1 .1. ... . :. if. Etelzéié. 4.2.2. . ._ . 2.? _ _. _. _ 7 v. .2 ..._V__..._:..... T . . _ _ . , an}. . ., _ .3. .3, its .5 . . _.1 z ._.;.._ r... 331?? ._..3:_ 2... .31.. .2 5R1 .F. This is to certify that the thesis entitled SELECTED SIGMATROPIC REARRANGEMENTS IN NORBORNYL SYSTEMS presented by DONALD G . FARNUM has been accepted towards fulfillment of the requirements for ”fag degeein M Major professor Date q/zq / H7, 0-7639 ABSTRACT SELECTED SIGMATROPIC REARRANGEMENTS IN NORBORNYL SYSTEMS By Glenn Richard Carlson Two specially designed norbornyl—fused sigmatropic systems were synthesized to determine the consequences of incorporating a norbornyl bond into a Woodward—Hoffmann allowed sigmatropic rearrangement. The two systems investigated were l-hydroxy—Z-norbornanone l and l,2,3,4,4a, 6,7,8,9,9a—decahydro—§yfl;2,4a,7,9a-dimethanophenazine g, both prepared from a common precursor, l-amino-Z-norbornanone hydrochloride g. 65” @G) W o 1 EZfii1;NH3Cl ’\J 0 NGOH \ N 3 @025 ’b \\ N 2 'b l—hydroxy-Z—norbornanone l was expected to undergo a thermally allowed a-tertiary ketol rearrangement. The detection of base catalyzed deuterium scrambling in l by mass spectral analysis and nuclear magnetic resonance spectroscopy verified this prediction and showed the rearrangement to be a relatively facile one. H NaOD 020 E 0 l ’h l,2,3,4,4a,6,7,8,9,9a-decahydro-syn-2,4,4a,7,9a—dimethanophenazine g was predicted to undergo a totally new type of sigmatropic rearrangement involving the simultaneous migration of two norbornyl bonds. Accordingly, the preparation of optically active g and the subsequent photochemically induced racemization of this material established the occurrence of the desired rearrangement. Other related dihydropyrazine derivatives, however, did not react analogously, probably due to the stringent stereoelectronic requirements of this rearrangement. N V e-fi” ESE )3 if N/ 2 w The synthesis of a third norbornyl system g, designed to undergo a thermally allowed l,5 sigmatropic shift, was not successful. Such isoimidazolones appear to be rather reactive towards nucleophiles. N /N @ZP—QH gs: 6E9. 0H 4, it N H i is age H Nu SELECTED SIGMATROPIC REARRANGEMENTS IN NORBORNYL SYSTEMS By Glenn Richard Carlson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry l97l 7%”? J}? To Kathy and Lisa 1'1 ACKNOWLEDGMENT The author would like to thank Professor Donald G. Farnum for his expert and imaginative professional guidance, his understanding, and his sincere friendship throughout the course of this study. Thanks also to the members of "the group” for their invaluable technical assistance and much appreciated comradeship. Finally, the author gratefully acknowledges the National Science Foundation and the Lubrizol Foundation for generous financial support received in the summer of 1969 and the academic year l970-7l. TABLE OF CONTENTS Page INTRODUCTION ............................ 1 RESULTS AND DISCUSSION ....................... 20 Section A .......................... . 20 Section B ........................... 38 Section C ........................... 58 EXPERIMENTAL. . .......................... 70 The preparation of 2-gflggfnorbornane carboxylic acid . . . . . 70 The preparation and hydrolysis of 2-engbromonorbornane-l— carboxylic acid ........................ 70 The preparation and purification of 2-keto-l-norbornane carboxylic acid Q. ............. p ......... 71 The preparation of l-amino-Z-norbornanone hydrochloride 2. . . 72 The preparation of l-hydroxy-Z—norbornanone lg ........ 73 General procedures for mass spectral recovery and analysis . . 74 Deuterium exchange in NaOD-D20 at room temperature . . . . . . 74 Sealed tube NaOD—D20 exchange. The preparation of tetra— deuterated ketol ll. . . . . . . . . . ............ 74 The preparation of dideuterated ketol lg in a pH = 9.8 buffer. 75 Hydrogen exchange of the tetradeuterated ketol with pH = 9.8 buffer. The preparation of dideuterated ketol lg. . . . . 75 TABLE OF CONTENTS (Continued) Kinetics of deuterium exchange in pH = 9.8 buffer ....... Acid catalyzed deuterium exchange using CF COZD-D 0 3 2 (sealed tube) ..................... The preparation of l,2,3,4,4a,6,7,8,9,9a-decahydro-2,4a,7,9a— dimethanophenazine (sxn- and anti- isomeric mixture) ..... The isolation of the anti—isomer ll .............. The isolation of the d,Z-§yfl;isomer ll. The preparation of the l-amino—Z-norboranone d—lO-camphor— sulfonate salt lg (diastereomic mixture) ........ The resolution of the d—lO-camphorsulfonate salts ....... The generation of optically active syn-decahydrophenazine ll. Photolysis: general procedure . . . . ............ The photochemical racemization in methanol .......... The recovery of photoracemized material ............ Kinetic run of the photoracemization in methanol ....... The attempted acid catalyzed racemization in ethylene dichloride .......................... Thermolysis runs on optically active decahydrophenazine. . A. heptane solutions with MgO .......... B. methanol solutions with MgO . . . . Page 75 76 76 76 78 78 79 79 79 8O 80 81 82 82 82 TABLE OF CONTENTS (Continued) The attempted self-condensation of l—amino-7,7-dimethyl-2- norbornanone in aqueous base ................. Other condensation attempts ................. The reaction of alkaline potassium ferricyanide with ketones (general procedure) ..................... The preparation of hexamethyl-2,5—dihydropyrazine lg ..... The preparation of 2,5-dispiropentane—3,6—dimethyl-2,5- dihydropyrazine ll ...................... The preparation of l,2,3,4,4a,6,7,8,9,9a—decahydro-4a,9a- dimethylphenazine lg (isomer mixture) . . .......... The preparation of 2,5-diphenyl—2,3,6,6-tetramethyl—2,5— dihydropyrazine g8 ..................... The preparation of hexadeuterated hexamethyl—2,5-dihydro- pyrazine £93 ......................... The attempted photolyses of hexadeuterated hexamethyl—2,5— dihydropyrazine ....................... The attempted photolyses of l,2,3,4,4a,6,7,8,9,9a—decahydro— 4a,9a-dimethylphenazine 23 and 2,5—dispiropentane—3,6- dimethyl-2,5-dihydropyrazine lg ................ Dihydropyrazine carbonium ions in FSO3H at probe temperature (general procedure) . .............. vi Page 83 83 84 84 85 85 86 86 87 88 88 TABLE OF CONTENTS (Continued) The reaction of aqueous potassium cyanate with l—amino-Z- norbornanone hydrochloride ................... The preparation of l-ureido—Z—norbornanone yia_a modified Curtius reaction ........................ Attempts to close 88 ...................... The preparation of l-ureido-7,7-dimethyl-2—norbornanone ll. . . The ring closure of l-ureido-7,7-dimethyl-2-norbornanone with NaH in ether ....................... The dehydration of £2 with thionyl chloride in chloroform— pyridine. The preparation of ll ................ The hydrolysis of ll ...................... The reaction of £8 with ethanol stabilized chloroform. The preparation of Q; ....................... The reaction of pyridine salt 81 with triethylamine. The preparation of ll ........................ REFERENCES ............................. Page 89 9O 90 91 92 93 94 94 95 96 LIST OF TABLES TABLE Page 1 Selected Norbornyl-fused Cyclic Sigmatropic Systems. . . . 12 2 Thermally Allowed Base Catalyzed l,2 Sigmatropic Shifts. . l4 3 Selection Rules for Synchronous Sigmatropic Shifts . . . . l7 4 Norbornyl Systems Predicted to Undergo Simultaneous Sigmatropic Shifts .................... l8 5 The Preparation of 2,5-Dihydropyrazines by Alkaline Ferricyanide ....................... 52 6 Dihydropyrazines in Fluorosulfonic Acid at 25° ...... 57 viii INTRODUCTION The simple carbonium ion l,2 alkyl shift, commonly called the Wagner- 'Meerwein rearrangement, is probably one of the most famous reactions in organic chemistry. This type of alkyl migration was originally discovered in 1920 by German workers investigating the camphene hydrochloride — isobornyl chloride isomerization.1a2 9 ““ Cl Since then, literally thousands of additional cyclic and acyclic systems have been shown to rearrange similarly, but to this day the Wagner-Meerwein rearrangement still remains most closely associated with norbornyl systems. In sharp contrast to the great number of carbonium ion migrations that occur in norbornyl systems, other types of norbornyl alkyl shifts analogous to the Wagner—Meerwein rearrangement are virtually unknown. Furthermore, until quite recently mechanistic theory had not progressed to the state where chemists might consider trying to coax new kinds of norbornyl 1,2 shifts in specially designed systems. It was only after the advent of modern molecular orbital theory and particularly the Woodward—Hoffmann rules that all types of alkyl shifts could be conceptually related and a rational battle plan could be drawn up for inducing a wide variety of totally new norbornyl rearrangements. The major premise of the Woodward-Hoffmann rules is tantalizingly simple: molecular orbital interactions can exert a profound influence on the direction and dynamics of chemical reactivity.3 Based partly on a correlation diagram approach in cases where elements of symmetry are present, and partly on a frontier orbital approach where such elements are absent, the Noodward-Hoffmann rules apply to a surprising number of thermally and photochemically induced transformations which can be classified into three main groups: electrocyclic reactions, cycloaddition reactions and sigmatropic reactions. A sigmatropic shift is usually defined as an intramolecular rearrangement where a a bond migrates from one end of an allylic or polyenylic chain to the other}+ R\ /R c-(c=c)n fi (c=c)n-c Several representative examples of sigmatropic shifts and their systematic names are given below. l,3 sigmatropic shift> // <__—_ l,5 sigmatropic shift “w <<_________.____.d 1,6 sigmatropic shift ®=®,®0r0 Many sigmatropic shifts do not possess any elements of symmetry on which to base a relevant correlation diagram. Thus in order to examine the molecular orbital interactions that are involved in these rearrangements it is neccesary to resort to frontier orbital theory, developed by Fukui.5 Simply stated, frontier orbital theory assumes that as two reactants approach one another through space they soon begin to interact in such a way as to mutually perturb their respective molecular orbitals. As this process proceeds a significant amount of charge transfer begins to occur, primarily through the flow of electrons out of the highest occupied molecular orbital (HOMO) of one reactant, the donor, into the lowest unoccupied molecular orbital (LUMO) of the other reactant, the acceptor. The nature of this HOMO—LUMO overlap during charge transfer determines the stereoselection rules for the reaction in question. In practice frontier orbital theory is rather easy to apply to sigmatropic rearrangements. It can be done quickly and Systematically without having to resort to sophisticated or time consuming calculations. The first problem using this approach is to clearly identify the two ”reactants“. Secondly, the assignments of which reactant is to play the role of the HOMO and which the LUMO must be made. Finally, and most importantly, the stereoselection rules are determined by the presence or absence of bonding interactions developing across the HOMO-LUMO boundary surface as the reaction progresses. Consider two reprensentative sigmatropic shifts as illustrative examples: the 1,6 shift in the heptamethylbenzenonium ion and the 1,4 shift in the pentamethylcyclobutenium ion. f; :— Li <————-——> At first glance it seems rather unlikely that there should be much difference between these two reactions. After all, both are formally Wagner-Meerwein shifts in closely related systems. As it turns out, however, there are significant kinetic differences that are successfully predicted by frontier orbital theory. The HOMO—LUMO assignments for these two rearrangements are depicted below. The nodal properties of the respective molecular orbitals are represented by shaded and unshaded regions. The 1,6 methyl shift in the heptamethylbenzenonium ion, as the HOMO- LUMO boundary surface suggests, might be viewed simply as a sigma bond reacting with a pentadienyl cation. In an analogous manner the cyclobutenium l,4-methyl shift might be conceived as the interaction of a sigma bond with an allyl cation. In both cases the cation portion is assigned as the LUMO simply because the excited state (LUMO) of the cation is more thermally accessible (lower energy) than the alternative 0* state of the sigma bond. molecular orbital levels molecular orbital levels for the pentadienyl cation for the allyl cation LUMO Now looking for maximum bonding overlap across the boundary surface, it is evident that the 1,6 shift in the benzenonium ion gives such a favorable correlation (shaded—shaded or unshaded-unshaded) when the migrating group shifts with retention of configuration. If the methyl group shifts with inversion of configuration, one antibonding (shaded-unshaded) correlation results. Since any reaction proceeds in the direction of maximum bonding overlap, the migration with retention is "allowed" while the migration with inversion is ”forbidden”. migration with retention thermally allowed migration with inversion thermally forbidden The l,4 sigmatropic shift in the cyclobutenium ion is quite the opposite. Migration with retention, a thermally allowed process in the benzenonium ion, is now a forbidden process due to the development of one shaded-unshaded correlation. Only migration with inversion is allowed. migration with retention thermally forbidden migration with inversion thermally allowed Selection rules for photochemical reactions are worked out similarly. The only difference is a photochemically-induced first excited state or ”singly occupied molecular orbital" (SOMO) takes the place of the HOMO, and thus the rules are derived from SOMO-LUMO orbital correlations. The stereoselection rules for photochemical reactions derived in this manner are usually the direct opposite of those derived for the corresponding thermal reactions. Thus the Noodward-Hoffmann rules predict that any two typical sigmatropic shifts, although being at least superficially related, may in fact proceed by quite different paths. Quite literally thousands of sigmatropic shifts can be imagined and the Woodward-Hoffmann rules can be worked out in each case using the frontier orbital approach described above or another method to be illustrated later. However, this does not mean that all these reactions do in fact exist in nature. There are exceedingly important stereoelectronic and stereochemical considerations that make many of these Woodward-Hoffmann allowed sigmatropic shifts clearly impossible. Although it was not mentioned earlier there are really two stereotopically different types of sigmatropic shifts: suprafacial shifts and antarafacial shifts. A suprafacial shift is one where the migrating group always remains associated with the same face of a particular bonding system. An antarafacial shift is one where the migrating group ”flips" from one side of a sigmatropic system to the other. All the migrations discussed so far have been supra- facial, but reconstructing the frontier orbital diagram of the thermal l,4 shift in the pentamethylcyclobutenium cation, for instance, it becomes obvious that a l,4 antarafacial shift with retention is also formally possible. \ antarafacial migration with retention thermally allowed Unfortunately this type of migration is clearly impossible in such a small ring system since the transition state for this reaction must be incredibly distorted in order to insure proper overlap. Thus this reaction would require a prohibitively high energy of activation. In fact, antarafacial sigmatropic shifts are usually quite imaginary except in certain very large floppy systems where the length of the carbon chain allows the molecule to assume a corkscrew configuration. This, however, still leaves two types of suprafacial shifts to be considered: those that proceed with inversion and those with retention. Migrations that occur with retention such as the Wagner-Meerwein rearrangement or the l,6 methyl shift in the heptamethylbenzenonium ion allow the migrating group to shift from one atom to the next while maintaining continuous bonding overlap at every stage of the reaction. In fact the transition state could be considered as a three—centered bond where the positive charge is distributed among three nuclei. The result is considerable delocalization stabilization in the transition state with a concomitant lowering of the activation energy for this process. Furthermore, retention of configuration allows the substituents on the migrating carbon to remain well out of the way during the course of the reaction, thereby making steric compression in the transition state almost negligible. All this is well in accord with existing experimental evidence which indicates that l,6 sigmatropic shifts in benzenonium ions are very fast.6a7’8 Further- more,many Wagner-Meerwein rearrangements are so facile that they require activation energies of less than 2-3 kcals/mole.9 Migrations with inversion like the one predicted for the pentamethyl- cyclobutenium cation l,4 shift do not exhibit any of the energetic advantages possessed by reactions that occur with retention. Continuous bonding overlap is severely hindered by steric interaction between the substituents on the migrating group and the face of the sigmatropic systemfh10 Thus, this reaction and other related shifts where inversion is required should have higher activation energies than shifts where retention is possible. This is in agreement with the experimental fact that methyl scrambling in the pentamethylcyclobutenium cation is uncommonly slow.11 Stereoelectronic factors can kinetically accelerate or retard a Woodward-Hoffmann allowed process. Is it possible then to construct a special sigmatropic system where bonds and bond angles are adjusted in such a way as to obtain maximum stereoelectronic efficiency? In other words, is it possible to superaccelerate an already facile sigmatropic process through structural modification? The norbornyl skeleton, so famous for its Wagner-Meerwein rearrangements, provides a unique framework that probably comes closer than any other system to maximizing stereoelectronic factors for l,2 migrations which occur with retention of configuration. Maximum overlap for this type of migration is available when reactant orbitals are parallel and coplanar; the closer to 7 this ideal situation the better. An inspection of molecular models estimates the deviation from coplanarity in the rigid norbornyl system is less than 20° if sp2 hybridization at C-2 is assumed. Likewise the bonds are very nearly parallel, this deviation being approximately l7-Zl°.12 H // // / /L_,/ H / 1/ l x ml7°-2l° For stereoelectronic reasons then the norbornyl bond is one that should be exceptionally well suited for migration. Thus if it were possible to fuse a norbornyl skeleton into a previously known sigmatropic system where migration with retention is predicted the result should be a considerable rate enhancement. Likewise such a system might make an excellent model for investigating totally new types of sigmatropic shifts since the predicted rate enhancement in norbornyl systems ought to make such new rearrangements easily detectable under relatively mild conditions. A few examples of such fused norbornyl systems and their respective Woodward-Hoffmann selection rules derived using the frontier orbital approach described earlier (assuming retention of configuration) are summarized in Table 1, Notice that practically any neutral, carbonium ion or carbanion cyclic sigmatropic shift, either thermally or photochemically allowed, can be readily incorporated into a norbornyl skeleton. None of these modified ’ a. l2 norbornyl compounds has been synthesized to date, providing a potentially fruitful area for exploratory research. TABLE 1_ Selected Norbornyl-fused Cyclic Sigmatropic Systems Rearrangement type l,4 sigmatropic shifts l,5 sigmatropic shifts l,6 sigmatropic shifts l,7 sigmatropic shifts Compound @QQQQ‘.’ EQe9@ Noodward-Hoffmann selection rules (retention) _AL. _IDL_ forbidden allowed allowed forbidden allowed forbidden allowed forbidden forbidden allowed forbidden allowed l3 Other types of sigmatropic shifts, besides the cyclic systems already mentioned, are amenable to the norbornyl system. The well known a-tertiary ketol rearrangement is a typical example. This rearrangement proceeds through an enolate anion and, quite unlike other base catalyzed carbon to carbon migrations, it has many examples in the literature.13 A : M The corresponding norbornyl system, unknown at the inception of this investigation“+ is as follows: 00 This rearrangement is thermally allowed with retention of configuration C) at the migrating center according to the Woodward—Hoffmann rules.15 This explains why this process appears to be a relatively facile one, and suggests that the norbornyl-incorporated a—tertiary ketol rearrangement might be especially fast. migration with retention thermally allowed 14 Other potential base catalyzed rearrangements of this class are listed in Table g, All are thermally allowed processes, predicted to occur with retention of configuration.1“ Like the cyclic sigmatropic systems mentioned earlier, investigation of these rearrangements might yield some interesting experimental results. TABLE g Thermally Allowed Base Catalyzed l,2 Sigmatropic Shifts Rearrangement Type Compound a—tertiary ketol rearrangement a-aminoketone rearrangement NH 0 G R/RR “.L. O H H 5% .33. 6—.— R R’ homoallyl carbanion rearrangement With a little imagination still more exotic norbornyl sigmatropic systems might be synthesized; ones that are constructed in such a way as to allow two simultaneous sigmatropic shifts. ‘_>\ \W X This type of rearrangement would certainly have more stringent stereo- electronic requirements than any of the simpler rearrangements discussed so far since many more bonds must be properly aligned before reaction is possible. Thus the norbornyl system whould be especially well suited to this new type of simultaneous shift relative to its open chain cousins. Needless to say the derivation of the Woodward—Hoffmann rules for such a rearrangement type becomes rather complex using a frontier orbital approach since so many bonds are being formed and destroyed simultaneously. Fortunately, however, there exists a simple rule of thumb for predicting the stereoselection rules for such complicated systems. The rule is as follows: thermally allowed reactions must possess an odd number of supra- facial additions to the bonds in question; photochemically allowed reactions must possess zero or an even number of suprafacial additions. In order to understand this type of approach it is necessary to slightly redefine the terms ”suprafacial" and ”antarafacial" additions as applied to single n or 0 bonds. A suprafacial addition upon a n bond means the electrons "enter“ and "leave” from the same side of the bond. Suprafacial additions to sigma bonds can arise in two ways: either by retaining configuration at both atoms (r, r) or by inverting configuration at both atoms (i, i). A supra— facial addition to a n bond is denoted n , a suprafacial addition to a o S bond, as. I 7§<:::::2::::5:§-§3 ”I, ‘\ \\\ Likewise an antarafacial addition on a n bond, denoted na, signifies electrons "entering" and "leaving" from opposite sides of the bond. In a similar manner an antarafacial addition to a sigma bond can only arise when one atom inverts and the other retains configuration (i, r). Before the selection rules for synchronous shifts can be derived using this approach, however, there is a complicating factor that must be considered. The fact that two bonds are simultaneously migrating makes three different reaction types possible: (l) the migrations may both occur with retention of configuration; (2) one may go with retention and one with inversion; (3) both may invert. The path where both migrating groups retain configuration should be the lowest in energy but, for the sake of thorough— ness, all cases except those involving impossible antarafacial migrations will be derived. These are given in Table 3: 17 TABLE 3 Selection Rules for Synchronous Sigmatropic Shifts R x E X .T;leVR ‘EE_—_____—-_ R'V’Eixj.R Photochemically allowed Thermally allowed Photochemically allowed double retention retention, inversion double inversion Notice that Table §_considers only the §yp_isomer. An analogous approach to the apti isomer leads to the same basic prediction: the photochemically allowed process should involve either double retention or double inversion at the migrating carbons while the thermally allowed reaction should occur with one retention and one inversion. Further elaboration of this reaction type leads to several new and interesting molecular rearrangements, all involving simultaneous sigmatropic shifts. A few examples of such norbornyl-fused systems are given in Table 3, TABLE 5 Norbornyl Systems Predicted to Undergo Simultaneous Sigmatropic Shifts Selection rules Compound (double retention) § L _h" ’ F forbidden allowed v: F" allowed forbidden l‘iiii T e ' F forbidden allowed g a 9 allowed forbidden \ i % forbidden allowed V fl. Thus it would appear that norbornyl systems can theoretically be incorporated into all sorts of novel sigmatropic rearrangements, not just the simple Wagner-Meerwein variety. The initial purpose of the following study was very simply to synthesize and investigate the chemical properties of three representative compounds discussed in this introduction. The three systems chosen were l-hydroxy-2-norbornanone, a compound expected to undergo a base catalyzed a—tertiary ketol rearrangement, _yp;l,2,3,4,4a, 6,7,8,9,9a-decahydo-2,4a,7,9a-dimethanophenazine, a compound that might exhibit a photochemically allowed "double—barreled" pair of sigmatropic shifts, and finally a norbornyl derivative of 2-isoimidazolone, a system that is predicted to undergo a thermal l,5 sigmatropic shift. N Q} N/ \ Kg ab” C: ‘< 0R RESULTS AND DISCUSSION Section A Although l—hydroxy—Z—norbornanone is a new compound, and one of significant synthetic challenge due to the nature of its bridgehead substitution, a quick survey of the literature reveals that certain bridgehead a—ketols structurally related to l—hydroxy—2-norbornanone are already known. Ishidate, for instance, has successfully synthesized l-hydroxy—7,7—dimethyl-2—norbornanone l starting from a readily available camphor derivative, ketopinic acid 2.16 According to this procedure ketopinic acid is converted to the respective amino—ketone l by a standard Curthm degradation. This is followed by a nitrous acid deamination of the amine to give the desired bridgehead alcohol. fiflsoc12fi%)ww3fi% coc1 )HC] NH3Cl % __.. dig Unfortunately Ishidate does not mention whether or not l rearranges in basic media. Heims able, however, to isolate this compound as a sharp 20 21 melting crystalline solid. Applequistl7 used a similar Curtius degradation sequence in the synthesis of l—hydroxy—7—norbornanone hydrate l. Curtius HONO )HCl NH3 Cl 84> Using the syntheses of these two compounds as a guide, the preparation of l-hydroxy-2-norbornanone turnedout to be relatively straightforward. The synthesis of the key keto—acid intermediate Q M5 accomplished by following established literature procedures. . Pd/C Q + L ‘1'” > EtOH COZH Br2 NaOH “mi—'9 3. ‘FSIEH 3 acid COZHO 5 '\J The Diels-Alder reaction18 and subsequent hydrogenation of the olefinic bond19 are both simple reactions that can be conveniently carried out using kilogram quantities of starting materials. The following bromination— rearrangement step, first described by Kwart,20 proceeds smoothly and can also be scaled up to handle large amounts of material. Hydrolysis21 of this bromo-acid yields three compounds: expf2—hydroxy-l-norbornane carboxylic acid Q,ex972-hydroxy-gpdp;2-norbornane carboxylic acid l and nortricyclene—l-carboxylic acid 8 in the ratio ll.5:4:l. CO H Z B Oxidation of this hydrolysis mixture using Brown's22 chromic acid method yields the desired keto—acid 8 along with 2-norbornanone and unreacted nortricyclene-l-carboxylic acid. These products can readily be separated using silica gel column chromatography. [0X] > OH co H 0 co H 5 2 I 0X ] H ——> fi+ 602 O [ 0X ] no reaction 5% NaOH Y‘ C02H COZH c02H Q _._..'-~ . 23 The resulting 2-keto-l-norbornane carboxylic acid can be converted to l-amino-Z-norbornanone hydrochloride 3 in what amounts to a one pot Curtius reaction. First,keto-acid Q is converted to the corresponding acid chloride by treatment with thionyl chloride in refluxing benzene. The crude acid chloride is then treated with sodium azide in refluxing toluene and the resulting isocyanate is hydrolyzed with concentrated hydrochloric acid to yield the amino-ketone hydrochloride in 60-65% yields. SOCl2 (mm 0 NdFO O NH§TO 2 The final step in the preparation of l-hydroxy-2-norbornanone is a nitrous acid deamination of g. This transformation can be accomplished using a large excess of nitrous acid at 60-65° -rather vigorous conditions. HONO 3 60-65° Nmo 0H 0 2 It The volatile ketol lg can be isolated in 69% yield after ether extraction and subsequent sublimation (50° at 760 mm). It is a fragrant waxy solid, m.p. (sealed tube) 142-4°, analyzing correctly as a C7H1002 compound. Its infrared spectrum is consistent with the assigned structure, having a norbornyl carbonyl at 5.68 H and a strong alcohol absorption at 2.85 p. The nuclear magnetic resonance spectrum of lg in CCl4 shows an alcohol proton singlet at 6.56 ppm (r), a bridgehead proton appearing as a broad band at 7.60 ppm (T), a two proton signal at 7.88 ppm (1) assigned as the two a-ketone hydrogens, and finally a complex six proton signal at 8.15 - 8.55 ppm (1) corresponding to the remaining protons in the system. The nuclear magnetic resonance spectrum of lg in 020 is very similar, the only differences being in the fine structure of the six proton signal at 8.15 - 8.55 ppm (1). This spectrum is reproduced below. 0H W 7.60 7.88 8.25 l l I ppm (1) 25 The easiest way to determine whether or not the Woodward—Hoffmann allowed a-hydroxy ketol rearrangement is occurring in l-hydroxy-Z— norbornanone is to determine the extent of base catalyzed deuterium incorporation. If no rearrangement occurs, lg can exchange a maximum of only two protons. DZO-base no rearrangement H 0 OH 0 Notice that the a-tertiary ketol rearrangement, however, causes the two oxygens, and thus carbons 3 and 7, to interchange roles. This makes the exchange of four deuteriums possible. 1 DZO-base \p D a-tertiary ketol OH 0 rearrangement 0H 0 The only remaining base catalyzed process that could possibly result in the exchange of more than two protons would be homoenolization, a phenomenon discovered by Nickon23 in another bicyclic ketone, 3,3-dimethyl— norbornanone. 26 strong D+ base 5 9 A H (P This process involves exo proton abstraction from carbon 6 by strong base such as potassium tertiary butoxide to form a “homoenolate" anion which can then be subsequently deuterated with inversion of configurationZ” to provide the expf6—deuterium isomer. Homoenolization in l-hydroxy—Z—norbornanone would result in the exchange of at least five deuteriums; two on carbon 3, two on carbon 7 and one on the exo position on carbon 6. Such a deuteration scheme involves a rapid prototropic shift in the homoenolate anion. One version of this scheme is represented below. rap id prototropic D we homoenolization giHé— goéj Shift D D D + a-ketone ——D——> D — w . H D D\ 0 H There is no reason to expect l-hydroxy—Z—norbornanone to exchange only four deuteriums if a homoenolization process is operative. Thus it is clearly 27 distinguishable from the desired a-tertiary ketol rearrangement. Happily, lg does in fact exchange a maximum of four protons, clearly indicating an a—tertiary ketol rearrangement is taking place. If this compound is placed in a 0.15 N NaOD—D20 solution at room temperature for a few hours the resulting recovered ketol, after washing with H20, contains mono-, di-, tri- and tetradeuterated material as determined by mass spectroscopy. Furthermore, if this exchange is carried out in a sealed tube heated at 100° for six days 84% of the ketol incorporates four deuterbms.1“ Pentadeuterated or hexadeuterated ketol can not be detected, eliminating any significant homoenolization process. The tetradeuterated ketol ll is assigned the following structure as assumed by the presence of an a-tertiary ketol rearrangement and as verified by nuclear magnetic resonance spectroscopy. The nuclear magnetic resonance spectrum of tetradeuterated l—hydroxy-2- norbornanone ll in D20 shows that the two proton signal centered at 7.88 ppm (T), corresponding to the a—ketone protons in the undeuterated ketol, is almost completely wiped out. A two proton signal centered at 8.25 ppm (T) is also destroyed, presumably that of the C-7 hydrogens. The spectrum of the tetradeuterated ketol is reproduced below. 28 rm D 7.60 7.88 8.25 I l I ppm (I) As might be expected l—hydroxy-Z—norbornanone does not exchange in neutral D20 at room temperature. Furthermore no measurable exchange takes place in 10% NaDCO3—D20 or a NaDC03-Na2CO3—D20 buffer adjusted to a pH of 9.0. A similar but slightly more basic buffer, one adjusted to pH 9.8, does, however, initiate slow deuterium exchange at room temperature (half life = 4 hours). After 288 hours under these conditions the isolated ketol is 13.9% monodeuterated, 81.9% dideuterated and significantly only 4.2% trideuterated. Thus under these milder conditions 29 a slower more selective exchange takes place, one where a maximum of only two deuteriums is incorporated into the ketol. Interestingly, all evidence to date indicates that the deuteriums in this dideuterated ketol lg are actually incorporated into the 3—gxg and 7-§yp positions, establishing, among other things, that the a-tertiary ketol rearrangement is still proceeding under these mild conditions. H D D 020 n H OH 0 OH O 1% The nuclear magnetic resonance spectrum of ketol lg in 020 indicates that only one of the two original a-ketone hydrogens has been exchanged for deuterium. A single remaining a—ketone hydrogen can readily be observed as a multiplet centered at 8.01 ppm (T). The fact that the other exchangable proton resides on C—7 is established by the loss of a one proton signal near 8.25 ppm (1), the region assigned to the C-7 methylene hydrogens, as determined from the tetradeuterated ketol. The remaining C-7 hydrogen is buried in a complex signal at 8.25 ppm (T). The exact expf§yp_deuterium stereochemistry of lg can not be absolutely verified by nuclear magnetic resonance spectroscopy alone due to the complexity of the spectrum. However, there is a great deal of circumstantial evidence in the literature to back up this assignment. The fact that norbornyl ketones preferentially exchange exo hydrogens is well known.25s26927 Tidwell28 has shown that the exo hydrogen in norbornanone exchanges 715 times -——a-. 30 Till H OH 7.60 8.91 8.25 I I ppm (1) faster than the corresponding endo proton when treated with NaOD in 2:1 dioxane-water at 25°C. In benzonorbornenone this factor increases to 1187 to 1, in norbornenone it is 593 to 1. H1 H1 H1 31 Von Schleyer29 has attributed this kinetic behavior to non-bonded interactions between hydrogens on carbons 3 and 4 which develop during protonation or deuteration of the enolate anion. H 0 “36-20 Ill When the deuteron attacks from the endo side of the molecule the C-3 proton is pushed up into an unfavorable eclipsing configuration with the bridgehead proton. The result is a relatively high energy transition state. Q) H (13)“ \De Deuteration from the exp side presumably is much faster since these two hydrogens now move away from one another in the development of the transition state, lowering the energy of activation relative to endo deuteration. 32 There is no reason to suppose the behavior of l-hydroxy-Z—norbornanone should be any different from the other bicyclo-[2,2,1] heptane systems previously mentioned in regard to the relative ease of exp exchange. Thus, this fact could easily rationalize the formation of the 3-gxp_7—§yp_dideutero isomer lg if the a-tertiary ketol rearrangement is also assumed to occur under these conditions. {I QXQ rearrangement exchange ____________€> OH H 0 O OH D exo D exchange 3 0 0H 1% The corresponding ppgpf3-d gptif7-d isomer lg can also be prepared by treating a sample of the tetradeuterated ketol with a NaHC03-Na2C03—H20 buffer adjusted to pH 9.8 at room temperature. After two weeks a mass spectral analysis of the recovered ketol indicates 88.7% dideuteration and 11.3% trideuteration. D D H D H D H20 ND OH lg As expected the nuclear magnetic resonance spectrum of this material is quite different from that of the pxp:3-d §ypf7-d isomer lg. The newly appearing exp;3 hydrogen shows up as a multiplet centered at 7.81 ppm (T) 33 while the 7-§yfl proton is observed as a multiplet at 8.39 ppm (1). This spectrum in combination with the spectra of the undeuterated, tetradeuterated and 259:3 §ypf7 dideuterated ketol enable a chemical shift assignment of all C-3 and C-7 hydrogens. H 7.81 T H 8.01 1 OH 0 The spectrum of the 3-endo-d 7-anti-d dideuterated ketol lg in D20 is J... reproduced below. m D N 7.60 7.81 8.39 l l 1 ppm (r) Fit. in. smear-us J‘ihi: [whim-aid 2. Means [om "' "m «NW1!!! to minute we HID _. 34 Thus the structure of the dideuterated ketols lg and lg eliminates at least one conceivable mechanism for tetradeuteration; one involving rapid preliminary exchange of the two a-ketone hydrogens followed by a slower a-tertiary ketol rearrangement. D D fast :> D __£EU§1;> H OH 0 H 10 'VV D D fast D ———6 D 0H 0 The establishment of the dideuterated ketol structures lg and lg, however, still does not provide evidence of the relative rates of exo deuterium exchange versus a-tertiary ketol rearrangement, a potentially valuable piece of information. In order to obtain these data it is necessary to resort to a simple kinetic experiment. A 20% ketol—pH 9.8 020 buffer solution was prepared and placed in a nmr tube. This sample was maintained at room temperature and subjected to periodic nmr integrations using the bridgehead proton at 7.59 ppm (T) as an internal standard. Importantly these integrations indicate that the signals at 7.81 and 8.39 ppm (T) decrease at the same rate within experimental error. Both protons exchange with a pseudo first -1 order rate constant of 2 x 10.3 min and a half life of approximately 35 four hours. This of course strongly implies that the a-tertiary ketol rearrangement is kinetically faster than deuterium exchange. If this were not the case, the a-ketone hydrogen signal at 7.81 ppm (1) would decrease at a measurably faster rate than that of the C-7 hydrogen. This then suggests that the following scheme for the eventual incorporation of four deuteriums into l-hydroxy-2-norbornanone is probably 0 erative. p D fast slow 0H 0 ’\/\z D slow very \ slow D“ 0 OH 12.0“ D D D fast ver slow OH Unfortunately the numerical value of the rate of the a-tertiary ketol rearrangement at pH 9.8 can not be directly measured using deuterium exchange as a probe. The rate of deuterium uptake will always be slower than rearrangement and thus will always be a limiting factor. One thing is reasonably certain however: this rearrangement is not occurring at what 36 might be termed a superaccelerated rate, at least under the relatively mild conditions employed so far. High temperature nuclear magnetic resonance spectroscopy of the undeuterated ketol lg in 10% NaOH-H20 at 100°C does not show any line broadening that would indicate a reaction that is fast on an nmr time scale. This of course does not eliminate the possibility that the a-tertiary ketol rearrangement in l-hydroxy-2—norbornanone might still be proceeding slowly in even milder media than the pH 9.8 buffer required to detect deuterium exchange. Indeed the fact that rearrangement is faster than deuterium exchange under such mild conditions implies that this reaction is still occuring at pH values where deuterium exchange would be prohibitively slow. Perhaps the rearrangement of lg does not require any base catalysis at all, but can occur by means of an uncatalyzed proton shift, which is incidentally also a Noodward—Hoffmann allowed process (20a + 20S + Zna). --a 01 Hf 0 At present it is difficult to detect such a rearrangement due to the limitations on deuterium incorporation in neutral media. Furthermore, all the available mono-, di—, tri— and tetradeuterated ketol isomers have their deuteriums symmetrically disposed among carbons 3 and 7, making them worthless as probes. l-hydroxy—2-norbornanone, however, can be shown to rearrange in acidic media; but this process probably involves a totally different reaction mechanism involving classical Nagner-Meerwein shifts induced through protonation of the ketone. 37 Wagner #3 Meerwein h'ft :> )1 (3 s 1 ii OH 0H OH H 0 lg The detection of this rearrangement is made possible through the onset of acid catalyzed deuterium incorporation. Thus, if a small sample of lg is placed in a sealed tube with CF3C020-DZO at 140° for four days the resultant ketol is 66.8% tetradeuterated. Furthermore, this tetradeuterated material is identical to the base catalyzed tetradeuterated ketol ll, as determined by nuclear magnetic resonance spectroscopy in D20. CF3CO 2D—DZO 140° OH OH 0 e—a e._l RESULTS AND DISCUSSION Section B l-hydroxy—2-norbornanone is not the only product produced in the nitrous acid deamination of l-amino-2-norbornanone hydrochloride. A trace amount, less than 2%, of another bicyclic ketonic product is formed (ir: 5.69 p), probably l—chloro—2-norbornanone. This product results if the bridgehead carbonium ion intermediate captures a chloride ion instead of a water molecule. HONO D \ H 0 G) NH3C1 1129/ /7 flfi 10 _ ’VV C1 _g§ [ As Cl 0 2&0 The best way to eliminate this bothersome contaminant is either to substitute a less nucleophilic anion such as sulfate or perchlorate for the chloride ion or to eliminate the counterion altogether and use the free amino-ketone. Ishidate16 reports that free l-amino—7,7—dimethyl-2-norbornanone is a perfectly stable compound, thus the preparation of l—amino—2-norbornanone from its hydrochloride salt g should be a simple matter. 38 39 When the amine hydrochloride is placed h1a10% aqueous sodium hydroxide solution at room temperature a rapid reaction ensues. Extraction of this solution with chloroform gives, not l-amino-2-norbornanone as originally expected, but a mixture of two stereoisomeric self—condensation products, §yp7 and £33171,2,3,4,4a,6,7,8,9,9a-decahydro-2,4a,7,9a-dimethanophenazine, lg and lg in 90-95% yield. 9 IO 1 “”3“ 10% NaOH fiNiEZ '———> @0 H20 7 6 53\N 4a L. 3 It 12 The crystalline mixture of diastereomers (mp l40-l70°) has an infrared spectrum which indicates the absence of NH or norbornyl carbonyl absorptions, showing instead a weak C=N absorption at 5.95 p. The nuclear magnetic resonance spectrum of the mixture in CCl4 shows a complex signal at 7.5—8.8 ppm (T). Only certain of these signals can be assigned with reasonable certainty: a 2H multiplet at 7.50 ppm (T) (the two bridgehead protons) and a 4H multiplet at 7.73 ppm (r) (the four a C=N protons). This spectrum is reproduced below. Fractional sublimation of this mixture at 110—130° provides a small amount of pure gptj:dihydropyrazine lg, m.p. (sealed tube) l90-2°. This material analyzes correctly as a Cl4H18N2 compound and has spectral properties almost identical to those of the diastereomeric mixture 4O ppm (i) Further sublimation at 130-150° gives the §yp_isomer 1% as the major product, apparently still contaminated with the gpti_derivative due to its melting point range of l60—170°. This material also analyzes as a C14H18N2 compound and again has spectral properties almost identical to the total isomeric mixture. The purest sample of this derivative isolated to date has a melting point of l64-70°. 41 Compounds ll and lg are not the only examples of norbornyl-fused dihydropyrazine derivatives. Applequist30 has prepared an isomeric compound lg from 1-amino-7-norbornanone hydrate hydrochloride. N _ agueous NaOH 3 ” / 16 ’\I'\; In addition Meinwald31 has recently reported the synthesis of a closely H OH NHng related norbornenyl system ll of unkwown stereochemistry by the following N basic hydrolysis ‘\ \ N 2CH3 lg reaction. NHCO Careful scrutiny of structures lg and lg reveals that these compounds are both capable of undergoing two ”double-barreled" simultaneous norbornyl sigmatropic shifts as discussed in the introduction. In the case of the meso compound lg this rearrangement is totally degenerate (15a 2 15b). However, a similar rearrangement in chiral lg, while being structurally degenerate, leads to complete racemization (14a and 14b are enantiomers). //N .—_____{;> N\\ N/ < \N 122 15a ’VVV 42 N I N $03 ‘ i @315 <—.—— \ N/ ' N Hi We The automerization of meso lg can be detected only by resorting to a labeling experiment where specifically labeled lg would rearrange resulting in deuterium scrambling. @Z3e&p 15b 'VVV The enantiomerization of lg can also be detected in this manner but a more convenient way to observe this rearrangement is to measure the racemization of optically active lg. ©UZ§ em. ”32%: $03 Fortunately the problem of preparing optically active lg and the difficulties of isomer separation and purification can be solved simultaneously. Notice that meso lg can only arise from the combination 43 of one d— and one Z— form of the amino-ketone. 0n the other hand lg can only be formed by a combination of a 3% with another d- or likewise an 1- with another Z-. Since a; and Z- forms are equally probable the net result is a normal product ratio of lg and lg, as determined by their relative rates of formation or alternately their relative thermodynamic stabilities. On the other hand if one of the enantiomers of the amino-ketone is in excess, a product distribution favoring lg results. Furthermore the dihydropyrazine lg formed is optically active. €11 bflfilfi Z meso lg N \ ::]QEEEEfi] LEEEEEi]::: :::JQEEEEED] ‘\\ H2 N Z Z optically active lg The amino-ketone can be resolved as its d-lO-camphorsulfonic acid salt lg. This material can be prepared by evaporating an aqueous solution of one equivalent of amino-ketone hydrochloride and one equivalent of dJO-camphorsulfonic acid to dryness. Alternately, this salt can be prepared by adding two equivalents of d-lO-camphorsulfonic acid to an aqueous solution of one equivalent of dihydropyrazine, again followed by evaporation. 44 (D 03SCH “=69 CHZSO3H @Nfi + d. g) H20 N In either case the resulting crude syrupy dFlO-camphorsulfonate salt can be purified by recrystallization from boiling acetone. The mixture of diastereomeric salts can then be fractionally recrystallized from acetone until materials of constant rotation are obtained. The salt with the lower optical rotation, m.p. 208—9°, [a]578 (CH3OH) + 23.9, [a]365 (CH30H) + 173°, is more convenient to work with since it is the less soluble of the two diastereomeric salts and thus is easier to obtain in a state of high optical purity. This material can be treated with 10% aqueous sodium hydroxide to yield optically active dihydropyrazine lg, [a]578 (MeOH) - 107°, [a]365 (MeOH) — 353°, exhibiting all the properties of the nonnal dihydropyrazine. This material can be sublimed as usual to give a product with a melting inint range of l60—7°, indicating the lingering presence of gagi- lg and therefore incomplete resolution of the d—lO—camphorsulfonate salt. Fortunately the presence of lg in this optically enriched §yp_dihydro— 45 pyrazine does not effect the validity of subsequent experiments and can virtually be ignored. Optically active lg is dissolved in spectral grade methanol (50-90 mg/100 ml) and irradiated with a ZOO—wattHanovia mercury lamp under a positive pressure of nitrogen at 25°. Using quartz optics, loss of optical activity in lg is greater than 90% complete after thirty hours when the sample is placed 1% inches from the middle of the mercury lamp. No loss of optical activity occurs with Pyrex or Corex filters after a similar irradiation time. Furthermore the use of a Vycor filter results in a trivial loss of activity. This would imply that ultraviolet light with wavelengths somewhat less than 260 nm is required to induce reaction. The observed loss of optical activity implies, but does not alone prove, that racemization has occurred. Several processes other than racemization can give the same result. Certain photoreactions that are known to take place with C=N bonds such as photodimerization or photo— reduction32 could easily account for the observed loss of activity. Even complete photodestruction of lg is possible. Such processes, however, can be eliminated by two simple experiments. Firstly, if the characteristic shape and intensity of the ultraviolet spectrum of lg in methanol (Amax 208 shoulder, e = 2.2 x 103; A 250, 2) max 9 = 2.5 x 10 is continuously monitored as the quartz tube photolysis proceeds, no changes in shape or intensity can be observed, indicating the absence of any significant photoreactions other than racemization. Secondly, and more importantly, partially racemized lg can be recovered after incomplete reaction. Following normal purification by sublimation this material exhibits an optical rotation essentially identical to that of the total photolyzed 46 solution. Furthermore this material is in every other way identical to the starting dihydropyrazine. Thus it would appear that optically active lg is actually racemizing during photolysis. But is this racemization actually an excited state reaction as predicted? Are there alternate ground state mechanisms that might also result in racemization? Formally at least there are alternate explanations, but these too can be effectively eliminated by further experimentation. Blank solutions of lg can be prepared and subjected to every experimental condition short of photolysis itself; the result being absolutely no loss of optical activity. This rules out any simple solvent or thermally induced rearrangement process. Ground state acid catalysis resulting from the possible photogeneration of carboxylic acids during the photolysis can also be discarded as a viable possibility. The racemization of lg still proceeds at its normal rate in the presence of acid scavengers like sodium carbonate or magnesium oxide. Furthermore lg does not rearrange in the presence of toluenesulfonic acid. Solutions of lg in ethylene dichloride in the presence of one or two equivalents of toluenesulfonic acid give a small initial loss of activity at 50° but then remain unchanged. This result can be attributed entirely to trace amounts of water entering the system causing hydrolysis of lg, giving the respective amino-ketone toluenesulfonate salt. Since this salt has a considerably lower specific rotation than its parent dihydropyrazine the result is an apparent loss of optical activity. This salt can be detected at the end of this partial hydrolysis as a crystalline precipitate (1:: 5.7 u)- Furthermore if extensive precautions are taken to exclude all traces of water, absolutely no loss of activity is detectable. we i—e 643$ 44 me +> obs Rearrangement of lg might possibly be accomplished through chemical interaction with other types of photogenerated products besides carboxylic acids. However, this hypothesis can also be rejected since the photo- chemical racemization is cleanly first order in starting material over several half lives. If any catalysis were occurring the resultant kinetics would exhibit a dependence on catalyst concentration and thus deviate from a pure first order plot since catalyst concentration would almost certainly be time dependent. Thus it would appear that lg actually is undergoing a photochemically induced "double-barreled” sigmatropic shift as predicted by the Woodward- Hoffmann rules. As is always the case in such situations, however, this does not necessarily mean that this rearrangement is a concerted (one step) process. Certainly the same rearrangement might also be accomplished in several steps through a photochemically induced triplet intermediate. 43 > —-> ('1 Assuming that this reaction is concerted and does obey the Woodward- Hoffmann rules, however, there still remain six different photochemically allowed paths, either with double retention at the migrating centers or double inversion. Unfortuantely it is impossible to tell whether migration with inversion or retention is actually occurring in lg, since the migrating norbornyl bond is symmetrical. For stereoelectronic reasons it might be assumed that one of the double retention mechanisms discussed in the introduction would be more probable, but this is difficult to believe with absolute certainty when dealing with such a high energy photochemical process. Dihydropyrazine lg does not racemize thermally. Optically active lg can be routinely sublimed without racemization at temperatures between 110° and 150°. Methanolic solutions of lg can be stored indefinitely at room temperatures. Finally, solutions of optically active lg can be placed in a nitrogen filled sealed tube heated to 110° for four days without any appreciable loss of activity, and what small loss there is can be attributed soley to the normal thermal decomposition of this material. Prolonged thermolyses at higher temperatures results in the total destruction of lg. 49 The addition of acid scavengers such as magnesium oxide retards but does not halt this thermal decomposition. Presumably the thermal rearrangement does not occur at these temperatures because all the thermally allowed Woodward-Hoffmann processes as described in the introduction involve one migration with inversion and the other with retention, a stereoelectronically and energetically unfavorable situation. At this point it is important to ask whether or not other dihydro- pyrazine derivatives might also undergo this new type of photochemical rearrangement. The gppifnorbornyl dihydropyrazine lg, as already mentioned, is also predicted to rearrage by the Woodward-Hoffmann rules and would thus make an interesting comparative compound. Unfortunately certain practical difficulties in synthesizing and accumulating pure deuterium- labeled lg make this compound extraordinarily elusive. Perhaps then the gp317dihydropyrazine lg derived from l-amino-7,7-dimethy1~2-norbornanone would make an attractive substitute compound. Unlike lg, it is logical to assume that the gppirisomer of this dihydropyrazine w0uld predominate over the pyp_isomer for simple steric reasons. Furthermore this compound would presumably possess the additional advantages of being readily accessible in gram quantities and having its own built-in label, the two gem dimethyl groups. " NH2 A / ‘l‘illll O major product 12 50 ea 2 at" )< .3 N N 12 The only problem with this idea is that 1-amino—7,7-dimethy1-2- norbornanone absolutely refuses to self-condense. Even extraordinarily vigorous conditions such as the use of excess sodium hydride in refluxing toluene are utterly useless in effecting this conversion. Presumably l-amino-7,7-dimethy1~2—norbornanone is unreactive toward self-condensation because of the presence of the large gem-dimethyl groups which effectively increase the bulkiness of the attacking amine nucleophile while simultaneously hindering nucleophilic attack on the ketone. Other more simple 2,5-dihydropyrazine derivatives that might also make good model systems are known. Most often these compounds are synthesized by the sequence summarized below.33,3” 0 R H 0 BY‘ 0 H2 ——> ———> - R R' R ' R R' N R / R / 51 Most of these preparations, however, suffer from severe practical limitations, especially in the case of unsymmetrical ketones where tedious isomer separations and low yields are common. As a result a—amino-a,a-dialky1 ketones and their reSpective hexaalkyl-2,5—dihydropyrazine derivatives are reasonably rare compounds in spite of their structural simplicity.35 Fortunately several valuable dihydropyrazine derivatives can be eagiiy synthesized by modifying an old and curious reaction discovered by Conant36 in 1928. This reaction involves a one-step base catalyzed oxidation of a,a-dialkyl ketones with warm excess aqueous potassium ferricyanide. The single case reported by Conant involves the oxidation of 3-methy1-2-butanone to hexamethyl-2,5-dihydropyrazine gg. CH 0 K3[F€(CN)6] CH N CH3 H NaOH-H 0 3 / 2 H3 0 / 80 CH3 // CH3 €53 Substituting concentrated ammonium hydroxide for the sodium hydroxide used by Conant and working with the strict exclusion of atmospheric oxygen significantly improves the yields and makes this reaction applicable to the synthesis of a wide variety of a,a-dialkyl-a-amino ketones and their derived dihydropyrazines. A partial list of compounds that can be synthesized using this reaction is given in Table §3 52 TABLE g The Preparation of 2,5-Dihydropyrazines by Alkaline Ferricyanide Ketone Dihydropyrazine Yield N 0 // L_ 377 > /\ ° N 20 'V'b N OIL / // 57% N 21 ’Lr’h N // 42% // N 22 ’VU Ph N 0 // Ph _.IL_< 51% / N h 23 '\/\I 53 Notice that only tertiary sites a to the ketone are attacked. Primary and secondary sites are essentially inert under these reaction conditions. It is difficult to comment on the mechanism of this transformation simply because relevant data are not available. The immediate source of the amine nitrogen is particularly puzzling, although the ferricyanide ion itself is surely the ultimate nitrogen source since the reaction still proceeds, albiet in lower yields, in the absence of any other potential nitrogen donors such as ammonium hydroxide or atmospheric nitrogen. Since this reaction is markedly base concentration dependent37 and ferricyanide usually oxidizes by abstracting a single electron from the most electron rich site in a substrate molecule,37 it is reasonable to assume this reagent would attack the most highly alkyl substituted enolate anion as the first step in the reaction. After that almost anything could happen. The question of mechanism aside, this reaction provides some interesting model compounds for detecting possible simultaneous sigmatropic shifts. Among these is hexamethyl-2,5-dihydropyrazine itself. The rearrangement of this 54 compound, if it occurs at all, should be detectable by measuring deuterium scrambling. CD N CH3 CH3 N H3 3 //' \\ /' firm hv 3 co3j: :CH CH < \ 3 IN» CD CH3 N CD 3 CH3 3 3 gm?» With this in mind, hexadeuterohexamethyl-2,5-dihydropyrazine hexahydrate is prepared by refluxing g9 in a 0.15 N NaOD-020 solution. Anhydrous $93 is then obtained by storing the hexahydrate over P205. CH CH N CH3 CD 'N 3 3 // CH 3 // H 3 NaOD > 3 ~6020 CH / D 0 CH3 / 3 2 N D N CH3 CH 3 CH3 3 32 \/ CD3 /N “3 H3 CH3 / \ N CD CH3 3 20a ’\/\J’\) Photolysis of the anhydrous hexadeuterated material under several types of irradiation conditions does not induce deuterium scrambling. The use of a 550-watt mercury lamp for periods of up to 90 hours only results in extensvie decomposition of the starting material as judged by nuclear magnetic resonance. Photosensitization with acetone or benzophenone does 55 not help. Thin layer chromatography of the total photolyzed product shows at least five products in addition to unreacted starting material. Similarly, the photolysis of the other dihydropyrazines listed in Table §_also fails to promote rearrangement. In fact these compounds are almost completely inert to ultraviolet radiation, except under the most vigorous conditions, where they simply decompose. N N /\_/‘ hv \ \Nj 22 El m N Ph N Ph \\ N// Ph 2% This result, while certainly being a disappointing one, is not totally unexpected considering the extreme stereoelectronic requirements of such an extensive rearrangement. Presumably of all the dihydropyrazines mentioned only the norbornyl system has just the right rigid stereochemistry to allow this rather exotic sigmatropic shift to occur. The dihydropyrazines in Table §_also refuse to rearrange by thermal activation or by acid catalysis, a result to be expected viewing the similar behavior of Ti. Hexadeuterohexamethyl—2,5-dihydropyrazine £93 for instance can be sublimed unchanged through a thermal barrier heated to approximately 225°. In addition all the dihydropyrazines listed in Table 5 including lg 56 form stable carbonium ions in fluorosulfOnic acid at 25° which surprisingly show no signs of rearrangement. These spectra involve either static diprotonated or rapidly exchanging monoprotonated dihydropyrazines since the carbonium ion spectra are consistent only with symmetrical structures. N69 H N N é..— / / @/ 1‘6) N N H l H CUP ab (1ij The nuclear magnetic resonance spectra of these dihydropyrazines in fluorosulfonic acid at 25° are given in Table g, 57 TABLE §_ Dihydropyrazines in Fluorosulfonic Acid at 25° nmr Compound of unprotonated species (1) nmr of carbonium ion (r) /N 12H singlet (8.69) 12H 6H singlet (7.97) 6H / N 6H singlet (8.01) 6H 16H multiplet (8.20) l6H 3H singlets (8.69 & 8.62) 3H 16H multiplet (7.50-8.50) 4H 12H (syn and anti) Ph / l2H singlet (8.46) l2H / 10H singlet (2.60) 10H N Ph 23 ’L’b 2H multiplet (7. 50 ) 6H 4H multiplet (7 7) l2H multiplet (7. 8- 8. 8) l2H l + 1 (mixture) WW ’1; singlet (8.04) singlet (7.24) singlet (7.27) multiplet (7.4-7.8l) singlets (8.06 and 7.99) multiplet (6.97) multiplet (7.55 - 7.90) singlet (7.79) multiplet (2.l8) pseudo-singlet (7.00) multiplet (7.40-8.40) RESULTS AND DISCUSSION Section C The preparation of a norbornyl-fused cyclopentadienyl system which is expected to undergo a Woodward-Hoffmann allowed thermal l,5 sigmatropic shift with retention of configuration represents a significant synthetic challenge. Perhaps a more convenient system to start with, considering the availability of certain intermediates already accumulated during the preparation of l—hydroxy-2-norbornanone l9 and dihydropyrazine 1%, would be a heterocyclic analog of this compound such as 2%. 63 :3 83 Although such isoimidazolones are rather rare,39 and no examples of thermal l,5 alkyl shifts have been observed in any of the few known derivatives of this heterocyclic system, there is good reason to assume that the desired degenerate l,5 sigmatropic shift will occur since the isoimidazolone ring is isoelectronic with cyclopentadiene and thus subject to the same Woodward-Hoffmann considerations as the hydrocarbon.3 58 59 One way to construct 23 utilizing available starting materials would be as follows: Cthfl NHC0NH2 H % (it -—-a fife \\O 0 0H 66» 62 H N -——9 @>:o :2 6E?” ENE 62% This scheme involves the preparation of cyclic hydroxy urea 25 as a key intermediate which, after dehydration, would give the parent heterocycle (shown in its two tautomeric forms). Using the classical method for the preparation of ureas from amines, l—amino—Z-norbornanone hydrochloride is slowly added to an aqueous solution of potassium cyanate at room temperature. Surprisingly, two isomeric ureas are formed under these conditions: both 25 and 26. The majority of amino— ketone hydrochloride, however, reacts to form dihydropyrazines 1% and 15, since potassium cyanate is considerably basic, allowing the rapid self- condensation of the free amino—ketone to successfully compete with urea formation. (3 GD NHC0NH NH3Cl 2 H __Jg!fll_3> + :>::0 + 14 + 15 H20 H '\/b «A. 0 0H 8 £8 E 60 l-ureido-2-norbornanone 2g exhibits a norbornyl carbonyl in its infrared spectrum at 5.68 H as well as a very complex signal between 6.00 H and 6.50 N characteristic of mono-alkyl ureas.“° Its nuclear magnetic resonance spectrum in dmso-d6 shows two sets of NH protons in the ratio of 2:l at 4.57 ppm (r) and 3.87 ppm (r) respectively. The closed urea 22 does not show a norbornyl carbonyl. Instead, it exhibits two peaks at 5.82 H and 6.04 p. In addition, the nuclear magnetic resonance spectrum of this compound in dmso-d6 clearly shows three one proton signals at 3.10 ppm (T), 3.45 ppm (1) and 4.66 ppm (r), presumably being the two NH and the one OH protons. Unfortunately, the reaction of potassium cyanate with l-amino-2- norbornanone hydrochloride is not a reliable source of ureas 22 and 22. Yields under a variety of conditions could never be elevated above 25-30%. Furthermore, the ratio of 22 to 22 is rather variable. Sometimes the closed urea represents almost half of the total urea fraction and other times it can not be isolated at all. A much better way of preparing l-ureido-2—norbornanone 2g is through a modified version of the Curtius reaction used to prepare the amino—ketone hydrochloride. Thus, if gaseous ammonia is slowly bubbled through a benzene solution of l-isocyanato-Z-norbornanone, an exothermic reaction yields 22 in 76% yield. =C=0 NH NHC0NH2 With reasonably large amounts of 26 available, the next problem is to find a way to convert this compound into its ring-chain tautomer. Interestingly, 22 does not close under the reaction conditions where 22 61 is formed. This would seem to indicate that the closed urea isolated in the potassium cyanate reaction does not arise from the open-chain isomer, but from some other intermediate. One mechanism that might account for this involves attack by cyanic acid or cyanate on the carbonyl group of l-amino- 2-norbornanone rather than on the amino group as expected.‘+1 Subsequent intramolecular nucleophilic attack on the intermediate hydroxy isocyanate 0—. @310; 25 ’\/\1 by the amino group would provide 22. CE <—> 0 This brings us to the interesting question of the stereochemistry of om- 22. Only one of the two possible epimers of 22 is formed; the question is which one. Although conclusive spectral evidence is not forthcoming, it seems likely that the closed hydroxy urea possesses structure 222. 0 '> ///A\\\ SHHE’ 53 NH H H \ / H H0 25a N’D’b H H H NH /N : ‘ H __ HN 25b 0 ’b'Vb 62 The reasons for this assumption are twofold. Firstly, norbornyl ketones are known to be attacked by nucleophiles and reducing agentszs’fi‘2 predominantly from the egg side of the molecule to yield gng9;alcohols. Secondly, 222 possesses a less strained gisrfused cyclopentane-—heterocycle ring junction, making this epimer both more thermodynamically favorable and more kinetically accessible than the transyfused 222. More vigorous attempts at closing l-ureido-2-norbornanone, both acid and base catalyzed, result only in the quantitative recovery of starting material. Thus 22, while being a very attractive compound for the preparation of a norbornyl-fused isoimidazolone system, is too difficult to obtain in large enough quantities to be synthetically practical. A suitable model compound was therefore sought in order to circumvent some of these difficulties. Luckily, such a model compound is available in l—ureido-7,7-dimethyl- 2-norbornanone 22. Easily prepared from camphor in several steps via ketopinic acid 2, this compound can be synthesized in gram quantities in a matter of days.”3a”“aN5 HCONH =C=0 2 C02“ Curtius N , NH3 > se uence q E 0 0 Z (Z 22 possesses many of the spectral properties exhibited by its norbornyl counterpart. The infrared spectrum of this material shows a norbornyl carbonyl at 5.70 p, and a typical complex signal from 6.00 u to 6.50 N 63 attributable to the mono-alkyl urea moiety. The nuclear magnetic resonance spectrum of 22 in CDCl3 shows two non-equivalent methyl groups at 8.75 ppm (T) and 9.l9 ppm (r) and two sets of NH protons in the ratio of 2:1 at 4.87 ppm (r) and 4.32 ppm (T). l-Ureido-7,7-dimethyl-2—norbornanone, unlike its norbornyl cousin, can be successfully closed. This can be conveniently accomplished in 55—80% yields by treatment with sodium hydride in refluxing ether. Presumably this reaction involves intramolecular nucleophillic attack of the ureido anion on the carbonyl group.”6 An analogous reaction appears to be the ring closure of a related sulfonamide 22 with sodium ethoxide in ethanol.‘+7 NHCONH2 l) NaH/ether If 0 2) H20 / 0 on H 27 29 '\/\J ’Vb HZSOZNHZ NaOEt/EtOH \> /’ 02 —H20 N RN 800 The structure of the closed urea 22 is verified by spectral evidence. Its infrared spectrum gives a carbonyl, slightly split, at 5.90 N and 5.95 H. A nuclear magnetic resonance spectrum of 22 in dmso—d6 shows a 6H singlet at 9.03 ppm (r) corresponding to the two methyl groups. In addition three distinct one hydrogen signals at 3.22 ppm (T), 3.32 ppm (T) and 4.69 ppm (T) indicate the presence of two non-equivalent NH and one 0H protons. 64 Interestingly, if pyridine is used as an nmr solvent the methyl groups are resolved into two three hydrogen singlets at 8.80 and 8.99 ppm (1). The direct relationship between 22 and 22 is confirmed by the fact that 22 opens in dmso-d6 when mildly heated to give 22. This observation tends to rule out the occurrance of any extensive rearrangement during the base catalyzed ring closure reaction. Presumably 22 possesses the same egggfhydroxy stereochemistry as that of its norbornyl relative, but such an assumption is less convincing in this case since nucleophiles”8 and reducing agents“2 prefer to attack 7,7—dimethyl-2-norbornanones from the sterically less hindered eng9_side of the molecule, yielding engalcohols. Fortunately, however, the question of stereochemistry is not crucial from a synthetic standpoint. Finding a way to dehydrate 22 is the major concern. fine ——>? tithe 2 0H 30 ’\/\1 '\;’\J Dicyclohexylcarbodiimide in refluxing pyridine, an extremely potent but rather bulky dehydrating agent, fails to promote the desired reaction. Under these conditions 22 slowly opens up to give 22 which is inert to further reaction. Treatment of 22 with thionyl chloride in chloroform- pyridine at 0°, on the other hand, results in quantitative dehydration. The isolated product, however, is not a neutral species as expected but a pyridine salt 22. 65 soc12 >: o —————% Pyridine salt H CHC13/Pyr‘idine 31 mm 29 «A. This extremely water sensitive material shows a carbonyl at 5.84 H and several peaks indicating the presence of a pyridine nucleus (6.00 and 6.55 H)- The extreme insolubility of 22 in all the usual nmr solvents except dmso—d6 made obtaining a satisfactory nuclear magnetic resonance spectrum impossible. Using dmso-d6 solves the solubility problem, but, despite considerable care to exclude moisture, the pyridine salt is rapidly hydrolyzed in this hydroscopic solvent. Interestingly, the mass spectrum of 22 shows a parent peak at m/e 179, which is a cationic species representing the loss of pyridine and chloride ion. Several satisfactory structures can be drawn for 22, but the simplest and most likely structures are 222 or 222, related to one another by a norbornyl l,2 shift. [;g] CN3 3] 31 b ’Vb ’VVL Whatever the true identity of 22, it is fairly obvious that only one major pyridine salt is produced. The formation of significant mixture;of 222 and 222 can be ruled out by the observation that crude 22 has a fairly sharp 66 melting point (l4l-5°). Likewise the possibility of a coincidental 1:1 mixture of 2g and pyridine hydrochloride can be discarded on the same grounds. As is implied by the difficulty encountered in the preparation of nmr samples of 22 in dmso—d6, this salt rapidly hydrolyses in H20 to give a mixture of three cyclic hydroxy ureas and pyridine hydrochloride. To date, only the minor product has been separated and positively identified. This product is 22. The infrared spectra of the remaining two products verify that they too are closed ureas (1:: 5.90 u). Since three products are formed at least one of these must arise from some sort of skeletal rearrangement during the reaction, possibly a l,2 Wagner-Meenwein shift in the intermediate carbonium ion. 2 H 0 H 2 closed 0 3L :> g>==0 + urea + [::«:] N N isomers N LO Q I I ('3 —l @NH Apparently, the anticipated isoimidazolone 2g is very susceptible to nucleophilic attack, readily adding pyridine to the carbon-nitrogen double bond during attempts to dehydrate 22. This observation is not without 67 precedent. Biltz“9 has observed that a related heterocycle, 22, is readily attacked by ethanol. 30> mag; n\ NH EtOH 3 HN NH OEt Et OEt Ph Ph Isoimidazolone 22 also behaves similarly as judged by a reaction taking place in the related cyclic norbornyl urea. If 25 is recrystallized from chloroform, a new compound, 22, is occasionally isolated, which appears to result from the elimination of water from 22 followed by the re—addition of ethanol, present as a stabilizer in the chloroform, to the resulting dehydrated urea. If the arguments used to assign the stereochemistry of ureas g5 and 22 are correct, 2% should possess an exg_ethoxy group. 68 2% EtOH CHCl 3 0 ————> o , A \ NH N 24 mm EtOH E > 0 N Finally, the attempted reaction of the pyridine salt 22 with triethylamine in the vain hope that this bulkier amine would force the elimination of pyridine hydrochloride while being too sterically hindered to re-attack the generated isoimidazolone gives only quantitative amine exchange to form a new salt 22 and free pyridine. The spectral properties of 22 are quite reminiscent of the pyridine salt. The infrared spectrum of this compound shows a carbonyl absorption at 5.85 p, while the nuclear magnetic resonance spectrum in dmso-d6 shows two non—equivalent NH signals at 3.3l ppm (T) and 4.58 ppm (r) as well as methyl signals at 9.00 ppm (1) and 9.ll ppm (r). Absorptions corresponding to the triethylamine moiety are also present. Like 22, the exact structural assignment of 22 can not be made based solely on existing data. Thus, the question of the thermal l,5 sigmatropic shift in heterocycles 22 or 22 remains an open one. The tendency of such norbornyl—fused isoimidazolones, to undergo nucleophilic attack presently represents a 69 significant stumbling block in the successful syntheses of these compounds as stable species. Furthermore, there does not appear to be enough evidence at present to decide whether or not the desired thermal rearrangement is occurring in reactions where 22 or 22 are believed to be transient intermediates. Ultimately, these compounds will have to be isolated or at least be capable of direct observation before any conclusive statement about their rearrangement rates can be made. EXPERIMENTAL The preparation of Z-endgrnorbornane carboxylic acid Z-ggggfnorbornene carboxylic acid was prepared from cyclopentadiene and acrylic acid according to the procedure described by K. Alder and J. Stein, 522:, §lfl, l97 (l934). It was distilled between 92-6° at 0.5 mn (stench). 2-eflggfnorbornane carboxylic acid was prepared by catalytic hydrogenation of the cyclopentadiene-acrylic acid adduct. Using a modified procedure described by 0. Diels and K. Alder, Ann,, ggg, 98 (l928), l50 mls of unsaturated acid was dissolved in an equal volume of 95% ethanol. This solution was hydrogenated over 3 grams of 5% Pd/C at hydrogen pressures of 20-50 p.s.i. Filtration and evaporation of the ethanol yielded pure 2-eng9; norbornane carboxylic acid in quantitative yield. This material could be used directly in the next reaction. The preparation and hydrolysis of Z-exgrbromonorbornane-l-carboxylic acid. 2—engbromonorbornane—l-carboxylic acid was prepared following the procedure of H. Kwart and G. Null, J. Amer. Chem. Soc., g1, 2765 (l959). Using this procedure 400 to 600 gms of product could be produced from 640 gms of starting material (40-60% yield). 70 71 This bromo acid was then be hydrolyzed in 5% NaOH according to the method developed by Vaughan et.al., J. Amer. Chem. Soc., gl, 2204 (1965). Normally 285 gms (l.3 moles) of bromo acid was dissolved in 2000 mls of 5% NaOH solution and placed on a steam bath for 22—24 hours. At this point the solution was cooled, acidified and-extracted with ether to yield between l75 and 200 gms of product mixture (87-99% yield based on C8H1203)‘ The preparation and purification of 2-keto—l-norbornane carboxylic acid 2. Using Brown's method for alcohol oxidation (H. C. Brown, J. Amer. Chem. §gg,, §§, 2952 (l96l)), 3l2 gms (2.0 moles) of hydrolysis mixture was dissolved in 800 mls of anhydrous ether. A chromic acid solution, prepared by mixing 200 gms of NaZCr207-2H20 with 150 mls H2504 and diluting to l000 mls, was slowly drop added while maintaining a temperature of less than 35° with an ice bath. The whole process was carried out in a three—necked flask fitted with an efficient reflux condenser, mechanical stirrer and thermometer. After the addition of the chromic acid was complete the reaction mixture was allowed to stand at room temperature for two hours, then extracted with ether to provide 230 gms of product mixture. The product mixture was separated into three components as follows: 2-norbornanone was first removed by simple basification of the product mixture and subsequent ether extraction. The aqueous layer was then acidified and extracted with ether to yield a mixture of 2-keto-l—norbornane carboxylic acid 2 and nortricylene—l—carboxylic acid which was readily separated on a silica gel column eluted with chloroform. The keto acid was the second fraction and could be recrystallized from hexane-ether to yield 6l.3 gms of pure material. 72 The preparation of 1-amino-2-norbornanone hydrochloride 2. 10.0 gms (0.065 moles) of keto-acid 2 was dissolved in 20 mls benzene. 10 ml of thionyl chloride was added and the mixture refluxed for four hours. The benzene and excess thionyl chloride was distilled off and the acid chloride used without further purification. If purification of the acid chloride was desired, however, simple distillation at 75° under a pressure of 0.5 mm yielded 10.5 gms (93.7% yield) of water white oil. j__(neat) 5.55 H (C001) and 5.73 H ( C=0); nmg (CC14) T 7.27 ppm, br. singlet, 1H, (bridgehead); remaining signals as a very complex multiplet 8.0-8.7 ppm. The acid chloride was dissolved in 30-50 mls of benzene or toluene and placed in a 200 ml round-bottomed flask. 6.0 gms of NaN3 freshly activated with hydrazine-hydrate and recrystallized from acetone-water (Org. Reactions, V01. 3, pp. 382 (1946)) was slowly added behind a safety shield at room temperature with rapid magnetic stirring. After the addition was complete the mixture was cautiously heated until nitrogen evolution started. The reaction mixture was then allowed to reflux overnight. The excess NaN3 and NaCl was filtered off leaving a solution of the isocyanate behind. Nfis solution was then placed in a 200 ml round—bottomed flask, 40 mls of concentrated HCl was added and the two phases were rapidly stirred with subsequent evolution of carbon dioxide. The layers were separated, the aqueous layer washed several times with ether and then evaporated leaving 4.0-7.3 gms of reasonably pure 1-amino-2-norbornanone hydrochloride. Average yield = 67% based on the keto—acid. This material could be recrystallized from large volumes of boiling acetonitrile or sublimed at 150° under aspirator pressures to give a white 73 powder with a melting point (sealed tube) of 225°. jr_(Nujol) 3.5-4.4 u, (Nch), 4.96 a (Nch) and 5.66 p ( :c=o); fig; (020) r 7.71 ppm, br. singlet, 1H,(bridgehead); 7.93—8.62 ppm, complex signal, 8H. Anal. Calcd for 07H12N001: C, 52.05; H, 7.31; N, 8.67; Cl, 21.93. Found: C, 52.34, H, 7.33; N, 8.59; C1, 22.11. The preparatimiof 1-hydroxy-2-norbornanone 22. 5.0 gms (0.03 moles) of the amine hydrochloride was dissolved in 75 m1 of 2.5N H2504 and heated to 60-65°. 10 gms of NaNO2 was dissolved in 26 m1 of H20 and very slowly drop-added with vigorous magnetic stirring. At this point a considerable amount of bubbling and evolution of brown nitrogen oxide occurred, requiring a gas trap or hood. The addition of the NaNO2 solution required two hours. At this point the reaction was allowed to stir at 60° for five hours. The excess nitrous acid was destroyed and the solution was adjusted to pH 8 after being saturated with NaCl. The solution was then extracted with alcohol free anhydrous ether, dried with M9504 and cautiously evaporated to yield a yellowish oil which could be sublimed at 50°C at atmospheric pressure. The product, obtained in 68% yield, was a waxy fragrant solid; E;E- (sealed tube) 142-4°; in (CHC13) 2.85 H (OH) and 5.68 n ( ;C=0); flm£_(CCl4) T 6.56 ppm, singlet, 1H, (0H); 7.60 ppm, broad singlet, 1H, (bridgehead); 7.95 ppm, multiplet, 2H, (/CHi7¢’O ); 8.1-8.6 ppm, multiplet, 6H; mass. spec. m/e 126 (parent peak). Anal. Calcd for C H 0 C, 66.64; H, 7.99. 10 2‘ Found: c, 66.83; H, 7.89. 7 74 Ganeral procedures for mass spectral recovery and analysis. The solution containing the exchanged ketol was saturated with NaCl and extracted with alcohol free anhydrous ether. The ethereal layer was washed several times with small amounts of NaCl-saturated H20 to wash out the deuterium on the alcohol oxygen. The ether was then dried with M9504, evaporated carefully and the ketol sublimed. A11 mass spectral calculations include corrections for a relative isotope abundance of 10% (the normal P + 1 intensity in the undeuterated ketol). Deuterium exchange in NaOD-D20 at room temperature. 65 mg of ketol 22 was dissolved in 0.5 m1 of 0.15N NaOD-D20 solution prepared by cautiously adding the calculated amount of sodium metal to D20. After 15 minutes under these conditions this material was recovered and showed the following deuterium incorporation after being washed with water: 22.2% dO’ 38.6% d], 27.7% d2, 11.5% d3. After one week the deuterium incorporation reached an equilibrium value: 0% d0’ 6.9% d], 16.3% d2, 39.4% d3, 38.2% d4. Sealed tube NaOD-D20 exchange. The preparation of tetradeuterated ketol 22. 375 mg of undeuterated ketol 22 was dissolved in 5.0 ml of 0.15N NaOD-020 solution. This was added to a pyrex test tube which was blown out with nitrogen and sealed. The sealed tube was maintained at 100° for six days. Mass spectroscopy indicated the following deuterium incorporation: 0% d0’ 0% d 7.5% d2, 8.1% d3, 84.4% d4. The nuclear magnetic resonance spectrum ‘I, of 11 in D 0 is given in the text. mm 2 75 The preparation of dideuterated ketol 22 in a pH = 9.8 buffer. The buffer solution was prepared by dissolving 473.8 mg of oven—dried NaHCO3 and 212.0 mg of oven-dried Na2C03 in 20 m1 of D20. 230 mg of undeuterated ketol was dissolved in 1.8 m1 of this buffer at room temperature. After 288 hours, mass spectroscopy indicated the following deuterium incorporation: 0% d0, 13.9% d], 81.9% d2, 4.2% d3, 0% d4. The nuclear magnetic resonance spectrum of 22 is reproduced in the text. Hydrogen exchange of the tetradeuterated ketol with pH = 9.8 buffer. The preparation of dideuterated ketol 22. The H20 buffer solution was prepared as described earlier for the corresponding D20 solution. 60 mg of 84.4% tetradeuterated ketol was dissolved in 0.5 ml of this H20 buffer. After two weeks at room temperature the isotope distribution was as follows: 0% d0’ 0% d], 88.7% d 11.3% d 2, 0% d4. The nuclear magnetic resonance spectrum of this material is 3’ reproduced in the text. Kinetics of deuterium exchange in pH = 9.8 buffer. A 20% ketol - 020 buffer solution was prepared by dissolving 100 mg of undeuterated ketol in 0.5 m1 of buffer solution at 25°C. Nuclear magnetic resonance integrations were taken periodically using the bridgehead proton absorption at 7.59 ppm (T) as a 1H internal standard. Averages of 5-10 integrations yielded the experimental values which provided an approximate straight line when plotted as log[intt - inttm] versus time. Both signals at 7.81 and 8.39 yielded the same pseudo first order rate constants (1.2 x 10'] hrs-1). 76 Acid catalyzed deuterium exchange using CF3002D-D20 (sealed tube). 3(202D and 0.55 ml of D20. This solution was placed in a pyrex tube, degassed with a nitrogen 70 mg of ketol 19 was dissolved in 0.08 ml of CF flow and sealed. The sealed tube was then heated at 140° for four days. Recovery of the ketol indicated considerable deuterium incorporation: 0% d0’ 0% dl, 6.9% d2, 26.3% d3, 66.8% d4. The nuclear magnetic resonance spectrum of this material was virtually identical to that of the base catalyzed tetradeuterated ketol 11. The preparation of l,2,3,4,4a,6,7,8,9,9a-decahydro-2,4a,7,9a-dimethano- henazine (s n— and anti— isomeric mixture . A 10% sodium hydroxide solution (30 ml) was prepared and put in a round-bottomed flask fitted with a magnetic stirrer. The amine hydrochloride (3.0 9, 0.0186 moles) was added. The reaction was allowed to stir at room temperature for two hours. Extraction with several volumes of chloroform, drying with magnesium sulfate and subsequent evaporation of solvent left 1.9 g (95%) of a white highly crystalline solid. This material could be sublimed at 130°: mp_(sealed tube) l45—75°C (variable); i_ (Nujol) 5.96 p (:C=N), no C=0 or NH; nmr (CC14) T 7.5—8.8 ppm, multiplet, 18H; mass spec. m/e 214 (parent peak); 2;!: (MeOH) A 2) 3). x 250 (e = 2.5 X 10 and Xmax 208 ma (8 = 2.2 X 10 The isolation of the anti—isomer lg. Sublimation at llO-120° (760 mm) in a home—made fractional sublimation apparatus provides pure anti-dihydropyrazine 15. A four-foot horizontal length of l l/2'I diameter pyrex tubing was wrapped with heating tape (see illustration). 77 A slightly smaller diameter tube, sealed at one end and wrapped with asbestos string, was inserted into the 1 1/2" tube. The dihydropyrazine sample (2-4 gms) was placed in the closed end of the smaller tube and the heating m110° ~130° M .4/ .—I: A-l.’ nv—n- H-l- '- oI'Q"; LII. «(A/unm- nmr-smug sample heating tape tape set so that a temperature gradient of mllO-l30° was set up. The more volatile gptjfdihydropyrazine traveled furthest along the tube. It constituted less than 10% of the total reaction product. Resublimation at 120° gave a crystalline material: mp (sealed tube) 190-2°; jg) pmp, mggg. ppgp. were virtually identical to that of the mixture. Anal. Calcd for C C, 78.46; H, 8.47; N, 13.07. 14H18N2: Found: c, 78.57; H, 8.42; N, 13.01. The isolation of the d,Z-§yp;isomer 14. The less volatile gypfisomer did not sublime very far along the heated tube. It was the major product: mp_(sealed tube) l60—70°, 1p, pmp, and mp§§, §p§g, again were almost identical to the isomeric mixture. Further sublimation did not appreciably improve the melting point. Apgl. Calcd for C14H18N2: C, 78.46; H, 8.47; N, 13.07. Found: C, 78.53; H, 8.46; N, 13.01. 78 The preparation of the 1—amino-2-norborannone d-lO-cgmphorsulfonate Salt 18 (diastereomic mixture). Method A: The amine hydrochloride 2 (3.24 9, 0.0201 moles) was dissolved in 50 ml of water. 4.64 9 (0.0200 moles) of d-lO-camphorsulfonic acid was added and the mixture placed on a steam bath until most of the water evaporated off (evolution of HCl). This process was repeated twice more. Removal of the water jp_!§gpp_1eft a light brown syrup that could be recrystallized from acetone to give the sulfonate salt (6.5 g, 91%). Method B: 0.680 9 (0.00318 moles) of the decahydrophenazine mixture (14 + 15) was dissolved in 15 m1 of water. 1.475 g of d-lO—camphorsulfonic acid (0.00636 moles) was added and the solution was gently warmed on a steam bath for several hours until the solution was neutral to pH paper. Evaporation of the water jp_vg£pp_left 2.30 g (100%) of diastereomeric salts which could be recrystallized from acetone as before. The resolution of the d-10-camphorsu1fonate Salts. The diastereomeric salt mixture was fractionally recrystallized a total of five times from acetone. At this point further recrystallization did not substantially change the optical rotation. Two fractions were ultimately obtained; a ”more soluble” fraction and a "less soluble“ fraction. The "more soluble” fraction (tiny spars, mp_201-2°) had a higher specific rotation, [a]578 (MeOH) + 28.1. The "less soluble” fraction (fluffy micro- needles, pp_208-10°) had a lower specific rotation, [a1578 + 23.9. Since the "less soluble“ fraction was more convenient to obtain in relatively pure form it was used in subsequent experiments. 79 The generation of optically active gyprdecahydrolphenazine 14. The amino-ketone camphorsulfonate salt with the lower rotation (1.08 9, 0.00302 moles) was dissolved in 10 m1 of a 10% NaOH solution and stirred for three hours. The usual workup yielded 0.219 g (68%) of optically active dihydropyrazine 14 after sublimation at 120°; [a]578 (MeOH) -107°, [a]365 (MeOH) -3531 mp_(sealed tube) l60-7°; jr_identical to that of the previously isolated Exp;decahydrophenazine. Photolysis: general procedure. A stock solution of sublimed optically active 14 in spectral grade methanol was prepared and stored for all experiments (56 mg/100 ml). In a typical photolysis, 20 m1 of this solution was placed in an 8-inch quartz test tube 1 1/2 inches from a Hanovia 654A-36 200—watt mercury lamp in a water cooled quartz jacket. All solutions were bubbled with nitrogen for ten minutes before photolysis and kept under a positive pressure of nitrogen for the duration of the experiment. All photolyses were carried out at room temperature. Nhen M90 or Na2C03 were used as acid scavengers these solids (generally 3 mg per 20 ml) were suspended by magnetic stirring. In all cases these solids were removed by filtration before any physical measurements were made, as suspended particles interferred with optical rotation and ultraviolet studies. Where there was solvent loss due to evaporation the solutions were carefully readjusted to the initial volume after filtration. The photochemical racemization in methanol. Runs in the presence of acid scavengers did not alter the rate of racemization by more than 4%. Blanks employing virtually every manipulation 80 (i.e. M90 or Na2C03 addition, filtration and redilution) except actual ultraviolet radiation established that loss of activity was not due to absorption on the solid phase or loss of material during filtration. Essentially no loss of activity could be detected in blank runs. Racemization did not occur using Pyrex or Corex filters, but proceeded readily using quartz and very slowly using Vycor. Photolyses in hydrocarbon solvents also resulted in racemization but considerable dihydropyrazine decomposition made them less desirable than methanol. The recovery of photoracemized material 36 m1 of the photolyzed decahydrophenazine solution([a]365 = -l38) was filtered and evaporated jp_v§£pp, The slightly tan solid residue was sublimed at 120° to yield 9.6 mg (50% recovery) of decahydrophenazine: mp_(sea1ed tube) 156-60°; 1: identical to the starting material. This material was redissolved in methanol and gave a specific rotation essentially identical to that of the photolyzed solution, [a]365 = -126. Kinetic run of the photoracemization in methanol. 30 ml of the methanol stock solution (56 mg/100 ml) was photolyzed with 3 mg M90. The quartz tube was situated 1 1/2 inches from the middle of the mercury lamp. Each time the solution was filtered twice and then diluted up to the original 30 m1 before the optical rotation was taken. After each reading fresh MgO was added and the solution was photolyzed again. This process was repeated for every point until a total of 5 points were obtained (29 3/4 hours). Readings were taken at several wavelengths but the ones at 365 mu were used to calculate the rate constant. A plot of log [a]365 vs. 81 time showed excellent linearity indicating clean first order kinetics. The rate constant was calculated from the slope of this line. time (hrs) [a]365 log[a]365 0 -346 2.533 5.5 —220 2.342 17.5 — 79 1.898 25.0 — 38 1.581 29.75 - 25 1.398 . -k -0.68 s10pe = _r_a£ = 2.303 18 hrs _ —2 -l krac — 8 7 x 10 hr The attempted acid catalyzed racemization in ethylene dichloride. A new stock solution of optically active decahydropyrazine (0.00869 N) was prepared in dry ethylene dichloride which was distilled from P205. A stock solution of dry toluenesulfonic acid was prepared by dissolving toluene sulfonic acid monohydrate in dry ethylene dichloride and azeotropically distilling off the water. The acid concentration was determined to be 0.0144N using a standard NaOH titration. Both stock solutions were then stored in septum-fitted two—neck flasks. All manipulations were carried out with a syringe to minimize contact with atmospheric moisture. Two runs were attempted at 50°. In the first run the volumes were adjusted 82 so that the acid to decahydrophenazine ratio was 1:1, in the second run the ratio was 2:1. In both cases there was slow initial loss of optical activity lasting for about four hours but this soon stopped completely. If special care was not taken to remove all traces of water, loss of activity due to hydrolysis was rather extensive. Thermolysis runs on optically active decahydrophenazine. A; heptane solutions with MgO A stock solution of decahydrophenazine (42 mg/100 ml) was prepared in spectral grade heptane. M90 (5 mg) was added to 6.5 ml of this solution. The suspension was syringed into a Pyrex tube, degassed with nitrogen and sealed. This sealed tube was heated in the dark at 110° for four days. The tube was then cooled, the M90 was filtered and the rotation taken. There was no significant loss of optical activity (only 1.7%). Thermolysis of solutions not blown out with nitrogen resulted in extensive decomposition of the sample. Decomposition was also more rapid if no MgO was present. p; methanol solutions with MgO 8.5 ml of a methanol stock solution (56 mg/100 ml) of optically active decahydrophenazine was placed in a nitrogen filled sealed tube along with 3 mg of M90. The temperature was held at 110-5° for nine days. After cooling, filtration and adjustment of the volume, the rotation indicated a loss of only 11% activity. A brown ring which formed inside the tube at the methanol-nitrogen interface indicated at least part of this loss was due to thermal decomposition. F 83 The attempted self-condensation of l-amino-7,7-dimethyl-2-norbornanone in agueous base. 9.0 9 (0.0476 moles) of l-amino-7,7-dimethy1-2-norbornanone hydrochloride was dissolved in 90 ml of 10% NaOH. The reaction mixture immediately became hot and a white precipitate formed. The reaction mixture was allowed to stand overnight and was then extracted with CHCl3. Evaporation left a white semi-solid which could be sublimed at 50-80° (0.5 mm) to give 4.2 g (57%) of free l-amino-7,7-dimethy1—2-norbornanone: mp_(sealed tube) 197—8°; i: (Nujol) 5.73 p ( C=0), 3.00 H (NH2); pm: (CDC13) T 9.18 ppm, singlet, 3H (CH3); 8.88 ppm, singlet, 3H, (CH3); 8.60 ppm, singlet, 2H exchangeable, (NH2); 7.68—8.60 ppm multiplet; 7H; mgég, ppgp, m/e 153 (parent peak). No dihydropyrazine could be detected. Other condensation attempts A; 1 g of amino—ketone was refluxed in benzene (35 ml) with a Dean—Stark water separator. No reaction took place after 48 h0urs, as determined by infrared spectroscopy. B: l g of amino—ketone was refluxed in 35 ml of mesitylene for 48 hours. No reaction. C: 1 g of amino-ketone was refluxed in 20 m1 of dry toluene along with one equivalent of sodium hydride. Quenching with H20 after 24 hours led to a quantitative recovery of the starting material (0.97 g). [:3 l g of amino-ketone was refluxed in toluene with a spatula-tip of toluenesulfonic acid for 24 hours. Again no reaction could be observed by infrared spectroscopy. 84 The reaction of alkaline potassium ferricyanide with ketones (general procedure). 666 ml of water was placed in a one 1. three-neck flask fitted with an efficient magnetic stirrer, reflux condenser, thermometer and fritted glass nitrogen inlet tube. The water was heated while degassing with a slow nitrogen stream. 78.2 g of potassium ferricyanide (0.237 moles) was added when the water temperature reached 80°. After a minute or two of additional degassing the nitrogen bubbler was removed and replaced with a 125 m1 addition funnel (without a pressure equilibrating sidearm). While a static nitrogen atmosphere was maintained 0.0526 moles of ketone was added all at once followed by the dropwise addition of 40 ml of degassed concentrated ammonium hydroxide. The drop rate was adjusted so that the total addition required about 30 minutes while maintaining a temperature of 80-85°. This temperature was maintained for an additional 10—14 hours. The reaction was then cooled and continuously extracted with chloroform. The chloroform was subsequently washed with NaCl saturated water and extracted with 6% H01. This layer could then be made basic with NaOH to yield the corresponding dihydropyrazine which could be extracted and sublimed at reduced pressure. The preparation of hexamethyl-2,5-dihydropyrazine 20. Following the general reaction procedure, 6.7 g (0.078 moles) of isopropyl methyl ketone was added to 800 ml of degassed water containing 116 g (0.351 moles) of potassium ferricyanide. 52 ml of concentrated ammonium hydroxide was drop-added. The reaction was allowed to run overnight at 85° and the product, in the form of the hexahydrate (mp 88-9°), was steam distilled 85 out of the reaction mixture. The yield of pure material was 37% (3.0 gms); j§_(Nujol) 6.00 n (>C=N); pmp_(pyridine) r 8.65 ppm. singlet, 12H, (CH3-C-CH3); H 7.96 ppm, singlet, 6H, (N2; 3); 4.74 ppm, singlet, 12H, (H20). The preparation of 2,5-dispiropentane-3,6-dimethy1-2,5-dihydropyrazine 21. 3.0 9 (0.0263 moles) of cyclopentyl methyl ketone was reacted with 39 g (0.118 moles) of potassium ferricyanide dissolved in 333 ml of water and 20 m1 of concentrated ammonium hydroxide. After 16 hours the reaction was continuously extracted with chloroform yielding 1.66 g of crude crystalline product (57% yield). Recrystallization from acetonitrile at low temperature followed by sublimation at 65° (1 mm) gave 1.51 g of pure material as white flakes: mp102-3°; 15(Nujoi) 6.02 p (:C=N); p_rn_r (c0013) 1 8.01 ppm, singlet, 6H, (CH3); 8.20 ppm, multiplet, 16H, (cyclopentyl); mass. spec. m/e 218 (parent peak). Appl, Calcd for C14H22N2: C, 77.01; H, 10.16; N, 12.83. Found: C, 76.71; H, 9.66; N, 12.97. The preparation of l,2,3,4,4a,6,7,8,9,9a-decahydro-4a,9a-dimethy1phenazine 22 (isomer mixture). 10 g of 2-methylcyclohexanone (0.0876 moles) was added to 130 g of potassium ferricyanide (0.394 moles) in 1250 m1 of water. 60 m1 of concentrated ammonium hydroxide was added and the reaction allowed to run overnight. Continuous extraction with chloroform yielded 4.1 g (42% yield) of a tan oil which crystallized in a day or two. Recrystallization of the solidified material from acetonitrile followed by sublimation at 120° (aspirator pressure) gave a white waxy solid (40-52%). ppp, 68-82°; ir_(Nujol) 6.03 p 86 ( C=N-); pm[_(CDC13) r 8.69 and 8.62 ppm, singlets, (CH3); 7.50-8.50 ppm, multiplet, 16H, (cyclohexyl); mass. spec. m/e 218 (parent peak). Anal. Calcd for 014H22N2: c, 77.01; H, 10.16; N, 12.03. Found: 0, 76.78; H, 10.24; N, 12.66. The preparation of 2,5-diphenyl-3,3,6,6-tetramethy1-2,5-dihydropyrazine 23. 7.8 g of isobutyrophenone (0.0526 moles), 78.2 g of potassium ferricyanide (0.237 moles) in 500 ml of water, and 200 m1 of concentrated ammonium hydroxide were reacted at 85° for 16 hours. Continuous extraction with chloroform yielded a yellow oil which gave 5.2 g of a-amino-isobutrophenone hydrochloride when treated with dilute HCl (mp 180-5°). The free amino-ketone, liberated from this hydrochloride by treatment with 10% NaOH, did not readily self-condense but did so after storage in a P205 dessicator for two weeks. The dihydropyrazine 23 was then recrystallized from chloroform at low temperatures, followed by sublimation at 110° (1 mm) to give clear prisms in 51% overall yield: mp_127-8°; ig_(Nujol) 6.11 p (:C=N—); pmp_(CDC13) T 8.46 ppm, singlet, 12H, (CH3); 2.60 ppm, pseudo—singlet, 10H, (phenyl); mass. spec. m/e 290 (parent peak). Apgl, Calcd for C H N ' C, 82.72; H, 7.64; N, 9.65. 20 22 2' Found: C, 82.92; H, 7.73; N, 9.61. The preparation of hexadeuterated hexamethyl-2,5-dihydropyrazine 20a. Undeuterated hexamethyl-2,5-dihydropyrazine hexahydrate was stored in a P205-dessicator for one week yielding the corresponding anhydrous dihydropyrazine (mp_69°). 1.28 g of this material was placed in 25 ml of 0.1 N NaOD-020 solution and refluxed for four days yielding 1.37 g of the 86 ( C=N-); pm: (CDC13) r 8.69 and 8.62 ppm, singlets, (CH3); 7.50—8.50 ppm, multiplet, 16H, (cyclohexyl); mass. spec. m/e 218 (parent peak). Anal. Calcd for C14H22N2: C, 77.01; H, 10.16; N, 12.83. Found: C, 76.78; H, 10.24; N, 12.66. The preparation of 2,5-diphenyl-3,3,6,6-tetramethy1-2,5—dihydropyrazine 23. 7.8 g of isobutyrophenone (0.0526 moles), 78.2 g of potassium ferricyanide (0.237 moles) in 500 ml of water, and 200 m1 of concentrated ammonium hydroxide were reacted at 85° for 16 hours. Continuous extraction with chlorofonn yielded a yellow oil which gave 5.2 g of a-amino-isobutrophenone hydrochloride when treated with dilute HCl (mp l80-5°). The free amino-ketone, liberated from this hydrochloride by treatment with 10% NaOH, did not readily self-condense but did so after storage in a P205 dessicator for two weeks. The dihydropyrazine 23 was then recrystallized from chloroform at low temperatures, followed by sublimation at 110° (1 mn) to give clear prisms in 51% overall yield: mp_127-8°; ip_(Nujol) 6.11 p (:C=N-); pmp_(CDCl3) T 8.46 ppm, singlet, 12H, (CH3); 2.60 ppm, pseudo-singlet, 10H, (phenyl); mass. spec. m/e 290 (parent peak). H N C, 82.72; H, 7.64; N, 9.65. 22 2‘ Found: 0, 82.92; H, 7.73; N, 9.61. Anal. Calcd for C20 The preparation of hexadeuterated hexamethyl-2,5—dihydropyrazine 20a. Undeuterated hexamethyl-2,5-dihydropyrazine hexahydrate was stored in a P205-dessicator for one week yielding the corresponding anhydrous dihydropyrazine (mp 69°). 1.28 g of this material was placed in 25 ml of 0.1 N NaOD-D20 solution and refluxed for four days yielding 1.37 g of the 87 hexadeuterated dihydropyrazine in the form of a hexahydrate. This material was again placed in a P205 dessicator for a week yielding the anhydrous product which was essentially 100% hexadeuterated as determined by nmr integration. The attempted photolyses of hexadeuterated hexamethyl—Z,5-dihydropyrazine. A; 40 mg of the hexadeuterated dihydropyrazine 20a was placed in a quartz test tube with 20 m1 of spectral grade methanol and irradiated for 3 days 1 1/2 inches from a ZOO-watt mercury lamp. No change was observed. A similar solution when irradiated with a 550-watt lamp for long periods of time gave several reaction products which remain unknown. The only recognized product was ethylene glycol. Nuclear magnetic resonance spectro- scopy showed no evidence of deuterium scrambling in the starting material recovered. .8: sensitized attempts 62.3 mg of hexadeuterated dihydropyrazine was dissolved in 25 ml of 0.25% ethereal acetone solution and photolyzed with a ZOO—watt mercury lamp (1 1/2 inches away) for three days. Again no change was evident. Likewise 60 mg of dihydropyrazine was dissolved in 25 m1 of ether solution containing 3 mg of benzophenone. This solution was placed in a pyrex vessel and photolyzed for two days with a 550-watt mercury lamp without change. Other unsuccessful attempts were made in cyclohexane and benzene solvents. In extreme cases the only result was extensive decomposition of the dihydropyrazine, without scrambling in recovered starting material. 88 The attempted photolyses of l,2,3,4,4a,6,7,8,9,9a-decahydro-4a,9a-dimethy1- phenazine 22 and 2,5-dispiropentane-3,6-dimethyl-2,5-dihydropyrazine 22. 3M). Methanolic stock solutions of both compounds were prepared (1.4 x 10- Both were photolyzed using quartz optics 1 1/2 inches away from a ZOO-watt mercury lamp for 72 hours. The solutions turned yellow and showed new absorptions in their ultraviolet spectra (325 mu and 225 mu). Nmr analysis, however, showed no signs of rearrangement. Thin layer chromatography on silica gel and alumina verified this observation. More concentrated solutions (1.4 x 10'2 M) also failed to rearrange, as did photosensitized attempts with benzophenone in methanol. Ethylene glycol was often observed as a side product, especially when the 550-watt lamp was used. Dihydropyrazine carbonium ions in FSO3H at probe temperature (general procedure). 100 mg of the dihydropyrazine was dissolved in 1 m1 of Freon-114B. This solution was slowly drop-added to 1 ml of FSO3H which was cooled in a Dry Ice-acetone bath at —78° under a dry nitrogen atmosphere. Rapid magnetic stirring was maintained throughout. After the addition was completed the cooling bath was allowed to slowly warm until it reached room temperature. By this time the flow of nitrogen had evaporated off the Freon-114B. The resulting clear white FSO3H solution was then transferred to an nmr tube and the spectrum was taken at probe temperature. Tetramethylammonium tetra- fluoroborate (TMA) was used as an internal standard. Individual spectra are reported in Table 6, 89 The reaction of agueous pptassium gyanate with 1-amino-2-norbornanone hydrodfizjgp In a typical run, 250 mg (3.09 in moles) of potassium cyanate was dissolved in 5 ml of water. 500 mg (3.09 nmoles) of amino-ketone hydrochloride 2 was added very slowly as a solid. This addition required approximately one hour. The solution was allowed to stir at room temperature. After a few minutes a white precipitate was formed. Two hours later this precipitate was filtered yielding 110 mg of 22 (22%). 22 was recrystallized from either chloroform or acetonitrile to give 107 mg of tiny colorless needles: mp 158-60°; jp_(Nujol) 2.84 u (0H?), 3.09 u (NH), 5.82 and 6.04 u (>C=0); pmp_(dmso-d6) T 3.10 ppm, singlet, 1H, (NH); 3.45 ppm, singlet, 1H, (NH); 4.66 ppm, singlet, 1H, (0H); 7.9-8.5 ppm, complex multiplet, 9H, (norbornyl CH); mass. spec. m/e 168 (parent peak). Apgl. Calcd for C8H12N202: C, 57.13; H, 7.19; N, 16.66. Found: C, 56.54; H, 7.14; N, 16.59. The filtered aqueous solution was extracted with ether several times. After drying with M9804 and evaporation, a mixture of dihydropyrazines 22 and 22 was isolated as a crystalline solid (180 mg, 55%). The remaining aqueous solution was then evaporated to dryness ip_y§gpp_and the resulting solid was boiled with acetonitrile, the solvent filtered, dried and evaporated yielding 130 mg (26%) of 22. 22 was recrystallized from chloroform to yield colorless needles: mp_l8l—2°, i:_(Nujol) 2.89, 3.00 and 3.10 u (NH), 5.68 u (>C=0), 6.00-6.50 u (mono-alkyl urea); pm: (dmso-d6) T 3.87 ppm, singlet, 1H, (NH); 4.57 ppm, singlet, 2H, (NHZ); 7.8—8.8 ppm multiplet, 9H, (norbornyl CH); mass. S ec. m/e 168 (parent peak). 90 Anal. Calcd for C8H12N202: C, 57.13; H, 7.19; N, 16.66. Found: C, 57.15; H, 7.14; N, 16.68. The preparation of l-ureido-2-norbornanone via a nodified Curtius reaction. 10 g (0.065 moleS) of Z-ketonorbornanone carboxylic acid was converted into its acid chloride as previously described. The acid chloride was then dissolved in 20 ml of xylene and treated with 6.0 g of acetone—damp activated sodium azide and reacted as before. The resulting solution of l-isocyanato- 2—norbornanone was filtered and diluted with an additional 50 ml benzene. This solution was placed in a 150 m1 flask fitted with an efficient magnetic stirrer and a glass tube that reached just below the surface of the benzene solution. Gaseous ammonia was slowly bubbled through the solution. After a few minutes the mixture became quite hot and deposited 22 as a white precipitate. The reaction was continued for an hour until the flask returned to room temperature. At this point, the benzene suspension of 22 was warmed on a steam bath and filtered. The resulting solid was recrystallized from acetonitrile after decolorizing with Norite. The yield was 8.31 g (76%). Attempts to close 26. 'Vb 3—Na2C03 buffer (pH = 9.8) for seven hours. Evaporation and infrared analysis A; 500 mg (3.09 mmoles) of 22 was refluxed in 20 ml of a NaHCO indicated only starting material. 8; 500 mg (3.09 nmoles) of 22 was heated on a steam bath in 20 ml of 95% ethanol containing two drops at concentrated sulfuric acid. No ring closure occurred. 91 g; 500 mg (3.09 mmoles) of 22 was treated at room temperature with a 5% sodium methoxide-methanol solution prepared by adding the calculated amount of sodium to dry methanol. After 12 hours no signs of reaction were apparent. Q; 256 mg (1.57 mmoles) of 22 was suspended in anhydrous ether and drop-added onto 104 mg (2.3 mmoles) of 57% NaH oil dispersion. After refluxing 12-14 hours under a nitrogen atmosphere and cautious quenching with water, only starting material could be identified. g; 500 mg (3.09 mmoles) of 22 was slowly added as a solid to 250 mg (3.09 mmoles) of potassium cyanate dissolved in 5 ml of water. After stirring overnight the reaction was worked up in the usual manner. Only starting material was recovered. The preparation of l-ureido-7,7-dimethy1-2-norbornanone 22. 15.2 gms (0.0835 moles) of ketopinic acid 293.94.45 was converted into its acid chloride by refluxing it with 15 ml thionyl chloride in 30 ml dry benzene for five hours. The excess thionyl chloride was distilled off and the acid chloride dissolved in 30 m1 of xylene. 9.0 g (0.137 moles) of activated sodium azide was cautiously added to the rapidly stirred xylene solution. The reaction mixture was then slowly warmed until the evolution of nitrogen proceeded. After 14 hours at reflux, the solution was cooled, filtered and diluted with 75 ml of benzene. Following the procedure described for the preparation of l-ureido-Z-norbornanone, gaseous ammonia was bubbled through this solution for two hours. The resulting precipitate was filtered yielding 8.0 g (49%) of 22. 92 22 was recrystallized from chloroform or benzene to give white crystals: mp170-171.5°; _i_r; (Nujol) 2.90 and 3.01 n (NH), 5.70 6 00:0), 6.00-6.50 u (mono-alkyl urea); pmp_(CDCl3) T 4.32 ppm, singlet, 1H, (NH); 4.87 ppm, singlet, 2H, (NHZ); 8.75 ppm, singlet, 3H, (CH3); 9.19 ppm, singlet, 3H, (CH3); 6.8-8.6 ppm, complex multiplet, 7H, (norbornyl CH); mass. spec. m/e 196 (parent peak). N 0 ' C, 61.20; H, 8.22; N, 14.28. 16 2 2' Found: C, 61.19; H, 8.18; N, 14.20. Anal. Calcd for C H 10 The ring closure of 1-Ureido-7,7-dimethy1-2-norbornanone with NaH in ether. A 50 ml three-neck flask was fitted with a magnetic stirrer, 25 m1 sidearm addition funnel, reflux condenser and nitrogen inlet tube. The flask was charged with nitrogen and 130 mg (2.9 mmoles) of a 57% NaH oil dispersion was placed in the bottom of the flask. 400 mg 22 (2.04 mmoles) was pulverized and placed in the addition funnel. 20 ml of anhydrous ether was syringed into the addition funnel, partially dissolving 22. This suspension was then drop added to the solid NaH over the period of five or ten minutes. The last bits of 22 were washed out with a few ml of fresh ether. The reaction mixture was brought to reflux under a nitrogen atmosphere and maintained for 16 hours. At this time the resulting suspension was very cautiously quenched with 5 m1 of water (safety shield) while maintaining a rapid nitrogen flow and rapid magnetic stirring. After a few minutes a white precipitate formed which after cooling and filtration provided crude 22 in yields of 215-315 mg (SS-80%). 93 22 was recrystallized from acetonitrile to give tiny colorless needles: mp186-8°, i_r (Nujol), 3.05-3.15 u, 5.90-5.95 u (>c=0); n_mr_ (dmso-d T 3.22 6) ppm, singlet, 1H, (NH); 3.22 ppm, singlet, 1H, (NH); 4.69 ppm, singlet, 1H (0H); 9.03 ppm, singlet, 6H, (CH3-C-CH3); 7.7-8.8 ppm, multiplet, 7H, (norbornyl CH); pm[_(pyridine) r 8.80 ppm, singlet, 3H, (CH3); 8.99 ppm, singlet, 3H, (CH3); 7.4-8.4 ppm, complex multiplet, 7H, (norbornyl CH); mass. §pec. m/e 196 (parent peak). Anal. Calcd for C10H16N202: C, 61.20; H, 8.22; N, 14.28. Found: C, 61.02; H, 8.05; N, 14.37. The dehydration of 22 with thionyl chloride in chloroform-pyridine. The preparation of 22. 186 mg (0.95 mmoles) of 22 was dissolved in 3 ml of dry pyridine and cooled to 0° in a 10 ml two-neck flask fitted with a rubber septum, CaCl2 drying tube and a magnetic stirrer. 140 mg (1.18 mmoles) of thionyl chloride was dissolved in 2 ml of chloroform and slowly injected Nun Umecold pyridine solution. The ice bath was removed after the addition of the thionyl chloride was complete. The reaction was then allowed to stir for approximately 12 hours. At this point, the white precipitate that formed was filtered and dried under high vacuum to yield 240 mg (87%) of 22. 22 was very water sensitive which made recrystallization and chemical analysis impossible. The crude reaction product, however, possessed the following properties: mp_l41-5° (dec.); j§_(Nujol) 3.10 p (NH), 5.84 u (:C=0), 6.00 and 6.55 u (pyridine); mass. spec. m/e 179 (p - pyridine and chloride). 94 The hydrolysis of 22. 78.3 mg (0.267 mmoles) of the pyridine salt was dissolved in 2 m1 of water and filtered. The water was taken off jp_gggpp_to yield 80.6 mg of solid which was fractioned according to the following scheme. The crude solid was suspended in 0.5 m1 of water and the insoluble fraction was filtered, providing 9.6 mg (18%) of urea 22 as identified by melting point and infrared analysis. The aqueous fraction was evaporated to dryness yielding 66.1 mg of crude solid. This in turn was treated with a dilute Na2C03-NaHC03 buffer solution (pH = 9.8). The evolution of pyridine was detected by its characteristic odor. The nmr of the remaining solid (32.2 mg) in dmso-d6 indicated the presence of two separate compounds with methyl groups at 9.00, 9.10 and 9.17 ppm (1). The infrared spectrum of this product mixture showed a carbonyl at 5.90 u characteristic of closed hydroxy ureas. The reaction of 22 with ethanol stabilized chloroform. The preparation of 22. 1.2 g (.0715 mole) of the crude water insoluble product from the reaction of potassium cyanate with 1-amino-2-norbornanone hydrochloride was filtered and boiled with approximately 100 m1 of chloroform. The chloroform solution was then dried with M9504 and evaporated to dryness yielding 1.2 g (85%) a white crystalline material, 22. This reaction did not always work. Perhaps traces of acid or base catalysts were necessary. 95 22 was recrystallized nicely from acetonitrile to give long thin colorless needles: mp 18l—2°; jp_(Nujol) 3.15 and 3.27 u (NH), 5.88 H (:C=0); pm: (dmso-d6) r 2.72 ppm, singlet, 1H, (NH); 3.47 ppm, singlet, 1H, (NH); 6.63 ppm, quartet (J = 7 cps), 2H, (OCHZ); 8.96 ppm, triplet (J = 7 cps), 3H, (CH3); 7.98-8.67 ppm, complex multiplet, (norbornyl CH): mass. spec. m/e 196 (parent peak). Appl. Calcd for C H N 0 : C, 61.20; H, 8.22; N, 14.28. 10 16 2 2 Found: C, 61.16; H, 8.28; N, 14.26. The reaction of pyridine salt 22 with triethylamine. The preparation of 22. 28.4 mg (0.097 mmole) of the pyridine salt was placed in a small test tube along with 2 ml of freshly distilled triethylamine. The suspension was brought to a boil with constant stirring. After a few mintues the crystalline 22 became gummy and then resolidified. This new solid was filtered yielding 25 mg (81%) of 22. 22 like its pyridine counterpart was a rather moisture sensitive white powder: mp_203-6 (dec.); 13 (Nujol) 3.10 u (NH), 5.84 u (:C=0); ppp_(dmso-d6) r 3.31 ppm, singlet, 1H, (NH); 4.58 ppm, singlet, 1H, (NH); 9.00 ppm, singlet, 3H, (CH3); 9.11 ppm, singlet, 3H, (CH3); 6.92 ppm, quartet (J = 7 cps), 6H, (NCHZ); 8.80 ppm, triplet (J = 7 cps), 9H, (CH CH ), 7.7—8.3 ppm, complex multiplet, 7H, (norbornyl CH); mass. spec. 2 3 m/e 179 (P — NEt3 and chloride). 10. 11. 12. 13. 14. 15. 16. 17. #00“) C4 :0 I I I O O @637? REFERENCES Meerwein and K. van Emster, 225,, 22, 1815 (1920). Meerwein and K. van Emster, 293,, 22, 2500 (1922). . B. Woodward and R. Hoffmann, Angew. Chem. I.D., 2, 781 (1969). A. Berson, Accs. Chem. Res., 2, 152 (1968). Fukui, AccstChem. Res., 2, 57 (1971). A. Olah, R. H. Schlosberg, D. P. Kelly and G. D. Mateescu, M. . Amer. Chem. Soc., 22, 2256 (1970). Saunders, "Magnetic Resonance in Biological Systems", Permagon Press, New York, N. Y., 1967, p. 85. C. >496) U3 . Applequist and J. Klieman, J. Org. Chem., 22, 2178 (1961 . MacLean and E. L. Mackor, Discussions Faraday Soc., 22, 165 (1962). . A. Olah and J. Lukas, J. Amer. Chem. Soc., 22, 4739 (1967). A. Berson and G. L. Nelson, J. Amer. Chem. Soc., 22, 1096 (1970). J. Katz and E. H. Gold, J. Amer. Chem. Soc., 22, 1600 (1964). C. MacDonald and J. Trotter, Acta. Cryst., 22, 456 (1965). . Selman and J. Eastham, Quart. Revs., 22, 221 (1960). Nickon, T. Nishida and Y. Lin, J. Amer. Chem. Soc., 22, 6860 (1969). w. Alder, Tetrahedron Letters, 193 (1971). Ishidate and A. Kawada, Pharm. Bull,, 2, 483 (1956); CA 14625f 51 ’VD ) 96 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 97 REFERENCES (Continued) Alder and J. Stein, App., 222, 197 (1934). Diels and K. Alder, App,, 222, 98 (1928). Kwart and G. Null, J. Amer. Chem. Soc., 22, 2765 (1959). SICK . Vaughan, R. Caple, J. Csapilla, P. Schneider, J. Amer. Chem. Soc., 22, 2204 (1965). H. C. Brown and C. Garg, J. Amer. Chem. Soc., 22, 2952 (1961). A. Nickon and J. Lambert, J. Amer. Chem. Soc., 22, 1905 (1966). A. Nickon, J. Hammond, J. Lambert, and R. Williams, J. Amer. Chem. §p§ , 22, 3713 (1963). A. Thomas and B. Wilhelm, Tett. Lett., 1309 (1965). J. Jerkunica, S. Borcic and D. Sunko, Tett. Lett., 1309 (1965). A. Thomas, R. Schneider and J. Meinwald, J. Amer. Chem. Soc., 22, 68 (1967). T. Tidwell, J. Amer. Chem. Soc., 22, 1448 (1970). P. von R. Schleyer, J. Amer. Chem. Soc., 22, 701 (1967). D. E. Applequist and J. P. Klieman, J. Org, Chem., 22, 2178 (1961). J. Meinwald and D. E. Putzig, J. Org. Chem., 22, 1891 (1970). M. Fischer, 235,, 222, 3599 (1967). C. Stevens, I. Klundt and M. Munk, J. Org. Chem., 22, 2967 (1965). S. Gabriel, 233,, 22, 57 (1911). M. Grimmett, Advances in Heterocyclic Chemistry, 22, pp. 115 (1970). J. Conant and J. Aston, J. Amer. Chem. Soc., 22, 2783 (1928). 37. 38. 39. 40. 41. 42. 44. 45. 46. 47. 48. 49. 98 REFERENCES (Continued) P. Speakman and W. Waters, J. Chem. Soc., 40 (1955). B. S. Thyagarajan, Chem. Revs., 22, 439 (1958). K. Hofmann, The Chemistry of Heterocyclic Compounds (Imidazole and Derivatives, Part 1), Interscience Publishers LTD., London, 1953 p. 55. P. J. L. Boivin and P. A. Boivin, Can. J. Chem., 22, 561 (1954). F. W. Hoover, H. B. Stevenson and H. S. Rothrock, J. Org. Chem., 22, 1825 (1963). H. C. Brown and J. Muzzio, J. Amer. Chem. Soc., 22, 2811 (1966). P. D. Bartlett and L. H. Knox, Organic Syntheses, 2222222, John Wiley and Sons, N. Y. (1965) p. 12. P. D. Bartlett and L. H. Knox, Organic Syntheses, 22 42, 1. «AAA. John Wiley and Sons, N. Y. (1965).p. 12. P. D. Bartlett and L. H. Knox, Organic Syntheses, 2222222, John Wiley and Sons, N. Y. (1965) p. 55. A. F. Hegarty and T. C. Bruice, J. Amer. Chem. Soc., 22, 6575 (1970). T. M. Lowry and E. H. Magson, J. Chem. Soc., 22, 1042 (1906). C. L. Capmau, W. Chodkiewicz and P. Cadiot, Tetrahedron Letters, 1619 (1965). H. Biltz, 533 , 222, 156 (1909). .. 5 :.1.. .. .3.. 1. :42,» 1! ... : .. ..3... :.: :95... Ltafums hr... 3.... 2.... . . . ..oorgv ..i 1! {6:53.}. . . . . . . . . .. . . .. . . .. . _ Sufi??? vamgtnix . ._.:....».....Z.... 1 , , , . , . . 2 . , . . :...::....:...... ....... ,1... :1. .t .. . .k..\....<.: .. ‘\LE1~‘\\.=.\‘\..\I s. .1... .115... r . . . . . . . . ..x.:.n|.......€.«\.. : . a: «..i... . . .\ u H.‘ ..v» .‘r...‘\ . _ \.n:1.. irkby . g: 63:1. .. _ .. ..4. ~V\ . .1 6;..727.»a . . 3...}... .r. ,. .. . 1aai: .. .. . . .1: . :3...:...?::%.:-:...... . .. . ............... .,.. .... ....3..........,. . ..........._.... . .... ._..... 1.... ....,........;.... .... ..-.' .. ........ ..3....