2-nitrobenzene reacted readily enough with potassium phthalimide in boiling ethanol 1231 should have been overlooked.

The vigorous conditions of acidic hydrolysis (Stage 2) used at this time prevented extension of the synthesis to compounds containing acid-sensitive groups, al- though preliminary alkaline hydrolysis to the N-sub- stituted phthalamic acid, followed by milder acidic hydrolysis to phthalic acid and the amine, permitted some variation. Again, the significant observation that hydrazine readily cleaved the phthaloyl group from ethyl phthalimidoacetate as phthalhydrazide under quite mild conditions (Radenhausen, 1895) [Z41 had either been missed or insufficiently appreciated. Ap- proximately 30 years were to elapse before hydrazinol- ysis was introduced as a technique for accomplishing Stage 2 of the Gabriel Synthesis 1253.

2. Stage 1: Formation of the N-Substituted Phthalimide

2.1. Displacement Reactions

2.1.1. Alkyl Ha l ides

The most commonly used alkylating agents in the Gabriel Synthesis are the alkyl halides. As indicated above, a wide range of other functional groups are tolerated as substituents within the molecule, although it may be necessary or desirabIe to protect some func- tional groups in order to facilitate reaction; e.g. it is standard practice to use halogenoalkanoic esters (not the acids), and it seems necessary to protect phenolic groups [261.

The consensus of opinion is that the alkylation stage is an sN2 reaction, the ambident phthalimide ion reacting at nitrogen, though this view rests largely on analogy with other nucleophilic reactions of alkyl halides 1271. Whilst in the main

[21] S . Gabriel, Ber. dtsch. chem. Ges. 38, 630 (1905). [22] S. Reich and A. Oganessian, Bull. SOC. chim. France (4) 21, 117 (1917). 1231 C. Beck, J. prakt. Chem. (2) 47, 397 (1893). [24] R. Radenhausen, J. prakt. Chem. (2) 52, 433 (1895). [25] H . R . Zng and R . F. H . Manske, J. chem. SOC. (London) 1926, 2348. [26] a) F. Turin, F. W. Caton, and A . C. 0. Hann, J. chem. Soc. (London) 95, 2113 (1909); b) F. Tutin, ibid. 97, 2495 (1910). [27] J . D . Roberts and M. C. Caserio: Basic Principles of organic Chemistry. Benjamin, New York 1964, p. 554.

920 Angew. Chem. internat. Edit. 1 Vol. 7 (1968) 1 No. 12

Page 3: The Gabriel Synthesis of Primary Amines

the products are those expected from the sN2 mechanism, it is clear that this mechanism cannot apply in all cases.

There are indications that primary chloroalkanes are less re- active than the corresponding bromides or iodides in the Gabriel Synthesis (see e.g. [28--331). However, since the reac- tions are carried out under such a wide range of conditions, the comparison must be interpreted cautiously. There is, nonetheless, a predisposition to accept this view of the rela- tive ease of reaction of halides, borne out by the choice of bromide or iodide rather than chloride 1341, the unreactivity of alkyl fluorides (illustrated by the preparation of 2-(fluoro- ethy1)phthalimide from 1-bromo-2-fluoroethane [351), and the occasional use of iodide ion as additiver361 (for examples of other additives see Ref. [30a, 37-39]), Recently, the reactions of 1-chloro-, 1-bromo-, and I-iodopropane with potassium phthalimide in dry dimethylformamide at 25 "C have been found to be first order, in halogenoalkane and in potassium phthalimide, the order of reactivity being RCI < RBr < RI; the rates of the first two reactions were not increased by addition of iodide ion r36Cl.

Primary halogenoalkanes not containing other func- tional groups-, except sterically hindered compounds of the neopentyl type, react reasonably well with po- tassium phthalimide. Thus (1) shows very marked difference in reactivity of the two bromine atoms; a competing elimination reaction was noted in this case, although the by-product was not identified 1401. Ex- amples in which this type of compound has been induced to react are provided by (2a) , (2b)r411 and

$133 N-CH2-CH2-y- CHzBr

H~C-C(CHZX)~ (2a), X = ~ - C H ~ - C G H ~ - S O ~ (Zb), X = Br

(CH3)2C(CHzX)z (3a), X = CsHsSO3 (361, X = Br

(CH3)nSi(CH2X)m (4a), n = 3, m = 1, X = C1 (4b), n = 2, m = 2, X = C1

[28] S. Gabriel, Ber. dtsch. chem. Ges. 21, 566 (1888). [29] S. Gabriel, Ber. dtsch. chem. Ges. 38, 3411 (1905). [30] a) L . H. Cretcher, J . A. Koch, and W. H. Pittenger, J. Amer. chem. SOC. 47, 1173 (1925); b) A. W. Baldwin and R. Robinson, J. chem. SOC. (London) 1934, 1264. [31 J a) 0. Seitz, Ber. dtsch. them.-Ges. 24, 2624 (1891); b) A. V. Belotsvetov and V. A. Izmail'skii, Z. obBE. Chim. 14, 216 (1944). [32] H. R. V. Arnstein, G. D. Hunter, H. M. Muir, and A. Neu- berger, J. chem. SOC. (London) 1952, 1329. (331 W. C. Austin, M. D. Potter, and E. P. Taylor, J. chem. SOC. (London) 1958, 1489. [34] D. B. Capps and C. S. Hamilton, J. Amer. chem. SOC. 75, 697 (1953). [35] A. F. Childs, L. J . Goldsworthy, G. F. Harding, F. E. King, A. W. Nineham, W. L. Norris, S. G. P. Plant, B. Selton, and A . L. L. Tompsett, J . chem. SOC. (London) 1948, 2174. [36] a) J . D. Roberts and R. H. Mazur, J. Amer. chem. SOC. 73, 2509 (1951); b) R . J . Collins, B. Ellis, S B. Hansen, H. S . Macken- zie, R. J. Moualim, V. Petrow, 0. Stephenson, and 3. Sturgeon, .I. Pharmacy Pharmacol. 4, 693 (1952); c) R. W. Bradshaw, un- published. 1371 F. P. Dwyer, N. S. Gill. E. C. Gyarfas and F.Lions, J. Amer. Chem. SOC. 75, 1526 (1953). 1381 G. R. Clemo, R. Raper, and C. R. S. Tenniswood, .I. chem. SOC. (London) 1931, 429. (391 R. Schoenheimer and S. Ratner, J. biol. Chemistry 127, 301 (1939). 1401 R . F. Brown and N. M. van Gulick, J. Amer. chem. SOC. 77, 1089 (1955). [41] H. Stetter and W. Buckmann, Chem. Ber. 84, 834 (1951).

(3a) and (36) 1421; in each case the reaction was car- ried out at high temperature. The analogous silicon compounds (4a) and (4b) have been successfully sub- stituted using phthalimide and potassium carbon- ate 143a1, and potassium phthalimide in dimethylform- amide r43bI.

Substitution has also been achieved with some secondary halogenoalkanes, though apparently with greater difficulty, and with attendant side reactions (elimination). 2-Bromo- propane and 2-bromobutane [441 give the substituted phthalirnide and apparently olefins; bromocyclohexane is reported to give cyclohexene [251. Bromodiphenylmethane gives according to conditions, the substituted phthalimide or mainly tetraphenylethylene [451.

Systems involving differently environed halogen atoms react preferentially at the carbon which is more SUS-

ceptible to s N 2 attack. An improved synthesis of the antimalarial drug primaquine, utilizes this idea in the treatment of 1,4-dibromopentane with potassium phthalimide in acetone to give 67 % of the desired N- (4-bromopenty1)phthalirnide [461. In the case of 1,2,3- tribromopropane, the product is N-(2-bromoally1)- phthalimide 1471. Elimination is also noted with 1,3- dichloro-3-methylbutane, which gives isopentenyl- phthalimide [48J, and with 1,4-dibromopentane in ethanol, which gives polymeric material [461.

There are a considerable number of applications to cases of halogeno compounds bearing a ketonic [491, ester [491, or nitrile [SO] function at the halogenated carbon atom; these relate to syntheses of amino- ketones, amino-acids, or derivatives thereof. A partic- ularly important case is that of diethyl bromomalonate, which gives the well-known diethyl phthalimido- malonate, starting point of the Sorensen amino-acid synthesisrsll. Scission of a benzoyl group has been reported during reaction of Z-bromo-l,3-diphenyl-l,3- propanedione with potassium phthalimide in ethanol (no cleavage is reported in benzene, or without solvent) [Q]. Substances such as 2-chloropropionitrile give acrylonitrile and a small amount of the substituted

1421 G. S. Skinner and P. R. Wurz, J. Amer. chem. SOC. 73, 3814 (1951). 1431 a) L. H. Sommer and J. Rockett, J. Amer. chem. SOC. 73, 5130 (1951); b) J . Goubeau and H. D. Fromm, Z. anorg. allg. Chem. 317, 41 (1962). [44] 0. Mumm and H. Richter, Ber. dtsch. chem. Ges. 73, 843 (1940). [45] a) C. K. lngold and C. L. Wilson, J. chern. SOC. (London) 1933, 1493; b) A . McKenzie and F. Barrow, ibid. 103, 1331 (1913).

[46] R. C. Elderfield, H. E. Mertel, R . T. Mitch, I. M. Wempen, and E. Werble, J. Amer. chem. SOC. 77, 4816 (1955). [47] J . A . Lamberton, Austral. J . Chem. 8, 289 (1955). I481 A. Gomory and L. Jeio, Chem. Zvesti 7, 41 (1953); Chem. Abstr. 49, 6185 (1955). [49] J. C. Sheehan and W. A. Bolhofer, J. Amer. chem. SOC. 72, 2786 (1950). [50] a) A. Sonn and S. Falkenheim, Ber. dtsch. chem. Ges. 55, 2975 (1922); b) J . Jenni, H. Kiihne, and 8. Prijs, Helv. chim. Acta 45, 1163 (1962). [51] a) S. P. L. Sorenson, Z . physiol. Chem. 44, 448 (1905); b) G. Barger and T. E. Weichselbaum, Biochem. J. 25, 997 (1931); c) R. N. Fink, T. Enns, C. P. KimbaU, H. E. Silberstein, W. F. Bale, S. C. Madden, and G. H . Whipple, J. exp. Medicine 80, 455 (1944). [52] J . Pascual and L. Rey , An. SOC. espan. Fisica Quim. 28, 632 (1930); Chem. Abstr. 25, 3645 (1931); b) see H. v. Enler, H. Has- selquist, and 0. Ceder, Liebigs Ann. Chem. 581, 198 (1953).

~ _ _

Angew. Chern. internat. Edit. / Vol. 7 (1968) / N o . 12 92 1

Page 4: The Gabriel Synthesis of Primary Amines

phthalimide when heated with potassium phthalimi- de [531; however, acrylonitrile adds to potassium phthalimide under appropriate conditions [543.

Allylic halides react reasonably we11 with potassium phthalimide to give the corresponding substituted phthalimides [36a, 551. N-(cis-2-butenyl)- and N-(trans- 2-buteny1)phthalimides have been made in this way 1561.

Meisenheimer and Link 1571 have investigated the allylic isomers (5 ) and (6) in this reaction, and obtained mixtures of virtually the same composition from each: the primary allylic phthalimide is the major compo- nent, and the overall yield of isomers was appreciably lower from the secondary halide. The reactions were conducted at 190-200 OC, and the absolute purity of the halides may perhaps be open to question. However,

0

the possibility of rearrangement remains, at least in some cases, and further investigation seems desirable. Propargylic halides have also been used in the Gabriel Synthesis 1581, and are reported to give excellent yields of the derivative.

Gabriel occasionally utilized solvents in some of his reactions of halogenoalkanes with potassium phthal- imide [9a, 591, although he evidently preferred the ab- sence of a diluent for most of his investigations in this field (see Ref. [9,23,25,45b,52a,55c,59-721 for examples of the solvents used and referred to in the older liter-

[53] J . F. Olin and T. B. Johnson, J. Amer. chem. SOC. 53, 1475 (1931). [54] A . Galat, J. Amer. chem. SOC. 67, 1414 (1945). 1551 a) E. Bergmann, J. chem. SOC. (London) 1935, 1361; b) W . J. Gender and J . C . Rockett, J. Amer. chem. SOC. 77, 3262 (1955); c) W . Langenbeck, W . Woltersdorf, and H. Blachnitzky, Ber. dtsch. chem. Ges. 72, 671 (1939); d) 0. Wichterle and M . Hudlick9, Collect. czechoslov. chem. Commun. 12, 101 (1947); e) A . Terada, J. chem. SOC. Japan, pure. Chem. Sect. (Nippon Kagaku Zassi) 81, 1773 (1960). 1561 a) A. Kjaer and K . Rubinstein, Acta chem. scand. 8, 1335 (1954); b) A . Kjaer, K . Rubinstein, and K . A . Jensen, ibid. 7, 518 (1953) . [57] J. Meisenheimer and J . Link, Liebigs Ann. Chem. 479, 21 1 (1930). 1581 a) M. M. Fraser and R. A . Raphael, J. chem. SOC. (London) 1952,226; b) M . G. Ettlinger and J . E. Hodgkins, J. Amer. chem. SOC. 77, 1831 (1955). [59] a) S. Gabriel and J. Colman, Ber. dtsch. chem. Ces. 35, 3805 (1902); b) S. Gabriel, ibid. 41, 1127 (1908); c) S. Gabriel and W . Gerhard, ibid. 54, 1067 (1921). 1601 H. Stephen and C . Weizmann, J. chem. SOC. (London) 105, 1046 (1914). 1611 E. Spath and W . Spitzy, Ber. dtsch. chem. Ges. 58, 2273 (1925); K. H. Slotfa and R. Tschesche, ibid. 62, 1398 (1929). [62] J. Loevenich, W . Becker, and T. Schroder, J. prakt. Chem. (2) 127, 248 (1930). [63] N . L. Drake and J. A . Garman, J. Amer. chem. SOC. 71,2425 (1949). I641 H. Rupe and F. Fehlmann, Helv. chim. Acta 9, 80 (1926).

ature). Dimethylformamide is now widely employed as a solvent for Stage 1 of the Gabriel Synthesis 1491 because potassium phthalimide is moderately soluble in this solvent, and relatively mild conditions suffice for the displacement of halogen by phthalimide ion. Moreover, yields are significantly enhanced [73,74J, though again the presence of water may adversely affect some reactions 1753. However, other solvents continue to be employed, e.g. petrol [761, kerosene 1421, acetamide [771, chlorobenzene 1781, acetonitrile C791, amyl alcohol [801, diglycol[811, ethylene glycol [36bl, and dimethylsulfoxide [821. However, on occasions the absence of a solvent may still be preferred where yields are critical, as in the synthesis of 13CH3NH2, which proceeds in higher yield using the original method [83].

A recent preliminary account of the use of hexamethyl- phosphoramide (hexametapol) 1841, indicates that Stage 1 of the Gabriel synthesis may be accomplished under conditions comparable with those reported for dimethylformamide. Indeed, it has been shown that sodium phthalimide can be generated in situ in this solvent, and that some reactions occur at room tem- perature with equally acceptable yields.

[65] 0. Widman and E. Wahlberg, Ber. dtsch. chem. Ges. 44, 2065 (1911). [66] N . Putochin, Trans. Inst. pure chem. Reagents, Sci. Tech. Dept. USSR 300, 119 (1929); Chem. Abstr. 24, 3756 (1930). [67] Brit. Pat. 600851 (1948), E. Lilly and Co., Chem. Abstr. 42, 8827 (1948). [68J J. H. Helberger, G. Manecke, and R. Heyden, Liebigs Ann. Chem. 565, 22 (1949). I691 B. R . Baker and M . V. Querry, I. org. Chemistry 15, 413 (1950). [70] W . Huber, R . 0. Clinton, W. Boehme, and M . Jackman, J. Amer. chem. SOC. 67, 1618 (1945). [71] A . Burger and G. E. Ullyot, J. org. Chemistry 12, 342 (1947). 1721 G. E. Hall and F. M . Umbertini, J. org. Chemistry 15, 715 (1950). [73] a) L . S. Hafner and R. Evans, J. Amer. chem. SOC. 79, 3783 (1957); b) L. Birkofer and K. Hempel, Chem. Ber. 93,2282 (1960); c) G. A . Alles, B. B. Wisegarver, N . B. Chapman, and A . J . Tompsett, J. org. Chemistry 22, 221 (1957); d) M . Sletzinger, W. A . Gaines, and W. V. Ruyle, Chem. and Ind. 1957, 1215; e) A. N. Nesmejanow, K. N . Anisimow, and Z. P. Valvera, Doklad. Akad. Nauk S S S R 162, 112 (1965). [74] H . K . Muller and G. Rieck, J. prakt. Chem. (4) 9, 3 0 (1959). 1751 J . H . Billman and R. V. Cash, Proc. Indiana Acad. Sci. 62, 158 (1953); A . N . Nesmejanow, K . G. Perevalova,L. S. Shilovtseva, and V . D . Tyurin, Izvest. Akad. nauk. SSSR, Otdel. chim. Nauk. 1962, 1997. [76] 0. Riobd and H. Cottin, C . R. hebd. SBances Acad. Sci. 240, 7783 (1955). 1771 W. Siedel and H . Nahm, German Pat. 928711 (June 10, 1955) Farbwerke Hoechst A.G.; Chem. Abstr. 52, 5471 (1958). [78J H . J. Barber, R. F. Fuller, M . B. Green, and H . T. Zwartouw, J. appl. Chem. 3, 266 (1953). [79] K. Wallenfels, F. Witzler, and K. Friedrich, Tetrahedron 23, 1845 (1967). [80] M . Ohara and T . Hattori, J. pharmac. SOC. Japan (Yakuga- kuzasshi) 71, 575 (1951); Chem. Abstr. 46, 4548 (1952). 1811 A . Weissberger, D. B. Glass, and P . W. Vittum, US-Pat. 2552240 (1951), Eastman Kodak Co.; Chem. Abstr. 45, 7900 (1951). [82] F. Sparatore and F. Pagani, Farmaco (Pavia), Ediz. sci. 20, 248 (1965); Chem. Abstr. 63, 8344 (1965). I831 J . D. Cox and R. J . Warne, J. chem. SOC. (London) 1951, 1896. [84] H. Normant and T. Cuvigny, Bull. SOC. chim. France 1965, 1866.

922 Angew. Chem. infernat. Edit. 1 Vol. 7 (1968) /No. I2

Page 5: The Gabriel Synthesis of Primary Amines

From the discussion in Section 2.1.4. it is clear that the first stage of the Gabriel synthesis does not always involve direct substitution of halogen by phthalimide ion in these reactions. It may be mentioned in passing, however, that the proposed sN2 mechanism does in- volve definite stereochemical implications. In a rare case of substitution at an asymmetric carbon atom, Fischer 1851 noted that (-)ethyl 2-bromopropionate, when treated with potassium phthalimide at 125 "C for 5 hours, gave the phthalimido derivative with approx- imately 42 % racemisation. Mild conditions are now available for carrying out the Gabriel Synthesis, and this point clearly requires investigation with optically stable materials.

2.1.2. Ary l H a l i d e s a n d Vinyl Ha l ides

In the early days of the Gabriel Synthesis, it was noted that picryl chloride reacted readily (70 "C) with potas- sium phthalimide to give N-picrylphthalimide; 2- chloroquinoline, however, was observed not to react even at temperatures as high as 2 3 0 " C r l z J . Later authors have expressed the view that picryl chloride is the only aromatic halide to react in the Gabriel Syn- thesis 1861, though this is not in fact the case.

1 -Chloroanthraquinone has been reported to react with phthalimide and sodium acetate in the presence of copper at temperatures near 200 "C, hydrolysis with sulfuric acid yielding 1-aminoanthraquinone. Similar results were obtained with 1,5-dichloroanthraquinone and other substituted anthraquinones [871. Again, 2- bromofluorene, 1 - and 2-chIoroanthracenes, and 9- bromoanthracene react with potassium phthalimide in nitrobenzene under similar conditions, though in indifferent yields [621.

A number of aromatic systems bearing halogeno substituents and capable of facilitating delocalization of the negative charge during attack by phthalimide ion (e.g. -NOz, -CN), have also been noted to react. Examples are provided by l-fluoro-2,4-dinitroben- zene 1881, and (less readily) l-chloro-2,4-dinitroben- zene 1881, 1 -chloro-2-nitrobenzene 1893 (although 1- chloro-4-methoxy-2-nitrobenzene is said not to react under the same conditions [891), 1,3,5-trichloro- and 1,3,5-trifluor0-2,4,6-tricyanobenzenes [791.

Cyanuric chloride reacts with potassium phthalimide (1 equivalent) in acetone to give N-(4,6-dichlorotri- azin-2-y1)phthalimide [903. Acetonylacetone condenses with N-aminophthalimide to give N-(2,5-dimethyl- pyrrol-1 -yl)phthalimide [911.

1851 E. Fischer, Ber. dtsch. chem. Ges. 40, 489 (1907) 1861 H. Krauch and W . Kunz, Organic Name Reactions. Wiley, New York 1964, p. 182. 1871 H . A . E . Drescher andJ . Thomas, US-Pat. 1528470 (March 3, 1925), Chem. Abstr. 20, 424 (1926). (881 A . L . Beckwith and J . Miller, Austral. J. sci. Res., Ser. A 5, 786 (1952). 1891 R. C. Elderfield, L. L. Smith, and E. Werble, J. Amer. chem. SOC. 75, 1245 (1953). 1901 W. F. Beech, J. chem. SOC. (London) (C) 1967, 466. 1911 R. Epton, Chem. and Ind. 196S, 425.

a-Bromostyrene (with copper catalysis) 1921 and tetra- fluorobenzoquinone 1931 provide examples of unsatu- rated halogeno compounds which react with potas- sium phthalimide. Nesmejanow and his co-workers have investigated the reaction of copper phthalimide with bromo- and chloroferrocene [941, when fairly good yields of N-ferrocenylphthalimide were obtained.

Recently, work has been performed on the direct phthalimidation of unsubstituted aromatic systems. For example N-chlorosulfonylphthalimide has been shown to react with biphenyl [95l; however, the prod- uct yield is low. Without the use of any catalyst, the major product of the reaction is N-(4-biphenylyl) phthalimide, whereas in the presence of cuprous chloride, the three isomeric monophthalimido- derivatives are produced in approximately equal pro- portions. This suggests that the reaction proceeds via a free radical mechanism, a view supported by the iso- lation of phthalimide itself from the reaction mixture. Since N-chlorophthalimide reacts with biphenyl to give only a trace of N-(4-biphenylyl)phthalimide, it i s thought that this possible intermediate is not formed. The reaction has been extended to halogenated ben- zenes using either cuprous chloride, or cuprous naph- thenate as catalyst.

2.1.3. O t h e r D i sp laceab le G r o u p s

Although the metathetical reaction of alkyl tosylates with phthalimide ion was discovered in 1928 1961, it was nearly 20 years later before Sakellarios [971 developed the use of these derivatives as alternatives to halogeno- alkanes in the Gabriel Synthesis. This author discussed the advantages of tosylates over halides with respect to their lower volatility and greater thermal stability. It may be added that these derivatives are also more accessible in the pure state than the corresponding halogenoalkanes. Coupled with the fact that yields are often enhanced when sulfonates rather than the bromides are employed [41,421, it is perhaps surprising that relatively few examples of this modification of the Gabriel Synthesis are reported. Moreover, the reaction with potassium phthalimide and 2-chloroethyl tos- ylate gives N-(2-chloroethyl)phthalimide in excellent yield 1971, at least suggesting that sulfonates are more reactive than the corresponding chlorides. Other typical exsmples include such tosylates as (7) (981 and (8) [821, the sultones (9) [68J and (10) (991, and the neopentyl sulfonates ( 2 ) and (3) [41,47J. A case of elimination has

1921 K. W . Rosenmund, M . Nothnagel, and H. Riesenfeldt, Ber. dtsch. chem. Ges. 60, 392 (1927). 1931 K . Wallenfets and W. Draber, Tetrahedron Letters 13, 24 (1959). 1941 A . N. Nesmejanow, V . A . Ssasonowa, and V. N. Drosd, Chem. Ber. 93, 2717 (1960). [95l R . A . Lidgett, E. R . Lynch, and E. B. McCall, J. chem. SOC. (London) 1965, 3754. [96] G. R . Clemo and E. Walton, J. chem. SOC. (London) 1928, 723. [97] E . J . Sakellarios, Helv. chim. Acta 29, 1675 (1946). I981 D . A . Shirley and J . R . Zeitz, J . org. Chemistry 18, 1591 (1 953). [991 J. H. Helberger, G . Manecke, and R . Heyden, German Pat. 901 054 (Jan. 7, 1954), Bohme Fettchemie GmbH.

Angew. Chem. internat. Edit. f Vol. 7 (1968) / N o . 12 923

Page 6: The Gabriel Synthesis of Primary Amines

been noted with 2-cyanoethyl tosylate, when the main prod- uct of reaction is the volatile acrylonitrile[971, as with the corresponding 2-chloropropiononitrile under similar condi- tions [531. Moreover, although N-(2-tosyloxyethylsuIfonyl-p- pheny1)acetamide appears t o react by displacement of the tosylate function [691, it is possible that this is.an elimination- addition reaction.

(CH2)sOTs

Although the reaction is usually conducted with phthalimide itself, it is evident that a small amount of phthalimide ion is required to initiate the process; pyridine can also be used as a base[lll,1121. If potas- sium phthalimide is employed as the reagent, further reaction may occur. For example, epichlorohydrin (13) gives rise to the epoxy compound (14) [1131, which may react further to give N,N'-(2-hydroxytrimethy1ene)- phthalimide as a by-product I l l , 1093. An analogous nitrogen heterocycle is the aziridine derivative (1 5 ) , which on treatment with phthalimide and a catalytic amount of potassium phthalimide affords the sulfon- amide (16) 11141.

Other types of ester have been occasionally employed. For example, potassium phthalimide and ethyl carbonate react to give N-(2-hydroxyethyl)phthalimide [1001, as do y-lac- tones [I01 9 1021, e.g. y-butyrolactone affords 3-phthalimido- butyric acid [1011; trimethyl phosphate gives N-methyl-4- nitrophthalimide on treatment with potassium 4-nitrophthal- imide [1031. N-Benzylphthalimide is formed from benzyl nitrate and potassium phthalimide in dimethylformamide [36CI.

Mannich compounds and derived salts have also been em- ployed in metathetical transformations involving the phthal- imide ion. For instance, phthalimidomethyltrimethylam- monium iodide and potassium phthalimide give diphthal- imidomethane [1041. Base catalyzed reactions of phthalimide with ( 1 Z) and (12) give the corresponding phthalimido deriv- atives [I043 1051. Although o-nitro-@-(dimethy1amino)propio- phenone and @-(dimethylamino)propiophenone[J061 may formally react by a displacement with phthalimide ion [1*71, the likelihood of an elimination-addition reaction still remains.

In 1917, Gabriel and Ohle demonstrated that epoxides could be used for direct alkylation of phthalimide itself [1081. In general, the product of reaction is the N-(2-hydroxyalkyl)phthalimide (see Scheme 2).

0 0

R = H, CH3 (1081, CHzCl [108,109], CHzOCzH,, CHzOCcH, [ l l O ] , C H , O C G H ~ N O ~ - ( ~ ) [111]

Scheme 2.

[loo] K. Yanagi and S . Akiyoshi, J . org. Chemistry 24, 1122 (1959). [loll M . Zaoral, Chem. Listy 52, 2338 (1958). [lo21 J . Bornstein, S . F. Bedell, P. E. Drummond, and C. L. Kosloski, J. Amer. chem. SOC. 78, 83 (1956). [lo31 f. H . Billman and R. V. Cash, Proc. Indiana Acad. Sci. 63, 108 (1954). [lo41 R. 0. Atkinson, J. chem. SOC. (London) 1954, 1329. [lo51 F. Poppeldorf and S. J . Holt, J. chem. SOC. (London) 1954, 4094. I1061 A . Butenandt and U. Renner, Z. Naturforsch. 8b, 454 (1953). [lo71 A . Butenandt, U. Renner, H . Henecka, and H. Timmler, German Pat. 933339 (Sept. 22,1955), Farbenfabriken Bayer AG. 11081 S. Gabriel and H . Ohle, Ber. dtsch. chem. Ges. 50, 819 (1917).

0

0

(Certain quaternary ammonium salts are also effective alkylating agents in the Gabriel Synthesis; these re- actions are discussed in Section 2.1.4. r l l 5 l . )

2.1.4. N e i g h b o r i n g G r o u p P a r t i c i p a t i o n

In general, very few examples of genuine neighboring group effects in the Gabriel reaction have been in- vestigated. Many cases where, in the light of present day knowledge [1161, a participation would be expected to occur, are either qualitatively expressed as being fast, or no data concerning the relative rate of reaction are recorded. For example, compounds of the type

[lo91 E. Cherbuliez, B. Baehler, A. Yazgi, and f. Rabinowitr, Helv. chim. Acta 43, 1158 (1960). [I101 W. R. Boon and T. Leigh, J. chem. SOC. (London) 1951, 1497. 11111 V. Petrow and 0. Stephenson, J. Pharmacy Pharmacol. 5, 359 (1953). [112] H . Ohle and E. Euler, Ber. dtsch. chem. Ges. 69, 1022 (1936). [113] a) M . Weizmann and S. Malkowa, Bull. SOC. chim. France (4) 47, 356 (1930); b) E. 0. Titus, L . C . Craig, C . Golumbic, H . R . Mighton, I . M . Wempen, and R . C . Elderfield, J. org. Chemistry 13, 39 (1948). 11141 V. I . Morkow and S . I . Burmistrow, 2. obSL Chem. 35, 158 (1965). [115] M . S . Kharasch and C. F. Fuchs, J. org. Chemistry 9, 359 (1944). [I161 B. Capon, Quart. Rev. (chem. Soc. London) 18, 45 (1964).

924 Angew. Chem. internat. Edit. Vol. 7 (1968) J No. 12

Page 7: The Gabriel Synthesis of Primary Amines

X - C H 2 C H 2 - Hal [where X is (substituted) nitro- gen 11171, sulfur L1181, or possibly oxygen 12911, might be expected to react rapidly with phthalimide ion, as in fact do N-(2-bromoethyl)-N-methylaniline [I191 and bis-[2-~hloroethyl)amine ‘1201. The observed rapidity of reaction in these instances could well mean the in- volvement of intermediates of the general formula (17), and the simple SNZ mechanism may not apply.

H C6H, -CO-&CH- (NC8H4O2) -

H3C6 @H3

Halonium ions are also formally possible intermedi- ates in certain cases of a,w-dihalides. Participation may also arise with some substituted benzyl halides. Whereas the above examples must rest largely upon analogy, the following systems leave little doubt that participation is in operation.

Gabriel and Ohle [I211 re-examined the product obtain- ed from 2-chloro-1-propanol and potassiuni phthal- imide, and found it to be identical with that from 1- chloro-2-propanol. The resulting material was identi- fied as N-(2-hydroxypropyl)phthalimide, and it was concluded that the reaction proceeded via propylene oxide as an intermediate. Moreover, it was separately shown that phthalimide ion reacts with epoxides under similar conditions ‘1081. Even in the absence of further

0

c H3 - c HOH - CH 2 - Nm 0

information it is clear that reaction (a) involves neigh- boring group participation; this is probably true for reaction (b) also. However, not every reaction of a hydroxyalkyl halide with potassium phthalimide in- volves rearrangement even in cases where a cyclic ether intermediate may be formed; e.g. no rearrange- ment occurs in the case of l-iodo-3-pentanolr1221.

Reaction may also proceed via a three-membered ring in the case of 2,2’-diiododiethyl ether 1291. An interest- ing related example is provided by 2-bromo-3,3-di- methoxy-1 -phenyl-1 -propanone, which reacts readily with potassium phthalimide to give (19) “231. Here, the oxonium salt (18) could function as intermediate.

[117] L. P. Walls, J. chem. SOC. (London) 1934, 104. [118] A. E. Cushnzore and H . MrCombie, J. chem. SOC. (London) 1923, 2884. [119] J. v. Bruun, K . Heider, and E. Miiffer, Ber. dtsch. chem. Ges. 51, 273 (1918). [120] F. G. Munn, J. chem. SOC. (London) 1934, 461. [121] S. Gabriel and H. Ohle, Ber. dtsch. chem. Ges. 50, 804 (1917). I1221 R. C. Elderfield and C. Ressler, J. Amer. chem. S O C . 72, 4059 (1950). 11231 H . Shir-uhurnu and T. Mutsuoto, Bull. chem. SOC. Japan 38, 1293 (1965).

-

(19)

Philippi and Seka 11241 examined the reaction between potassium phthalimide and 2,3-dibromopropyl acet- ate (20), and suggested that 2 moles of potassium phthalimide were consumed, to give rise to the meta- thetical product, which on acid hydrolysis afforded the diamine. A parallel sequence with 2,3-dibromo- propanol provided the same compound. However, Fairburn and Cowdrey 11251 later examined the product from 1,3-dichloro-2-propanol, which they found to be identical with the compound examined by Phitippi and Seka. In the light of present day knowledge, it is clear that the product is not the 1 ,Ldiamine, but is the 1,3- analog, produced via at least one cyclic five-membered intermediate (cf. also ‘126,73a]).

CHzBr

Similar effects are found in nitrogen- and sulfur-con- taining compounds; for instance, N, N-diethyl-N-(4- chloropenty1)amine gives only a trace of the meta- thetical product. The major component of the reaction with potassium phthalimide was found to be N(4-di- ethylaminopenty1)phthalimide [*151. Here, the reaction was shown to proceed vfa the pyrrofidinium salt (21).

Again, benzyl-2-chloro-2-cyanoethyl sulfide gave the metathetical product as the minor component of reaction, whereas the major product was found to be N-(2-benzylthio-2-cyanoethyl)phthalimide [1271.

From the above examples of neighboring group

[124] E . Philippi and R. Seku, Liebigs Ann. Chem. 433, 88 (1923). f.1251 A. Fairburn and C. W. Cowdrey, J. chem. SOC. (London) 1929, 129. [I261 G. W. Kilmer and H . McKennis, J. biol. Chemistry 152, 103

[1271 K. D. Gundermann, Liebigs Ann. Chem. 588, 167 (1954). (1944).

Angew. Chem. internat. Edit. Vol. 7 (1968) 1 No. I 2 925

Page 8: The Gabriel Synthesis of Primary Amines

participation, it can be seen that where such an effect is liable to occur, due caution must be placed in inter- preting the course of the phthalimidation reaction.

2.2. Addition Reactions

In the presence of basic catalysts, phthalimide has been found to add readily to a variety of a$-unsatu- rated systems [54,128-1341. Such cr,P-unsaturated com-

X-CH=CHz + CsH402NH --f X-CH2CH2(NC8H402)

X = CN, CHO, C02CH3, COCH3 CsH402NH = phthalimide

pounds may be intermediates in reactions of halogeno- compounds with potassium phthalimide. This elimina- tion addition mechanism seems certain to apply in the reactions of phthalimide ion with y-bromo-P-methoxy- butyrophenone 11351, and P-bromo-P,P-dinitroethyl- acetate 11361 which give respectively 1-(2-oxo-2-phenyl- ethyl)-N,N’-ethylenediphthalimide) and N-(2,2-di- nitroethy1)phthalimide as the sodium salt. A novel, but slow addition of N-bromophthalimide to isolated double bonds (which competes with allylic abstraction) has recently been reported 11371. Thus, 3,3-dimethyl-1 -butene (CCl4, reflux I20 hours) gives a 1:l adduct [1371. However, it is reported that ethyl vinyl ether adds to N-bromophthalimide at 50°C in two hours [1383.

2.3. Condensation Reactions

Whereas secondary aliphatic amines and primary and secondary aromatic amines condense with formalde- hyde and phthalimide to give N-alkyl (or N-aryl) aminomethylphthalimides [1391, primary aliphatic amines give the N,N‘-(alky1amino)dimethylenephthal-

[128] S. R. Buc, J. Amer. chem. SOC. 69, 254 (1947). [129] V . M. Rodionow and N . G . Yartsewa, Bull. Acad. Sci. URSS, C1. Sci. chim. 1948, 251; Chem. Abstr. 42, 4942 (1948). [130] K. D . Gundermann and G. Holtmann, Chem. Ber. 91, 160 (1958). 11311 0. A. Moe and D. T . Warner, J. Amer. chem. SOC. 71,1251 (1 949). [132] a) H . Irai, S . Shima, and N . Murata, J. chem. SOC. Japan, ind. Chem. Sect. (Kogy6 Kagaku Zassi) 62,82 (1959); 1955,3151 ; A. M. Islam and R . A . Raphael, J. chem. SOC. (London) 1955,3151. 11331 E. N. Alekseeva and V . A. Vuver, NauC. Doklady vysiei Skoly Chim. i Chim. Technol. 3, 545 (1958); Chem. Abstr. 53, 2149 (1959). (1341 F. K. Kirchner, J . R . McCormick, C . J . Cavallito, and L. C . Miller, J. or@. Chemistry 14, 388 (1949). 11351 I . S. Kao, P. C . Pan, S. H. Lu, C. H. Chen, and H . Y. Hsii, Sci. sinica 7, 738 (1958). [136] M. B. Frankel, J. org. Chemistry 23, 813 (1958). [137] A. Guillemonat, G. Peiffer, J . C . Traynard, and A. Leger, Bull. SOC. chim. France 1964, 1192. [138] J. C . Traynard and G. Peiffer, C.R. hebd. Seances Acad. Sci. 258, 3735 (1964). [139] a) H . W . Heine and M . B. Winstead, J. Amer. chem. SOC. 77, 1913 (1955); H . W . Heine, M . B. Winstead, and R. P. Blair, ibid. 78, 672 (1956).

imide. It is somewhat surprising that primary and secondary alcohols react with phthalimide to give condensation products, with the elimination of wa- ter [1401. A typical example is provided by the reaction of 3-morpholino-1-propanol, which gives a good yield of the corresponding phthalimido derivative. The high pressure condensation of phthalimide with acetone in the presence of hydrogen and an A1203/Pt catalyst, to give N-isopropylphthalimide [1411 is pre- sumably an example of reductive alkylation, and as such could be a new route to this type of primary amine.

3. Stage 2: Formation of the Primary Amine

3.1. Hydrolytic Methods

3.1.1. Ac id ic Hydro lys i s

Hydrolysis of N-substituted phthalimides by hydro- chloric acid has been widely practised in the Gabriel Synthesis. At first, fuming or concentrated acid was used, often at temperatures as high as ZOO “C in sealed tubes (e.g. r4,6,7a,9b,221). Such extreme conditions are probably unnecessary in many cases, and more ac- ceptable conditions involve the use of concentrated or 20% acid at reflux temperature for several hours (e.g., [16,21,55d,68,83,113,115]). In Some cases, acetic acid is employed as co-solvent (e.g., [8, 17,29,58al). Typical examples of acid hydrolyses are provided by (2- mercaptoethy1)phthalimide r16.1421, N-acetonylphthal- imide 11431, N-(2-diethylaminoethy1)phthalimide [I 171, N,N’-iminodiethylenephthalimide [1201, a variety of phthalimidoalkyl heterocycles 182,1441, tetraethyloxydi- ethyl-2,2-diphthalimido-2,2,2’2’- tetracarboxylate 11451.

Similar hydrolytic procedures involve the use of hy- drobromic (e.g., [*4,15919,40,97, 1469, hydriodic (e.g., [26,47bJ), and sulfuric (e.g., [13b,14,231) acids. The choice of acid is perhaps significant in the hydrolysis of N-(halogenoalkyl)phthalimides, when it is prefer- able to use the halogeno acid corresponding to the halogen atom in the side chain, so as to avoid the possibility of exchange [113a, 147-1491. The conversion

[1401 K. Nakajima, J . chem. SOC. Japan, pure Chem. Sect. (Nippon Kagaku Zassi) 81, 1476 (1960). [141] L. Schmerling, US-Pat. 3274211 (Sept. 20,1966), Universal Oil Products Co.; Chem. Abstr. 66, 2229 (1967). 11421 S. Gabriel, Ber. dtsch. chem. Ges. 22, 1137 (1889). [143] S. Gabriel and G. Pinkus, Ber. dtsch. chem. Ges. 26, 2197 (1893); L. P. Ellinger and A . A . Goldberg, J. chem. SOC. (London) 1949, 263. [144] a) R. G. Jones and M . J. Mann, J. Amer. chem. SOC. 75, 4048 (1953); b) C . Ainsworth and R. G. Jones, ibid. 75,4915(1953). [145] H. Zahn, R. Dietrich, and W . Gerstner, Chem. Ber. 88, 1737 (1955). [1461 D. Y. Curtin and S. Schmukler, J. Amer. chem. SOC. 77, 1105 (1955). [147] M. Frankel, Ber. dtsch. chern. Ges. 30, 2497 (1897). 11481 A. Miiller and P. Kruus, Mh. Chem. 61, 219 (1932). [149] L . Ruzicka, G. Salomon, and K. E. Meyer, Helv. chim. Acta 17, 882 (1934); 20, 109 (1937).

926 Angew. Chem. internat. Edit. Vol. 7 (1968) /No. 12

Page 9: The Gabriel Synthesis of Primary Amines

of N-(4-bromobutyl)phthalimide to 4-iodobutylamine hydriodide by 55 % hydriodic acid illustrates such ex- change [1501. Sulfuric acid may also promote hydrolysis of the halogeno function in these cases [141.

A common complication is the hydrolysis of other functional

though y-lactone rings [1531, and a number of ether bondst1451 survive. Moreover, electrophilic displacement of aryl iodine during acidic hydrolysis occurred in a n unsuccessful attempt t o prepare p-iodobenzylamine by the Gabriel method [*611, and esterification of alcoholic functions may take place when concentrated halogen acids are used [73a, 108,1621. Finally, some phthalimides resist acidic hydrolysis, or a re decomposed under the reaction conditions (e .g . , [59C, 158% 1631). Whilst in such cases as amino-acid synthesis, the concomitant hy- drolysis of e.g. ester and phthaloyl groups is convenient. the usefulness of acidic hydrolysis is otherwise limited by the susceptibility of other groups to attack.

groups, 113,18-20,36b, 39, S l c , S4,69,77,124,15 I , 152,154-160]; al-

3.1.2. T w o S t a g e H y d r o l y s i s

In This method the imide is hydrolyzed under alkaline conditions to the phthalamic acid, and the latter under acidic conditions to the amine. The first stage occurs very readily 1641, and vigorous alkaline treatment is unnecessary unless other groups are to be hydrolyz- ed[51aI. Indeed, opening of the imide ring occurs to some extent under such mild conditions as are obtain- ed in Schotten-Baumann acylations 11651, and oxida- tions with [Fe(CN)&- f1601.

The second stage can usually be accomplished under milder acid conditions than those used for the sub- stituted phthalimide [17,1661, a feature relevant where some of the more acid-sensitive groups are concern- ed 1159a, 164a1. A major drawback is the loss of material

[150] W. S . Fones, R. S . Stander, and J . White, J. org. Chemistry 16, 708 (1951). 11511 D. P . Tschudy and A. C o l h s , J. org. Chemistry 24, 556 (1959). [152] A. 0. Jackson and C. S. Marvel, J . biol. Chemistry 103, 191 (1933). I1531 A. N. Dey, J. chem. SOC. (London) 1937, 1166. [1541 S. Gabriel and J . Colman, Ber. dtsch. chem. Ges. 41, 513 (1 908). [155] J . B. Cloke, E. Stehr, T. R. Steadman, and L . C. Westcott, J. Amer. chem. SOC. 67, 1587 (1945).

[156] M. Fields. D . E. Walz, and S. Rothchild, J. Amer. chem. SOC. 73, 1000 (1951); B. Heggediis, Helv. chim. Acta 38, 22 (1955).

11571 J . Michalskj and A . Podpr'rowa, Publ. Fac. Sci. Unw. Masaryk 395, 279 (1958); Chem. Abstr. 53, 11404 (1959). [1581 S. Gabriel and T. Posner, Ber. dtsch. chem. Ges. 27, 1037 (1894).

[1591 a) J . Lohmann, Ber. dtsch. chem. Ges. 24, 2631 (1891); b) S. Gabriel and T. A . Maass, ibid. 32, 1266 (1899). [1601 T. R . Seshadri, J. chem. Soc. (London) 1929, 2952. I1611 C. W. Shoppee, J. chem. SOC. (London) 1931, 1225. 11621 S. Gabriel, Ber. dtsch. chem. Ges. 22, 224 (1889). 11631 K . Matejka and R. Robinson, J. chem. SOC. (London) 1934, 1322. [164J a) T. Posner, Ber. dtsch. chem. Ges. 26, 1856 (1893); b) J . H . Speer and A . J . Hill, J. org. Chemistry 2, 139 (1937); C) F. E. Lehmann, A. Bretsher, H. Kiihne, E. Sorkin, M . Erne, and H . Erlenmeyer, Helv. chim. Acta 33, 1217 (1950); d) R. M . Perk, J. org. Chemistry 27, 2677 (1962). [1651 J. C. Sheehan and lf. S. Frank, J. Amer. chem. Soc. 71, 1856 (1949). [1661 W . Markwald, Ber. dtsch. chem. Ges. 37, 1038 (1904).

through re-cyclization to the imide in the second stage (e.g., [12,159a, 160,1671), though this can be recovered and re-cycled if necessary [1681.

3.1.3. A l k a l i n e Hydro lys i s

Complete alkaline hydrolysis was advocated by Pufochin 166,1691, particularly for the preparation of diaminoalkanes 11701 from or,w-diphthalimidoalkanes, and aminoalcohols [I711 from w-phthalimidoalkyl hal- ides (with the reported exception of N-(4-chlorobutyl)- phthalimide which gives pyrrolidine 11729; the method was independently used for the preparation of certain allylic amines 161,1731. More recent applications in- clude syntheses of neopentyl type amines '41,421. The method has not been widely used, and is unsuitable for alkali sensitive molecules; for example, a Smiles re- arrangement occurs during alkaline hydrolysis of N- [2- hydroxy-3-(p-nitrophenoxy)propyl]phthalimide [1111, and its acetater1741.

3.2. Aminolytic and Hydrazinolytic Methods

3.2.1. Amino1 ysis

There were definite indications in early work on the Gabriel reaction that alkyl phthalimides reacted with primary amines with ring opening. Thus Seitz had noted N-phenylphthalimide as a product of reaction of N-(2-bromopropyl)phthalimide with aniline 13Ial, and Risfcrzpart had obtained the corresponding phthal- imides from methylamine and ethylamine with N42- bromoethy1)phthalimide 11751. Alternative ring closure of such phthalamides will (a) regenerate the original phthalimide and (b) lead to the phthalimide derived from the amine reactant [1761. Such aminolytic reac- tions have led to complications in much of the syn- thetic antimalarial work based on reactions of N- (halogenoalky1)phthalimides with primary amines [I771.

Aminolysis of phthalimides either as the major [1781 or

11671 S. Gabriel, Ber. dtsch. chem. Ges. 42, 4050 (1909). [168] W. Schneider, Liebigs Ann. Chem. 375, 207 (1910). [169] N . Putochin, Ber. dtsch. chem. Ges. 59, 625 (1926); Trans. Inst. pure chem. Reagents 6 , 10 (1927); Chem. Abstr. 23, 2938 (1929). 11701 F. Chambret and D. Joly, Bull. Soc. chim. France (5) 14, 1023 (1947). [171] C. B. Kremer, J. Amer. chem. SOC. 61, 1321 (1939). 11721 W. Keil, Ber. dtsch. chem. Ges. 63, 1614 (1930). 11731 M . Olomucki, G. Desvages, N . Van-Thoai, and J. Roch, C.R. hebd. Seances Acad. Sci. 260, 4519 (1965). [174] W. T. Caldwell and G . C. Srhweiker, J. Amer. chem. SOC. 74, 5187 (1952). 11751 E. Risfenpart, Ber. dtsch. chem. Ges. 29, 2526 (1896). [176] a) F. S . Spring and J. C. Woods, J. chem. SOC. (London) 1945, 625; Nature (London) 158, 754 (1946); b) H. J . Barber and W. R. Wragg, ibid. 158, 514 (1946); J. chem. SOC. (London) 1947, 1331; personal communication. [177] H . S . Mosher, J. Amer. chem. SOC. 68, 1565 (1946); and references cited. [178] 1. K . Kormendy, Acta chim. Acad. S C I . hung. 17, 255 (1958); N . K . Dalta and P . B. Talukdar, J. Indian chem. SOC. 1966, 461.

Angew. Chem. internat. Edit. 1 Vol. 7 (1968) /No. I 2 927

Page 10: The Gabriel Synthesis of Primary Amines

as the minor reaction11791 has been noted in other instances. Thalidomide (K-phthalimidoglutarimide) also apparently reacts with certain diamines, notably putrescine and spermidine, very readily [1801.

3.2.2. H y d r a z i n o l y s i s

Hydrazinolysis, introduced by Ing and Manske 125,1811

in 1926, has largely but not entirely superseded hy- drolytic methods for accomplishing the second stage of the Gabriel Synthesis. In the original method, the substituted phthalimide was treated with hydrazine in hot ethanolic solution, when an uncharacterized (and incorrectly formulated) intermediate was readily produced. This was decomposed by heating with hydrochloric acid, when phthalhydrazide was precip- itated and the amine hydrochloride usually remained in solution. The comparatively mild conditions of re- action and ease of separation of products presented obvious attractions and constituted an important advance, but it was not until nearly 20 years later that such technical innovations as conducting the hydra- zinolysis 11x41 or acidic decomposition 1701 stages at room temperature were employed in one or two cases, and that the intermediate was recognized as the amine salt of phthalhydrazide (22) ‘176bl.

Manske applied the technique to the Curtius and related reactions 11821. Thus, substituted urethanes and ureas react with phthalic anhydride to give the corresponding phthal- imides which can then be hydrazinolyzed; isocyanates can like- wise be converted in two steps into the primary amine[I831.

Inherent merits in the method were apparent in cases of difficult hydrolysis, for example W(3-p-acetamido- phenoxypropy1)phthalimide ‘1631, and in avoidance of side reactions r159a, 1851. In 1946 Mosherr1771 queried the necessity of boiling the hydrazinolysis intermediate with hydrochloric acid prior to isolation of the amine, and he illustrated an alternative procedure in which the intermediate was rendered alkaline and the expect- ed amine isolated directly by solvent extraction. In- dependently, Barber and Wragg 1176bJ had recognized

~ - [179] J. 0. Jiiek and M . Protiva, Collect. czechoslov. chem. Commun. 15, 659 (1950); S. Winter and H. Pracejus, Chem. Ber. 99, 151 (1966). [180] S. Farbo, R. L . Smith, and R . T. Williarns, Nature (London) 208, 1208 (1965). [181] R. F. H . Manske, W . H. Perkin, and R. Robinson, J. chem. Sac. (London) 1927, 1. 11821 R. F. H . Manske, 3 . Amer. chem. Sac. 51, 1202 (1929) and references cited; C. A . Grob, H. Kny, and A . Gagneux, Helv. chim. Acta 40, 130 (1957). [183] H . R. Snyder and E. L. Eliel, J. Amer. chem. SOC. 71, 663 (1949); T. Shono, T. Morikawa, R. Okayama, and R. Oda, Makro- molekulare Chem. 81, 142 (1965). [184] R. Adams and T. L. Cairns, J. Amer. chem. Sac. 61, 2464 (1939). [la51 D. C. Quin and Sir R. Robinson, J. chem. SOC. (London) 1943, 555.

the hydrazinolysis intermediate in these reactions as (22); formation of the amine or its hydrochloride, depending on whether (22) was treated with potas- sium hydroxide or hydrochloric acid, was thus im- mediately intelligible. The use of an excess of hydrazine and the separation of the expected amine as the free base has been used occasionally [144bl, and specifically recommended for some peptides where it would be otherwise difficult to remove residual hydrazine hydrochloride from the peptide hydrochloride 11561. However, the presence of residual hydrazine in some cases may pose problems of purification r187J, requiring the use of e.g. salicyl- aldehyde for its removal [35,73aI.

The conditions available for hydrazinolysis are suf- ficiently mild to allow the preservation of many func- tional groups (e.g. r56,73e, 189,191,1929, including most of those which are sensitive to hydrolysis. Thus, K,W-

aminoalkanenitriles are available through hydrazinol- ysis 11931. Carboxylic ester functions generally survive r49,73b, 1901, though there are indications that this group is somewhat sensitivell941, as indeed are xan- thates 11951.

Hydrazinolysis has frequently been practised i n boiling ethanol, and under such conditions, reaction is likely to be fast. Nevertheless, reported reaction times vary considerably, and some quoted periods are! probably excessive. Intermediate temperatures have also been employed 11881, and occasionally necessitated by the sensitivity of another par t of the molecule to attack, as for example with certain 2-aminothiazole deriv- atives 11891. Finally, hydrazinolysis has been accomplished for a considerable number of compounds, including peptides (which may be hydrazinolyzed in aqueous solution [19ol), a t room temperature r184,1911.

Moreover, the hydrazinolysis procedure cannot generally be applied to molecuies containing (a) halogenoalkyl[1961, or (b) aldehyde or ketonic groups, although the latter may often be protected by conversioninto acetals and thioacetals [lo22 1971, or ketals and thioketals [132b, 1981. For molecules containing

[186] J . C. Sheelinn, D . W . Chapman, and R. W . Roth, J. Amer. chem. SOC. 74, 3822 (1952). [187] J. Crum and Sir R. Robinson, J. chem. SOC. (London) 1943, 561. [188] a) A. J. Speziale and P. C. Hamm, J. Amer. chem. SOC. 78, 2556 (1956); b) M. Masaki and M. Ohta, Bull. chern. SOC. Japan 35, 1808 (1962). [189] R. Winterbottom, J . W. Clapp, W . H. Miller, J. P. English, and R. 0. Roblin, J. Amer. chem. SOC. 69, 1393 (1947). [190] a) F. E. King and D. A . A. Kidd, J. chem. Sac. (London) 1949, 3315; b) F. E. King, J. W . Clark-Lewis, D. A. A . Kidd, and G. R. Smith, ibid. 1954, 1039. [191] H . Bickel, B. Fechtig, G. E. Hall, W . Keller-Sehierlein. V. Prelog, and E. Vischer, Helv. chim. Acta 43, 901 (1960). [192] H. Sclimidand P. Karrer, Helv. chim. Acta 31, 1497(1948). [1931 A. A. Goldberg and W . Kelly, J. chem. Sac. (London) 1947, 1369; A . F. McKay, D. L. Gnrmaise, R. Gaudry, H. A . Baker, G. Y. Paris, R. W . Kay, G. E. Just, and R. Schwartr, J. Amer. chem. SOC. 81, 4328 (1959). [194] a) R. B. Barlow, J. chem. SOC. (London) 1951, 2225; b) E. E. van Tamelen and E . E. Smissman, J. Amer. chem. SOC. 75, 2031 (1953). [195] Th. Wieland and H. Hornig, Liebigs Ann. Chem. 600, 12 (1956). [196] J . G . N . Drewitt and D. P. Young, Brit. Pat. 581 153 (1946), Chem. Abstr. 41, 2069 (1947). [197] K . BalenoviC, N. Bregant, D. Cerar, D. Fl i i , and I . Jam- breiit, J. org. Chemistry 18, 297 (1953). [198] K. BalenoviC and N. Bregant, J. org. Chemistry 17, 1328 (1952).

.-

928 Angew. Chem. internat. Edit. / Vol. 7 (1968) 1 No. 12

Page 11: The Gabriel Synthesis of Primary Amines

acid sensitive groups (e.g. acetal), isolation of the amine by basification may be necessary for the group to be preserv- ed[lozl. In this connection, use of a n excess of hydrazine to permit direct isolation of amines may be used[l4lbl. If the expected amine is base-sensitive, a carefully controlled hydra- zinolysis followed by acidic work-up may provide the only possibility of isolation of the hydrochloride 11111 (for examples ~ee[ lh5 , IY91).

3.2.3. P he n y 1 h y d r a z i n o 1 y s i s

Phenylhydrazinolysis was introduced for removing phthaloyl groups in one stage, particularly from pep- tide derivatives 12001, so avoiding the final acid treat- ment of the Ing-Manske hydrazinolysis procedure. The technique has not been widely used and has been reported as unsuccessful when applied to cis-2-butenyl- phthal im ide I N - (5 - sulfamoyl - 2 - thieny1)- phthalimide r2011; the last compound is reported to suffer decomposition on attempted hydrazinolysis.

and

3.3. Neighboring Group Participation

Neighboring group participation effects are observable in the alkaline hydrolysis of N-(2-bromoethyl)- and N-(3-bromopropyl)phthalimide, in which the N - nitroso compounds (24a) and (246) resp. are form- ed 12031. Gabriel postulated the formation of inter- mediate products having 8- and 9-membered rings 12021

respectively. These reactions have recently been re- investigated, and the intermediate products recognized as the oxazolene (23a) and 1,3-oxazine derivative (236) respectively, presumably in the zwitterionic form. Similar considerations probably apply for example, in the reaction of N-(2-bromoethyl)phthalimide with

(a), n = 1; (b), n = 2

[199] J. C. Sheehan and J . J . Ryan, J. Amer. chem. S O C . 73, 1204 (1951). [200] I . Schun~ann and R. A . Borssonnas, Helv. chlm. Acta 35, 2235, 2237 (1952). [201] J . Cymerman-Craig and D. Willis, J. chem. Soc. (London) 1955, 1071. [202] S. Gabriel, Ber. dtsch. chem. Ges. 38, 2389, 2405 (1905); b) P. M . Bnrtholdy, ibd. 40, 4400 (1907).

potassium pyrrole with subsequent hydrazinolysis, which leads not to N-(2-aminoethyl)pyrrole 1204' but to pyrrole itself and a product that may be equated with that subsequently obtained by hydrazinolysis of (23u) 1203aj. This apparently anomalous result is readily explained by assuming that pyrrole anion, like hydrox- ide ion, reacts with the phthalimide to give an inter- mediate product, which then reacts with hydrazine in the expected way fcf. also IzOSJ).

Artefacts are frequently obtained in hydrolysis and hydrazinolysis reactions where the generated primary amino group can interact with other functional groups in the molecule. Examples are provided by the formation of pyrrolene derivatives in (a) the acidic hy- drolysis of N-(4-oxo-4-phenylbutyl)phthalimider1~41

and (b) the hydrazinolysis of methyl 3-phthalimido- crotonate 12061; formation of analogous six-membered rings has also been noted 12071. The involvement of cyclic transition states may also lead to rearranged products. Thus the alkaline hydrolysis of N-[Z-hy- droxy-3-(p-nitrophenoxy)propyl]phthalimide involves an 0->N aryl migrationl111.1741; 0 - N acetyl migra- tion has been noted during hydrazinolysis and work- up of 3-methoxy-l - ( p - nitrophenyl)-2-phthalimido- propyl acetatelzoxl. An example of N-tN migration is afforded by the reaction (25) -t(26) 1164dl.

4. Conclusions

Finally a few limitations of the Gabriel Synthesis should be noted. For example, sym-dichloroacetone is reported to react with potassium phthalimide to give phthal- imide itself 12091. Difficulties have also been noted with certain pentaerythritol derivatives such as (27), which, with potassium phthalimide also produces phthalimide and probably an oxetane 173aI. Further (28) gives an unidentified compound rather than the metathetical product [ZIOJ.

(YCH&C(CH2X)z (27): X=Br, Y=OH (28): X=Y=Br, I

~ ..__

[203] a) K . Korrnendy and J . Voljord, Acta chim. Acad. Sci. hung. 32, 115 (1962); b) S. Hiiriig and L. Geldern, Chem. Ber. 96, 3105 (1963); c) I . T. Bornish, R. W. Brad.sslraw, M . S. Gibson, and G. W. Prenton, unpublished. [204] 0. Klamnierrh, Chem. Ber. 84, 254 (1951). [205] a) S . Gabriel and J . Coinzrm, Ber. dtsch. chem. Ges. 45, 1643 (1912); b) W. Myliirs, ihid. 49, 1091 (1916). [206] W. Lrrrigenbecli and K. €loser, Chern. Ber. 84, 526 (1951). [207] S. Gabriel, Ber. dtsch. chem. Ges. 42, 1238 (1909). [ZOS] M . C. Rebstock, J . org. Chemistry 19, 851 (1954). [209] T. Posner and K . Rolrde, Ber. dtsch. chem. Ges. 42, 3233 (1909). [210] J. A. Larrrberlon, Austral. J. Chem. I,?, 106 (1959).

Angew. Chem. internnt. Edit. / Vol. 7 (1968j J No. I 2 929

Page 12: The Gabriel Synthesis of Primary Amines

Most of the reported failures in the second stage of the synthesis are largely concerned with the inherent in- stability of either the generated amine, or some other sensitive part of the molecule. Examples of the former category are encountered in systems of the type (29) - (31), which do not survive hydrolysis or hydrazinol- ysis, and usually decompose inter alia to formaldehyde and ammonia. The latter group includes some examples of N-thienylphthalimides, in which the heterocyclic ring decomposes, and gives rise to hydrogen sulfide 12011.

X(CHz-NC8H402)n (29) X = RS-; n = 1[211] (30) X = N; n = 312121 (31) X = NO2; n = 112131

Despite these minor limitations, the Gabriel Synthesis will doubtless continue to be exploited for the prepara- tion of primary amines, but more complete understand- ing of the reactions involved must await further mechanistic studies.

The authors wish to thank Dr. E. M . Howells for read- ing the manuscript.

Received: February 16, 1968 [A 670 IEI German version: Angew. Chem. 80, 986 (1968)

12111 W. Schneider, Liebigs Ann. Chem. 386, 332 (1912). [212] F. B. Kipping and F. G. Munn, J. chem. SOC. (London) 1927, 528. [213] L. W. Kissinger and H . E. Ungnade, J . org. Chemistry 23, 815 (1958).

Kinetic Studies on Substitution Reactions of Carbonylmetal Complexes

BY H. WERNER[*]

Kinetic studies of substitutiort reactions of carbonylmetals and of carbonylmetal derivatives in which other ligands (n or n donors) besides CO are bound to the metal atom have recently attracted great interest, and the results of these investigations throw a new light on theoretical and preparative problems. The substituction mechanism is determined mainly by steric and electronic factors.

1. Introduction

It is known from the classic work of Hieber and his school that metal carbonyls undergo substitution re- actions of type (1) when thermal or photochemical

M(C0)X + yL + M(C0)zLy + (x-z)CO (1 )

energy is supplied. The ligand L may be varied within wide limits, and may be either an n donor (e.g. amine, phosphine, arsine, stibine, isocyanide, thioether) or a 7t donor (e.g. olefin, aromatic compound). Carbonyl- metal derivatives in which one or more other groups X (e.g. NO, H, halogen, alkyl) are bound to the metal atom besides CO also react readily with donors L. The course of the substitution is in many cases deter- mined by the nature of the non-CO ligands, their donor and acceptor properties being particularly im- portant. General observations of this type were until recently based almost exclusively on the results of preparative studiesur. Kinetic studies on the mechanism of sub-

[*] Prof. Dr. H. Werner Anorganisch-Chemisches Institut der Universitat CH-8001 Zurich, Ramistrasse 76 (Switzerland)

[ l ] H. Werner, Habilitationsschrift, Technische Hochschule Munchen, 1966. [2] For a review, see T. A. Manuel, Advances organometallic Chem. 3, 181 (1965).

stitutions of carbonylmetai compounds have only recently been initiated, and are now being carried out in rapidly growing numbers. The first study in this field was reported in 1959 and was concerned with the kinetics of the reactions of carbonylphos- phinenickel complexes which a re important in industry [31. This report referred to the use of spectrophotometric methods for the determination of the concentration of carbonylmetals, and so provided a decisive stimulus t o further development.

Photochemical reactions will not be discussed in the present article, since a review of this topic has recently been published [41.

2. Possible Reaction Courses

There are two possible courses for the reaction [eq. (l)] of a carbonylmetal M(CO), with a donor L ( y =

1, z = x-1).

(a) The first step is the cleavage of an M-CO bond to form an intermediate in which the coordination num- ber has been reduced by I , which then reacts with L.

[3] L. S. Meriwether and M . L . Fiene, J. Amer. chem. SOC. 81, 4200 (1959). [4] W. Strohmeier, Angew. Chem. 76,873 (1964); Angew. Chem. internat. Edit. 3,730 (1964); see also E.O. Fischer and M . Herber- hold, Experientia, Suppl. 1964, 259.

930 Angew. Chem. internat. Edit. 1 Vol. 7 (1968) / N o . 12