I. Electrophilic Reactions of p-Toluenesulfonyl Azide. II. ¹⁵N and ¹³C Nuclear Magnetic Studies of Aryldiazonium Compounds Effect of Substituent, Solvent and 18-Crown-6 Citation Casewit, Carla Jutta (1981) I. Electrophilic Reactions of p-Toluenesulfonyl Azide. II. ¹⁵N and ¹³C Nuclear Magnetic Studies of Aryldiazonium Compounds Effect of Substituent, Solvent and 18-Crown-6. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/nyft-yr23. https://resolver.caltech.edu/CaltechETD:etd-01112005-143434 Abstract Part I Electrophilic Reactions of p -Toluenesulfonyl Azide Section 1. Review of Electrophilic Reactions of p -Toluenesulfonyl Azide The electrophilic reactions of p -toluenesulfonyl azide are reviewed using the principles of hard and soft acids and bases (HSAB). Section 2. The Reaction of p -Toluenesulfonyl Azide with the Sodium Salt of p -Toluenesulfonamide A number of concurrent reactions of p -toluenesulfonyl azide-3- 15 N (I-3- 15 N) with the sodium salt of p -toluenesulfonamide were followed by 15 N NMR. I-2- 15 N is formed as a result of a degenerate diazo transfer by I-3- 15 N to p -toluenesulfonamide anion. p -Toluenesulfonamide anion also reacts with 1-3- 15 N to give di- p -toluenesulfonamide and azide ion. The 15 -N-labeled azide ion exchanges with I to give I-1- 15 N. I also reacts with azide ion, yielding dinitrogen and p -toluenesulfinate anion. The sulfinate salt reacts readily and reversibly with I to give 1,3-di- p -toluenesulfontriazene anion, which provides another pathway for interconversion of 1-3- 15 N and I-1- 15 N. Section 3. The Reaction of p -Toluenesulfonyl Azide with Potassium Azide The reaction of p -toluenesulfonyl azide with potassium azide-1- 15 N has been examined in toluene and dichloromethane by 15 N NMR. In addition to azide-ion exchange leading to 1-1- 15 N and I-3- 15 N, the formation of 1-2- 15 N is indicated. Two mechanisms for this novel scrambling are proposed. Azide-ion metathesis involving reversible formation of an N -Pentazole derivative from I and azide ion, followed by azide exchange could account for the formation of 1-2- 15 N. Alternatively, a scrambling route involving the reversible addition of p -toluenesulfonylnitrene to 1-3- 15 N can be envisioned. The inhibition of scrambling in dichloromethane by addition of iodide ion suggests that a discrete p -toluenesulfonyl azide – azide ion intermediate is involved in any case. Part II 15 N and 13 C Nuclear Magnetic Resonance Studies of Aryldiazonium Compounds. Effect of Substituent, Solvent and 18-Crown-6. 15 N and 13 C shifts induced by addition of one equivalent of 18-crown-6 have been determined for several para-substituted aryldiazonium fluoborates in dimethylformamide. The α-nitrogen (N1) and para carbon (C4) shift upfield and the β-nitrogen (N2) and C1 shift downfield on complexation. The effect of solvent on the positions of the 15 N and 13 C resonances of p -( n -butyl)benzenediazonium fluoborate are small. The influence of substituents on the 15 N chemical shifts is relatively large and comparable to the effect of 18-crown-6. These 13 C and 15 N results, in conjunction with previous spectroscopic studies indicate that the crown ether complexed aryldiazonium cation is electronically more diazonium-like and less diazo-like than the uncomplexed form. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Dervan, Peter B. Thesis Committee: Dervan, Peter B. (chair) Roberts, John D. Goddard, William A., III Grubbs, Robert H. Defense Date: 1 November 1980 Funders: Funding Agency Grant Number IBM UNSPECIFIED Caltech UNSPECIFIED Record Number: CaltechETD:etd-01112005-143434 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-01112005-143434 DOI: 10.7907/nyft-yr23 Related URLs: URL URL Type Description https://resolver.caltech.edu/CaltechAUTHORS:20151006-073240179 Related Item Article in CaltechAUTHORS Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 122 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 14 Jan 2005 Last Modified: 23 Jan 2025 21:26 Thesis Files Preview PDF - Final Version See Usage Policy. 3MB Repository Staff Only: item control page

I. II. ELECTROPHILIC REACTIONS OF .E_-TOLUENESULFONYL AZIDE 15 N AND 13c NUCLEAR MAGNETIC STUDIES OF ARYLDIAZONIUM COMPOUNDS EFFECT OF SUBSTITUENT~ SOLVENT AND 18-CROWN-6 Thesis by Carla Jutta Casewit In Partial Fulfillment of the Requirements for the Degree of .Doctor of Philosophy California Institute of Technology Pasadena, California 1981 (Submitted November 1, 1980) ;; To Tante Renate iii ACKNOWLEDGMENTS I thank John D. Roberts for his guidance, encouragement and most of all his imagination and honesty. I thank the members of the Roberts group, past and present, for the many long discussions and excursions. I am especially grateful to com- rades in arms Mike Nee, Nancy Becker, and Nina Gonnella. I am indebted to Amy Abe for her generous help in getting started. I am grateful to David Horowitz for the two years of unmatched wit and humor. I thank Bill Goddard for giving me the opportunity to "go where my scientific curiousity drove me", and the members of his group for their warm and lively companionship. I owe much to the faculty of the University of Colorado at Denver. I am especially grateful to Bob and Lennie Damrauer for their encouragement and friendship then and now. IBM and Caltech Fellowships are gratefully acknowledged. Most of all I thank my family for their support and interest, and Tony for his insight, encouragement and love. iv ABSTRACT PART I ELECTROPHILIC REACTIONS OF .E_-TOLUENESULFONYL AZIDE Section 1. Review of Electrophilic Reactions of £.-Toluenesulfonyl Azide The electrophilic reactions of .E_-toluenesulfonyl azide are reviewed using the principle of hard and soft acids and bases (HSAB). Section 2. The Reaction of .e_-Toluenesulfonyl Azide with the Sodium Salt of .E_-Toluenesulfonamide A number of concurrent reactions of _g_-toluenesulfonyl azide-3l 5N (I-3! 5N) with the sodium salt of .e_-toluenesulfonamide were followed by 15 N NMR. I-2- 15 N is fonned as a result of a degenerate diazo transfer by I-3- 15N to .e_-toluenesulfonamide anion. £_-Toluenesulfonamide anion also reacts with I-3- 15 N to give di-.e_-toluenesulfonamide and azide 15 . ion. The N-labeled azide ion exchanges with I to give I-1- 15N. I also reacts with azide ion, yielding dinitrogen and £.-toluenesulfin ate anion. The sulfinate salt reacts readily and reversibly with I to give 1 ,3-di-.e_-toluenesulfontriazene anion, which provides another pathway for interconversion of I-3- 15 N and 1-1- 15 N. Section 3. The Reaction of £_-Toluenesulfonyl Azide with Potassium Azide The reaction of .e.-toluenesulfonyl azide with potassium azide-1- 15 N has been examined in toluene and dichloromethane by 15 N NMR. In addition v 15 15 to azide-ion exchange leading to I-1- N and I-3- N, the fo·rmation of I-2- 15 N is indicated. Two mechanisms for this novel scrambling are proposed. Azide-ion metathesis involving reversible formation of an !!_-pentazole derivative from I and azide ion, followed by azide exchange 15 could account for the formation of I-2- N. Alternatively, a scrambling route involving the reversible addition of E_-toluenesulfonylnitrene to I-3- 15 N can be envisioned. The inhibition of scrambling in dichloromethane by addition of iodide ion suggests that a discrete £toluenesulfonyl azide - azide ion intermediate is involved in any case. vi ABSTRACT PART II 15 N and 13 c NUCLEAR MAGNETIC RESONANCE STUDIES OF ARYLDIAZONIUM COMPOUNDS. AND 18-CROWN-6. EFFECT OF SUBSTITUENT, SOLVENT 15 N and 13 c shifts induced by addition of one equivalent of 18crown-6 have been determined for several para-substituted aryldiazonium fluoborates in dimethylformamide. The a-nitrogen (Nl) and para carbon (C4) shift upfield and the 6-nitrogen (N2) and Cl shift downfield on complexation. The effect of solvent on the positions of the 15 N and 13c resonances of E_-(!!_-butyl)benzenediazonium fluoborate are small. The infl,uence of substituents on the 15 N chemical shifts is relatively large and comparable to the effect of 18-crown-6, These 13 c and 15 N results, in conjunction with previous spectroscopic studies indicate that the crown ether complexed aryldiazonium cation is electronically more diazonium-like and less diazo-like than the uncomplexed form. vii TABLE OF CONTENTS PART I ELECTROPHILIC REACTIONS OF Q-TOLUENESULFONYL AZIDE Section 1. Review of Electrophilic Reactions of Q-Toluenesulfonyl Azide INTRODUCT10N THE ELECTRONIC STRUCTURE OF Q-TOLUENESULFONYL AZIDE Page 3 4 NUCLEOPHILIC ATTACK AT THE TERMINAL NITROGEN OF .E_-TOLUENESULFONYL AZIDE 6 Azido Transfers 6 Diazo Transfers 8 Loss of Molecular Nitrogen 11 NUCLEOPHILIC ATTACK AT THE SULFONYL SULFUR OF .E_-TOLUENESULFONYL AZIDE The Mechanism of Nuc~eophilic Substitution 12 13 CONCLUSION 14 REFERENCES AND NOTES 16 viii Section 2. The Reaction of £,:-Toluenesu1fonyl Azide with the Sodium Salt of £_-Toluenesulfonamide Page INTRODUCTION 21 RESULTS AND DISCUSSION 21 Preparation of £_-Toluenesulfonyl Azide3~15N 21 The Reaction of ,E_-Toluenesulfonyl Azide-3- 15 N with Azide Anion 22 Formation of £_-Toluenesulfonyl Azide-2- 15 N 23 Related Reactions and Tetrazene Intermediates 23 Consequences of £_-Toluenesulfonyl Azide-3- 15N Scrambling 29 Formation of di-£_-To1uenesulfonamide 30 Formation of 15 N-labeled £_-Toluenesulfonamide 32 Exchange of Azide Ion 32 Azido Transfer to Azide Ion 33 Formation of l,2-di(£_-Toluenesulfonyl) Triazenyl Anion 35 Triazenyl Anions and Reaction of Azides with Sulfinate Salts 37 Formation and Reaction of £.-Toluenesulfonyl Sulfiny1 Anhydride 40 ix Cleavage of 1,3-di(.Q_~Toluenesulfonyl) Triazenyl Anion in Acid 42 CONCLUSION 44 EXPERIMENTAL 45 REFERENCES AND NOTES 47 x Section 3. The Reaction of Q.-Toluenesulfonyl Azide with Potassium Azide Page INTRODUCTION 53 RESULTS AND DISCUSSION 54 The Reaction of Q.-Toluenesulfonyl Azide with Potassium Azide-1- 15N in Dimethyl Sulfoxide The Reaction of Q-Toluenesulfonyl Azide with Potassium Azide-1- 15 N in Toluene 54 56 The Reaction of Q-Toluenesulfonyl Azide with Potassium Azide-1- 15 N in Dichloromethane 59 15 The Reaction of Potassium Azide-1- N with Dichloromethane 61 The Reaction of Q-To·luenesulfinate Ion with Chloromethyl Azide in Dichloromethane 63 The Reaction of Q-Toluenesulfonyl Azide with Potassium Azide-1-15 N in the Presence of Potassium Iodide 66 Mechanism of the Formation of Q_~Toluenesulfonyl Azide-2- 15 N: Azide-Ion Metathesis 68 Mechanism of the Formation of Q-Toluenesulfonyl Azide-2- 15 N: Q.-Toluenesulfonylnitrene 72 xi CONCLUSION 74 EXPERIMENTAL SECTION 75 REFERENCES AND NOTES 76 xii PART II 15 N AND 13 c NUCLEAR MAGNETIC RESONANCE STUDIES OF ARYLDIAZONIUM COMPOUNDS. EFFECT OF SUBSTITUENT, SOLVENT AND 18-CROWN-6 Page INTRODUCTION 82 RESULTS 84 Sol vent Sh if ts 84 Benzonitrile Solvent Shifts 86 Counterion Shifts 88 Substituent Effects 90 Crown Shifts 95 Crown Complexation of Related Systems 98 DISCUSSION 100 EXPERIMENTAL SECTION 100 REFERENCES AND NOTES 103 1 PART I ELECTROPHILIC REACTIONS OF £_-TOLUENESULFONYL AZIDE 2 SECTION 1 REVIEW OF ELECTROPHILIC REACTIONS OF Q-TOLUENESULFONYL AZIDE. 3 INTRODUCTION Q_-Toluenesulfonyl azide (I) is a highly versatile and useful reagent 1 , which, depending on the conditions can act as an electrophile, nucleophile 2 , 1,3 dipole 3 or source of Q_-toluenesulfonylnitrene 4 H~ _F\_~ + - -~~-N=N=N 0 I The chemistry of I with nucleophiles is particularly interesting because I is an ambident electrophile which can react both at the sulfur In the tenninology of the theory of hard and soft acids and bases. (HSAB) 5 , the terminal nitrogen of I is a soft and at the terminal nitrogen, acid and the sulfonyl center is hard. This leads to a duality of reaction paths$ where the nature of the nucleophile determines which site is preferentially attacked. Soft nucleophiles tend to react at the tenninal nitrogen of I to form a triazenyl anion (II) and hard nucleophiles at the sulfonyl to displace azide ion (scheme 1) 4 Scheme l H~ -0- + ?.1-N=N=N 0 I H3 C -0- ?i ~-8 0 n THE ELECTRONIC STRUCTURE OF I No one resonance structure adequately accounts for the diverse chemistry of the azido group. Such groups are best described as resonance hybrids of several contributing structures, IIIa-IIId: + R/ - N=N=N ma .. .. - + ,.......N-N=N R mb .. .. - R/ + N-N=N me .. .. . . ,.......N-N=N R IDd 5 The J!-diazonium resonance structures, Illb and Ille are useful in illustrating the electrophilic and nucleophilic character of the terminal and a-nitrogens, respectively, The relative contribution of a particular structure to the hybrid is expected to be greatly affected by the nature of R. Indeed, there are examples of azides with electron-withdrawing R groups which should be regarded as es- sentially the equivalent of 1!_-diazonium compounds 6 , I should have at least some N-diazonium character because of the ability of the sulfonyl group to delocalize the negative charge on the a-nitrogen. One would expect that along with an increased contribution of resonance structure !Vb, structure IVc would become less important. .. H 5C-o--~~N-N: ., .. 0 ISz:a Illb Dre Evidence for the decreased importance of IVc is provided by the 13 c NMR of I and related compounds. The 13 c NMR of the carbon para to the S0 2N group of I shows a 6 ppm upfield shift relative to the para 3 carbon of p_-toluenesulfonamide or p_-toluenesulfonyl chloride 7 Because para 13 c chemical shifts are dominated by mesomeric effects 8 the 13c results indicate that the so2N3 group increases the electron density of the para carbon relative to the S0 2NH 2 or so 2c1 group. 6 NUCLEOPHILIC ATTACK AT THE TERMINAL NITROGEN OF I The triazenyl anions formed by the nucleopholic attack of soft bases on the terminal nitrogen of I have been isolated in some cases. Triazenyl anions have been obtained from the reaction of I with 1-norbornyl anion and several other 10 9 11 . , triphenyl phosphine , aryl grignard reagents bases 12 . These adducts may decompose either by the loss of .e_-toluenesulfinate (Eq. 1), loss of .e_-toluenesulfonamide anion (Eq. 2) or by extrusion of molecular nitrogen (Eq, 3). TsN 3 + s- .. Ts-N-N=N-B -E Ts- + N3 B ( 1) Ts NH + {2 ) Ts-N=B The nature of the bqse determines the decomposition route. N28 + N2 ( 3) Each path· way is discussed separately and in more detail below, Azido-trans fers When the anionic nucleophile does not possess a labile a-hydrogen, decomposition of the triazenyl anion yields a new azide and the excellent leaving group .e_-toluenesulfinate anion. The azido-trans fer 7 reaction is il1ustrated for the reaction of 1-norbornyl anion with I ( Eq. 4). I + 9 ~ cb Li+ N II A ~ + (4) Tsli N3 N I N- u+ I Ts Use of I as an azido-transfer reagent is a relatively new development. In most synthetic procedures, the triazene salt is not isolated, but · t t o g1· ve the az1· de 13 . · ~ · t ea d genera t e d I!!_ ms I has been found by Anselme and coworkers l 4 ,l 5 to also transfer an azido group to nitrogen anions with no a-hydrogen (Eqs, 5-6). Ph 2 C== N- + TsN 3 The detailed mechanism of these novel reactions is unknown, (5) Anselme suggests that the initially formed triazenyl anion fragments to give an 8 N-azide. Subsequent loss of N2 from the N-azide accounts for the ob- served products. Azido transfer by I to an ambident nucleophile has been reported by Sakai and Anselme 16 The products of azido transfer to 2,4,5triphenyl imidazole anion (V) can only be explained if C-azidation is favored over N-azidation almost 2:1: Ph'!===-( Ph eN y N Ph Va Vb C-azidation was in fact unexpected~ because such a reaction would disrupt the aromaticity of the heterocycle. However, this observation can be readily interpreted in tenns of HSAB concepts. The products are a consequence of the preference for the soft terminal nitrogen of I to attack the softer carbon site rather than the harder nitrogen site of v. Diazo-transfers When I reacts with an anionic nucleophile possessing a labile a-hydrogen, azido transfer does not occur. is thought to tautomerize. Instead, the triazenyl anion The resulting triazene then decomposes into a diazo compound and the stabilized £_-toluenesulfonamide anion, These 9 steps are illustrated for the reaction of I with cyclopentpdienyl anion {Eq. 7). (7) Si nee the pi vota 1 work by Doering and DePuy 17 i. n which I was employed to synthesize diazocyclopentadiene, I has been used to transfer diazo groups to a wide variety of active methylene groups in the presence . base 18 of an organic Indeed, use of I as a diazo-transfer reagent has stimulated the most synthetic interest, and considerable effort has been directed at improving isolated yields of diazo-transfers. Variations of the reagent and procedure include modification of the aryl nucleus of I l 9 , 2ob, phase-transfer catalysis 20 and immobilization of I on a polymer support 21 . Diazo group transfer has been extended to the anions of primary 22 . The diazo transfer amines 19 •22 , hydrazines 23 •24 and hydrazones by I to amine anions has been developed into a useful synthesis of ali22 phatic and aromatic azides 19 • (Eq. 8). 10 H ( }-~- + - + N=N=N-Ts {8) ~N) ~ H H ~-Ts + I -N-Ts 'N=N0 Anselme and Fischer 22 have reported the isolation of an adduct from the reaction of the chloromagnesium salt of aniline with I~ which decomposed on heating to give phenyl azide. This yellowish chloromagnesium salt is thought to be VI., A novel reaction of I with sodium hydride (which acts as a soft base) has been reported by Lee and Closson 25 . This prototype diazo transfer yields Q_-toluene sulfonamide anion and molecular nitrogen: ( Eq. 9) + Ts-N=r~=N \ - Ts-NH- + N2 + H- Ts-N-N=N / H H I r- / Ts -N~N=N (9) 11 Loss of Molecular Nitrogen l also torms triazene complexes with soft neutral bases. These triazenes decompose on heating to give a nitrene and molecular nitrogen. The Staudinger reaction is a classic example 11 •26 . The phosphazide complex formed by reaction of I with triphenylphosphine loses molecular nitrogen to give a phosphine imine (Eq. 10): + Ts-N-N=N-PR 3 ( 10) A similar process is thought to occur in the activation of benzenesulfonyl azide by metallic copper. Kwart and Kahn 27 have investigated the copper-catalyzed decomposition of benzenesulfonyl azide and propose the formation of an arylsulfonyl azide-copper complex (VII): ___ ., [ Cu=N-S0 Ph] 2 + N 2 ( 11) lZlI There are many examples of activation of I by low-valent transition-metal complexes to give Q_-toluenesulfonylnitrene derivatives. The generated sulfonylnitrene is either trapped by co-ordinated ligands 28 on the metal or by added reagents 29 . These reactions may involve triazene-like intermediates, or more likely intermediates that bridge 12 the gap between Lewis acid-base complexes 30 and 1,3-dipolar addition complexes. NUCLEOPHILIC ATTACK AT THE SULFONYL SULFUR OF I Nucleophilic attack occurs at the sulfonyl sulfur of I when the base is hard or borderline. Hard oxygen anions such as hydroxide, alkoxide, phenoxide and carboxylate 19 are known to displace azide ions f ram I ( Eq. 12) . + + ( 12) There have been no reports of the hard oxy bases attacking the soft terminal nitrogen of I. The nitrogen bases studied by Anselme and coworkers 13 - 16 ~ 22 • 23 are somewhat softer than the oxygen bases discussed above and so nucleophilic attack occurs at both the terminal nitrogen and sulfonyl sulfur of I. For example, in the reaction of benzylamine anion with I; ( Eq. 13) ~ PhCH 2N3 + TsNH- //,/. ~~ ( 13) 13 both benzyl azide and !!_-benzyltosylamide were isolated 22 The soft carbanions used in azido and diazo transfer procedures do not appear to attack the sulfonyl group. The Mechanism of Nucleophilic Substitution The mechanism of the nucleophilic substitution at the sulfonyl center of I is unknown. However, other sulfonyl systems have been studied extensively, and the substitution mechanism is still a matter of considerable controversy. Depending on the system and solvent, dis- sociative, associative-concerted and associative-stepwise mechanisms have all been proposed. The dissociative or SNl mechanism (Eq. 14) is thought to be followed in the hydrolysis of electron rich arylsulfonyl chlorides ... . 31 ( 14) Most substitutions at sulfonyl sulfur occur by an addition-elimination process, but there is disagreement on the details of the mechanism, The substitution may occur synchronously (Eq. 15) or in two steps through a discrete penta-coordinate intennediate (Eq. 16). 14 l r ( 15) CJ,,,,O · X--!:-Y 0 II Ar-S-X II + y- 0 II Ar -S- Y + x- II 0 - 0 ,a Q, -Y [ x-Y'~ ( 16) Ar For most sulfonyl systems the existing evidence favors the concerted sub33 , however. Most recently, stitution (15) 32 . There are exceptions Engberts and coworkers 33 a provided evidence for an associative stepwise mechanism (Eq. 16) operating in the neighboring carbonyl-group catalyzed hydrolysis of arylsulfonamide derivatives. Unfortunately, a penta- coordinate sulfonyl addition intermediate has yet to be isolated by anyone. Because of the marked dependence of the substitution mechanism on the system and conditions, no attempt will be made to predict the mechanism of azide ion displacement from I. CONCLUSION The concepts of the hard and soft acid and base theory can be 15 used successfully to understand and predict the reactions of .I with nucleophiles. Whereas soft nucleophiles preferentially attack the terminal nitrogen, hard nucleophiles tend to attack the sulfonyl sulfur to displace azide ion. both centers. Borderline nucleophiles appear to react at 16 REFERENCES AND NOTES 1. For a review of azides, see S., Patai, (ed.). 11 The Chemistry of the Azido Group"; Interscience: New York~ N.Y. 1971. 2. For example, see, R. Kreher and G. Jager, Z. Naturforsch. B. Anorg. Chem. Org. Chem. 31B, 126 (1976). 3. G. L'Abbe, Chem. Rev. 69, 345 (1969). 4. D. S. Breslow in "Nitrenes", W. Lwowski, ed.; Interscience: New York, N.Y. 1970, Chapter 8. 5. (a) R. G. Pearson, J. Songstad, J. Am . Chem. Soc. 89, 1827 ( 1967). (b) For an excellent discussion of the application of the HSAB principle to organic chemistry, see: T.-L Ho. "Hard and Soft Acids and Bases Principle in Organic Chemistry"; Academic: New York, N.Y. 1977. 6. Most notably, the heterocyclic azidium salts investigated by Balli and coworkers, and the triflyl azide studied by Shiner and coworkers. H. Balli, R. Low, V. MUller and H. Rempfler, Sezen-Gezgin,Helv. Chim. Acta. §1_, 97 (1978) and references therein., 7. C. J. Cavender and V. J, Shiner~ J .. Org, Chem, 37, 3567 (1972). The decoupled 13 c spectra of I, p-toluenesulfonamide and p-toluenesulfonyl chloride were taken in acetonitrile solvent. The para carbon chemica1 shifts were 135,9, 141.2 and 141,9 ppm downfield of TMS, respectively. 8. J. Bromilow, R. T. C. Brownlee, V. 0. Lopez and R. W. Taft, J. Org. Chem. 44, 4766 (1979). 17 9. J. 0. Reed and W. Lwowski, J. Org, Chem. 36, 2864 (1971), 10. P.A.S. Smith, C, D. Rowe and L, B. Bruner, J. Org. Chem. 34, 3430 ( 1969). 11. H. Bock and W. Wiegrabe, Angew. Chem, 75~ 789 (1963), 12. (a) P. Spagnolo and P. Zanirato, J. Org. Chem, 43, 3539 (1978). (b) R. Gompper, J. Schelble and C. S. Schneider, Tetrahedron Lett. 3897 (1978). (c) 13. K. KUhlein, H. Jensen Justus Liebigs Ann. Chem. 369 (1974). S. J. Weininger, S. Kohen, S. Mataka, G. Koga and J.-P. Anselme, J. Org. Chem. 39, 1591 (1974). 14. J.-P. Anselme, W. Fischer and N. Koga, Tetrahedron~' 89 (1969). 15. M. Nakajima and J.-P. Anselme, Tetrahedron Lett. 4421 (1976). 16. K. Sakai and J.-P. Anselme, Bull. Chem. Soc. Jpn. 45, 306 (1972). 17. W. van E. Doering and C. H. DePuy, J. Am. Chem. Soc. I§_, 5955 (1953). 18. For a review, see, M. Regitz, Synthesis, 351 (1972). 19, J. B. Hendrickson and W. A. Wolf, J. Org. Chem. 33, 3610 (1968). 20. (a) H. Ledon, Synthesis, 347 (1974). (b) L. Lombardo and L. N. Mander, Synthesis, 368 (1980). 21. W. R. Roush, D. Feitler and J. Rebek, Tetrahedron Lett., 1391 (1974). 22. J.-P. Anselme and W, Fischer, Tetrahedron 25, 855 (1969). 23. G. Koga and J.-P. Anselme, J. Org. Chem. 35, 960 (1970). 24. B. Stanovnik, M. Tisler, M. Kunaver, D. Gabrijelcic and M. Kocevar, Tetrahedron Lett. 3059 {1978). 18 25. Y,-J. Lee and W, D, Closson, Tetrahedron Lett, 381 (1974). 26, M. P. Ponomarchuk, L. F. Kasukhin, I, Y. Budilova and Y. G. Goblobov~ 27. (a) J. Gen, Chem. USSR (Engl. Transl.) 48, 1379 (1978). H. Kwart and A. A, Kahn, J. Am. Chem, Soc. 89, 1950 (1967). 28. (b) H, Kwart and A. A. Kahn, J. Am. Chem. Soc. 89, 1951 (1967). (a} S. Cenini, M. Pizzotti, F. Porta and G. La Monica, J. Organomet. Chem. 88, 237 (1975). (b) W. Beck, W.Rieber, S. Cenini, F. Porta and G. La Monica, J. Chem. Soc. Dalton Trans. 298 (197 4). (c) R. A. Abramovitch, G. N, Knaus and R. W. Stowe, J. Org. Chem. 39, 2513 (1974}. 29. (a) I. Yamamoto, H, Tokanou, H.-A, Uemura and H. Gotoh, J. Chem. Soc. Perkin Trans. 1, 1241 {_1977}. (b) 30. P. Svoronos and V. Horak, Synthesis, 596 (1979). Bonding of Q_-toluenesulfonylazide to the transition-metal should be similar to the bonding in the known doubly bent aryldiazonium cation complexes of transition metals recently reviewed by Sutton. D. Sutton, Chem. Soc. Rev . .1_, 443 (1975}. 31. R. V. Vizgert and E, K. Savchuk, J. Gen. Chem. USSR (Engl. Transl.) 34, 3437 (1964). 32. (a) S. Thea, M. G, Harun and A. Williams, J, Chem, Soc. Chem. Commun. 717 (1979). (b) B. Giese and K. Heuck, Chem. Ber. ll!_, 1384 (1978). 19 33. (c) R. Rogne, J. Chem, Soc. Perkin Trans. 2, 1486 (1975), (d) A. R. Haughton, J. Chem. Soc, Perkin Trans. 2,637 (1975). (a) T. Graafland, A. ~Jagenaar, A. J. Kirby and J.B. F. N. Engberts, J. Am, Chem. Soc. 101, 6981 (_1979). (b) E. Ciuffarin, L. Senatore and M. Isola, J. Chem. Soc. Perkin Trans. 2, 468 (1972). 20 SECTION 2 THE REACTION OF Q.-TOLUENESULFONYL AZIDE WITH THE SODIUM SALT OF .2_-TOLUENESULFONAMIDE 21 INTRODUCTION To increase understanding of the reactions of £_-toluenesulfonyl azide (I) with nucleophiles, the reaction of I with the sodium salt of £_-toluenesulfonamide in dimethyl sulfoxide was examined. 15 The reaction of I ( l abe 1ed with N on the te~i na 1 nitrogen), and the sodium salt of £_-toluenesulfonamide (generated ..:!..!!_situ from excess E_-toluenesulfonamide) was observed by 15 N NMR. Strong evidence was obtained by 15 N NMR for each of the reactions shown by Eqs. 1-3. Ts-N=N-JS N + Ts-NH Ts-N= 15 N=N +Ts-NH ( 1a) ~ Ts-NH-Ts + N=N= 15 N- (lb) ~ ,,/ Ts-N=N= 15 N + N=N= 15 N- , Ts-+ 315N ~ ·~Ts- 15 N=N=N Ts-N=N= 15 N + Ts - r--- (2a) 2 + N=N= 15 N- [ Ts·-~:-: N:::15 _N-Ts J- (2b) (3) RESULTS AND DISCUSSION Preparation of I-3- 15 N I-3- 15 N was prepared by diazotization of E_-toluenesulfonyl hydrazide with 30.2% enriched sodium nitrite, following the method of Curtius l 22 The report by Clusius and Weisser 2 that a similar preparation of phenyl azide-3- 15 N resulted in the formation of 2-7% of phenyl azide2-15N made it necessary to detennine if the labeling of I was completely specific. A 15 N spectrum of I-3- 15 N in dimethyl sulfoxide indicated that within the limits of detection (~0.5%) only N3 was 15 N-labeled. It was found that I-3- 15 N maintains its isotopic integrity in dimethyl sulfoxide for at least 34 h. However, on longer standing in dimethyl sulfoxide, isotope scrambling of I occurs. Approximately 20% I-1- 15 N was observed by 15 N NMR after two weeks. Presumably a slow reversible ionization (perhaps solvent assisted) of I to £_-toluenesulfonyl cation and azide ion takes place (4). Ts-N=N=15N + Ts-15 N=N=N (4) Control experiments showed that unionized £_-toluenesulfonamide does not react with I. The 15 N NMR spectrum of p-toluenesulfonamide and I-3- 15 N in dimethyl sulfoxide showed that p-toluenesulfonamide does not scramble the 15 N label in £_-toluenesulfonyl azide, even after 3 days. The Reaction of I-3- 15 N with p-Toluenesulfonamide Anion The 15 N spectra of I-3- 15 N in the presence of the sodium salt of p-toluenesulfonamide, in dimethyl sulfoxide, show that within two hours I is completely scrambled and, in addition, many new 15 N-labeled 23 products are fonned (Fig. 1). The 15 N signal assignments and chemical shifts are collected in Table I. Formation of I-2- 15 N Work by Anselme and coworkers 3 on the diazo transfers by I to the magnesium salts of primary amines led us to expect that I would react with p-toluenesulfonamide anion in a similar manner. The diazo transfer by I to £_-toluenesulfonamide ion is a degenerate reaction, and cannot be detected unless labeling techniques are employed. The diazo transfer by I-3- 15 N to p-toluenesulfonamide anion, would be expected to result in the formation of I-2- 15 N, through the intermediate tetrazenyl anion (II). - 111i+ + N•N•N-Ts I \ - (5) H I Ts-N N-Ts 16 'N= N.,.... n D Fig. l clearly shows a signal due to I-2- 15 N at 142.0 ppm. The I-3- 15 N equilibrates with I-2- 15 N within a few hours indicating that the degenerate diazo transfer is fast 4 Related Reactions and Tetrazene Intermediates The reaction of Eq. 5 is related to the reactions of aryldiazonium 24 e Ts-N=N=N-Ts. c: c 0 Ts.~T.s 100 300 200 PP"" rel. to MNO, Figure 1. 15 N spectra of 5.0xl0- 3 mol sodium salt of Q_-toluenesulfona3 mide, 5.0xl0- 3 mol p-toluenesulfonamide and 8.lxl0- mol 15 15 Q_-toluenesulfonyl azide-3- N (I-3- N) in 20 ml of dry dimethyl sulfoxide; 20.Jlsec pulse angle, 10-sec repetition rate. Spectrum of sample (a) 10 min after preparation, 410 transients; (b) 5 h after preparation, 1015 transients; (c) 78 h after preparation, 1654 transients. 25 TABLE I. 15 N Signals in the Reaction of I-3- 15 N with p-Tolue'nesulfonamide Anion 15 N Chemical Shift a Assignment 24. 1 Nl of 1,3-di(g_-tolue nesulfonyl) triazenyl anion 64. 1 di nitrogen 125.8 N2 of azide ion 132. 1 N3 of I 142.0 N2 of I 211 .6 di-g_-toluenesulfonamide anion 234.2 Nl of I 271. 5 Nl of azide ion 279.0 g_-toluenesulfonamide and its anion b a) ppm relative to l!i external HN0 3 b) averaged because of fast proton exchange 26 salts with amines, but because diazonium salts are more nuclepphilic 5 than I, the amine need not be ionized in order to react . The diazo migration 6 (Eq. 6) ArN + + H NAr 2 2 1 + ArNH 2 + N2Ar 1 H+/ ~+ ArN 2NHAr' (6) ArNHN 2Ar' and the coupling of an aryldiazonium salt with an aryl hydrazine 7 (7) are examples of such reactions. + Ar-NH 2 + N=N=N-Ar (7) Ar-N=N-N-N-Ar I l HH Ar-N-N=N-N-Ar I H I H In contrast to the 1 ,4-di (E._-toluenesulfonyl) tetrazenyl anion, II, 1 ,4-di(aryl)tetrazene intermediates appear to be relatively stable. 8 Recently Evanochko and Shevlin succeeded in isolating and characterizing 1 ,4-di(~_-nitrophenyl)tetrazene (Ill) generated from the reaction of E._-nitrophenylhydrazine with £_-nitrobenzenediazonium tetrafl uoroborate. 27 HH 0 N-·~l=N-~-~_/~NO 2 ~I ~/ 2 I II In a somewhat similar reaction, the oxidation of Q_-toluenesulfonylhydrazide is suggested to involve an intermediate 1,4-di(Q_-toluenesulfonyl)tetrazene 9 (8) Ts-N=N-H Ts-NH-NH 2 ----'I (8) Ts-NH-N=N-NH-Ts Ts-NH-NH-Ts La Monica and coworkers lO have isolated 1,4-bis(Q_-toluenesulfonyl)tetrazene complexes (IV) from the reaction of Rh(NO}(PPh ) 3 3 with I. The tetrazene derivatives react with HCl to give p-toluene- sulfonyl azide and p-toluenesulfonamide (91 Ts 1 L, ,..N'N M ~ 'N,........N HCl (9) II I Ts IV The authors suggest that IV is in equilibrium with an open form (Vl which reacts with HCl, lO 28 Ts \ L, iJ=N:::N M:-.. L_/ "N I Ts v However, it is not necessary to invoke reaction of the open form to explain Equation 9, because this reaction follows naturally from Equation 5. Indeed, it is quite possible that the tetrazenyl anion (II) of Equation 5 is also an intermediate in Equation 9. Interestingly, McDonald and coworkers lla have recently investigated the gas-phase reaction of phenylnitrene anion radical with phenyl azide by flowing afterglow. Anions are known to react at the nucleophili<: y-nitrogen of phenyl azide llb and nitrenes give products consistent with attack on the electrophilic a-nitrogen llc Because the anion radical is ambiphilic, both sites of phenyl azide should be attacked. Indeed, products corresponding to addition at both they and a-nitrogen are observed (Eq. 10-11). Ph-N-N=N-N-Ph PhN~ + N Ph 3 ~ ~ Ph-N=N-Ph' (10) ( 11 ) 29 Reaction 10 is quite analogous to the formation of III by the reaction of I with £_-toluenesulfonarnide anion. Consequences of I-3- 15 N Scrambling 15 The formation of I-2- N by the reaction of I-3- 15 N with£_toluenesulfonarnide anion may impose limitations on the utility of this reagent to specifically 15 N label diazo compounds. Recently, I-3- 15 N was used to synthesize labeled diazocyclopentadiene, and, contrary to expectations, the 15 N NMR spectrum of the product showed that, in addition to VI about 5% of VII was formed 12 111N- N- HOHI. ~ II II 1sN+ Et2 NH + 15- Ts-N•N= N N+ 0 0 + JZI + TsNHz (12) EI Because diethylamine catalyzes the ionization of cyclopentadiene (p~a~l5) 13 in the first step of the diazo transfer 14 •15 , it must 16 certainly be able to convert E_-toluenesulfonamide (pKa 10.3) to some extent to its conjugate base. The £_-toluenesulfanamide anion so generated will scramble I-3- 15 N by reaction 5. The reaction of the resulting I-2- 15 N with cyclopentadienyl anion then accounts for the small amount of VII formed. Although I-2- 15 N is likely to be formed under the diazo-transfer 12, this may not be the on1y pathway 1eading to VII. observation by Farnum and Yates 17 that a-diazoacetophenone transfers conditions of Eq. The 30 a diazo group to the sodium salt of methyl phenylacetate suggests that a degenerate diazo transfer by diazocyclopentadiene to cyc1opentdienyl anion could occur (13). ~N=N + . © © \ / J + (13) The bimolecular scrambling of I by £_-toluenesulfonamide anion (Eq. 5) can be readily eliminated by employing a polymer support carrying sulfonylazido groups. This 15 N-labeled sulfonyl azide polymer could be prepared by diazotization of the poly(styrenesu1fony1hydrazine) reported by Emerson and coworkers 18 + 15 ~\ \.:_)- S0 N=N= N 2 ( 14) Using this sulfonyl azide polymer as a diazo-transfer reagent may provide a delicate probe for the importance of reaction 13. Formation of di-p-1oluenesulfonamide As discussed in Section l, the hard and soft acids and bases 19 is very useful in understanding the reactions of (HSAB) principle I with nucleophiles. The borderline base £_-toluenesulfonamide anion reacts with both the soft terminal nitrogen and the hard su1fony1 31 sulfur of I. TsN 3 + TsNH TsNHTs + NJ (15) Anselme 3 has noted a similar dichotomy in the reaction of magnesium salts of primary amines with I. The azide ion generated in reaction 15 is visible in the 15 N spectra of Fig. l at 125.8 and 271.5 ppm 20 . Control experiments in dimethyl sulfoxide show that Eq. 15 is not appreciably reversible. It is suspected that strongly acidic 21 di-.E_-toluenesulfonamide protonates the azide ion and thereby inhibits the reverse reaction. In the reaction mixture of Fig. l, however, the strongest base is p_-toluenesulfonamide anion, and the following proton exchange occurs. 16 has been verified by 15 N NMR. TsNHTs + TsNH The reaction of Eq. TsNTs + TsNH 2 (16) Eqs. 15 and 16 can account for the gradual transformation of the singlet at 279 ppm (Fig. la) to the triplet at 279 ppm 22 {Fig. le). The singlet in Fig. la arises from p_-toluenesulfonamide in equilibrium with a small amount of p_-toluenesulfonamide anion (only one peak is observed because of fast proton exchange). In Fig. le ( 78 h after sample preparation) all of the anion is consumed and only p_-toluenesulfonamide remains so that a triplet results. The resonance of the anion of the strongest acid, di-B_-toluenesulfonamide, is now visible 32 at 211.6 ppm, and this anion is not basic enough to exchange rapid1y with .e._-toluenesulfonamide. Formation of 15 N-labeled p-Toluenesulfonamide The resonances of .e._-toluenesulfonamide and its anion in Fig. l are observed only because these materials have become enriched with 15 N. No 15 N signal was observed with a natural-abundance sample of .e._-toluenesulfonamide in dimethyl sulfoxide at a concentration and with an acquisition time comparable to the spectrum in Fig. le. Formation of 15 N-labeled .e._-toluenesulfonamide and its anion is not consistent with the equilibrium of Eq. 5, although the same sequence of reactions with I-1- 15 N would lead to labeled .e._-toluenesu1fonamide (Eq. 17). Ts-"N=N=N + TsNW I \ The two pathways leading to the fonnation of I-l- ( 17) 15 N will be discussed in the following paragraphs. Exchange of Azide Ion The first route for I-1- 15 N fonnation is associated with the 15 direct nucleophilic attack of azide ion on I. A N spectrum of sodium azide and I-3- 15 N in dimethyl sulfoxide showed signals at 132.l 33 ppm (I-3- 15 N), 234.6 ppm (I-l- 15 N) and 271.5 ppm (sodium azide-1- 15 N), indicating fast azide exchange (Eq. 18). + + 15 - Ts-N=N= N + N=N=N-~ Ts-N=N=N + N=N= Ts- 15 15 N- - - - + - ( 18) N=N=N + N=N=N The reaction of Eq. 18 must occur by an addition-elimination mechanism rather than by dissociation-recombina tion because the 1 ,3-scrambling is greatly accelerated by azide ion. Azido Transfer to Azide Ion In Fig. 1 two small resonances appear at 64. l and 24,l ppm. These 15 signals are also observed in solutions of I-3- N and sodium azide in 15 dimethyl sulfoxide. The resonance at 64. l ppm is attributed to N15 l abeled dinitrogen. This assignment is based on the reported N shift 23 . In addition, this 64.1 of dinitrogen of 66.5±1 ppm in benzene ppm signal disappears on purging solutions of .e_-toluenesulfonamide anion and I-3- 15 N with dry unlabeled nitrogen and then slowly reappears over several hours. We propose that dinitrogen is evolved in a novel reaction in 24 •25 (Eq. which I formally transfers an azido group to azide ion 19) (19) 34 Though reaction 19 was unexpected, HSAB theory predicts that the borderline base, azide ion, should react both at the sulfonyl sulfur and the tenninal nitrogen of I. Several mechanisms for this slow transfonnation may be envisioned. The simplest involves a straightforward nucleophilic coupling of I and azide ion to give the unknown free pseudohalogen (N ) 26 (Eq. 20). 3 2 Ts-N=N=N + N=N=N---+ Ts - + [ N=N=N-N=N=N] (20) The azide dimer 27 could decompose directly to 3N or else cyclize to 2 28 hexazine (VIII), before decomposing (21}. (21) -J + + [ N=N-N-N=N-N Alternatively, intermediates similar to those proposed in the coupling of azide ion with aryldiazonium salts could be involved 29 + - Ts-N=N=N + N=N=N- [ Ts-N-N==N-N-N=N I \ IX J (22) 35 Distinction between the mechanisms of Eqs. 21 and 22 by 15 N labeling is possible if the fonnation of IX were reversible (Eq. 23). I + I!! N=N= N - ... .,.,.1-- ... However, the 15 N spectrum of potassium azide-1- 15 N and I in dimethyl sulfoxide after 8 h showed no scrambling of the potassium azide. This result therefore excludes pentazole reversibility but not irreversible formation of IX. The reaction of I with azide ion in other solvents was investigated in order to probe the mechanism of this unique reaction. The results of this study are reported in Section 3 of this thesis. Formation of 1,2-di (p-Toluenesulfonyl) Triazenyl Anion The second route which interconverts I-3- 15 N to I-1- 15 N involves £_-toluenesulfinate anion, formed in the reaction of Eq. 19 by way of a degenerate azido transfer with I through the intermediate triazenyl anion X (Eq. 24). The 15 N NMR of I-3- 15 N in the presence of small concentrations of the sodium salt of p-toluenesulfinic acid (dihydrate) in dimethyl sulfoxide 36 shows that the reversible complexation of the soft terminal nitrogen of I by the soft sulfur of the sulfinate anion occurs readily. Both I-3- 15 N and I-l- 15 N are visible in about equimolar amounts within 4 h. Thus, reaction 24 as well as reaction 18 allow for the formation of I-1- 15 N from I-3- 15 N. If an excess of the sulfinate salt is added to I-3- 15 N in dimethyl sulfoxide, a concentration builds up of the triazene salt. The 15 N NMR of multilabeled I with an excess of the sulfinate salt (as the dihydrate) in dimethyl sulfoxide shows that no I remains, and only 15 N signals attributed to Nl (24.2 ppm) and N2 (-160.3 ppm) of X are observed. 15 N chemical shifts have been reported for triazenes XI 30a XII 30b, and XIII 3oc-d (Table II). The nitrogens of triazenyl anion are substantially deshielded relative 37 to the triazenes listed i.n Table II. The paramagnetic shift of X is probably primarily a consequence of increased N-lone pair delocalization 31 because of ionization. The 15 N spectrum of a concentrated solution of multilabeled I and sodium azide in dimethyl sulfoxide after the gas evolution has stopped is shown in Figure 2. The resonances of X are visible, the azide ion becomes 15 N-labeled in accord with Eq. 18, and, in agreement with Eq. 19, the resonances of a small amount of labeled dinitrogen is observed, Triazenyl Anions and Reaction of Azides with Sulfinate Salts 32 b The formation of diary1 32 a, dia1ky1 33 and su1fony1triazeny1 anions by deprotonation of the parent triazene is well documented. In addition, triazenyl anions can be fanned by the reaction of a 34 . As noted carbanion with an alkyl llb, aryl llb or sulfonyl azide in Section 1, sulfonyltriazene salts decompose readily into the sul34 finate salt and azide (Scheme 1). Scheme 1 ? The reverse reaction, addition of sulfinate salts to organic azides, 38 TABLE II. 15 N NMR Chemical Shifts of Triazenes a Compound oNl oN2 oN3 reference XI +78.2 ~118.3 -56.4 30a XII +220.8 - 76.0 +16.0 30b XIII +200 + 9 30c XIII b + 96.5 - 65.6 +96.5 30d x + 24.2 -160.3 +24.2 a} ppm relative to HN0 3 b) Nl and N3 are averaged because of fast proton exchange 39 e N=N=f'il Ts-N=~=N-Ts -100 -200 0 100 200 JOO ppm rel. to MNOJ Figure 2. -2 N spectrum of l.2xl0 mol scrambled Q_-toluenesulfonyl . . az1de az1.d e (l - 15 N, 2- 15 N, 3- 15 N} an d 1.2xl 0-3 mol sodium 15 in 25 ml dry dimethyl sulfoxide. The spectrum was taken 53h after sample preparation using 200 µsec pulse angle, lO~sec repetition rate, 1949 transients, 40 It should~ however, be quite has not been systematically i.nvestigp.ted~ The first reported example appears to be the reaction of carbamoyl azides with arylsulfinate salts 35 a (Eq. 25). general. 0n ArS0 2NHCNHR (25) Whether the sulfonylureas are derived from the decomposition of a sulfonyltriazenyl intermediate or from the reaction of a nitrene is not known. The azidinium salts studied by Balli 25 have also been found to form complexes with sulfinate salts: (26) Considering the reaction of azidinium salts with arylsulfinate salts (Eq. 26), it is not surprising to find a counterpart to reaction . t ry 35 b (27) . c hem1s . ary ld"iazon1um . . 24 in 0 " ArN + 2 + HOSAr 0 Ar-N=N-~-Ar 0 (27) Formation and Reaction of p-Toluenesulfonyl Sulfinyl Anhydride An 15 N spectrum of a mixture of multilabelled I in dimethyl sulfoxide treated with an equimolar amount of the sodium salt of p- 41 toluenesulfinate (dihydrate) shows two signals at 128 and 250 ppm after one week. A resonance at about 250 ppm is also visible in the 15 N spectrum of a solution of I-3~ 15 N treated in the same manner. The signals at 128 and 250 ppm are attributed to 15 N label at the N2 and Nl of a mixture of hydrazoic acid and azide ion 36 The source of the azide ion is very likely the slow reaction of the sodium salt of p~toluenesulfinate at the sulfonyl sulfur of I. In this case, because the sulfonyl of I is a hard acid, the harder oxygen of the ambident anion is the attacking center 37 (Eq. 28). -_F\_?. ~ _F\_ H5C~-s-o-~ ~-CH3 (28) + N,- 0 A precedent for reaction 28 is found in the studies by Corson and Pews 38 of the reaction of the sodium salt of p-toluenesulfinate with ptoluenesulfonyl chloride. The initial product, which was not isolated but instead reacted further under their reaction conditions, was thought to be the sulfonyl sulfinyl anhydride (XIV). Under our experimental conditions, the sulfonyl sulfinyl anhydride would be expected to be hydrolyzed readily by the ambient water in solution (Eq. 29). (29) 42 The E_-to1uenesulfinic and sulfonic acids should protonate the azide ion generated in reaction 28 to some extent, in agreement with the observed 15 N shifts. Reaction 28 is far slower than formation of the triazenyl anion (X). However, because formation of Xis reversible and reaction 28 is not (under hydrolysis conditions), formation of hydrazoic acid and azide ion should result. When the sulfinate salt is generated in situ, as in reaction of I with azide ion, the equilibrium of reaction 28 should lie far to the left because of the large excess of azide ion, In addition, the absence of water in solution precludes reaction 26 from driving reaction 28 to the right. As expected, Fig. 2 shows no evidence for the formation of hydrazoic acid. Cleavage of 1 ,3-di(p-Toluenesulfonyl} Triazenyl Anion in Acid Both diaryl and l-arylsulfonyl-3-aryltriazene are cleaved by acids to an aryldiazonium salt and an amine 32 {_Eqs. 30-31). H+ + ArNH N Ar ArNHN Ar 2 22 H+ ArS0 NHN Ar 2 2 + ArS0 2NH 2N Ar 2 (30) ( 31) These acid-induced decompositions are the reverse of the coupling procedures for making the triazenes. The 15 N NMR of a solution of X (generated in situ by addition of 43 the sodium salt of .e_-toluenesulfinic acid to 1-3- 15 N in dimethyl sulfoxide), acidified with trifluoroacetic acid or acetic acid, shows that X decomposes into I-3- 15 N,I-1- 15 N and presumably sulfinic acid ( Eq. 32). (32) There is no evidence for the intermediate formation of 1 ,3-di(E_toluenesulfonyl)triazene (XV): ArS0 2-NH-N=N-S0 2Ar xv Instead, it appears that the highly delocalized anion (X} is preferentially protonated at the hard oxygen site of the sulfonyl. oxygen-protonated tautomer of XV then decomposes into I and Rtol uenesul fini c acid. This 44 This mode of decomposition is consistent wi.th the principles of HSAB, CONCLUSION The reactions of I-3- 15 N with several nucleophiles in dimethyl sulfoxide studied here by 15 N NMR enlarge the scope of £_toluenesulfonyl azide reactivity. The principles of HSAB are very useful in rationalizing the course of these reactions. Soft nucleophiles. such as the sulfur of E_-toluenesulfinate, preferentially attack the terminal nitrogen of I, giving adducts similar to those arising from the reaction of aryldiazonium salts with those nucleophiles. Hard nucleophiles, such as the oxygen of E_-toluenesulfinate, appear only to react with the sulfonyl of I to displace azide ion. Borderline nucleo- philes, such as azide ion or .e_-toluenesulfonamide anion, show both types of reactivity. These investigations also impose limitations on the utility of this reagent. The 2,3-nitrogen scrambling of I by E_-toluenesulfonamide anion indicates that completely specific 15 N labeling of diazo compounds with I-3- 15 N may be difficult, if not impossible. Indeed, if I-3- 15 N had been employed to synthesize diazocyclopentadiene by the classic route of Doering and DePuy 39 • using the lithium salt of cyclopentadiene, completely 15 N..,scrambled diazocyclopentadiene would surely have resulted, Furthermore, the sequestering of E_-toluenesulfonyl azide by .e_-toluenesulfinate anion requires that the yields in azido-transfer reactions using I must be poor, as has been observed. 45 EXPERIMENTAL The 15 N NMR spectra were taken at 18.25 MHz with a Bruker WH-180 spectrometer, using 15-25 m1 samples in 25-mrn sample tubes. A 5-mm concentric tube containing a solution prepared by dissolving sufficient 15 H No 3 in o2o to give a 1-M acid concentration provided lock and reference signals. Chemical shifts are reported in ppm upfield from external HN0 3 . All spectra were proton-coupled, and observations were made at ambient probe temperature, which was approximately 22°c. Unless otherwise specified, dimethyl sulfoxide was dried before use by distillation over calcium hydride, £_-toluenesulfonyl azide was obtained from £_-to1uenesulfonyl chloride and sodium azide in ethanol, following the procedure of Curtius 1 . Di-E_-toluenesulfonamide and its sodium salt was prepared according to the method of Dykhanov 21 with .e._-toluenesulfonamide and E_-toluenesulfonyl chloride in aqueous sodium hydroxide. Sodium Salt of p-Toluenesulfonamide A mixture of Q_-toluenesulfonamide and the sodium salt of Q_toluenesulfonamide in dimethyl sulfoxide was generated in situ by the slow addition of an excess of Q_-toluenesulfonamide in dimethyl su1foxide to a stirred solution of dimsyl anion under nitrogen, Dimsyl anion was prepared by treating dimethyl sulfoxide with sodium hydride {50% oil suspension) as described by Corey and Chaykovsky 40 . After addition of .e._-toluenesulfonamide to the dimsyl anion the solution was stirred for 2.5 h at 60°c under nitrogen. temperature before further use, This solution was cooled to room 46 Isolation of a Mixture of 1 ,2 and 3-15 N-Labeled I Dimethyl sulfoxide solutions of 15 N-labeled 2_-toluenesulfony1 azide scrambled by 2_-toluenesulfonamide anion or azide ion were combined. After addition of water to the solution, it was extracted three times with pentane. The combined pentane extracts were washed three times with water, dried first over anhydrous sodium sulfate, then over calcium sulfate, The solvent was removed under reduced pressure, and the 15 N spectrum of the residual mixture of 15 N-labeled Q_-toluenesulfonyl azides showed that no other 15N-labeled compounds were present. Reactions of di-p-Toluenesulfonamide with Sodium Azide Di-2_-toluenesulfonamide (5.5 g. 0.017 moles) and 3.0 g (0.46 moles) of sodium azide were added to 80 ml of dimethyl su1foxide. was stirred at room temperature for 4 days. The mixture After addition of 400 ml water, the solution was extracted with pentane, The combined extracts were washed twice with water. dried over anhydrous sodium sulfate, and the solvent removed under reduced pressure. The remaining faint-yellow residue was dissolved in ether and an ethereal solution of triphenylphosphine was added. No 2_-toluenesulfonyl azide triphenylphosphine adduct could be isolated 41 47 REFERENCES AND NOTES 1. T. Curtius and G. Kraemer, J. Prakt, Chem 125, 323 (1930). 2. K. Clusius and H, R. Weisser, Helv. Chim, Acta 35, 1548 (1952}. 3. J.-P. Anselme and W. Fischer, Tetrahedron 2~, 855 (1969), 4. This scrambling is essentially complete after 1.5 h. The rate would be difficult to measure by 15 N NMR because the signal-tonoise ratio is poor, unless relaxation times are short to permit rapid pulsing. With azide systems, as used here. the reaction half-life has to be on the order of a few weeks for accurate kinetics, 5. The highly nucleophilic trifluoromethanesulfonyl azide is reported to transfer a diazo group directly to a primary amine without base catalysis. C. J. Cavender and V. J, Shiner, J. Org. Chem, 37, 3567 (1972). Clusius and H. R. ~!eisser~ Helv. Chim. Acta~. 1524 (1952). 6. K. 7. K. Clusius and H. Craubner, Helv. Chim. Acta 38, 1060 (1955). 8. W. T. Evanochko and P. B. Shevlin, J. Am. Chem. Soc. 101, 4668 (1979), 9. R. L. Jacobs, J. Org. Chem. 42,571 {1977), 10. G. La Monica, P. Sandrini, F. Zingales ands. Cenni, J, Organomet. Chem. 50, 287 (1973). 11. Cal R. N. McDonald and A. K, Chowdhury, J. Am, Chem. Soc, 102, 5118 (1980). (b) 0. Dimroth, Chem. Ber, 38, 670 (1905). 48 (c) P. A. Lehman, R. S. Berry, J. Am. Chem. Soc. ~. 8614 (1973). 12. R. O. Duthaler, H, G. Forster and J. D, Roberts, J, Am. Chem, Soc. 100, 4974 (_19781. 13. 0. W. Webster, J, Am. Chem. Soc. 88, 3046 {1966). 14. (a) M. Regitz, Synthesis 351 (1972}. (b) T. Weil and M. Cais, J, Org, Chem, 28, 2472 (_1963), 15. L1oyd and Wasson do not believe that the amine catalyst used in this diazo transfer is a strong enough base to appreciably ionize the cyc1opentadiene, and they suggest a different mechanism for diazo transfer. 408 (1966). D. L1oyd and F. I, Wasson, J. Chem, Soc, C. However, the diazo-transfer reaction is slow, and rate-limiting ionization of cyclopentadiene by base cannot be ru1ed out. Indeed, cyclopentadiene has been found to be deuterated in o2o in the presence of N,N-dimethylpyridonimine; N. Kursanov and Z. N. Parnes, Dokl. Chem. (_Engl. Transl.) 106, 385 (1956). 16. I. H. Pitman, H. S. Dawn. T. Higuchi and A. Hussain. J, Chem. Soc. B, 1230 (1969). 17. D. G. Farnum and P. Yates, Proc. Chem, Soc. 224 (1960). 18. D. W. Emerson, R. R. Emerson, S. C. Joshi, E. M. Sorensen and J. E. Turek, J. Org. Chem. 44, 4634 (1979). 19. 20. R. G. Pearson and J. Songstad, J. Am. Chem. Soc. 89, 1827 (1967). 15 The 15 N chemical shift of potassium azide-1- N (96.8%) in dimethyl 49 sulfoxide is 270,0 ppm. A natural~abundance 15 N spectrum of sodium azide in water gives chemical shifts for Nl of 275,8 ppm and for N2 of 127.0 ppm. These 15 N chemical shifts are in general agreement with the reported 14 N chemical shifts. W, Beck, W. Becker; K. F. Cnew; W. Derbyshire; N. Logan, D. M. Revitt, D. B. Sowerby, J. Chem. Soc. Dalton Trans. 245 (1972). Dykhanov, J. Gen. Chem. USSR (Engl. Transl.) 29, 3563 (1959). 21, N. N. 22, The 15 N NMR of natural-abundance E_-toluenesulfonamide in undried dimethyl sulfoxide shows a triplet at 279.3 ppm with an 81-Hz coupling constant, in agreement with the triplet seen in Fig. le. The 15 N NMR of an equimolar mixture of £_-toluenesulfonamide and its anion in dimethyl sulfoxide shows a singlet at 273.l ppm. This indicates that very little of the £_-toluenesulfonamide anion remains in Fig. la. 23, N. A. Porter, J. G. Green, and G. R. Dubay, Tetrahedron Lett., 3363 (1975}. 24. An inorganic analogue of reaction 19 has been reported, Br + 3N 2 although the detailed mechanism is unknown. R. 0, Griffi.th and R. Irving, Trans, Faraday Soc. 45, 563 (1949). 25 . Somewhat similar azide couplings and decompositons have been reported for a heterocyclic azidinium salt, Justus Liebigs Ann. Chem. 647, 11 (1961) H. Balli, 50 26. So far, all attempts to isolate the hypothetical N6 have failed. A. W. Browne and G, E. F. Lundell, J, Am, Chem, Soc . ll_, 435 (1909). 27. The reaction of two azide radicals to give dinitrogen, proposed in the decomposition of solid ionic heavy-metal azides, might involve azide dimers, C. B. Colburn (Ed.), "Developments in Inorganic Nitrogen Chemistry", Vol. 1, Elsevier: Amsterdam, 1966, 73. 28. The possible aromaticity of VIII and its instability relative to 3 N2 has been probed by ab initio calculations. See, for example, J. S. Wright, J. Am. Chem. Soc. 96, 4753 (1974). 29. See I. Ugi, Tetrahedron .J.2., 1801 (1963) and references cited therein. 30. (a) E. Fanghanel. R. Radeglia, D. Hauptmann, B, Tyszkiewicz and M. Tyszkiewicz, J. Prakt. Chem . 320, 618 (1978). (b) T. Axenrod, P. Mangiaracina and P. S. Pregosin, Helv. Chim. Acta 59, 1655 (1976). (c) A. N. Nesmeyanov, E. V. Borisov, A. S. Peregudov, D. N. Kravtsov, L, A. Fedorov, Z, I. Fedin, and S. A. Postovoi, Dokl. Chem. (Engl. Transl.) 247, 380 (1980). (d) E. Lippmaa, T. Saluvere, T. Pehk and A. Olivson~ Org. Magn, Reson. ~. 429 (1973). 31. 32. J. P. Gouesnard and G. J. Martin, Org. Magn, Res on, Jl 263 ( 1979), ..., , , (a) J, Bene~, V. Beranek, J. Zimprich and P. Vetesnik, Collect. Czech. Chem. Commun. 42, 702 {1977). 51 (b) A. Key and P, K, Dutt, J, Chem, Soc. 2035 (_1928). , 33, D. H. Sieh, D. J, Wilbur and C, J, Michejda, J. Am, Chem, Soc. 102 ' 3383 ( 1980) . 34, See, for example, J. 0. Reed and W. Lwowski, J, Org, Chem, 36, 2864 (1971). 35, (a) V. J. Bauer, W. J. Fanshawe and S. R, Safir, J. Org, Chem. n_, 3440 (1966) . (b) C. D. Ritchie , J. D. Sa1thJe1 and E, S, Lewis~ J. Am. Chem. Soc. 83, 4601 (1961), 36. The 15 N chemical shifts of HN 3 in ether are 321.,0, 130._6 and 175.0 ppm relative to NOj. z. Naturforsch.~ B:Anorg. Chem., Org, Chem. 33b, 993 (1978), The 14 N chemical shifts of rapidly ex- changing HN 3 in ether/water are 240 and 129 ppm relative to No3. 19 37. Studies on the alkylation of the sodium salt of £~toluenesulfinic acid show similar effects. Hard alkylating agents give rise to the sulfinic ester, whereas soft alkylating agents give rise to the sulfone. See, for example, J. S. Meek and J, S. Fowler, J. Org. Chem. ll· 3422 (1968), 38, F. P. Corson and R. G, Pews~ J. Org. Chem, 36, 1654 (1971), 39, W. von E. Doering and C, H, DePuy, J. Am, Chem, Soc. 75, 5955 (1953). 40, E. J. Corey and M. Chaykovsky, J, /lm. Chem, Soc.~. 866 {1962). 41. H. Bock and W. Wiegrabe, Angew. Chem, 75, 789 (1963). 52 SECTION 3 THE REACTION OF .12_-TOLUENESULFONYL AZIDE WITH POTASSIUM AZIDE 53 INTRODUCTION In Section 2 it was reported that E_-toluenesulfonyl azide (I) formally transfers an azido group to azide ion in dimethyl sulfoxide to give E_-toluenesulfinate anion and dinitrogen {Eq. 1). (1) This reaction bears a strong formal resemblance to halogen-exchange 1 reactions , where sulfinate ion and azide are regarded as pseudohalides (Eq. 2). (2) 2 In this case, the pseudohalogen (N 3) 2 , is not stable and decomposes to dinitrogen. 15 The reaction of I with potassium azide-1- N has been reexamined 15 in toluene and dichloromethane by N NMR. Azide ion exchange (3), 15 leading to I-l- 15 N, and I-3- N, occurs in these solvents, just as it does in dimethyl sulfoxide. Curiously, in dichloromethane and toluene, the formation of I-2is also indicated (4). 15 N 54 . Ts-Nd'N=N + N3- (4) Two mechanisms for this scrambling are proposed in this section. Azide-ion metathesis (Eq. 5) involving reversible formation of an N-pentazole derivative (II) followed by reaction 3 could account for the 15 N-label at the central nitrogen of I, {5) . TsN 3 5 + N::::! N=N- n Alternatively, a scrambling route involving reversible addition of .E_-toluenesulfonylnitrene to I can be envisioned {6) TsN: + 15 N=N=N-Ts - Ts-N N-Ts '°'N•N,..... - Ts-N:J'N•N + :NTs (6) These mechanisms, and the evidence for them will be described in detail 15 later. In addition, the reaction of I with potassium azide-1- N in dichloromethane in the presence of iodide ion is examined. RESULTS AND DISCUSSION The Reaction of I with Potassium Azide-1- 15 N in Dimethyl Sulfoxide Although reaction 3 occurs in relatively dilute dimethyl sulfoxide 15 solutions, reaction 4 does not. The N NMR of a more concentrated 15 solution of unlabeled I and potassium azide-1- N in dimethyl sulfoxide 15 15 is shewn in Fig. 1. the signals of I-3- N at 132.2 ppm and I-l- N at 234.0 ?Pm are visible in equimolar amounts, in accord with azide ion exchange (Eq. 3). Terminally labeled azide ion is visible at 270.5 55 ' I -'-.. . . . . _ _......_..--...,_. .__,. ___. __l .. l_. 0 100 200 I \ 300 ppm 15 N spectra of 4.Bxlo- 3 mol potassium azide-1- 15 N~ Figure 1. 2 l.Oxlo- mol Q_-toluenesulfonyl azide in 20 ml dry dimethyl sulfoxide~ 20 µsec pulse, 10-sec repetition rate, Spectrum of sample (a} immediately after preparation, 2109 transients; (b) 12 hours after preparation, 3719 transients;(c) two weeks after preparation, 1955 transients, 56 ppm. Resonances due to 15 N-labeled dinitrogen (64.2 ppm) and di(Q_- toluenesulfonyl) triazenyl anion (24.2 ppm) are visible in Fig. la, consistent with reaction l followed by reaction 7. ... There are no signals corresponding to Ts-N-N=N-Ts (7) 15 N-label in the central nitrogen of I (142.2 ppm) or azide ion (125.8 ppm) in Fig. 1. Thus the scrambling of Equation 4 does not occur in this more concentrated solution, even after two weeks (Fig. le). The Reaction of I with Potassium Azide-1- 15 N in Toluene The 15 N NMR of a mixture of I and potassium azide-1- 15 N in A large excess of 18-crown-6 was added 3 to the sample to partially solubilize the solid potassium azide. a The 15 N signals due to label at Nl and N3 of Q_-toluenesulfonyl azide toluene is shown in Fig. 2. are visible at 236.2 and 133.5 ppm, in accord with fast azide exchange (Eq. 3). 15 The concentration of dissolved potassium azide-1- N is small and 15 dissolution is slow. The N signal assigned to dissolved potassium azide-1- 15 N at 275.0 ppm is visible only after 8 hours (Fig. 2b). Curiously, an 15 N NMR of a potassium azide-l- 15 N/18-crown-6 mixture 3b shows two signals: one at 256.6 ppm and the other at 274.2 ppm. After several days the signal at 256.6 moves to 254.4 ppm and the 274.2 ppm signal disappears 3b 57 0 100 200 300 ppm 15 3 Figure 2. 15 N spectra of 5.2xlo- mol potassium azide~1~ N 2 1.0xl0- 2 mol I and 6xl0- mol 18-crown-6 in 20 ml toluene; 20 µsec pulse. 10-sec repetition rate. Spectrum of sample (al immediately after preparation, 2554 transients;(b) 8 hours after preparation, 3611 transients;(c) 30 days after preparation~ 4802 transients, 58 The 15 N signal at approximately 275 ppm in toluene has ,a chemical shift closer to that of azide ion in pure water (275.8) than in a nonhydrogen-bonding solvent like dimethyl sulfoxide (270.5 ppm). This may indicate that the 275 ppm signal is due to azide ion-1- 15 N associated with water impurity present in solution. The compound tentatively assigned to the water-associated azide ion appears to 15 be converted to another symmetrical species with an N signal at about 255 ppm. The observation that, as the 255 ppm signal grows the 275 ppm signal disappears even though there is an excess of solid potassium azide present, may indicate that the new compound is also associated with water. Possible candidates for this new species contact ion pairs (III) 4 and an un- include fast-exchanging precedented 5 crown-azide ion complex (IV). N=fr=N m m: The chemical shifts of III and IV might be expected to be dependent on the relative concentration of water in solution. The spectrum of .e_-toluenesulfonyl azide (I) and potassium azide-1- 75 N in toluene-crown ether after 30 days (Fig. 2c) snows no 275 ppm resonance. A rather broad signal at 263.6 ppm is visible, 59 which appears to move to 257.3 ppm after 20 more days. This roaming signal is assigned to the unknown azide ion complex. The formation of I-2- 15 N occurs readily in this solvent mixture. After approximately 8 hours (Fig. 2b) about 5% of I is 15 N The potassium azide-2~ 15 N concentration is too low to permit observation of its 15 N signal. labeled at N2 (141.8 ppm). Azido transfer by I to azide ion (Eq. 1) is indicated by the presence of 15 N-labeled molecular nitrogen (65.2 ppm) in Fig. 2b and 2c. A signal at 26.1 ppm appears in Fig. 2c which can be assigned to di{Q_-toluenesulfonyl) triazenyl anion, fanned by reaction 7. The signals at 158.6, 165.6 and 289.3 ppm present in Fig. 2c are unassigned. The Reaction of I with Potassium Azide-1- 15 N in Dichloromethane The reaction of .e_-toluenesulfonyl azide (I) with azide ion was investigated in dichloromethane. The potassium azide-1- 15 N was completely dissolved by complexation with 18-crown-6. The 15 N spectra of this sample are shown in Fig. 3. Equimolar amounts of 1-1- 15 N (236.5 ppm) and I-3- 15 N (133.3 ppm) are visible in Fig. which indicates fast azide exchange (reaction 3). of azide ion is visible at 275.8 ppm. 3a, The terminal nitrogen The signals at 154.0 and 295.6 ppm result from a product of a side reaction of the azide ion with solvent and will be discussed later. Within five days (Fig. 3b) no azide ion remains and about 7% of I is labeled at the central nitrogen. 60 0 100 200 300 ppm 15 Figure 3. N spectra of 5.9xl0- 3 mol potassium azide-1- 15 N, 6.lxlO -3 mol I, 6.7x10 -3 mol 18-crown-6 in 30 ml dichloromethane; 20 µ-sec pulse, 10-sec repetition rate. (a) Spectrum of sample immediately after preparation, 3748 transients;(b) five days after preparation, 5213 transients. 61 Surprisingly, no labeled molecular nitrogen or di(E_~toluene~ulfonyl)­ triazenyl anion signals can be seen in Fig. 3. Thus it appears that reaction 1, the azido transfer to azide ion to give £_-toluenesulfinate and dinitrogen, does not occur in this solvent system. The Reaction of Potassium Azide-1- 15 N with Dichloromethane Potassium azide-1- 15 N solubilized by comp1exation with 18-crown-6 reacts with dichloromethane to give a product with 15 N signals at 154.0 and 295.l ppm (Fig, 4a). These chemical shifts are similar to the 15 N shifts recently reported for methyl azide (166.9, 317.1}. 6 The 15 N chemical shifts along with a positive test for free chloride ion suggest that azide ion displaces chloride ion from the solvent to give a chloromethyl azide or diazidomethane (Eq. 8-9). ... (8) (9) Neither of these azides appears to have been previously reported, but the diazidomethane formed in reaction 9 is undoubtedly highly . 7 exp 1os1ve Considering the large excess of dichloromethane, reaction 9 seems somewhat unlikely, although the chlorine of chloromethyl azide is expected to be displaced in an SN2-type reaction more easily than a chlorine of dichloromethane. The chloromethyl azide does not undergo fast azide exchange. 62 100 0 Figure 4. (a) 200 300 ppm 15 N spe~trum of a solution of chloromethyl azide 15 3 generated in situ from 3.4xl0- mol potassium azide-1- N, 2.5 g 18crown-6 in 20 ml dichloromethane, 20 µsec pulse, 10-sec repetition 15 rate, 1831 transients. (b) N spectrum of a solution of chloromethyl . . az1'd e- l - 15 N, . from 3 .4xl0 - 3 mo l potassium azi de ( generated in situ 3.4xl0- 3 mol potassium azide, 5.0 g 18-crown-6 in 20 ml dichloromethane) after addition of 3xl0- 3 mol of the sodium salt of ~-toluenesulfinic acid, dihydrate and 5.0 g 18-crown-6. rate, 4059 transients. 20 µsec pulse, 10-sec repetition 63 This was shown by the addition of natural abundance potassium azide to a solution of 15N-labeled chloromethyl azide (10) ... ,, /I { 10) No 15 N-1abeled potassium azide was visible in the 15N NMR spectrum of this solution. This explains why there is no resonance assignable to 15 N-label at the central nitrogen of chloromethyl azide in Fig. 3b. The Reaction of £_-i.oluenesulfinate Ion with Chloromethyl Azide in Dichloromethane Having generated an 15 N-labeled alkyl azide in situ, the possibility of its reaction with .e._-toluenesulfinate anion to give a triazenyl anion was investigated. To do this, 0.5 equivalents of the sodium salt of p-toluenesulfinic acid (dihydrate) was added to the chloromethyl azide in dichloromethane (Fig. 4b). Gas was visibly evolved. The expected reaction (Eq. 11) did not appear to occur. H I Cf-C-N3 1 H + Ts- . H I Cl-C-N=N-N-Ts I H ( 11 ) The 15 N NMR of the solution {Fig. 4b) showed resonances at 65.2, 157.6, 271.2 and 308.4 ppm, in addition to the signals of the chloromethyl azide (154.7 and 295.6 ppm). The 15 N signals at 65.2 64 and 271.2 ppm are assigned to 15 N-1abeled dinitrogen and£_toluenesulfonamide anion 8 , respectively. The signals at 157.6 and 308.4 ppm are assigned to 15 N-1abeled y and a nitrogens of a new alkyl azide. A mechanistic scheme which is consistent with these assignments is outlined below. The first step involves the reaction of .E_- toluenesulfinate at the carbon of chloromethyl azide rather than the terminal nitrogen (Eq, 12). Presumably the azido group of ch1oro methyl azide is not electrophilic enough to couple with £_-toluenesulfinate as in reaction 11, or reaction 11 is sufficiently reversible that the irreversible SN2 reaction diverts the triazenyl anion to other products. H3 C --0- 0 1 H I I& 11 I 0 H (12) 15 S-C- N=N= N + c1- Based on hard and soft acid-base arguments, .e.-toluenesulfinate is expected to be alkylated at sulfur 9 to give the known a-azido sulfone10 (V), rather than at oxygen to give a sulfinate ester. The signals at 157.6 and 308.4 in Fig. 4b are reasonably assigned to the expected sulfone V. In contrast to the triazenyl anion formed by reaction 11, the triazenyl anion (VI) formed by the coupling of V with .e_-toluenesulfinate, possesses a relatively acidic a-hydrogen 11 65 H I IS IS- Ts -C- N= N-N-Ts I H VI is expected to tautomerize~ and the resulting triazene should decompose to give a diazo compound and .e_-toluenesulfonamide anion (Eq. 13) lZI . .. Because .e_-toluenesulfinate was added in the form of the dihydrate, the known .e_-toluenesulfonyldiazomethane 12 will hydrolyze to give the a-hydroxy and a-chloro sulfone and dinitrogen 12 b. These reactions explain the presence of 15 N-labeled .e_-toluenesulfonamide anion and molecular nit~ogen in Fig. 4b. VI is identical to the intermediate expected from the diazotransfer by .e_-toluenesulfonyl azide (I) to the carbanion of E_-toluenesulfonyl methane. Although di(g_-toluenesulfonyl) diazomethanes can be prepared by the reaction of di(.e_-toluenesulfonyl) methanes with I and alkali, monosulfonyldiazomethanes cannot 13 . These preliminary studies suggest that the addition of E_-toluenesulfinate anion to electrophilic azides possessing an a-hydrogen might be a new and useful entry onto the standard diazo-transfer surface. 66 The Reaction of I with Potassium Azide-1- 15 N in the Presence of Potassium Iodide The 15 N spectrum of an equimolar solution of I, potassium azide1-15N, potassium iodide in dichloromethane/18-crown-6, is markedly Azide exchange and formation of chloromethyl 15 azide is observed, but in the presence of iodide ion, 1-2- N is not formed. In contrast to 15 N-spectra in the absence of iodide ion, a different from Fig. 3. strong signal due to 15 N-labeled molecular nitrogen (65.3 ppm) is visible inmediately after sample preapration. Molecular iodine is also formed under these conditions. We propose that the iodine and molecular nitrogen are products from the reactions of an intermediate iodoazide 14 Control experiments indicate that in dichloromethane, iodine is not formed from a direct reaction of I and iodide. Azide ion is apparently required as a catalyst (Eq. 14). TsN 3 + N3- + r- (14) The iodoazide formed in reaction 14 is expected to react with azide ion to give dinitrogen (Eq. 15): (15) or iodide to give iodine (Eq. 16): {16) 67 Analogous reactions have been observed with bromoazide. 15 Dubgen and Dehnicke16 have recently reported the low temperature isolation of iodoazide complexes from the reaction of ~odoazide with tetramethylaJJJnonium iodide or azide (Eq. 17-18). (17) (18) Interestingly, the reaction of 12 with azide ion, thought at one time only to occur in the presence of certain sulfur-containing catalysts 17 was found by DUbgen and Dehnicke 16 to yield the iodine-iodoazide complex. Thus reaction 16 may be reversible. Reaction 14, if it occurs, requires the formation of £_-toluenesulfinate anion. Just as observed for the conditions of the spectrum shown in Fig. 3, no di(E._-toluenesulfonyl) triazenyl anion or the products from the reaction of the sulfinate with chloromethyl azide are visible in the presence of iodide. Control experiments show that di(Q_--toluenesulfonyl} triazenyl anion is formed on treatment of I with the sulfinate salt {_dihydrate) in dichloromethane/crown ether. The formation of the triazenyl anion, however, is expected to be reversible. Thus either reaction 14 does not occur, or the£_- toluenesulfinate anion is consumed in one or more fast and irreversible reactions. A likely route for the disappearance of E_-toluenesulfinate anion 68 is its known reaction with iodine (Eq. 19) 18 Tsl . + 1- This reaction alone is not enough to account for the lack of (19) 15 N- labe1ed products derived from E_-toluenesulfinate. · But reaction (19) coupled with the nucleophilic attack by E_-toluenesulfinate on E_19 should be sufficient to keep the £_toluenesulfonyl iodide toluene sulfinate anion concentration low. Mechanism of the Fonnation of I-2- 15 N: Azide-Ion Metathesis Reversible coupling of I with terminally 15 N-labeled azide ion to give an intermediate E_-toluenesulfonylpentazole anion (II) could account for formation of scrambled azide ion (Eq. 5). Azide exchange 15 with I (Eq. 3} would then afford 1-2- N. There are many cases of hetereocumnulenes reacting with azide ion or hydrazoic acid to form stable five-membered ring hetereocycles. Examples include the 20 and thioketenes 21 with hydrazoic reactions of isothiocyanates acid to give thiatriazoles, the reactions of hydrazoic acid with 23 to give tetrazoles, and the carbodiimides 22 and ketenimines coupling of carbon disulfide 24 with azide ion to give 1 ,2,3,4-thia- triazolinethion ate anion. The inhibition of I-2- 15 N formation by iodide in dichloromethane, 69 however, suggests that the !:!_-pentazole (II) derivative is actually formed by the cyclization of a linear E_-toluenesulfonylhexazenyl anion (VII) 25 . The iodide ion appears to intercept the hexazine derivative, inducing decomposition to E_-toluenesulfinate anion, azide ion and iodoazide, before cyc1ization can occur. The suggested mechanism is outlined in Scheme 1. Scheme 1 Ts-N•N•N ~ + N5- Ts-N-NsN-N•N•N lZI[ - Ts- + N1 - 3N 2 n The decomposition route of VII is strongly dependent on solvent. In dimethyl sulfoxide, VII decomposes to E_-toluenesulfinate anion and molecular nitrogen. In toluene, the substituted hexazene (VII) seems long-lived enough for reversible cyclization to II to compete with decomposition. In dichloromethane, however, only the reversible cyclization appears to occur. In this solvent the cyclization is effectively prevented by addition of potassium iodide. The origin of the large solvent effect on the mechanism is not understood, but 70 polarity and basicity of the solvent may be i.nvolved, The reactions of Scheme 1 are related to the reactions of aryldiazonium salts with azide ion, The existence of a similar ring-chain equilibrium has been established by 15N-1abeling and kinetic evidence (Scheme 2) 26 Scheme 2 ... . Ar-N-N-N=N=N u It is not clear if the arylpentazole (VIII) is formed directly from the aryldiazonium salt and azide ion, or by cyclization of the arylpentazene. Arylpentazole (VIII) i.ntermediates have been isolated and characterized spectroscopically 27 . There have also been several studies concerning the reaction of azides with azide ion. Both azide exchange 28 and reaction at the terminal nitrogen have been observed. Balli has examined the reaction of hetereocycltc azidi.nium Sf:llts 29 with azide ion. Not surprisingly, the li-diazonium salts react wi.th azide ion to give !!-azides, presumably via a substituted hexazene (Eq. 20) 3o 71 (Jr\ //C-N 3 +N I C2H5 N5- (}rs~C=N-N5 . N I C2H5 IX (20) 1-N, o:s'c:+N // ~ (rs~C=N-N3 -2N 2 N I I C2H5 C2H5 x The N-azide (Xl is stable at low temperature, but loses two molecules of nitrogen at room temperature to afford an ylide 30, Here 1 l5N, labeling experiments were used to demonstrate that azide exchange with IX also occurs 30 . Inorganic examp1es of azido transfer to azide ion have been reported. Most notable is the reaction of bromoazide with azide ion in aqueous solution (Eq. 21) 15 . (21) 72 Mechanism of the Formation of r~2- 15 N: £_-Toluenesu1fony1nitrene Another way in which I-2- 15 N might be fanned is via a chain reaction of E_-toluenesulfonylnitrene with I-3~ 15 N_ (Eq. 6) 31 through the intermediacy of a di-(E._-toluene)su1fonyltetrazadiene (XI) 32 . Ts-N N-Ts ~N-N~ llo Transition metal complexes of XI are in fact known 33 and are thought to be formed by the addition of a metal-bound £_-toluenesulfonylnitrene to £_-toluenesulfony1 azide (I). One possible source of £_-toluenesulfonylnitrene in the presence of I and azide ion, is the decomposition of the !i_-pentazole derivative {II} discussed in Scheme 1. II resembles ylides XII 34 and XIII 35 , which have both been investigated as potential £_-toluenesulfonylnitrene precursors. ©>~-N-Ts n xrr XIII The decomposition of II could give E._-to1uenesulfonylnitrene and the unknown pentazole anion (Eq. 22 ). 27 73 n . -N N1" \ I - ~ N N'-" -N I + :N-Ts (22) Once formed, the pentazole anion could quickly decompose to azide ion and molecular nitrogen by a symmetry allowed cycloreversion 36 . Another potential £_-toluenesulfonylnitrene precursor is the !!_-azide derivative (XIV) fanned by loss of nitrogen from £_-toluenesulfonylhexazenyl anion (VII) Ts-N-N=N-N=N=N . (23) :mz: This reaction is related to the stepwise decomposition of X (Eq. 20) 30 . Elimination of azide ion from XIV would clearly give £_-toluenesulfonylnitrene. The proposed attack of £_-toluene~ulfonylni.trene on the te.rminal nitrogen of £_-tol uenesulfonyl azide (I) to give XII I is somewhat unusual, because singlet £_-toluenesulfonyl nitrenes are hi'ghly electrophilic 31 and should prefer the a-nitrogen of I, Indeed, Gibson and coworkers 37 have isolated phenylazocarboxylate from the reaction of phenyl azide with singlet ethoxycarbonylnitrene, a product consistent with a-attack by the nitrene (Eq. 24), (24) 74 It is, however, quite possible that the anionic nitrene precursors (II,XIV} are somewhat more nucleophilic than the free nitrene and that it is these complexes that attack the terminal nitrogen of I. The .nitrenoid chloramine-T, a reasonable model for XIV, does in fact 38 , behave as a nucleophile in some cases If reaction 6 occurs, the reaction of R-toluenesulfo nylnitrene 15 (or a complex thereof) with I-l- N (.formed in reaction 3) will afford the 15 N-labeled nitrene (25}. Ts~N=N=N + :N-Ts ... Ts~N: + N=N=N-Ts Thus, products from the insertion of the singlet nitrene (25) 39 into the 15 C-H bonds of the solvent or crown ether should be visible in the N 15 NMR. Examination of a sample of I and potassium azide-1- N in toluene/ 15 18-crown-6 fifty days after preparation by N NMR, reveals a small 1H-decoupling doublet at 283.9 ppm (J=86 Hz) and a singlet at 289.3 ppm. experiments show that these nitrogens possess a strong negative NOE. These signals are too far upfield to correspond to the sulfonamides 40 expected from nitrene insertion into the toluene C-H bonds, However, 15 because of a lack of N-shifts for suitable model compounds, the product of nitrene insertion into the C-H bond of 18-crown-6 cannot be ruled out. CONCLUSION In surmJary, evidence for the addition of azide ion to £_-toluenesulfonyl azide (I) fonning a di.screte addition intermediate has been presented. In dimethyl sulfoxide this intermediate, the £_-toluenesulfonylhexazenyl 75 anion (VII), decomposes to .E_-toluenesulfinate anion and molecular 15 nitrogen, In dichloromethane and toluene, I . . 2... N is fonned from the 15 reaction of I wi. th terminally N- labeled azi de ion sol ubi 1i zed by 18-crown-6. Two mechanisms for this scrambling have been proposed. In the first mechanism, VII is reversibly formed, and can reversibly cyclize to an ,ri. . pentazole derivative, The 15N-label at N2 of I is derived from azide exchange with azide ion scrambled in this way. 15 In the second mechanism, I . . 2- N is formed by chain reaction of 15 p_-toluenesulfonylnitrene with I-3- N. The anionic nitrene precursors are derived from the initially formed £_. . toluenesulfonylhexazenyl anion. 15 Iodide ion prevents the formation of I-2- N in dichloromethane by inducing the decomposition of the hexazene derivative before further reaction can occur. More study is needed to unravel the mechanistic details of this novel reaction of azide ion with I. EXPERIMENTAL The 15N NMR spectra were taken at 18.25 MHz with a Bruker WH-180 spectrometer, using 15-25 ml samples in 25-mm sample tubes. A 5-mm concentric tube containing a solution prepared by dissolving sufficient H15 No 3 in o2o to give a 1-M acid concentration provided lock and reference signals. Chemical shifts are reported in ppm upfield from external HN0 3 . All spectra were proton-coupled, and observations were made at ambient probe temperature, which was approximate1y 22°c. E_-Toluenesulfonyl azide. was synthesized following the procedure of Potassium azide-1- 15 N (97.2-99.5%) was obtained from Curtius 41 76 Prochem.. 18-crown-6 was purchased from Aldrich Chemi.cal Co. and used without purification. Dimethyl sulfoxide was sequentially dried over 0 3 A molecular sieves and toluene fractionally distilled before use. Commercially available spectrophotometric grade dichloromethane was employed in these studies, REFERENCES AND NOTES 1. M. F.A. Dove, D, B. Sowerby, 11 Ha 1ogen Chemi stry'1, V, Gutmann, ed. , Academic Press: New York, 1967, pp. 41-132. 2. (a) The trimer of dinitrogen (N 6 ) is also thought to be involved in the oxidation of azide ion to dinitrogen, Both a cyclic and 15 linear form of N6 are consistent with the N labeling studies. K. Clusius, H. Schumacher, Helv . Chim. Acta. 11_, 2264 (1958); R. K. Murmann, J. C. Sullivan, R. C. Thompson, Inorg. Chem. 7_, 1876 (1968). (b) Interestingly~. the weakly been experimentally observed. bound dimer of dinitrogen (N 4) has C. A. Long, G. Henderson and G. E. Ewing, Chem. Phys. 2, 485 (1973). 3. (a) For review on the use of crown ethers as solid-liquid phase transfer reagents, see: G. W. Gokel and H. D. Durst, Aldrichimica Acta~' 3 (1976). 15 (b) The 15 N spectra of potassium azide-1- N in toluene/crown ether were obtained with a Bruker WM500 (Southern California Regional NMR Facility} by Dr. W. M. Croasman and Mr, T. A, 77 Perkins, 4. Their help ls gratefully acknowledged, For review of soli_d and soluti.on structures of crown ether complexes, see; A. L Popovi J,-M, Lehn, 11 Coordi.nation Chemistry of Macrocyc 1i c Compounds 11 ~ G, A. Me 1son, ed, , Pl en urn; New York~ 1979, Chapter 9, 5. Evidence for a symmetrical 18-crown-6 complex with the isoelectronic, but positively charged, nitronium ion has recently been reported, I L. Elsenbaumer and E. Wasserman,presented in part at the Second Congress of the North American Continent~ Las Vegas~ August~ 1980. 6. J. MiJller, Z. Naturforsch B: Anorg. Org. Chem. 34b~ 437 (1979). 7. Diazidoethane is highly explosive. M. O. Forster, H, E, Fierz and W. P. Joshua~ J. Chem. Soc. 93, 1070 (1908} . 8. The 15 N shift of the potassium salt of .e_-toluenesulfonamide is 266.8 ppm in dimethyl sulfoxide, If .e_-toluenesulfinate is added to a solution of chloromethyl azide in dichloromethane in the pre15 sence of I, the N·spectrum after 1 month shows a triplet (J=83 Hz) centered at 281.4 ppm. This signal probably arises from p- toluenesulfonamide. 9. J. S. Meek and J, S. Fowler, J, Org, Chem. 33~ 3422 (1968) . 10, K. Hovius and J, B. F. N. Engberts. Tetrahedron Lett. 2477 (1972). 11. Azido sulfones are deprotonated by weak bases. B,. B, Jarvis and P. E. Nicholas, J, Org, Chem. 44, 2951 (1979). 12. (a) B. Zwanenburg and J, B. F. N, Engberts, Rec, Trav, Chim, Pays-Bas 84, 165 (1965), 78 (b) A. M. Van Leusen and J. Strating, Rec, Tr~v, Chim, Pays-Bas 84. 151 (1965). gz_, 735 (1964). 13. F, Klages and K, Bott, Ber, 14. For review of the reactions of iodoazideJ seei K, Dehnicke~ Angew. Chem, Int, Ed . Engl, ,lg_~ 507 (1979), 15. ·r: 0. Griffith and R. Irving, Trans, Faraday Soc. 45, 563 (1949L 16. R. DUbgen and K. Dehnicke, Naturwissensch aften 65, 535 (1978). 17. E. Friedman~ J. Prakt, Chem, 146, 179 (1936) .. 18. P. S. Skell and J~ H, McNamara, J. Am. Chem .. Soc~ 79. 85 ~, (1957) . 19. In analogy with the reaction of .E_-toluenesulfi nate with gtoluenesulfonyl chloride perhaps the following occurs: 0 ArS02I + Arso2 ~ ArS020S-Ar 0 0 Arso os-Ar + ArS02 ..____, Arso 2 ~-Ar + Arso3 2 F. P. Corson and R. G. Pews, J. Org, Chem. 36, 1654 (1971). 20. E. Lieber and J. Ramachandran, Can. J. Chem. I!_, 101 (1959). 21. M. S. Raasch, J. Org. Chem. 35, 3470 (1970). 22. D. F. Percival and R. M. Herbst, J. Org. Chem. 22, 925 (1957). 23. G. L 1 Abb€, J.-P. Dekerkt A, Verbruggen, S. Tappet, J. P. Declercq, G. Germain and M. 24. Van Meerssche 1 J. Org. Chem. 43. 3042 (1978). E. Lieber, E. Oftedahl and C. N. R. Rao. J. Org. Chem. 28, 194 ( 1963). 79 25. Formation of the previousl_y mentioned five-membered ring heterocycles may also occur in two steps, through the cyclization of an azide intermediate, 26. P. J. Smith and K. C, Westaway~ The Chemistry of Diazonium and 11 Diazo Group~',s. Patai. ed., Interscience: New York~ 1971, p. 716 27. For review of pentazoles, see I. Ugi, Adv, Heterocycl. Chem. 3, 373 (1964)' 28. A. Holm, L. Carlsen and E, Larsen, J. Org. Chem, 43, 4816 (1978}. 29. H. Balli, Justus Liebi_gs Ann. Chem._ 647, 11 (1961). 30. H. Balli, Helv, Chim._ Acta §l_, 1912 (1974). 31. For general review of sulfonylnitrene chemistry see D. S, Breslow, 11 32. Nitrenes 1' , W. Lwowskil ed ., Interscience, New York: 1970~ Ch. 8, Free tetrazadienes (R-N=N-N=N-R) are unknown, but X-ray structures of transition-metal bound tetrazadiene complexes (R=Ar) indicate that, depending on the environment, complexes corresponding to either Xla or Xlb can be formed, P. Overbosch, G. van Koten and 0, Overbeek, J. Am. Chem._ Soc. l 02, 2091 (1980). 33. G. La Monica~ P, Sandrini, F. Zingales and S. Cenini 1 J, Organomet. Chem, 50, 287 (1973L 34. R. A. Abramovitch and T. Takaya, J, Org. Chem, 2.§_. 3311 (1973). 35. R. A. Abramovitch, T. D. Bailey, T, Takaya and V, Uma~ J, Org. Chem._ 39, 340 (1974). 36, H. Meier and K-P Ze11er, Angew. Chem. Int. Ed. Engl, 16. 835 (1977). 80 37. H. H. Gibson, Jr., H, R, Gaddy III and C, S, Blankenship. J. Org, Chem, 42, 2443 (J 977}. 38. M. M. Campbell and G. Johnson, Chem. Rev, 78, 65 [1978), 39. p_-Toluenesulfonamide, a product derived from hydrogen abstraction by triplet nitrene,should not be formed to any great extent because singlet to triplet intersystem crossing is probably inhibited by crown ether or dichloromethane, Oxygen or chlorine containing solvents have been reported to stabilize singlet ethoxycarbonylnitrene by complexing the nitrene. See~ P, Casagrande, L. Pellacani and P. A, Tardella, J, Org. Chem . 43, 2725 (1978); H. Takeuchi, Y. Kasamatsu, M. Mitani~ T. Tsuchida and K, Koyama, J. Chem. Soc. Perkin Trans,2, 780 (_1978), 40. I. Schuster, S. H, Doss and J, D, Roberts, J, Org, Chem, 4~, 4693 (1978). 41. T, Curtius and G. Kraemer, J, Prakt, Chem . 125, 323 (1930). 81 PART II 15 N AND 13 c NUCLEAR MAGNETIC RESONANCE STUDIES OF ARYLDIAZONIUM COMPOUNDS. EFFECT OF SUBSTITUENT, SOLVENT, AND 18-CROWN-6. 82 INTRODUCTION Just as macrocyclic polyethers bind a wide variety of metal cations la crown ethers have been found to complex several organic cations lb A particular interesting example is the 18-crown-6 complexation of aryldiazonium cations in which the linear N~ group is thought to insert into the hole of the macrocyle (I). 2 (0) 0 0 (c::>)( N )=N (o..J I 3 Evidence for I includes the importance of cavity size of the crown , 1 4 steric effects of ortho ring substituents , and H NMR shift changes of Recent thermodynamic work on the the ortho hydrogens 5 on comp.lexation. stability of such complexes as a function of para substituent indicates 5 that the interaction between N; and crown is predominantly electrostatic . In tact, the structure of crystalline benzenediazonium chloride might 83 serve as a model for the crown complex, The N; group is surrounded 6 by four chloride ions in a planar arrangement normal to the NN axis The cation-oxygen interaction between aryldiazonium salt and crown is strong enough to solubilize the diazoniurn salt in non-polar solvents such as chloroform, Crown ethers have recently been exploited in phase transfer of so1id ary1diazonium sa1ts to non-po1ar so7utions 7- 9 In addition to enhancing the solubility of aryldiazonium salts in non-polar solventst crown ethers also exert a strong effect on the electronic structure of the cations. This has been clearly demonstrated by alterations in the spectral and chemical properties of the complexed Bartsch and coworkers have shown that the crown~complexed 70 and photochemical aryldiazonium salts enjoy increased thermal cations. stability 11 Spectroscopic changes observed by 1H NMR 5, 13 c relative to the uncomplexed form. in the cation on complexation have been NMR 12 , infrared 12 and UV lO,l 3, This section reports on 15 N and 13 c studies of the complexation 15 of aryldiazonium salts by 18-crown-6, The N resonances of the a(Nl) and B(N2) nitrogens of .e_-tert-butylphenyldiazonium fluoborate in dichloromethane shift 5.1 ppm upfield and 1.5 ppm downfield, res- pectively, on addition of one equivalent of 18-crown-6, Addition of four more equivalents of 18-crown-6 shifts Nl upfield another 1.0 ppm and N2 downfield .4 ppm. In order to understand the origin of these crown~induced 15N changes the effects of para substituents and solvent on the uncomplexed aryldiazonium salts were a·lso examined. 84 The 15 N solvent shifts are found to be small, and the substi~uent effects are comparable to the shifts induced by 18-crown-6. In con13 15 junction with previous spectroscopic studies, the N and c results indicate that the compiexed aryldiazonium salt has more pure diazonium character than the uncomplexed diazonium salt. RESULTS Solvent Shifts The results of the 13 c and 15 N solvent study on ,!!.-butylphenyl- diazonium fluoborate (II) are presented in Table I. _ _ .A__ I 2 ~~=N I[ The effects of solvent changes on nitrogen and carbon chemical shifts of Il are not large. This is in accord with the reported NN stretching frequencies of aryldiazonium chlorides, which are also relatively 16 . Penton and Zollinger 17 have insensitive to solvent changes postulated that the aryldiazonium cation is only weakly solvated to explain the small influence of solvent changes on the rates of diazo coupling of £_-toluenediazonium cation with !i~N-dimethylaniline. The solvent effect on the rate of hetereolytic dediazoniation of 18 benzenediazonium fluoborate is also small 128 166 159.5 161. 0 112. 2 112.0 59.9 60.3 144.4 143.7 1 ,4-dioxane dimethylformamide c a Lewis basicity parameter, reference 15. 123 159.8 111. 9 59_7- 144.0 acetone d 103 160,6 111. 2 58.8 143.5 acetonitril e b 59 160.9 110.4 58,8 143.6 nitromethane Corrected for diamagnetic susceptibil ity, reference 14, 15 Upfield from external 1 t!_ H No 3 in o2o Downfield from external Me 4si. 43 160.4 110. 1 58. l 143.8 Bd dichloromethane 6C4c 6Clc 6N2b oNlb 15 N and 13 c NMR Solvent Shifts of Ila Solvent TABLE I. co U1 86 Despite the small range of chemical shifts a qualitative correlation 15 between the Ci and N2 chemical shifts of II and s , an empirical measure of solvent Lewis basicity, is observed. The terminal nitrogen shifts upfield and the Cl shifts downfield with increasing basicity of solvent. These shift changes seem to indicate some acid-base type diazonium cation-solvent interaction. The upfield shift of N2 with basic solvents is probably due to diamagnetic shielding of the terminal nitrogen by interaction with the basic solvent. The downfie1d shift of Cl is then very likely a consequence of reduced resonance interaction of Cl with the N; group because of the stabilization of positive 19 The smaller C4 and Nl shift charge on the S nitrogen by solvent changes do not show a correlation with basicity of solvent. Similarly, the NN stretching frequencies are not related to solvent basicity 16 Benzonitrile Solvent Shifts For comparison, solvent studies on the isoelectronic benzonitrile were undertaken. Whereas the chemical shift of the terminal nitrogen of II correlates roughly with the basicity of the solvent, the chemical shift of the cyano nitrogen of benzonitrile is a function of acceptor number (AN)~ an empirical measure of Lewis acidity, of the solvent 0 (Table II). The upfield shifts observed in the more acidic solvents are consistent with the formation of a hydrogen-bond between the cyano 21 lone-pair and solvent (III). 87 TABLE II. 15 N NMR Solvent Shifts of Benzonitrile a AN c Solvent p-dioxane 117 .2 10.8 acetone 117 .8 12.5 dimethylformamide 118. 5 16.0 acetonitrile 119.3 18.9 dichloromethane 119. 9 20.4 nitromethane 121. 2 20.5 methanol 121 .8 41.3 a b c Corrected for diamagnetic susceptibility, reference 14. Upfield from external 1 ~ H15 No 3 in o2o. Acceptor number, reference 20. 88 Q--cet(D········H-S m Counterion Shifts Replacement of the fluoborate counterion of a 1.2 Msolution of II in dichloromethane by chloride ion induces downfield shifts of both nitrogens. The N2 resonance changes from 58.2 to 57,4 ppm and Nl moves from 143.9 to 142.5 ppm. Infrared studies of aryldiazonium salts 1n acetone show that a change of counterion exerts a small effect on the NN stretching frequencies 22 . The NN stretching frequencies for the fluoborate are about 5 cm- 1 higher than the chloride. Larger effects on the stretching frequency have been noted for other anions 23 . The decrease in NN bond order with more nucleophilic counterions has lead several workers 22 , 23 to propose some covalent bonding between 15 anion and cation in salts other than fluoborates. The N shifts of 24 with some covalent the aryldiazonium chloride is al~o consistent interaction between diazoniurn cation and chloride ion (IV). Ar-N~6+ .s- ·°:'J=N .. ···CI 89 Evidence for the formation of transitory iodine and astatine aryldia25 zonium complexes has, in fact, recently been reported . The effect of counterion on reactivity and stability of aryldiazonium salts has been widely debated and appears to be dependent on solvent and counterion. The anion of the diazonium salt in dilute solutions of H2so 4 and HCl has no appreciable influence on the rate of thermal decomposition 26 • The counterion also has no significant effect on the rate of azo coupling in aprotic polar solvents 17 Evidence for ion pairing of aryldiazonium salts in non-polar solutions has recently been provided by Juri and Bartsch 5. Stronger ion-pairing between BF4- and p_-.!-buty1benzened1azonium cation relative to PF~ was demonstrated by kinetic and 19 F NMR studies 5• This is particularly remarkable considering the similarity between the two counterions. It appears that the interaction between an aryldiazonium cation and non-nucleophilic counterions or solvents is fundamentally different than that of nucleophilic counterions or highly nucleophilic solvents. For nucleophilic anions (e.g., chloride) the interaction appears to involve a slight geometry change of the diazonium cation (1): .. ( 1) In the same way, but to a larger degree, highly basic solvents tend to form covalent complexes with ary1diazonium salts, Spectral evidence for pyridine and dimethylsulfoxide aryldiazonium complexes have been reported by Zollinger and coworkers 27 Interestingly, aryldiazonium 90 salts decompose by a hemolytic mechanism in highly nucleo'philic solvents, 18 . Halide presumably by cleavage of the covalent complex (2) counterions also appear to induce the hemolytic decomposition of aryldiazonium salts 28 Ar-N .. ~N-X ·X Ar· ·X (2) Substituent Effects of In contrast to the small solvent effect observed, the influence isubstituents on 13 C and 0 N shifts of aryldiazonium salts is sub- stantial. 13 c studies by Olah and Grant 29 on substituted aryldiazonium salts have shown that resonance forms Vb and Ve contribute significantly to the ground state structure of aryldiazonium salts, particularly those with an electron donating para substituent. NII N+ NII N+ x x x• 'll:o 'lib :SZ:c N Ill N+ ¢ .. .. ¢ .. . ¢ The importance of the diazo resonance structures (Vb,Vc) in delocalizing 15 the positive charge of the N~ group is also demonstrated by N NMR. 15 N NMR shifts of five para-substituted aryldiazonium fluoborates in acetonitrile are listed in Table III, The Nl resonance shifts downfield with electron-donating substituents ~ as expected for an increase in diazo character (Vb,Vc). 30 91 TABLE III. 15 N NMR of Para-Substituted Diazonium Salts in Acetonitrilea Concentration(M) oN2 oNl MeO- .8 51. 9 140.5 (!!_-butyl)- .5 57.6 142.4 H- 1.1 60. 1 143.6 H0 2C- .5 59.6 144.6 02N .3 59.4 146. 1 Substituent a Upfield from external l!i H15 No 3 in o2o. 92 In addition, the Nl chemical shifts correlate well with o;, indicating strong resonance interaction of the N; group with the para substituents 31 of (Fig. 1). Both the frequency 16 and logarithm of the intensity the NN stretch of aryldiazonium salts in solution correlate with the op+ substituent constant. The increase in intensity and decrease in frequency of the NN stretch with electron-donating substituents has also been interpreted as an increase in diazo character of the N+ 2 group 16,31. The electronic character of N2 is also influenced by para substituents. N2 and A linear relationship between the n~electron population of a; has been obtained by Oja and Nielsen from a Townes-Dailey analysis of 14N NQR results 32 . The chemical reactivity of aryldiazonium salts is also strongly dependent on ring substituent. The equilibrium and rate of coupling of N2 with nucleophiles gives good Hammett plots against ordinary o values 33 • The half-wave potential for the one-electron reduction + of substituted aryldiazonium salts also correlates well with ap values Curiously, the chemical shift of N2 does not correlate with or op (Fig. 2). 34 o; The N2 resonances of salts with para groups more electron donating than H move downfield with increasing electron donation, 30 , represented by as expected for an increase in diazo character structures Vb and Ve. In fact, the N2 resonances for these aryldia- zonium salts are more sensitive to substituent than the Nl resonances of the same compounds. This can be taken as evidence for greater positive charge on the terminal nitrogen, in accord with the calculated 93 146 -0.6 Figure 1. -0.4 -0.2 0 0.2 0.4 0.6 0.8 Plot of the Nl 15 N shifts of substituted aryldiazonium fluoborates against cr+. (slope 3.83, intercept 143.2 ppm, correlation coefficient 0.979). 94 Figure 2. Plot of the N2 15 N shifts of substituted aryldiazonium fluoborates against o+. 95 electron populations of gas-phase benzenediazonium ion (ST0-3G) 35 and the X-ray structure of benzenediazonium chloride 6 and tribromide 31 The arydiazonium salts with para groups more electron accepting than H seem to lie on a line of opposite slope relative to the electrondonating substituents. As the substituents become more electron with- drawing, the N2 resonances shift downfield - a trend which is not easily reconciled with the simple picture indicated by resonance structures Va-c. The curvature in Fig. 2 must be due to another factor increasing in importance as the substituent becomes more electron withdrawing. As the para substituent becomes more electron-withdrawing it is expected that resonance structures Vb and Ve wi11 become less important and the positive charge on N2 will increase. Diamagnetic deshielding of N2 due to the increased positive charge might provide an explanation for the curvature in Fig. 2. Crown Shifts 15N and 13c shifts fnduced by crown ether complexation have been determined for several para-substituted aryldiazanium salts in dimethylformamide. The crown shifts for II have also been measured in dichloromethane (Table IV). Addition of one equ1va1ent of 18-crown-6 to the aryldiazonium salt solutions induces relatively small changes the 13 c resonances. The para carbon (C4) shifts upfield and the Cl shifts downffeld on comp1exation by 18-crown-6 for all salts examined. The crown induced i 3c changes for the !!,-butyl salt (II) in in 96 TABLE IV. 13 c and 15 N Crown Shifts for Aryldiazonium Flu~borates a oNlb oClb oC4b .5 - .5 3.2 1.2 -1.1 -4.5 2.3 1.8 -1.8 CH 2c1 2 -5.8 2.0 3.9 -2.0 DMF c -4.5 1.5 2.6 -1.3 Substituent Solvent o2N- DMF c H- DMF c -5.7 (~-butyl)- DMF c (~-butyl)- Meo- oN2b a Addition of one equivalent 18-crown-6. b A positive crown shift indicates a downfield shift on complexation. c Dimethylformamide. 97 dichloromethane are similar to those reported for the !-butyl salt 12 in the same solvent These 13 c crown shifts are consistent with a decrease in the importance of diazo-like resonance structures (Vb,Vc) upon complexation. The diminished delocalization of the positive charge onto the ring by compiexation with crown ether is also evident in the chemistry 7 of complexed aryldiazonium salts. Gokel and coworkers have shown that the nucleophilic aromatic substitution of .E_-bromobenzenediazonium fluoborate by chloride ion is stronly affected by complexation (3). ( 3) Br Cl In fact, these workers find that reaction 3 is completely suppressed in the presence of ten equivalents of 18-crown-6. Both the reactivity and spectral changes on complexation point to reduced positive charge on the para carbon in the complexed diazonium cation, The larger upfield shift of Nl on complexation also indicates a reduction of the diazo character in the complexed salt, The N2 resonance moves slightly downfield on complexation~ a. direction not consistent with increased diazonium character, unless, along with it, N2 experiences a strong diamagnetic deshielding due to the increased 98 positive charge. Additional factors responsible for the downfield shift of the N2 resonance may be the displacement of solvent or counterion by the bulky crown ether, Izatt and coworkers have studied the complexation of aryldiazonium 4 cations wtth 18-.crown-6 methanol by calorimetric titration . The extent of comp1exation was found to be highest for the para-nitro substituent and an order of magnitude higher than for the para-methoxy. For this reason, until limiting shifts are obtained, using Table IV to determine trends in crown shift as a function of substituent is dangerous. However, it is notable that all of the Cl and some of the C4 crown shifts become larger as the para-substituent becomes more electron donating; the reverse of the expected equilibrium effect. This trend suggests that the more diazo-like the uncomp1exed aryldiazonium cation, the larger the electronic perturbation, and hence shift, on complexation. Crown Complexation of Reiated Systems Several unsuccessful attempts to demonstrate crown complexation of related systems by 15 N NMR are described below. Spectroscopic 37 and chemical studies 38 of phenyl azide indicate that the azido group has a resonance effect which is electron donating, in accord with an important contribution of resonance structure Vlb. . ... lZlo 0-N-N=N lZI b 99 It was thought that the diazonium-like terminus of phenyl azide might be complexed by 18-crown-6, and the electronic changes on complexation could be detected by 15 N NMR. The addition of one equivalent of 18-crown-6 to a 2.4 M solution of phenyl azide in chlorofonn did not 15 result in any changes in the N chemical shifts. The possibility of crown cornplexation of Q_-to"luenesulfonyl azide Addition .l M solution of 15 N- was also investigated because of its N-diazonium character. of up to five equivalents of 18-crown-6 to a labeled Q_-toluenesulfonyl azide in dimethyl sulfoxide did not change the 15 N resonance positions. No 15 N changes were observed after the addition of one equivalent of 18-crown-6 to a 1.4 M solution of diazoquinone tetrahydrate (VII) in dimethyl sulfoxide, also ruling out complexation. -F\_+ O~N=N The addition of one equivalent of 18-crown-6 to a 1.9 M solution of benzonitrile in cyclohexane induced a 1.1 ppm upfield shift. 15 The same change, however, was observed after addition of one equivalent of .E..-dioxanet indicating that the crown-benzonitrile interaction is not specific. It appears that the compound must be a cation in order for compiexation by 18-crown-6 to occur. N 100 DISCUSSION 2 The 15 N and 13 c results are consistent with the Gokel-Cram model of complexation, in which the N~ group inserts into the hole of the crown ether. The crown complexation reduces delocalization of the positive charge by diazo-like resonance structures. Thus, the crown complexed salt is electronically more diazonium-like and less diazo-like than the uncomplexed form. The infrared studies of~-.!­ butylbenzenediazonium fluoborate in dichloromethane reported by Korzeniowski 12 support this interpretation. The approximately 40 cm- 1 shift to higher NN stretching frequency upon complexation indicates a higher NN bond order in the complexed aryldiazonium salt. Infrared studies of solid l :1 aryldiazonium salt/crown ether adducts also show 12 39 such a frequency change relative to the solid uncomplexed salt , . 15 N Interestingly, the magnitude of the crown-induced infrared and Nl shifts are both comparable to the effects caused by a change in para substituent. EXPERIMENTAL SECTION Spectra 15 N NMR spectra were measured at 18.25 MHz with a Bruker WH-180 spectrometer using 15-25 ml samples in 25 mm sample tubes. Lock and reference signals were provided by a concentric 5-mm tube containing 1 M95% H15No in o2o solution. All spectra were proton coupled and 3 obtained at ambient probe temperature. The accumulation of 10004000 transients (20 µs pulse, lOS repetition rate) gave reasonable 101 15 N solvent effects were determined for 1.2-1.6 signal/noise rations, M solutions of I and 1.9 Msolutions of benzonitrile. Substituent effects for aryldiazonium fluoborates were obtained for .3~1.0 M solutions of the salt. Crown shifts were measured using .5-1.5 M solutions of aryldiazonium fluoborate. 13 c NMR spectra were taken at 45.28 MHz with a Bruker WH-180 spectrometer. A concentric 5-mm tube containing t-butyl alcohol in o2o provided lock and reference signals. been converted to the Me 4Si scale. The reported shifts have The accumulation of 1000-2000 transients using 30 µs pulses gave satisfactory signal/noise ratios. Spectra of aryldiazonium salts were obtained for 1.4 M solutions at 14-15°C. Materials Commercially available spectrophotometric grade dichloromethane, nitromethane, acetonitrile, acetone, chloroform and cyclohexane were used in this study. The other solvents and benzonitrile used were conmercial reagent grade. 18-crown-6 was purchased from Aldrich Chemical Co. and used without purification. The aryldiazonium fluoborates were synthesized by standard 40 . The £~(.!:!:"butyl)benzenediazonium and p-(!-butyl)techniques benzenediazonium floborates were purified by several recrystallizations from dichloromethane/diethyl ether and dichloromethane/pentane, respectively. ether. The other salts were recrystallized from acetonitrile/ 102 A dichlorometnane solution of £_-(!!_-butyl}benzenediazonium chloride was prepared by treating the fluoborate in dichloromethane with one equivalent of tetramethylammonium chloride 41 The filtered solution was used immediately. A mixture of 1-,2-, and 3- 15 N-labeled p-toluenesulfonyl azides . 1y 42 . Phenyl azide was synthesized was prepare d as descr1 bed previous 0 using the method of Lindsay and Allen. 43 £_-Diazoquinone was obtained from £_-hydroxybenzenediazonium chloride 44 by the procedure of Puza and Doetschman 45 103 REFERENCES AND NOTES l.(a) For review see: G. A. Melson, 11 Coordination Chemistry of Macrocyclic Compounds 11 ; Plenum Press: New York, 1979. (b) G. w. Gokel, H. D. Durst, Aldrichimica Acta 1976, ~. 3 (1976). 2. G. w. Gokel, D. J. Cram, J. Chem. Soc. Chem. Commun. 481 ' (1973). 3. R. A. Bartsch, P. N. Juri, J. Org. Chem, ~, 1011 ( 1980). 4. R. M. Izat.c, J. D. Lamb, C. s. Swain, J. J. Christensen, and B. L. Haymore, J. Am. Chem. Soc. 102, 3032 (1980). 5. P. N. Juri and R. A. Bartsch, J. Org. Chem, 45, 2028 (1980). 6. C. Ramming, Acta Chem. Scand. lZ_, 1444 (1963). 7. G. W. Gokel, S. H. Korzeniowsk, and L. Blum, Tetrahedron Lett. 1633 (1977). 8. G. D. Hartman and S, E. Biffar, J. Org, Chem. 42, 1468 (1977). 9. S. H. Korzeniowski~ L. Blum and G. W. Gokel, J. Org. Chem. 42, 1469 (1977). 10. R. A. Bartsch, H. Chen, N. F. Haddock and P. N. Juri~ J. Am, Chem. Soc. 98, 6753 (1976). 11. R. A. Bartsch, N. F. Haddock and D. W, Mccann, Tetrahedron Lett. 3779 (1977). 12. S.H.Korzeniowski, R. J. Petcavich, M. M, Coleman and G. W. Gokel, Tetrahedron Lett. 2647 (1977), 13. T. Kuokkanen and P.0.1. Virtanen, Acta Chem. Scand. 833, 725 (1979). 104 14. W. Gerger, U. Mayer and V. Gutmann, Monatsch. Chem. 108, 417 (1977). 15. A. G. Burden, G. Collier and J. Shorter, J. Chem. Soc. Perkin Trans. 2, 1627 (1976). 16. L. A. Kozitsyna, B. S. Kokot and L. D. Ashkinadze, Dokl. Chem. (Engl. Transl.), 151, 567 (1963). 17. J. R. Penton and H. Zollinger, Helv. Chim. Acta 54, 573 (1971). 18. I. Szele and H. Zollinger, Helv. Chim. Acta§]_, 1721 (1978). 19. R. O. Duthaler, H. G. Forster and J. D. Roberts, J. Am. Chem. Soc. 100, 4974 (1978). 20. W. B. Jensen, Chem. Rev. 78, l (1978). 21. Protonation of acetonitrile shifts the nitrogen resonance downfield by 99 ppm. 15 G. C. Levy and R. L. Lichter, 11 Nitrogen- N Nuclear Magnetic Resonance Spectroscopy 11 ; Wiley: New York, 1979; p. 67. 22. L.A. Kazitsyna, O. A. Reutov and Z. F. Buchkovskii, Russ. J. Phys. Chem. (Engl. Transl.) 34, 404 (1960). 23. R. H. Nuttal, E. R. Roberts, and D. W. A. Sharp, Spectrochim. 24. Acta Jl., 947 (1961). For instance, the 15 N shifts of the azo nitrogens of l-phenyl-3,3dimethyltriazene are significantly downfield of aryldiazonium 15 N resonances. T. Axenrod and P. Mangiaracina, Helv, Chim. Acta~. 1655 (1976). 25. G.-J. Meyer, K. Rossler and G. Stocklin, J., Arn., Chem •. Soc., 1£1~ 3121 (1979), 105 26. F. D. DeTar and A. R. Ballentine, J. Am. Chem. Soc. 78, 3916 ( 1956). 27. (a) B. Gloor, B. L. Kaul and H. Zollinger, Helv. Chim. Acta. 55' 1596 ( 1972) . (b) H. Loewenschuss, G. H. Wahl, Jr., and H. Zollinger, Helv. Chim. Acta 59, 1438 (1976). 28. G. A. Olah and W. S. Tolgyesi, J. Org. Chem. 26, 2053 (1961). 29. G. A. Olah and J. L. Grant, J. Am. Chem. Soc. 97, 1546 {1975). 30. For example, relative to the E_-methoxy benzenediazonium salt, the N1 and N2 of diazoquinone are shifted downfield by approximately 28 and 52 ppm, respectively. 31. L. A. Kazitsyna, L. D. Ashkinadze, A. V. Upadysheva, L, E. Vinogradova and O. A. Reutov, Dokl. Chem. (Engl. Transl.) 167, 397 (1966). 32. T. Oja and L. Nielsen, J, Magn. Reson. 28, 247 {1977). 33. A. F. Hegarty in The Chemistry of Diazonium and Diazo Groups 11 11 , s. Patai, Ed.~ Wiley;New York, 1978; P~rt 2~ Cn. 12. 34. R. M. Elofson and F. F. Gadallah, J. Org. Chem. 34, 854 (1969}. 35. M. A. Vincent and L. Radom, J. Am . Chem. Soc. 100, 3306 (1978). 36. O. Andresen and C. Rommi ng, Acta. Chem. Sc and. l§_, 1882 (l 962). 37. J. Kroner, W. Schneid, N. Wiberg, B. Wrackmeyer and G. Ziegleder, J. Chem. Soc. Faraday Trans, 2 74, 1909 (1978). 38. P.A.S. Smith, J. H. Hall and R, 0. Kan, J. Am. Chem, Soc. 84, 485 (1962). 39. B. L. Haymore, J. A. Ibers and D. W. Meek, Inorg. Chem, .!.±., 541 ( l 97 5). 106 40. (a} M.F.W. Dunker, E. B. Starkey and G. L. Jenkins, J. Am. Chem. Soc. 58, 2308 (1936). (b) E.B. Starkey, "Organic Synthesis", Collect. Vol. II, Wiley, New York, 1943, p. 225. (c) A. Roe, Org. React. N.Y. ~. 193 (1949). (d) C.G. Swain and R. J. Rogers, J. Am. Chem. Soc. 97, 799 (1975). (e) The gift of E_-(t-butyl)benzenediazonium fluoborate precursor from Professor R. A. Bartsch is gratefully acknowledged. s. H. Korzeniowski and S. W. Gokel, Tetrahedron Lett. 1637 (1977). 42. c. Casewit and J. D. Roberts, J. Am. Chem. Soc. 102, 2364 (1980). 41. 43. R. 0. Lindsay and C. F. H. Allen, Org. Synth. 22, 96 (1942). 44. L. 45. M. Puza and D. Doetschman, Synthesis, 481 (1971). c. Anderson and M. J. Roedel, J. Am. Chem. Soc. 67, 955 (1945).