Inaug.-diss.--Strassburg.

Inaug.-diss.--Strassburg.

La variante « glyoxylique » de la réaction de Perkin permet de lier entre eux deux fragments aromatiques par un pont maléique. La rigidification de cet intermédiaire flexible mène à des systèmes aromatiques polycycliques étendus par création des liaisons carbone-carbone manquantes. Cette stratégie requiert l'utilisation d'acides arylacétiques et arylglyoxyliques, et l’utilisation d'unités bifonctionnelles a auparavant permis la synthèse de cibles variées allant des phénacènes plans aux (poly)hélicènes très distordus, mais aussi des macrocycles conjugués. Afin d'étendre la taille et la variété des molécules obtenues à l'aide de cette stratégie, des méthodes générales de protection/déprotection sont développées. Une dissymétrisation efficace des unités bifonctionnelles mène à des nouveaux précurseurs monoprotégés qui sont ensuite assemblés par la réaction de Perkin en oligomères de taille contrôlée possédant encore des fonctions chimiques réactives aux extrémités après déprotection. L'utilisation de ces derniers lors d'une deuxième réaction de Perkin donne alors accès à de longs précurseurs flexibles, d’au moins cinq unités, qui donneront ensuite de très longs phénacènes, de grands macrocycles mais aussi des cyclo-tris[5]hélicènes qui présentent une géométrie de Möbius persistante et une aromaticité de Möbius. / The “glyoxylic” variant of the Perkin reaction allows to link two aromatic fragments by a maleic bridge. The stiffening of the obtained flexible intermediate by the creation of the missing carbon-carbon bonds leads to extended polycyclic aromatic systems. This strategy relies on the use of arylacetic and arylglyoxylic acids, and the use of bifunctional units has previously allowed the synthesis of a variety of targets spanning from flat phenacenes to strongly distorted (poly)helicenes, as well as of conjugated macrocycles. This approach is generalized here by developing protection/deprotection techniques in order to enhance the size and the variety of the molecules that can be obtained by this strategy. These techniques enable an efficient dissymetrization of bifunctional units and the recycling of symmetrical side products. The new monoprotected building blocks are connected by Perkin reactions to yield oligomeric intermediates with reactive functions at their extremities after deprotection. Several of these intermediates are assembled in a second Perkin reaction to obtain long phenacenes, large macrocycles and also cyclo-tris[5]helicenes with persistent Möbius geometry and Möbius aromaticity.

Les nanorubans de graphène (NRGs) sont des matériaux prometteurs pour l'organique électronique, à mi chemin entre polymères conjugués et nanotubes de carbone. Deux approches différentes pour la synthèse de nanorubans aromatiques sont développées et évaluées. La première est fondée sur la formation de céramidonines par cyclisation d'arylamino-anthraquinones en milieu acide. Plusieurs tétraaza-arènes incorporant deux de ces unités sont obtenus, mais l'approche s'est uniquement avérée appropriée dans le cas de courts substrats. La seconde approche repose sur la condensation d'acides aryle-acétiques avec des formylarènes ou acides aryle-glyoxyliques, suivie soit de cyclo-deshydrogénations en présence de quinone, soit de deshydrodebromation catalysée par le palladium, pour donner des arenes carboxy-substitués allongés. La méthode impliquant la quinone s'avère limitée à des substrats suffisamment réactifs tels que des thiophènes et laisse envisager des poly(arènodithiophènes) en partie rigidifiés et carboxy-substitués. La catalyse au palladium s'avère plus générale, ouvrant des perspectives d'obtention d'une grande variété de rubans aux propriétés électroniques ajustables. / Graphene nanoribbons (GNRs) are promising materials for organic electronics, as they bridge the gap betweensingle-stranded conjugated polymers and carbon nanotubes. Two different synthetic approaches to GNRs aredeveloped and evaluated. The first approach is based on the acid-promoted cyclisation of arylaminoanthraquinonesto ceramidonines. Tetraazaarenes with two ceramidonine units are obtained, but the approachis found to be appropriate only to such small systems. The second approach is based on the condensation ofarylacetic acids with arenecarboxaldehydes or arylglyoxylic acids, followed either by quinone-assistedoxidative cyclodehydrogenation or palladium-catalysed dehydrodebromination to yield carboxy-substitutedelongated arenes. The quinone-based variant is found to be limited to reactive substrates such as thiophenederivatives and offers the perspective of partially rigidified carboxy-substituted poly(arenodithiophenes). Thepalladium-based variant is found to be more general, opening the prospect of obtaining a variety of ribbontypestructures with tunable electronic properties.

La synthèse organique contrôlée de nanobagues de carbone est un challenge scientifique de longue date. Ces composés polycycliques aromatiques cylindriques peuvent être définis comme des sections de nanotubes de carbone plus larges qu’épaisses ; et la courbure de leur système pi pourrait leur conférer des propriétés électroniques intéressantes.Depuis quelques années, notre équipe développe une approche générale de synthèse de composés aromatiques polycycliques fonctionnalisés par des fonctions carboxyliques. Cette approche repose sur la réaction de Perkin entre des acides aryle-acétiques et des acides aryle-glyoxyliques, qui va permettre l’assemblage de ces briques élémentaires en longs précurseurs flexibles mais conjugués. Une dernière étape de cyclisation intramoléculaire, ou « graphitisation », pourra alors conjuguer complètement et donc rigidifier la molécule finale. Cette approche a permis la synthèse de plusieurs nouveaux composés aromatiques polycycliques linéaires.L’objectif de cette thèse est l’adaptation de l’approche de Perkin à la formation de nanobagues aromatiques. Un premier défi a été efficacement remporté avec l’obtention et la caractérisation complète de plusieurs macrocycles flexibles mais conjugués. Certains de ces macrocycles ont même été formés avec d’excellents rendements grâce à la mise en place d’une technique de haute dilution. Plusieurs tentatives de graphitisation ont été menées sur ces composés, impliquant différentes techniques de synthèse telles que la photochimie ou la catalyse au palladium, mais ne permirent malheureusement pas la formation des nanobagues aromatiques désirées. Néanmoins, en modifiant la structure initiale de certaines briques élémentaires nous avons pu obtenir d’autres macrocycles conjugués plus flexibles qui, après photocyclisation, ont abouti à la formation d’autres macrocycles conjugués présentant des structures rigides mais atypiques car non planes. / The controlled organic synthesis of carbon nanobelts has been scientific challenge for a long time. Such cylindrical polycyclic aromatic compounds can be defined as sections of carbon nanotubes that are larger than wide. Interesting electronic properties could result from the curvature of their pi system.These last years, our team has developed a general synthetic approach for the formation of carboxy-functionalised polycyclic aromatic compounds. This approach involves the Perkin reaction of arylacetic acids with arylglyoxylic acids, in order to form conjugated and flexible elongated precursors. The last step is an intramolecular cyclisation, or “graphitisation”, which rigidifies the precursor and yields a fully conjugated final molecule. Applying this approach, our team has synthesised several new linear polycyclic aromatic compounds.The aim of this thesis is to adapt the Perkin reaction for the formation of aromatic nanobelts. A first challenge has been solved by synthesising and fully characterising several flexible and conjugated macrocycles. Some of those macrocyclic compounds have been obtained with unexpectedly good yields using a high dilution addition technique. Graphitisations have been tried on some of those macrocycles by different synthetic methods, i.e. photochemistry and palladium catalysis, but none of them led to the formation of the desired aromatic nanobelt. However, by modifying the initial structure of some of the building blocks, we obtained more flexible conjugated macrocycles, which then reacted, by photocyclisation, to form conjugated and non-planar rigid macrocycles with atypical structures.

This thesis reports diverse synthetic and mechanistic studies in six chapters, as summarized below. Chapter 1. Revised mechanism and improved methodology for the perkin condensation.1 The generally accepted mechanism for the well-known Perkin condensation is unviable for at least two reasons: (1) the normally employed base, acetate ion, is too weak to deprotonate acetic anhydride (Ac2O, the substrate); and (2) even were Ac2O to be derprotonated , its anion would rapidly fragment to ketene and acetate ion at the high temperatures employed for the reaction. It has proved in this study that the Perkin condensation occurs most likely via the initial formation of a fem-diacetate (3, Scheme 1) from benzaldehyde (2) and acetic anhydride (1).1 The key nucleophile appears to be the enolate of 3 (and not of 1), which adds t the C=O group of the aldehyde 2 (present in equilibrium with 3). Thus cinnamic acid (4a) was formed in -75% yield with 3 as the substrate under the normal conditions of the Perkin reaction. The deprotonation of the diacetate appears to be electrophilically assisted by the neighbouring acetate group, the resulting enolate being also thermodynamically stabilized in form of an orthoester (I). The possibility that the diacetate 3 is the actual substrate in the Perkin reaction indicates that the reaction can be effected under far milder conditions, with a base much stronger than acetate ion. This was indeed realized with potassium t-butoxide in dioxane, which converted the gem-diacetates derived from a variety of aromatic aldehydes to the corresponding cinnamic acids (4), rapidly and in good yields at room temperature (Scheme 2). This represents a vast improvement in the synthetic protocol for the classical Perkin reaction, which remains an important carbon-carbon bond forming reaction to this day. Chapter 2. Aromaticity in azlactone anions and its sifnificance for the Erlenmeyer synthesis.2 The classical Erlenmeyer azlactone synthesis of amino acids occur via the formation of an intermediate azlactone, and its subsequent deprotonation by a relatively weak base(acetate ion),. The resulting azlactone anion (cf. II, Scheme 3) functions as a glycine enolate equilvalent, and is considered in situ with an aromatic aldehyde, subsequent dehydration leading to the 4-alkylidene oxazolone(analogously to the Perkin reaction). Interestingly, azlactone anions are possibly aromatic, as they possess 6π electrons in cyclic conjugation; this would explain their facile formation as also the overall success of the Erlenmeyer synthesis. The following studies evidence this possibility. The strategy involved studying the rates of base-catalyzed deprotonation in 2-phenyl-5(4H)-oxazolone (azlactone, 5) and its amide and ketone analogs, 3-methyl-2-phenyl-4(5H)-imidazolone (6), and 3,3-dimethyl-2-phenyl-493H)-pyrrolone (7) respectively.2 Two processes were studied, deuterium exchange and condensation with hexadeuteroacetone (Scheme3): both are presumably mediated by the anions II-IV, so their stabilities would govern the overall rates. These were followed by 1H NMR spectroscopy by monitoroing the disappearance of the resonance of the proton α to the carbonyl group. The order of deprotonation was found to be 6 > 5 > 7. However, the expected order based on pKa values would be ketone > ester > amide, i.e. 7 > 5 > 6. The inverted order observed strongly indicates the incursion of aromaticity, which would be enhanced by the electron-donor capabilities of the heteroatoms is 5 and 6. This is further substantiated by the greater reactivity in the case of the nitrogen analog 6 relative to the oxygen 5, which parallel the electronegativity order. (The aromaticity order would thus be: III > II > IV. The imidazole nucleus is indeed to be considerably more aromatic than the oxazole.) The synthesis of the analogs 6 and 7 was accomplished via an interesting intramolecular aza-Wittig reaction (Schemes 4 & 5) Chapter 3. Umpolung approach to the Erlenmeyer process in the synthesis of dehydro amino acids. These studies are based on the general observation that most of the strategies for the synthesis of α-amino acids introduce the side chain (or part was inverted in an umpolung sense. The key reaction studied was that of 2-phenyl-4-ethoxymethylne-5(4H)-oxazolone (11) with Grignard reagents: this resulted in the opening to yield a protected dehydro amino acid (12), in good to excellent yields (65-87%)(Scheme ^). As the azlactone reactant 11 is the ekectrophilic partner, this may be viewed as a partial umpolung version of the classical Erlenmeyer process. The readily available reactants, simple procedure and mild reaction conditions make this a very attractive method for the synthesis of a variety of α-dehydro amino acids. Chapter 4. The Erlenmeyer azlactone synthesis with aliphatic aldehydes under solvent-free microwave conditions. 3 A serious limitation to the classical Erlenmeyer reaction is that it generally fails in the case of aliphatic aldehydes. This chapter describes a convenient approach to this problem that extends the scope of the Erlenmeyer synthesis, via a novel microwave-induced, solvent-free process. This, it was observed that azlactones (5) react with aliphatic aldehydes (13) upon adsorption on neutral alumina and irradiation with microwaves (< 2 min), forming the corresponding Erlenmeyer products (14) in good yields (62-78%, Scheme 7). (The possible mechanistic basis of the procedure, which is presumably mediated by V , is discussed).3 Chapter 5. 2,4, 10-Trioxaadamantane as a carboxyl protecting group: application to the asymmetric synthesis of α-amino acids (umpolung approach).It is known that the 2,4,10-trioxaadamantane moiety is not only remarkably stable to nucleophilic attack, but can also be easily hydrolyzed to the corresponding carboxylic acid.4 It was of interest to apply this carboxyl protection strategy for designing a synthesis of α-amino acids, essentially by starting with a protected glyoxylic acid. The corresponding aldimine was expected to (nucleophilically) add organometallic reagents at the C=N moiety (cf. Shceme 8), the side chain of the amino acid being thus introduced in umpolung fashion. Also, a chiral aldimine would define an asymmetric synthesis of amino acids. Indeed, the chiral aldimine 17, derived from 2,4,10-troxaadamantane-3-carbaldehyde 15 and [(S)-(-)-1-phenylethylamine] 16, reacted with a variety of Grignard reagents to furnish the corresponding protected α-amino acids (18) in good yields, with moderate diastereometric excess (Scheme 8). Better yields and ‘de’ values were obtained with organolithium reagents. Chapter 6: possible one-pot oligopeptide synthesis with azlactones or amino acid N-carboxyanhydrides (NCAs). This chapter describes a novel approach to oligopeptide synthesis employing azlactones or NCA’s as amino acid equivalents which are simultaneously protected and activated (Scheme 9). Thus, the addition of the 4-substituted 2-benzyloxyazlactone (19) to an N-protected amino acid under basic conditions, was initially explored. The reaction was expected to yield a dipeptide (21) via the rearrangement of the mixed anhydride intermediate (VI) (Scheme 9). The subsequent addition of a different azlactone to the dipeptide (21) would analogously lead to the formation of a tripeptide (22). This may be performed repetitively to define a strategy for C-terminal extension of an oligopeptide chain, noting that no intervening deprotecting and activating steps are necessary. (In toto deprotection may be effected finally via the hydrogenolyis of the bvenzyloxy groups, to obtain 23.) A closely analogous strategy may also be envisaged by employing N.carboxyanhydrides (NCA’S, 24) instead of azlactones, as shown in Scheme 10 (forming dipeptide 26 and tripeptide 27). The main difference n this case is that the carbamic acid moiety of the intermediate mixed anhydride (VII) is expected to undergo decarboxylation to VIII (thus obviating the need for a deprotection step). However, this putative advantage is offset by the instability of NCA’s and their tendency toward polymerization. However, only partial success could be achieved in these attempts, although a variety of conditions were explored. The strategy and the experimental results have been analyzed in detail, as this interesting approach appears to be promising, and worth further study. (For structural formula pl refer the pdf file)