12.07.2015 Views

Contributions à l'Etude du Vertex Topologique en Théorie ... - Toubkal

Contributions à l'Etude du Vertex Topologique en Théorie ... - Toubkal

Contributions à l'Etude du Vertex Topologique en Théorie ... - Toubkal

SHOW MORE
SHOW LESS
  • No tags were found...

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

UNIVERSITÉ MOHAMMED V – AGDALFACULTÉ DES SCIENCESRabatN d’ordre : 2454THÈSE DE DOCTORATPrés<strong>en</strong>tée parHOUDA JEHJOUHDiscipline :Spécialité :PhysiquePhysique des Hautes EnergiesTitre :<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong>Théorie des CordesSout<strong>en</strong>ue le 27/06/2009,devant le juryPrésid<strong>en</strong>t :M. Daoud Professeur, Faculté des Sci<strong>en</strong>ces, AgadirExaminateurs :M. Ait B<strong>en</strong> Haddou Professeur, Faculté des Sci<strong>en</strong>ces, MeknèsA. Belhaj Dr Chercheur, CNESTEN, RabatE. H. El Kinani Professeur, Faculté des Sci<strong>en</strong>ces et Techniques, ErrachidiaA. Jellal Professeur, Faculté des Sci<strong>en</strong>ces, El JadidaT. Lhallabi Professeur, Faculté des Sci<strong>en</strong>ces, RabatE. H. Saidi Professeur, Faculté des Sci<strong>en</strong>ces, RabatJ. Zerouaoui Professeur, Faculté des Sci<strong>en</strong>ces, K<strong>en</strong>itraFaculté des Sci<strong>en</strong>ces, 4 Av<strong>en</strong>ue Ibn Battouta B.P. 1014 RP, Rabat – MarocTel +212 (0) 37 77 18 34/35/38, Fax : +212 (0) 37 77 42 61, http://www.fsr.ac.ma


Table des matières1 Intro<strong>du</strong>ction Générale 112 Variétés de CY et Géométrie Torique 232.1 Généralités sur les variétés de CY . . . . . . . . . . . . . . . . . . . . . . . 232.1.1 Variétés complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . 242.1.2 Variétés de Calabi-Yau . . . . . . . . . . . . . . . . . . . . . . . . . 262.1.3 Exemples des variétés de Calabi-Yau . . . . . . . . . . . . . . . . . 272.2 Conifold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302.2.1 Conifold singulier . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302.2.2 Conifold déformé et conifold résolu . . . . . . . . . . . . . . . . . . 312.2.3 Transitions géométriques . . . . . . . . . . . . . . . . . . . . . . . . 332.3 Variétés de CY toriques . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342.3.1 Variétés toriques . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.3.2 Variétés de Calabi-Yau toriques . . . . . . . . . . . . . . . . . . . . 362.3.3 Sous-variétés spéciales lagrangi<strong>en</strong>nes (SSL) . . . . . . . . . . . . . . 372.3.4 Diagrammes toriques . . . . . . . . . . . . . . . . . . . . . . . . . . 393 Cordes et Invariants <strong>Topologique</strong>s 473.1 Théorie des cordes topologiques . . . . . . . . . . . . . . . . . . . . . . . . 473.1.1 Modèle sigma non linéaire N = 2 . . . . . . . . . . . . . . . . . . . 483.1.2 Modèles A et B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523.1.3 Modèle sigma linéaire jaugé . . . . . . . . . . . . . . . . . . . . . . 573.1.4 Branes topologiques . . . . . . . . . . . . . . . . . . . . . . . . . . 583.2 Dualité corde ouverte / corde fermée . . . . . . . . . . . . . . . . . . . . . 593.2.1 Dualité jauge/ gravité . . . . . . . . . . . . . . . . . . . . . . . . . 603.2.2 Dualité de Gopakumar-Vafa . . . . . . . . . . . . . . . . . . . . . . 601


TABLE DES MATIÈRES3.3 Invariants topologiques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623.3.1 Invariants de Gromov-Witt<strong>en</strong> . . . . . . . . . . . . . . . . . . . . . 633.3.2 Invariants de Gopakumar-Vafa . . . . . . . . . . . . . . . . . . . . . 643.3.3 Boucle de Wilson, invariants de noeuds et d’<strong>en</strong>trelacs . . . . . . . . 663.4 Modèle B et espace twistoriel . . . . . . . . . . . . . . . . . . . . . . . . . 703.4.1 Espace Twistoriel . . . . . . . . . . . . . . . . . . . . . . . . . . . . 703.4.2 Contribution : Pure fermionic twistor like model . . . . . . . . . . . 723.4.3 Contribution : Théorie des cordes twistorielles . . . . . . . . . . . . 724 Modèle Cristallin <strong>du</strong> <strong>Vertex</strong> C 3 1054.1 Variétés de CY toriques et cristal fon<strong>du</strong> . . . . . . . . . . . . . . . . . . . . 1064.1.1 Variété de Calabi-Yau quantique . . . . . . . . . . . . . . . . . . . 1064.1.2 Cristal de CY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1074.1.3 Matrice de transfert . . . . . . . . . . . . . . . . . . . . . . . . . . 1084.2 Fonction de partition perp<strong>en</strong>diculaire . . . . . . . . . . . . . . . . . . . . . 1124.2.1 Formule P λ,µ,ν (q) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1124.2.2 Calcul de la fonction de partition perp<strong>en</strong>diculaire . . . . . . . . . . 1134.3 La fonction perp<strong>en</strong>diculaire raffinée . . . . . . . . . . . . . . . . . . . . . . 1154.3.1 Fonction de partition avec un nombre infini de paramètres . . . . . 1164.3.2 P λµν (q, t) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184.4 Modèle <strong>du</strong> cristal fon<strong>du</strong> et conifold résolu . . . . . . . . . . . . . . . . . . . 1214.4.1 Modèle cristallin avec un seul mur . . . . . . . . . . . . . . . . . . 1214.4.2 Modèle cubique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1224.5 Invariants topologiques dans le modèle cristallin . . . . . . . . . . . . . . . 1234.5.1 Invariant unknot . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1244.5.2 Invariant Entrelacs de Hopf . . . . . . . . . . . . . . . . . . . . . . 1254.6 Contribution : G<strong>en</strong>eralized MacMahon . . . . . . . . . . . . . . . . . . . . 1265 <strong>Vertex</strong> <strong>Topologique</strong> Usuel 1575.1 <strong>Vertex</strong> topologique I : formalisme . . . . . . . . . . . . . . . . . . . . . . . 1575.1.1 <strong>Vertex</strong> et amplitude de la corde ouverte . . . . . . . . . . . . . . . . 1585.1.2 Symétrie de vertex . . . . . . . . . . . . . . . . . . . . . . . . . . . 1595.1.3 Collage des vertex . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1605.1.4 Règles de collage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1612


TABLE DES MATIÈRES5.2 <strong>Vertex</strong> topologique II : Calcul des amplitudes . . . . . . . . . . . . . . . . 1635.2.1 L’espace complexe C 3 . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.2.2 Variété locale P 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.2.3 Les branes non-compactes . . . . . . . . . . . . . . . . . . . . . . . 1645.2.4 La ”casquette” et le ”pantalon” . . . . . . . . . . . . . . . . . . . . . 1675.3 <strong>Vertex</strong> <strong>Topologique</strong> et Théorie de Chern-Simons . . . . . . . . . . . . . . . 1715.3.1 Ingrédi<strong>en</strong>ts de Chern-Simons . . . . . . . . . . . . . . . . . . . . . . 1715.3.2 Expression de l’amplitude . . . . . . . . . . . . . . . . . . . . . . . 1725.4 <strong>Contributions</strong> . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1745.4.1 Topological String on Toric CY3s in Large Complex Structure Limit 1745.4.2 Non Planar Topological 3-<strong>Vertex</strong> Formalism . . . . . . . . . . . . . 1756 <strong>Vertex</strong> <strong>Topologique</strong> Raffiné 2106.1 Formalisme <strong>du</strong> vertex raffiné . . . . . . . . . . . . . . . . . . . . . . . . . . 2106.1.1 <strong>Vertex</strong> raffiné et fonction de partition de la corde ouverte . . . . . . 2126.2 Fonctions de partitions raffinées . . . . . . . . . . . . . . . . . . . . . . . . 2146.2.1 O(−1) ⊕ O(−1) ↦→ P 1 . . . . . . . . . . . . . . . . . . . . . . . . . 2146.2.2 O(0) ⊕ O(−2) → P 1 . . . . . . . . . . . . . . . . . . . . . . . . . . 2156.2.3 Variété P 1 × P 1 locale . . . . . . . . . . . . . . . . . . . . . . . . . 2176.3 <strong>Vertex</strong> topologique sous la transition de flop . . . . . . . . . . . . . . . . . 2186.4 <strong>Vertex</strong> raffiné et homologie des <strong>en</strong>trelacs . . . . . . . . . . . . . . . . . . . 2206.4.1 Unknot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2236.4.2 Entrelac de Hopf . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2246.5 Contribution : Refining the Shifted 3-<strong>Vertex</strong> . . . . . . . . . . . . . . . . . 2257 Conclusion et Perspectives 2388 Annexe : Fonctions de Schur et MacMahon 2438.1 Diagrammes et tableaux de Young . . . . . . . . . . . . . . . . . . . . . . 2438.1.1 Définitions et propriétés . . . . . . . . . . . . . . . . . . . . . . . . 2448.1.2 Diagrammes de Young et groupes de symétrie . . . . . . . . . . . . 2498.1.3 Notation de Frob<strong>en</strong>ius . . . . . . . . . . . . . . . . . . . . . . . . . 2508.1.4 Diagramme de Maya . . . . . . . . . . . . . . . . . . . . . . . . . . 2518.1.5 Partitions planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2528.1.6 Partitions strictes et diagramme de Young shifté . . . . . . . . . . . 2533


TABLE DES MATIÈRES8.2 Fonctions de Schur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2558.2.1 Propriétés de la fonction de Schur . . . . . . . . . . . . . . . . . . . 2588.2.2 Opétareurs vertex . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2598.3 Fonction de MacMahon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2638.3.1 Conjecture de MacMahon . . . . . . . . . . . . . . . . . . . . . . . 2648.3.2 Fonctions génératrices . . . . . . . . . . . . . . . . . . . . . . . . . 2654


Je dédie cette thèse àA mes très chers par<strong>en</strong>tsqui m’ont accompagnée tout au long de ce parcours périlleux ; qui ont toujours étélà dans mes mom<strong>en</strong>ts de détresse, qui se sont toujours dévoués et sacrifiés pourmoi. Merci pour leur amour inconditionnel et pour avoir toujours cru <strong>en</strong> moi. Je netrouverai jamais les mots suffisants pour dire merci. Mes réussites sont aussi lesleurs. Enfin ! Merci tout simplem<strong>en</strong>t d’être… mes par<strong>en</strong>ts.A mes très chères sœurs Mounia et Safaequi m’ont toujours aidée, écoutée, sout<strong>en</strong>ue et <strong>en</strong>couragée tout au long de monparcours ; qui ont toujours été prés<strong>en</strong>tes pour moi.


Avant ProposCe travail à été effectué au Laboratoire/ UFR de Physique des Hautes Energies à laFaculté des Sci<strong>en</strong>ces de Rabat et sous la direction de Mr le professeur El Hassan Saidi. Cetravail n’aurait pu être réalisé sans l’accord, le souti<strong>en</strong> et l’aide de plusieurs personnes.Mes plus sincères remerciem<strong>en</strong>ts vont à mon directeur de thèse, Mr El Hassan Saidi,professeur à la Faculté des sci<strong>en</strong>ces de Rabat. Je le remercie de m’avoir accueillie et intégréedans l’équipe de recherche au sein <strong>du</strong> Laboratoire/ UFR Physique des Hautes Energieset de m’avoir très vite <strong>en</strong>couragée à participer à des échanges sci<strong>en</strong>tifiques qui m’ont étébénéfiques. Ses conseils, sa disponibilité et ses connaissances ont largem<strong>en</strong>t contribué àla réussite de cette thèse. Je le remercie pour sa rigueur et son souci de clarté qui m’ontaidé à aller plus loin. J’ai particulièrem<strong>en</strong>t apprécié l’intérêt qu’il a porté à mes travauxet la réelle collaboration qu’il <strong>en</strong>treti<strong>en</strong>t avec l’équipe de recherche <strong>du</strong> Laboratoire/ UFRPhysique des Hautes Energies. Ces <strong>en</strong>couragem<strong>en</strong>ts constants m’ont donné suffisamm<strong>en</strong>tde confiance <strong>en</strong> moi pour m<strong>en</strong>er cette thèse à bout. Soyez assuré, Monsieur, de toute monestime et de mon profond respect.J’adresse mes vifs remerciem<strong>en</strong>ts à Monsieur M. Daoud Professeur à la Faculté desSci<strong>en</strong>ces, Agadir qui m’a fait l’honneur d’accepter la présid<strong>en</strong>ce <strong>du</strong> jury. Veuillez agréer,Monsieur, l’assurance de mon profond respect.Je ti<strong>en</strong>s à exprimer ma très vive reconnaissance <strong>en</strong>vers Madame T. Lhallabi Professeurà la Faculté des Sci<strong>en</strong>ces, Rabat. De me faire l’honneur de participer au jury de cette thèseet pour l’intérêt qu’elle a porté à ce travail. Sa grandeur sci<strong>en</strong>tifique et sa g<strong>en</strong>tillesse ont été6


Avant Propospour moi de véritables atouts. Veuillez recevoir, Madame, l’expression de mes respectueuxs<strong>en</strong>tim<strong>en</strong>ts.Un grand merci à M. Ait B<strong>en</strong> Haddou Professeur à la Faculté des Sci<strong>en</strong>ces, Meknès,qui a investi beaucoup de temps dans une lecture minutieuse de la thèse, ses remarques ontpermis de corriger des erreurs et d’apporter des précisions <strong>en</strong> vue d’améliorer la qualité dece docum<strong>en</strong>t. Veuillez recevoir, Professeur, l’assurance de ma considération distinguée.J’adresse mes remerciem<strong>en</strong>ts à Monsieur E. H. El Kinani Professeur à la Faculté desSci<strong>en</strong>ces et Techniques, Errachidia, qui a bi<strong>en</strong> voulu faire partie <strong>du</strong> jury et d’apporter leursvives contributions à l’<strong>en</strong>richissem<strong>en</strong>t de ce travail. Veuillez agréer, Monsieur, l’assurancede mon profond respect.Je ti<strong>en</strong>s égalem<strong>en</strong>t à adresser mes remerciem<strong>en</strong>ts à Monsieur A. Jellal Professeur à laFaculté des Sci<strong>en</strong>ces El Jadida, pour l’honneur et le privilège qu’il me fait de juger cemodeste travail. Veuillez recevoir, Monsieur, l’expression de mes respectueux s<strong>en</strong>tim<strong>en</strong>ts.Je ti<strong>en</strong>s égalem<strong>en</strong>t à remercier Monsieur J. Zerouaoui Professeur à la Faculté desSci<strong>en</strong>ces, K<strong>en</strong>itra. Pour l’intérêt qu’il a porté à ce travail et dont les remarques et lessuggestions m’ont permis d’améliorer le prés<strong>en</strong>t manuscrit. Veuillez agréer, Monsieur, l’expressionde mon respect et de ma profonde gratitude.J’adresse mes remerciem<strong>en</strong>ts aussi à Monsieur A. Belhaj Docteur Chercheur, CNES-TEN à Rabat, pour avoir accepté de faire partie de mon jury de thèse et d’avoir reluet comm<strong>en</strong>té des chapitres de ma thèse de manière si approfondie. Merci aussi pour sesconseils sci<strong>en</strong>tifiques, et son regard critique. Veuillez trouver ici l’expression de ma respectueuseconsidération.Je voudrais adresser un remerciem<strong>en</strong>t particulier à tous le corps professoral <strong>du</strong> DESA,pour leur souti<strong>en</strong> et pour leur pati<strong>en</strong>ce. A Prof M. Ait B<strong>en</strong> Haddou, Dr A. Belhaj, Dr L.BDrissi, Prof A. El K<strong>en</strong>z, Dr M. Kessabi, Prof T. Lhallabi, Dr A. Moujib, Prof E.H. Saidi,Prof M.B. Sedra, Prof J.Zerouaoui. Ils ont contribué à nous transmettre leur savoir pourassurer notre formation..7


Lab/UFR PHEUne thèse représ<strong>en</strong>te un parcours long et difficile, mais dans mon cas il fut agrém<strong>en</strong>tépar une atmosphère dét<strong>en</strong><strong>du</strong>e.A Lalla Btissam Drissi, je ti<strong>en</strong>s à te témoigner toute ma reconnaissance pour la qualitéde sa collaboration. C’est à cette occasion que je te remercier pour tout ce que tu asfait pour moi p<strong>en</strong>dant mon DESA, puis ma thèse. Je ti<strong>en</strong>s <strong>en</strong>fin à te remercier pour tonimplication dans ce travail de thèse, pour ton aide, ta disponibilité, tes nombreux conseilset ton souti<strong>en</strong> sans faille. Sois assuré, Btissam, de ma profonde gratitude et de toute monamitié.A Asmaâ Moujib, Je t’exprime ma reconnaissance pour le temps passé à corriger deserreurs et d’apporter des précisions <strong>en</strong> vue d’améliorer la qualité de ce docum<strong>en</strong>t. Je teremercie égalem<strong>en</strong>t pour tes nombreux conseils amicaux et pour tous ces bons mom<strong>en</strong>tspassés. Sois assuré, Asmaâ, de toute mon amitié..Si j’<strong>en</strong> suis arrivée là aujourd’hui, c’est sans doute parce que j’ai r<strong>en</strong>contré sur mon chemindes personnes qui m’ont apporté beaucoup. Je remercie l’<strong>en</strong>semble de l’équipe de recherche<strong>du</strong> Laboratoire/ UFR Physique des Hautes Energies pour les nombreuses et toujours fructueusesdiscussions. Travailler avec eux a été un réel plaisir.Plus particulièrem<strong>en</strong>t à Rachid Ahl Laamara, Adil Belhaj, Abdelouahed Jraifi, AzizRhalami, Hanane Sebatta, A. Soumili. Abdou qui ont égalem<strong>en</strong>t contribué à cette thèsepar des réflexions, des précisions et des appuis technique.Je ne peux pas oublier tous mes amies qui m’ont accompagnée tout au long de ceparcours et qui n’ont pas cessé de me sout<strong>en</strong>ir moralem<strong>en</strong>t et de m’<strong>en</strong>courager, et plusparticulièrem<strong>en</strong>t à Fatimazahra Elhassani, Nabila Mamouni, Souad Nekhlaoui.Je ti<strong>en</strong>s aussi à remercier l’<strong>en</strong>semble de l’équipe de recherche <strong>du</strong> Laboratoire de Magnétismeet Physique des Hautes Energies..Ne pouvant malheureusem<strong>en</strong>t pas citer toutes les personnes que j’ai r<strong>en</strong>contré <strong>du</strong>rant monparcours et qui ont contribué d’une façon ou d’une autre, de près ou de loin, à l’aboutissem<strong>en</strong>tde cette thèse, je leur dis à toutes merci d’avoir été là à cette instant précis où je8


Avant Proposles ai r<strong>en</strong>contrées et où ils m’ont apportée cette aide qui a surem<strong>en</strong>t contribuée à aller aubout de cette thèse.Je ti<strong>en</strong>s à remercier- Le programme de la bourse d’excell<strong>en</strong>ce, CNRST. Grâce à cet appui stimulant,j’ai été capable de me conc<strong>en</strong>trer davantage sur mes études.- Le programme Protars III D12/25 CNRST.- Le groupem<strong>en</strong>t national de Physique des Hautes Energies GNPHE.- Le c<strong>en</strong>tre international de Physique Théorique, ICTP.Je ne pourrais trouver un épanouissem<strong>en</strong>t complet dans le travail si ma vie privée n’étaitpas elle-même épanouie. Je ti<strong>en</strong>s donc à remercier mes chers par<strong>en</strong>ts qui r<strong>en</strong>d<strong>en</strong>t toutsimplem<strong>en</strong>t ma vie belle au quotidi<strong>en</strong>, pour leurs <strong>en</strong>couragem<strong>en</strong>ts inestimables, pour leurspati<strong>en</strong>ces, leurs prés<strong>en</strong>ces a mes côtes, leurs contributions à l’élaboration de ce manuscrit.Ma chère Safae, merci d’être toujours là, même quand le temps est plus maussade. Ungrand merci à Mounia et son mari Jamal, à mon frère Rachid. Et Enfin Merci Marwa etImane de me faire sourire chaque jour.Je suis seul à signer cette thèse et pourtant de nombreuses personnes ont contribué par leuraide, leur souti<strong>en</strong> et leurs conseils à la faire exister. Je ti<strong>en</strong>s à remercier très sincèrem<strong>en</strong>ttous ceux qui ont participé de près ou de loin à ce travail. Je terminerai <strong>en</strong> disant que“Sci<strong>en</strong>tific thought is the common heritage of mankind”Ab<strong>du</strong>s Salam9


10Lab/UFR PHE


Chapitre 1Intro<strong>du</strong>ction GénéraleL’une des grandes quêtes de la physique des hautes énergies est l’unification des quatreforces fondam<strong>en</strong>tales (électromagnétique, faible, forte et gravitationnelle) <strong>en</strong> une seule théorie.A l’heure actuelle, seule la théorie des supercordes propose un cadre assez large pourunifier de manière cohér<strong>en</strong>te ces quatre interactions de la nature [1]-[10]. La théorie dessupercordes, depuis son apparition dans le monde de la physique, ne cesse de surpr<strong>en</strong>drepar des applications dans des domaines très variés des sci<strong>en</strong>ces fondam<strong>en</strong>tales ; <strong>en</strong> particulier<strong>en</strong> physique des particules élém<strong>en</strong>taires, la physique mathématique, la physiquestatistique des phénomènes critiques [11, 12, 13], les théories des champs conformes à deuxdim<strong>en</strong>sions [14, 15], la physique des branes, la physique des trous noirs [16, 17] et la théoriedes noeuds [18, 19]. Avant de passer à la prés<strong>en</strong>tation <strong>du</strong> cont<strong>en</strong>u de ce mémoire de thèse,qui traite certains aspects des modèles des supercordes tout <strong>en</strong> s’appuyant sur la théoriedes cordes topologiques et <strong>en</strong> faisant usage de la méthode <strong>du</strong> 3-vertex topologique et sonli<strong>en</strong> avec des modèles critiques de la physique statistique, il est utile de comm<strong>en</strong>cer parrappeler brièvem<strong>en</strong>t des résultats fondam<strong>en</strong>taux <strong>en</strong> théorie de supercorde permettant à lafois d’intro<strong>du</strong>ire rapidem<strong>en</strong>t certains concepts facilitant la lecture de ce travail de rechercheet situant le cadre générale de cette étude.Théorie des supercordesGénéralem<strong>en</strong>t, on distingue differ<strong>en</strong>ts types de cordes dont certaines sont physiques etd’autres non physiques comme c’est le cas de la corde bosonique X µ (τ, σ) vivant dansun espace temps à 26 dim<strong>en</strong>sions (µ = 0, .., 25) [20]-[30] et la corde topologique sur desvariétés de Calabi-Yau 3-folds que nous aurons l’occasion d’analyser avec détails dans cetravail de thèse. Dans toutes ces théories, les cordes lors de leurs mouvem<strong>en</strong>ts dans l’espacetemps décriv<strong>en</strong>t une surface d’univers à deux dim<strong>en</strong>sions (dont la version euclidi<strong>en</strong>ne cor-11


<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong> Théorie des Cordesrespond à une surface de Riemann) paramétrisée par des coordonnées locales ξ α = (τ, σ).La dynamique d’une corde libre bosonique représ<strong>en</strong>tée par les champs X µ (ξ) est donnéepar l’action S NG de Nambu-Goto [2, 7]. Cette dernière est décrite de façon plus appropriéepar l’action de Polyakov S P olyakov faisant usage de champs auxiliaires portés par le champgravitationelle h αβ (ξ) à 2D [20, 21]. Les cordes bosoniques (ouvertes avec 0 ≤ σ ≤ π etfermées avec 0 ≤ σ ≤ 2π ) prés<strong>en</strong>t<strong>en</strong>t des modes de vibration α µ n qui donn<strong>en</strong>t naissanceà des particules similaires à celles qui transport<strong>en</strong>t les interactions de jauge A µ (x) oude gravitation G µν (x). Les cordes X µ (ξ) peuv<strong>en</strong>t être alors ouvertes avec des extrémitésX µ (ξ) | σ=0,π libres ou fermées lorsque les deux extrémités se boucl<strong>en</strong>t l’une sur l’autre ;c.à.d X µ (τ, σ = 0) = X µ (τ, σ = 2π).Dans le cas des cordes ouvertes, il est nécessaire de specifier la nature des conditions à leursextrémités. On distingue ess<strong>en</strong>tiellem<strong>en</strong>t deux types de conditions aux bords ; il s’agit : (i)condition de Dirichlet :X µ (τ, σ) |éxtrémités= constante,et (ii) condition de Neumann,∂ σ X µ (τ, σ) |éxtrémités= 0. (1.1)La quantification de la corde bosonique libre fait apparaître un état tachyonique (de massecarrée négative m 2 < 0). Ce problème est surmonté si des fermions ψ µ ± (τ, σ), conv<strong>en</strong>ablem<strong>en</strong>tassociés aux bosons par supersymétrie sont intro<strong>du</strong>its sur la surface d’univers de lacorde con<strong>du</strong>isant ainsi à la théorie des supercordes de dim<strong>en</strong>sion critique ré<strong>du</strong>ite à D = 10.Selon le nombre de charges supersymétriques, on distingue cinq théories de supercordescomme il est m<strong>en</strong>tionné dans le tableau suivant [1]-[22] :supercorde type IIA type IIB hétérotique hétérotiquetype ISO (32) E 8 × E 8fermé<strong>en</strong>ature fermée fermée fermée ferméeouverte# de supersymétries 32 32 16 16 16symétrie de jauge − − SO (32) E 8 × E 8 SO (32)Finalem<strong>en</strong>t, à cause de l’abs<strong>en</strong>ce de diverg<strong>en</strong>ces ultraviolettes et de l’exist<strong>en</strong>ce d’une particulede spin 2 et de masse nulle (graviton G µν ) dans leurs spectres perturbatifs, les théoriesdes supercordes sont actuellem<strong>en</strong>t les seules candidats pour une théorie quantique de gravitation.Ces théories, qui exist<strong>en</strong>t seulem<strong>en</strong>t à 10-dim<strong>en</strong>sions, mais pouvant être aussi12


Intro<strong>du</strong>ction Généraleré<strong>du</strong>ites vers des dim<strong>en</strong>sions inférieures par compactification, sont étroitem<strong>en</strong>t reliées lesunes aux autres par des relations de <strong>du</strong>alité [2]. Il existe ess<strong>en</strong>tiellem<strong>en</strong>t trois types derelations de <strong>du</strong>alité :– <strong>du</strong>alité S qui relie le régime de couplage faible d’une première théorie est au régimede couplage fort d’une seconde théorie. Cette symétrie S inverse la constante decouplage S : g ←→ 1/g,– <strong>du</strong>alité T : Deux théories de supercordes compactifiées sur un cercles de rayon Rsont dites T- <strong>du</strong>ales si elles sont invariantes sous la transformation R ←→ 1/R. La<strong>du</strong>alité T agit par inversion <strong>du</strong> rayon de compactification R.– <strong>du</strong>alité U : échange les <strong>du</strong>alités S et T, soit U : g ←→ 1/R.Les théories de supercordes type IIA et type IIB compactifiées sur un cercle sont T -<strong>du</strong>ales. C’est le cas aussi des supercordes hétérotiques E 8 × E 8 et SO(32) à 9 dim<strong>en</strong>sions.La supercorde hétérotique SO(32) et la supercorde de type I sont S <strong>du</strong>ales : la limite decouplage faible de la théorie effective de l’une correspond à la limite de couplage fort del’autre. La limite de couplage fort de la corde de type IIB s’id<strong>en</strong>tifie à elle-même.Fig. 1.1 – Quelques <strong>du</strong>alités <strong>en</strong>tre les cinq types de théories des supercordes à 10D et 9D.Dans le régime perturbatif des théories de supercordes de type IIA et type IIB, il existedans le spectre des états de masse nulle des (p + 1)-formes C p+1 = C µ0 .. µp dx µ 0 ∧ .. ∧ dxµ pqui généralis<strong>en</strong>t le champ de Maxwell et qui admett<strong>en</strong>t une interprétation remarquable.13


<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong> Théorie des CordesLes champs C µ0 .. µp sont antisymétriques <strong>en</strong> les (p + 1) indices et sont associés à des étatsnon perturbatifs <strong>en</strong> occur<strong>en</strong>ce les branes Dp auxquelles nous nous référons par Dp- branes.Ces objets solitoniques ont aussi la propriété d’être des états BPS de la théorie, c’est à direqu’elles préserv<strong>en</strong>t la moitié des charges supersymétriques de la théorie des supercordes[130]. Pour se faire une idée plus précise sur ces objets, nous donnons dans ce qui suitcertaines de leurs propriétés caractéristiques.D-branesUne D-brane, ou plus exactem<strong>en</strong>t, une Dp-brane avec p <strong>en</strong>tier positif pr<strong>en</strong>ant des valeurs<strong>en</strong>tre zero et 7 est un objet ét<strong>en</strong><strong>du</strong> apparaissant dans le spectre non perturbatif de lathéorie des supercordes type II. Le nombre p est le nombre de dim<strong>en</strong>sions spatiales danslesquelles la brane a des ext<strong>en</strong>sions. Il faut rajouter à ce nombre une dim<strong>en</strong>sion temporellepour obt<strong>en</strong>ir le nombre total de (p + 1) dim<strong>en</strong>sions. Une des propriétés remarquables desD-branes est qu’elles sont intimem<strong>en</strong>t liées aux supercordes ouvertes. Une D-brane est unebrane sur laquelle sont fixées les extrémités des supercordes ouvertes qui sont à l’origine dela matière qu’elle conti<strong>en</strong>t et et offre ainsi un li<strong>en</strong> avec le modèle standard de la physiquedes particules. Signalons au passage que le suffix "D" dans la terminologie D-brane réfèreà la condition de Dirichlet (1.1). Il est égalem<strong>en</strong>t possible que les deux extrémités d’unesupercorde ouverte soi<strong>en</strong>t fixées dans deux D-branes distinctes n’ayant pas forcém<strong>en</strong>t lemême nombre de dim<strong>en</strong>sions. Lorsqu’on veut préciser le nombre de dim<strong>en</strong>sions dans lequella D-brane a des ext<strong>en</strong>sions, on parle alors de Dp-brane qui <strong>en</strong>g<strong>en</strong>dre un volume d’universde dim<strong>en</strong>sion p + 1. Notons aussi que les D- branes dans la théorie des de supercordeIIA sont : D0, D2, D4, D6. Celles de la supercorde IIB sont : D1, D3, D5 et D7. Ilexiste d’autres types de branes ; <strong>en</strong> particulier celles admettant des dim<strong>en</strong>sions compactes<strong>en</strong>roulant des cycles de variétés de compactifications.Avec ces outils <strong>en</strong> main, nous sommes maint<strong>en</strong>ant <strong>en</strong> position d’aborder la théorie descordes topologiques constituant un des objectifs principaux de memoire de thèse.Théorie des cordes topologiquesA partir de la théorie des supercordes sur une variété Calabi-Yau non compacte et avec desD-branes qui <strong>en</strong>roul<strong>en</strong>t certaines p-cycles des variétés, les propriétés de la théorie de jaugesont alors exprimées dans la structure géométrique de la variété Calabi-Yau. A priori, onpeut calculer plusieurs quantités physiques à basses énergies dans le cadre de cette théorie; <strong>en</strong> particulier le superpot<strong>en</strong>tiel effectif qui décrit la structure des vides de la théoriede jauge. Il se trouve que ces calculs ont des interprétations remarquables dans le cadre dela théorie des cordes topologiques que nous décrivons rapidem<strong>en</strong>t dans ce qui suit.14


Intro<strong>du</strong>ction GénéraleLa théorie des cordes topologiques à été intro<strong>du</strong>ite par Witt<strong>en</strong> dans [34]- [37] comme unmodèle simplifié de la théorie des supercordes qui conti<strong>en</strong>t des informations topologiquesde l’espace cible. Elle est considérée comme un sous secteur de la théorie des supercordesmuni de plusieurs applications physiques et mathématiques. Les amplitudes de la théoriedes cordes topologiques sont profondém<strong>en</strong>t liées aux amplitudes de la théorie des supercordesde type II. La description d’espace-temps des cordes topologiques <strong>en</strong> termes dethéorie des champs se ré<strong>du</strong>it ess<strong>en</strong>tiellem<strong>en</strong>t à des théories de jauge.On peut obt<strong>en</strong>ir une théorie des cordes topologiques à partir d’un modèle sigma non linéairesupersymétrique N = 2 à deux dim<strong>en</strong>sions par le biais de twist topologique. Cedernier peut être effectué de deux manières différ<strong>en</strong>tes [34]- [36] et con<strong>du</strong>it à deux modèlesdiffér<strong>en</strong>ts à savoir le modèle topologique A et le modèle topologique B (<strong>en</strong> anglais topologicalA-model et topological B- model) [37]-[39].La richesse des domaines liés aux théories de cordes topologiques constitue un objet d’étudepassionnant. L’étude de la théorie des cordes topologiques a connu un grand intérêt aprèsla découverte de la conjecture de Maldac<strong>en</strong>a qui relie la théorie des supercordes type IIBvivant sur l’espace AdS 5 ×S 5 (AdS = Anti de Sitter) à la théorie de jauge supersymétriqueconforme N = 4 à 4 dim<strong>en</strong>sions d’espace temps [40]. On parle aussi de <strong>du</strong>alité AdS/CFT[41] qui ouvre des issues pour l’exam<strong>en</strong> d’autres conjectures au niveau des supercordes.Face aux difficultés majeures liées au tests de ces conjectures, un certain nombre de travauxse sont consacrés à l’étude des équival<strong>en</strong>ces <strong>en</strong>tre des théories de jauge topologiqueset des modèles de théorie des cordes topologiques [42]-[44]. Ces modèles sont plus simplesque les théories de supercordes usuelles [8]-[9] constituant ainsi un laboratoire pour testerces conjectures de <strong>du</strong>alités.L’étude des cordes topologiques a égalem<strong>en</strong>t permis de mieux compr<strong>en</strong>dre le calcul microscopiquede l’<strong>en</strong>tropie des trous noirs <strong>en</strong> théorie des supercordes [45]-[50]. Le succèsmajeur de l’étude de la théorie des cordes topologiques a été largem<strong>en</strong>t dû à la conjecturede Ooguri, Strominger, et Vafa [46] qui relie la fonction de partition <strong>du</strong> trou noir Z T R etcelle de la théorie des cordes topologiques Z top comme suit [46, 47]Z T R = |Z top | 2 .L’étude des amplitudes Z top des théories de cordes topologiques a connu des avancéesmajeures ces dix dernières années [51, 52, 53, 54]. Dans le cas de la théorie de cordetopologique type A, les amplitudes Z top sont utilisées pour calculer les prepot<strong>en</strong>tiels desthéories de jauge supersymétriques N = 2 à quatre et à cinq dim<strong>en</strong>sions. Tandis que les15


<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong> Théorie des Cordesamplitudes de la théorie des cordes topologiques <strong>du</strong> modèle B, avec des flux et/ou branes,sont utilisées pour calculer les superpot<strong>en</strong>tiels des théories de jauge supersymétriques N =1 à quatre dim<strong>en</strong>sions. Il se trouve qu’il y a un accord <strong>en</strong>tre la fonction de partition <strong>du</strong>modèle-A associé à la variété de Calabi-Yau torique et la fonction de partition de la théoriede jauge de Nekrasov N = 2 SU(N) [55]-[57].L’approche standard <strong>du</strong> calcul des amplitudes Z (A)top de la théorie des cordes topologiquesde type A qui donne accès au calcul des invariants N g,nA de Gromov-Witt<strong>en</strong> (voir eq(1.3))est définie sur une variété de Calabi-Yau 3-folds comme suit∞∑= e Ftop = exp 2g−2 F g (t i ) , (1.2)Z (A)topoù est le couplage de la corde, t i sont les paramètres de Kahler et F g est la fonction departition de la théorie des cordes topologiques de surface d’univers de g<strong>en</strong>re g. Notons aupassage que l’énergie libre F top de la corde topologique est exprimée <strong>en</strong> terme des invariantsde Gromov-Witt<strong>en</strong> commeF top = ∑ g≥0∑g=0n A ≥0∈H 2N g,nA g 2g−2top e −n.t . (1.3)L’utilisation de ces fonctions de partition qui conti<strong>en</strong>n<strong>en</strong>t toutes les informations sur laphysique <strong>du</strong> système, nous a permis de compter les états BPS de la théorie de supercordetype IIA sur la variété de Calabi-Yau 3-folds torique <strong>en</strong> prés<strong>en</strong>ce de graviphoton self <strong>du</strong>alF 12 = F 34 = .Bi<strong>en</strong> que cette procé<strong>du</strong>re soit très bi<strong>en</strong> connue, la complexité <strong>du</strong> calcul augm<strong>en</strong>te, ce quinous a permis d’intro<strong>du</strong>ire plusieurs <strong>du</strong>alités remarquables qui peuv<strong>en</strong>t aider à résoudrece problème. Au cours de la dernière déc<strong>en</strong>nie, Vafa et ses collaborateurs ont élaboré unetechnique puissante pour le calcul des amplitudes de la théorie des cordes topologiques.En utilisant les variétés de Calabi-Yau toriques [58, 59], ils ont exprimé les fonctions departition <strong>en</strong> termes des invariants topologiques. De cette manière, ils ont pu déterminerles invariants de Gromov-Witt<strong>en</strong>, Gopakumar-Vafa et Donaldson-Thomas de la variété deCalabi-Yau torique. Dans ce cadre, la fonction de partition <strong>du</strong> modèle de corde topologiqueA sur le conifold résolu n’est autre que la fonction de partition de la théorie deChern-Simons sur S 3 [60]. C’est une conséqu<strong>en</strong>ce de la transition géométrique <strong>en</strong>tre leconifold résolu et celui déformé [62]. Les amplitudes des théories des cordes topologiques,<strong>en</strong> particulier celle <strong>du</strong> modèle A, fourniss<strong>en</strong>t ainsi une nouvelle technique pour approcherles calculs complexes <strong>en</strong> théorie des supercordes. C’est dans cette optique que s’inscrit <strong>en</strong>16


Intro<strong>du</strong>ction Généralepartie le travail de ce mémoire de thèse qui consiste à étudier certains aspects de la théoriedes cordes topologiques sur des variétés de Calabi-Yau toriques à trois dim<strong>en</strong>sions avec unacc<strong>en</strong>t particulier sur la <strong>du</strong>alité <strong>en</strong>tre la théorie des cordes topologiques, le vertex topologiqueet le modèle statistique <strong>du</strong> cristal fon<strong>du</strong>. Rappelons au passage que le calcul de lafonction de partition Z top <strong>du</strong> modèle A peut être fait de deux façons ; soit par le biais <strong>du</strong>vertex topologique C R1 R 2 R 3[63, 64, 65] ou par usage <strong>du</strong> formalisme de la matrice de transfertT qui sera développée dans une de nos contributions. Le li<strong>en</strong> <strong>en</strong>tre ces deux méthodesest à l’origine d’une <strong>du</strong>alité <strong>en</strong>tre le modèle de la physique statistique et la théorie descordes topologiques [66, 69, 70]. Pour se faire une idée sur ces méthodes, nous rappelonsbrièvem<strong>en</strong>t dans ce qui suit les grandes lignes de ces méthodes.Du cristal fon<strong>du</strong> à la corde topologiqueDans une réc<strong>en</strong>te et belle contribution de A.Okounkov, N.Reshetikhin, et C.Vafa [91], il aété montré qu’il existe un li<strong>en</strong> remarquable <strong>en</strong>tre l’étude de la physique de la fusion descristaux et la théorie des cordes topologiques. Dans le premier lieu, le calcul de fonctionde partition Z cristal repose sur le comptage des différ<strong>en</strong>tes configurations {n} <strong>du</strong> systèmecanonique,Z cristal = ∑ ne − E nT (1.4)où E n est l’énergie de la configuration n. A.Okounkov, N.Reshetikhin et C.Vafa ont montréque les différ<strong>en</strong>tes amplitudes <strong>du</strong> modèle topologique type A sur l’espace complexe C 3pourront être exprimées <strong>en</strong> termes de la fonction Z cristal <strong>du</strong> cristal fon<strong>du</strong>. La relation avecle cristal repose sur le fait que la fusion se manifeste par des départs progressifs d’atomes<strong>du</strong> cristal.Fig. 1.2 – Cristal fon<strong>du</strong> et la partition plane 3d.17


<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong> Théorie des CordesUne configuration <strong>du</strong> cristal fon<strong>du</strong> est décrite par une partition π, représ<strong>en</strong>tée par undiagramme de Young à 3 dim<strong>en</strong>sions généralisant le diagramme de Young usuel. Chaquepartition contribue par un facteurq = e −µToù µ est le pot<strong>en</strong>tiel chimique. La fonction de partition génératrice Z cristal de la fusioncristal est alorsZ cristal = ∑ πq |π| . (1.5)Cette fonction, qui dép<strong>en</strong>d de la température et <strong>du</strong> pot<strong>en</strong>tiel chimique, est aussi celle dela corde topologique <strong>du</strong> modèle A con<strong>du</strong>isant ainsi à l’id<strong>en</strong>tification suivanteZ cristal (q) = Z top (q) , (1.6)où la constante de couplage de la corde g s est maint<strong>en</strong>ant interprétée <strong>en</strong> terme de températureg s = 1 T .Notons que le paramètre q qui satisfait la condition 0 < q < 1 joue le rôle <strong>du</strong> pot<strong>en</strong>tiel chimique; soit l’énergie de l’arrachem<strong>en</strong>t d’un atome <strong>du</strong> cristal. La fonction ∑ q |π| représ<strong>en</strong>tela somme de toutes les partitions planaires π. Une telle fonction est bi<strong>en</strong> connue dans lalittérature et est égale à la fonction de MacMahon à trois dim<strong>en</strong>sions∞∏Z 3D = M(q) = (1 − q n ) −n . (1.7)L’un des résultats principaux de la <strong>du</strong>alité cristal/corde (1.6) est l’emerg<strong>en</strong>ce <strong>du</strong> vertextopologique C R1 R 2 R 3comme objet fondam<strong>en</strong>tal à l’instar des vertex de la théorie quantiquedes champs. Les trois pattes externes suivant les directions x, y et z de ce vertex sont interprétées<strong>en</strong> termes de représ<strong>en</strong>tations R 1 , R 2 , R 3 <strong>du</strong> groupe SU (N) ; soit aussi <strong>en</strong> termesdes diagrammes de Young à 2 dim<strong>en</strong>sions. L’expression explicite <strong>du</strong> vertex topologiqueC R1 R 2 R 3peut être obt<strong>en</strong>u <strong>en</strong> utilisant le formalisme de la matrice de transfert [91]∑C R1 R 2 R 3= M(q) −1 q |π| . (1.8)n=1π→{R 1 R 2 R 3 }Dans le cas le plus simple <strong>du</strong> vertex topologique C ∅∅∅ , qui est aussi la fonction de partitionde C 3 , où les représ<strong>en</strong>tations portées par les pattes externes sont des représ<strong>en</strong>tations18


Intro<strong>du</strong>ction Généraletriviales c.à.d R 1 = R 2 = R 3 = ∅, on retrouve précisém<strong>en</strong>t la fonction de MacMahon quidécrit l’amplitude de la théorie de corde fermée <strong>du</strong> modèle A sur C 3 ,∞∏C ∅∅∅ = M 3 (q) = (1 − q n ) −n . (1.9)Cet objet, qui fut conjecturé par MacMahon <strong>en</strong> 1912 [92], est égalem<strong>en</strong>t la fonction génératricepour toutes les fonctions indexées par les partitions vides. En fait la relation M 3 (q)est un cas particulier de la généralisation M d (q) associée avec l’octant positif <strong>du</strong> cube àd-dim<strong>en</strong>sions et dont l’expression conjecturée est donnée par :n=1M d (q) = ∏ (1 − q i ) − (i+d−3)!(i−1)!(d−2)!i≥1(1.10)En ce qui est de la fonction de MacMahon à 3 dim<strong>en</strong>sions, elle intervi<strong>en</strong>t dans plusieursdomaines physiques et mathématiques. Elle apparaît comme la fonction génératrice dedegré zéro dans l’étude des invariants de Donaldson-Thomas de la variété de Calabi-YauX dont la fonction de partition est donnée parZ (q) = M(q) χ(X) (1.11)et où χ (X) est la caractéristique d’Euler classique. Les fonctions de partition <strong>du</strong> modèlecristallin à 1D et 2D sont [93, 94]Z 1D = (1 − q) −1 ,Z 2D =∏ ∞ (1 − q n ) −1 .n=1(1.12)Celle à 3D est donnée par la relation (1.9) mais les expressions de Z pD pour p > 4 rest<strong>en</strong>t<strong>en</strong>core une question ouverte. Concernant les relations conjecturées (1.10), nous avons démontrédans [93] l’origine de ces relations <strong>en</strong> utilisant la méthode de matrice de transfertet nous avons donné une interprétation de ces objets <strong>en</strong> théorie des champs conformes à2 dim<strong>en</strong>sions. Toutefois, il faut noter que ces fonctions ne peuv<strong>en</strong>t être des fonctions departition des ”hypersolides” de dim<strong>en</strong>sion supérieure.Avant de clôturer cette intro<strong>du</strong>ction générale, il est utile de dire quelques mots sur laméthode <strong>du</strong> vertex topologique C R1 R 2 R 3vu que c’est un outil très puissant pour étudierégalem<strong>en</strong>t l’équival<strong>en</strong>ce <strong>en</strong>tre la théorie de jauge N = 2 à 4D et la fonction de partitionde la corde topologique [71], [72]. Cette puissance découle de la ressemblance formelle avecla méthode des diagrammes de Feynman de la théorie quantique des champs. Le vertex19


<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong> Théorie des Cordestopologique permet, <strong>en</strong>semble avec le propagateur, de calculer toutes les amplitudes dela corde topologique <strong>en</strong> prés<strong>en</strong>ce des branes [73]-[76]. Ici les diagrammes de Feynman nesont autre que les diagrammes toriques des variétés locales de Calabi-Yau dont les lignesgénér<strong>en</strong>t les cycles de la fibration torique et dont les sommets sont des vertex trival<strong>en</strong>ts.Ces diagrammes toriques form<strong>en</strong>t la base de la variété de Calabi-Yau torique C 3 [77], [78].Le vertex C R1 R 2 R 3avec R i ≠ ∅ représ<strong>en</strong>te des cordes ouvertes qui se termin<strong>en</strong>t sur desempilem<strong>en</strong>ts de branes, dont les configurations sont codées par des diagrammes de YoungR i , i = 1, 2, 3. Le calcul des amplitudes topologiques utilise C R1 R 2 R 3qui est fonction deq = e −g set qui peut être exprimé sous plusieurs formes ; soit <strong>en</strong> terme de fonctions deSchur commeC λ,µ,ν (q) = q k(µ) s ν t(q −ρ ) ∑ ηs λ t /η(q −ν−ρ )s µ/η (q −νt −ρ ) , (1.13)définissant une fonction génératrice des invariants de Gromov-Witt<strong>en</strong>, soit <strong>en</strong> fonction desinvariants topologiques donnés par la relation suivante [80, 88] :C λµ t ∅ (q) = q − k(µ)2 Wλµ (q) , (1.14)où λ, µ et ν sont des partitions 2d et W λµ (q) est une expression combinatoire reliée àl’invariant d’<strong>en</strong>trelac de Hopf. L’éq(1.14) est remarquable dans le s<strong>en</strong>s où elle établit unerelation naturelle <strong>en</strong>tre les invariants de Gromov-Witt<strong>en</strong> [83, 84, 85, 86] et les invariantsde la théorie de Chern Simons [87].Signalons égalem<strong>en</strong>t que le vertex topologique C λ,µ,ν (q) peut être généralisé <strong>en</strong> une fonctiondép<strong>en</strong>dant de deux argum<strong>en</strong>ts q et t,q = e iɛ 1, t = e −iɛ 2,pour donner naissance au vertex topologique raffiné C λµν (q, t). La fonction de partition dedeux paramètres ɛ 1,2 avec ɛ 1 + ɛ 2 ≠ 0 est la fonction de partition raffiné de la corde topologiquefournissant à son tour un raffinem<strong>en</strong>t des théories de Gromov-Witt<strong>en</strong> et Donaldson-Thomas de la variété de CY 3 torique. Cet axe de recherche connaît actuellem<strong>en</strong>t une grandeimportance à la fois <strong>en</strong> physique théorique et <strong>en</strong> mathématique. Nous aurons donc affaireà différ<strong>en</strong>tes connections <strong>en</strong>tre des propriétés mathématiques et des applications physiquesoù s’<strong>en</strong>tremêl<strong>en</strong>t plusieurs objets tels que la fonction de partition, les fonctions de Schur,les diagrammes de Young à 2D et 3D, le cristal fon<strong>du</strong> et la théorie des cordes topologiques.Tous ces objets constitu<strong>en</strong>t un champ de recherche très prometteur.20


Intro<strong>du</strong>ction GénéralePlan de thèseCe mémoire de thèse est composé de cinq chapitres, une conclusion et une annexe. Descontributions sont reportées à la fin des chapitres. Dans le chapitre 2 nous prés<strong>en</strong>tons desoutils de base dont nous avons besoin. Nous comm<strong>en</strong>çons par une intro<strong>du</strong>ction sur lesvariétés de Calabi-Yau (CY) et la géométrie torique dont les propriétés ess<strong>en</strong>tielles sontillustrées sur des exemples explicites. C’est le cas de plusieurs notion telle que la conditionde CY ou <strong>en</strong>core la transition géométrique qui joue un rôle crucial dans l’obt<strong>en</strong>tionde plusieurs résultats et que nous illustrons sur l’exemple <strong>du</strong> conifold. Nous terminons lechapitre par intro<strong>du</strong>ire les variétés toriques qui permett<strong>en</strong>t le calcul explicite des fonctionsde partition et <strong>en</strong> général des amplitudes topologiques. Dans le chapitre 3, nous traitonsdans la première section la théorie des cordes topologiques avec ses deux versions A et Bet les D-branes topologiques. La <strong>du</strong>alité des cordes ouverte / fermée sont prés<strong>en</strong>tées dansla deuxième section. Ensuite, nous étudions de près les invariants topologiques interv<strong>en</strong>antdans la description des fonctions de partition des cordes topologiques. Nous terminons cechapitre par l’une de nos contributions intitulée ”Pure fermionic twistor like model & targetspace supersymmetry” dans laquelle nous avons étudié la théorie des cordes twistorielles.Dans le chapitre 4, nous étudions la <strong>du</strong>alité <strong>en</strong>tre le modèle statistique et la théorie descordes topologiques. Les propriétés <strong>du</strong> modèle cristallin sont exprimées dans la structuregéométrique de la variété de Calabi-Yau, d’où le nom de variété de Calabi-Yau cristalline (<strong>en</strong> anglais Calabi-Yau crystal). Nous nous cont<strong>en</strong>terons ici de montrer comm<strong>en</strong>t le modèle<strong>du</strong> cristal fon<strong>du</strong> peut être mis <strong>en</strong> correspondance avec les partitions à 3 dim<strong>en</strong>sions. Parla suite, nous explicitons le calcul de la fonction de partition <strong>en</strong> utilisant la méthode de lamatrice de transfert. Dans cette logique des choses, nous discuterons le modèle <strong>du</strong> cristalfon<strong>du</strong> <strong>en</strong> prés<strong>en</strong>ce de D-branes placées aux bords. Ce modèle permet ainsi d’obt<strong>en</strong>ir desrésultats qui serv<strong>en</strong>t à compr<strong>en</strong>dre le rôle que joue les modèles cristallins dans la descriptiondes invariants de la théorie topologique. A la fin <strong>du</strong> chapitre, nous prés<strong>en</strong>terons notrecontribution "G<strong>en</strong>eralized MacMahon G(q) as q-deformed CFT Correlation Function” oùnous avons réussi à démontrer la conjecture de MacMahon <strong>en</strong> utilisant la <strong>du</strong>alité <strong>en</strong>tre lemodèle cristallin et la théorie de corde topologique.Dans le chapitre 5, nous intro<strong>du</strong>isons une approche basée sur les variétés de Calabi-Yautoriques et les règles de collage. Cette approche originale part de la bonne connaissancedes espaces de Calabi-Yau toriques et vise à <strong>en</strong> extraire plus d’information à partir desdiagrammes toriques. Dans ce cas, nous calculons les amplitudes et les fonctions de partitiondes cordes topologiques sur les espaces de Calabi-Yau toriques <strong>en</strong> utilisant juste21


<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong> Théorie des Cordesle formalisme de vertex topologique. Ce chapitre se termine par nos deux contributions”Topological String on Toric CY3s in Large Complex Structure Limit” et ”Non Planar Topological3-<strong>Vertex</strong> Formalism”.Dans le chapitre 6, on montre comm<strong>en</strong>t le raffinem<strong>en</strong>t de vertex topologique permet d’établirles expressions des amplitudes et les fonctions de partitions à deux paramètres ainsi queles relations avec les invariants de la théorie de Chern-Simons. Finalem<strong>en</strong>t, nous intro<strong>du</strong>isonsnotre contribution ”Refining the Shifted Topological <strong>Vertex</strong>” qui a pour but d’étudierla version raffinée de vertex topologique shifté.Nous achevons ce mémoire de thèse par une conclusion où nous résumons nos principauxrésultats et où nous examinons quelques perspectives et des applications. Pour plus dedétails sur le calcul des amplitudes et les fonctions de partitions, nous avons juger utile decollecter quelques notions fondam<strong>en</strong>tales <strong>en</strong> annexe.Ce travail de thèse a donné lieu à plusieurs résultats dont certains sont déjà parus sousforme de publications de recherche. Dans la liste ci-dessous, nous prés<strong>en</strong>tons nos travauxoriginaux :- Topological String on Toric CY3s in Large Complex Structure Limit,Nuclear Physics B, Volume 813, Issue 3, 1 June 2009, Pages 315-348, arXiv :0812.0526.- Refining the Shifted Topological <strong>Vertex</strong>,J.Math.Phys.50 :013509,2009, arXiv : 0812.0513.- G<strong>en</strong>eralized MacMahon G(q) as q-deformed CFT Correlation Function,Nuclear Physics B, Volume 801, Issue 3, 1 October 2008, Pages 316-345, arXiv :0801.2661.- Non Planar Topological 3-<strong>Vertex</strong> Formalism,Nuclear Physics B, Volume 804, Issue 3, 1 December 2008, Pages 307-341, arXiv : 0712.4249.- De la théorie des cordes twistorielles à la Supergravité Conforme N=4, D=4African Journal of Mathematical physics Volume 4 Number (2007) pages 65-96.- Pure fermionic twistor like model & target space supersymmetry,http ://xxx.lanl.gov/hep-th/0605167.22


Chapitre 2Variétés de CY et Géométrie ToriqueL’étude des variétés de Calabi-Yau est dev<strong>en</strong>ue un domaine de recherche très passionnantet constitue une des plus importantes collaborations <strong>en</strong>tre physici<strong>en</strong>s théorici<strong>en</strong>s etmathématici<strong>en</strong>s. Non seulem<strong>en</strong>t ces variétés sont considérées comme un type particulierde variétés complexes interv<strong>en</strong>ant dans des domaines comme la géométrie algébrique [95]-[96], mais elles apparaiss<strong>en</strong>t égalem<strong>en</strong>t <strong>en</strong> physique des hautes énergies notamm<strong>en</strong>t dansla théorie des supercordes où elles jou<strong>en</strong>t le rôle d’espace de compactification [97]. Dansce chapitre, nous prés<strong>en</strong>tons différ<strong>en</strong>tes propriétés géométriques et homologiques sur lesvariétés de Calabi-Yau toriques. Dans la première section, nous considérons les variétés deCalabi-Yau comme étant des variétés complexes Kahleri<strong>en</strong>nes ayant un groupe d’holonomieSU(n) et une première classe de Chern nulle. Dans la deuxième section de ce chapitre,nous intro<strong>du</strong>isons la transition géométrique <strong>en</strong>tre le conifold résolu et celui déformé etdans la section 3, nous exposons les variétés de Calabi-Yau toriques qui form<strong>en</strong>t une classed’espaces jouant un grand rôle dans plusieurs domaines physiques notamm<strong>en</strong>t dans l’étude<strong>du</strong> vertex topologique qui sera considéré avec détails dans le chapitre 5 de ce mémoire.Nous terminons ce chapitre par la construction des diagrammes toriques tout <strong>en</strong> donnantdes exemples d’applications.2.1 Généralités sur les variétés de CYNous nous intéressons dans cette section aux structures géométriques sur les variétéscomplexes compactes admettant une métrique de Kähler à courbure de Ricci nulle [95].Mais avant d’aborder les propriétés remarquables de ces variétés, il est intéressant decomm<strong>en</strong>cer par rappeler quelques notions géométriques utiles.23


2.1 Généralités sur les variétés de CY2.1.1 Variétés complexesa) Généralités : Les variétés complexes M de dim<strong>en</strong>sion complexe n (dim C = n) sontavant tout des variétés de dim<strong>en</strong>sion réelle dim R = 2n où chaque point (x 1 , ..., x 2n ) peutêtre décrit par n coordonnées locales complexes (z 1 , . . . z n ) avec z k = x k + ix n+k . Sur cesvariétés complexes M, viv<strong>en</strong>t <strong>en</strong>tre autres, les objets suivants :i) Une structure presque complexe portée par un champ d’opérateur J vérifiant J 2 = −I d ;il définit une transformation linéaire globale dans l’espace tang<strong>en</strong>t à M <strong>en</strong> tout point P :J : T p M → T p M ∀ P ∈ M.ii) Une métrique hérmiti<strong>en</strong>ne G satisfait la condition :G(X, Y ) = G(JX, JY ),où X, Y sont des champs de vecteurs sur M. En coordonnées complexes (z 1 , . . . z n ), lamétrique G peut être réalisée comme suit :G = G j¯k dz j ∧ d¯z¯k.(2.1)Nous allons nous intéresser surtout aux variétés complexes compactes dont les plus étudiéesdans la littérature sont les variétés algébriques projectives X . Ces variétés sont des sousvariétés de l’espace projectif complexe à n dim<strong>en</strong>sions X ⊂ CP n avec :CP n = {droites vectorielles de C n }= { C n+1 \ {0} /C ∗} . (2.2)A titre d’exemple, signalons que les courbes complexes compactes, qui ne sont autres queles surfaces réelles de Riemann, sont des sous variétés compactes à une dim<strong>en</strong>sion complexequi se réalis<strong>en</strong>t comme des hypercourbes dans CP 2 .b) Variétés Kahléri<strong>en</strong>nes : Une variété Kahléri<strong>en</strong>ne K ou variété de Kahler tout courtest la donnée d’une structure presque complexe J sur une variété différ<strong>en</strong>tielle M et d’unemétrique hermiti<strong>en</strong>ne G. La 2- forme antisymétrique K associée à K définie par :K(X, Y ) = G(X, JY ), (2.3)vérifie la condition de fermeture suivante :dK = 0. (2.4)La forme de Kahler K, qui s’écrit <strong>en</strong> coordonnées réelles comme K = 1J 2 [µν]dx µ ∧ dx ν avecµ, ν = 1, · · · , 2n, peut être exprimée <strong>en</strong> coordonnées complexes comme :K = i 2 G j¯kdz j ∧ d¯z, j, k = 1, · · · , n, (2.5)24


Variétés de CY et Géométrie Toriqueoù la métrique G vérifiant (2.4) est définie localem<strong>en</strong>t <strong>en</strong> terme <strong>du</strong> pot<strong>en</strong>tiel de Kahler Kcomme suit :G i¯j = ∂ i ∂¯jK. (2.6)Les composantes G ij et leurs complexes conjuguées Gī¯j sont nulles.c) Formes différ<strong>en</strong>tielles sur M : Sur une variété complexe Kahleri<strong>en</strong>ne M, on définitgénéralem<strong>en</strong>t les formes différ<strong>en</strong>tielles réelles ω = ω [µ1···µ s ]dx µ 1 · · · dxµ s de degré s ≤ 2navec les (p, r)-formes de degré d’holomorphicité p ≤ n et d’antiholomorphicité r ≤ n sontdonnées par :ω = ω i1···i p j 1···j rdz i 1∧ · · · ∧ dz i p∧ d¯z¯j 1∧ · · · ∧ d¯z¯j r(2.7)Par conséqu<strong>en</strong>t, il est possible de décomposer l’espace Λ s (M) des s-formes differ<strong>en</strong>tiablesréelles <strong>en</strong> une somme directe de (p, r)- espaces Ω p,r (M) comme suit :Λ s (M) =⊕ Ω p,r (M), (2.8)s=p+rcon<strong>du</strong>isant à une décomposition <strong>du</strong> groupe de cohomologie H s (M) <strong>en</strong> une somme directedes groupes H p,r (M) [58, 59],et leurs dim<strong>en</strong>sions sont notées parH s (M) =⊕ H p,r (M) (2.9)s=p+rb s = dim H s (M) ,h p,r = dim H p,r (M) ,b s = ∑ s=p+r hp,r ,(2.10)avec b s connu sous le nom de nombre de Betti et vérifie les conditions :h p,r = h r,p , h p,r = h n−p,n−r , dim C M = n , (2.11)Pour le cas de H 3 (M, R), nous avonsH 3 (M, C) = H 3,0 ⊕ H 2,1 ⊕ H 1,2 ⊕ H 0,3 , (2.12)Une autre propriété importante de la variété M est la caractéristique d’Euler notée χ(M)qui est un nombre associé à tout espace topologique. C’est un invariant topologique donnépar la somme alternée des nombres de Betti :χ(M) = ∑ s(−1) s b s = ∑ p,r(−1) p+r h p,r (2.13)25


2.1 Généralités sur les variétés de CYLa caractéristique d’Euler χ(M) est assez facile à calculer si on exprime ce nombre pourles polyèdres selon la formule :χ(M) = V − E + F.Les nombres V , E et F sont respectivem<strong>en</strong>t le nombre de sommets (vertex), d’arêtes(”edges”) et de faces dans un polyèdre donné. Pour un cristal cubique avec huit sommets,douze arrêtes et six faces, la caractéristique d’Euler estχ(M) = V − E + F = 8 − 12 + 6 = 2Ce nombre peut être id<strong>en</strong>tifié avec la caractéristique d’Euler <strong>du</strong> cristal de Calabi-Yau [91].Nous terminons cette sous section par rappeler égalem<strong>en</strong>t les paramètres t i de Kähler quisont donnés par le volume des 2- cycles S i de la variété M :t i = ∫ S iK , i = 1, · · · b 2 (M) ,Les S i définiss<strong>en</strong>t une base <strong>du</strong> groupe d’homologie H 2 (M, Z) des 2-cycles réelles de M.2.1.2 Variétés de Calabi-YauLes variétés de Calabi-Yau sont particulièrem<strong>en</strong>t utilisées <strong>en</strong> théorie des supercordes[97]. Elles préserv<strong>en</strong>t une partie des charges supersymétriques de la théorie originale à10 dim<strong>en</strong>sions compactifiées vers des dim<strong>en</strong>sions d’espaces temps inférieures. Initialem<strong>en</strong>tintro<strong>du</strong>ite par le mathématici<strong>en</strong> E. Calabi qui a conjecturé <strong>en</strong> 1957 que ces variétés admett<strong>en</strong>tnécessairem<strong>en</strong>t une métrique avec un t<strong>en</strong>seur de Ricci nul. Cette conjecture futdémontrée par S.Yau <strong>en</strong> 1977 [98]. Dès lors, la définition précise de ces variétés est dev<strong>en</strong>uetrès technique dont les lignes de base sont comme suit :Les variétés de Calabi-Yau M de dim<strong>en</strong>sion n sont des espaces complexes, Kahléri<strong>en</strong>s,compacts ayant un t<strong>en</strong>seur de Ricci nul R i¯j = 0.Cette condition de courbure de Ricci nulle est équival<strong>en</strong>te aux équations suivantes :i) Le groupe d’holonomie de M est SU(n).ii) La première classe de Chern c 1 est nulle ; c 1 (T M) = 0.Les variétés de Calabi-Yau admett<strong>en</strong>t de types de déformations : déformations de KahlerK et déformation de la structure complexe Ω (n,0) . Les déformations de la structureKahleri<strong>en</strong>ne d’une variété de Calabi-Yau M de dim<strong>en</strong>sion n sont associées au groupe d’homologieH (1,1) de dim<strong>en</strong>sion h (1,1) qui classifie les formes avec un indice holomorphique et26


Variétés de CY et Géométrie Toriqueanti-holomorphique.Par ailleurs, sachant que les variétés de Calabi-Yau sont classifiées par les nombres de Bettidéterminant le nombre de formes harmoniques de divers rangs [98], on peut donc décrireexplicitem<strong>en</strong>t l’<strong>en</strong>semble des paramètres qui caractéris<strong>en</strong>t la structure cohomologique dela variété <strong>en</strong> considérant le diamant de Hodge :h 0,0h 1,0 h 0,1 . ..(2.14). ..h n,m−1h n,0 · · · h 0,nh n,n h n−1,nParmi les différ<strong>en</strong>tes classes des variétés de Calabi-Yau qui ont été utilisées <strong>en</strong> théorie dessupercordes, on trouve des variétés non compactes qui sont des variétés de Calabi-Yaulocales correspondant à la limite de découplage des degrés de liberté supergravitationnels.2.1.3 Exemples des variétés de Calabi-YauIl n’existe relativem<strong>en</strong>t que quelques exemples de variétés de Calabi-Yau connues. Leursconstructions sont généralem<strong>en</strong>t justifiées par des applications physiques notamm<strong>en</strong>t dansle cadre de la conjecture de symétrie miroir [37],[99] reliant les supercordes type IIA ettype IIB. Dans cette sous section, nous donnons trois exemples.1) Courbes complexesL’exemple le plus simple des variétés de Calabi-Yau compactes de dim<strong>en</strong>sion dim C = 1est donné par les courbes complexes (surfaces de Riemann) dont la ligne projective CP 1isomorphe à la sphère S 2 réelle et le tore T 2 = S 1 × S 1 constitu<strong>en</strong>t les deux premiersélém<strong>en</strong>ts. Le diagramme de Hodge <strong>du</strong> tore est donné par :11 11(2.15)Il a un seul paramètre de Kahler t puisque h 1,1 = 1 et un paramètre complexe τ.27


2.1 Généralités sur les variétés de CY2) Surfaces complexesAu dim<strong>en</strong>sion complexe 2, il n’existe que deux variétés de Calabi-Yau à isomorphismeprès. Il s’agit <strong>du</strong> tore T 4 et de l’espace K3. Sur ce dernier, aucune métrique explicite deRicci n’est connue. Il <strong>en</strong> va de même pour tous les Calabi-Yau de dim<strong>en</strong>sions supérieuresnon-triviales. Le diagramme de Hodge de la surface K3 est :10 01 20 10 01(2.16)A partir de ce tableau on appr<strong>en</strong>d que h 1,1 = 20 et le nombre de Betti b 2 = 22.3) Dim<strong>en</strong>sion complexe 3A partir de la dim<strong>en</strong>sion complexe 3 (6 dim<strong>en</strong>sions réelles), le nombre de variétés deCalabi-Yau devi<strong>en</strong>t infini et il n’existe pas <strong>en</strong>core une classification générale. On peutconstruire toutefois un grand nombre de variétés possédant, <strong>en</strong> plus, la propriété d’être desvariétés toriques.i) Variété locale CP 1De la même façon, la variété complexe locale CP 1 est représ<strong>en</strong>tée par quatre coordonnées(x 1 , x 2 , z 1 , z 2 ) soumises à la condition (z 1 , z 2 ) ≠ (0, 0) où C ∗ est une action torique qui agitsur C 2 de la façon suivante (x 1 , x 2 , z 1 , z 2 ) ∼ (λ −1 x 1 , λ −1 x 2 , λz 1 , λz 2 ).L’espace total résultant est donné par la fibration de la surface O(−1) ⊕ O(−1) au dessusde la base CP 1 ; soit : O(−1) ⊕ O(−1) → CP 1 et peut être vu comme la somme de deuxfibrés <strong>en</strong> ligne sur la courbe projective complexe CP 1 paramétrisée par les deux variables(z 1 , z 2 ) <strong>en</strong> ajoutant 4 directions non compactes portées par les variables (x 1 , x 2 ) . Signalonsau passage que cet exemple représ<strong>en</strong>te le conifold résolu qui sera examiné <strong>en</strong> détail lors dela prochaine section.ii) Variété locale CP 2C’est une variété de Calabi-Yau non compacte à trois dim<strong>en</strong>sions complexes pramétrisée28


Variétés de CY et Géométrie ToriqueFig. 2-1 – Une représ<strong>en</strong>tation géométrique de la courbe locale de base CP 1 et de fibreO(−1) ⊕ O(−1).par quatre coordonnées complexes (x, z 1 , z 2 , z 3 ) mo<strong>du</strong>lo la relation d’équival<strong>en</strong>ce :{(z1 , z 2 , z 3 ) ≠ (0, 0, 0) | (x, z 1 , z 2 , z 3 ) ∼ ( λ −3 x, λz 1 , λz 2 , λz 3), λ ∈ C∗ } (2.17)Fig. 2-2 – Représ<strong>en</strong>tation de la géométrie CP 2 locale.Cet espace peut être considéré mathématiquem<strong>en</strong>t comme l’espace total <strong>du</strong> fibré <strong>en</strong>ligne O(−3) → CP 2 . Autrem<strong>en</strong>t dit, la variété locale CP 2 peut être vu localem<strong>en</strong>t commeétant le pro<strong>du</strong>it de CP 2 ×C où la base CP 2 est paramétrisée par les coordonnées (z 1 , z 2 , z 3 )sur lesquelles vit une courbe complexe C paramétrisée par la coordonnée complexe x. Dansce cas, on a quatre directions réelles compactes et deux non compactes.iii) Variété locale CP 1 ×CP 1Pour décrire la variété locale CP 1 ×CP 1 , il faut utiliser cinq coordonnées complexes (x, y 1 , y 2 , z 1 , z 2 )29


2.2 Conifoldavec (y 1 , y 2 ) ≠ (0, 0), (z 1 , z 2 ) ≠ (0, 0) et l’id<strong>en</strong>tification(x, y 1 , y 2 , z 1 , z 2 ) ∼ (λ −1 µ −1 x, y 1 , y 2 , µz 1 , µz 2 ).L’espace total donné par O(−2, −2) → CP 1 ×CP 1 ; est paramétrisé par quatre directionscompactes et deux directions non-compactes dont le fibré normal est isomorphe àO(−2, −2).2.2 ConifoldUne autre classe des exemples concernant les variétés de Calabi-Yau est étalée dans[108, 111] dont les grandes lignes sont prés<strong>en</strong>tées comme suit. Nous prés<strong>en</strong>tons dans unpremier temps le conifold singulier et ses déformations de Kahler et complexe. Ensuite, nousconsidérons la transition géométrique qui permet de relier les deux versions <strong>du</strong> conifold.2.2.1 Conifold singulierLe conifold singulier est une variété algébrique définie par la relation suivante [111],z1 2 + z2 2 + z3 2 + z4 2 = 0. (2.18)C’est donc une hypersurface de l’espace complexe de C 4 ∼ R 8 . En décomposant z <strong>en</strong> partieréelle et partie imaginaire z i = x i + iy i , l’équation ci-dessus se scinde comme suit :∑ 4k=1 (x2 k − y2 k ) = 0 ,2i ∑ 4k=1 x ky k = 0 .(2.19)Pour déterminer la base <strong>du</strong> conifold, nous considérons l’intersection <strong>en</strong>tre l’équation (2.18)et la sphère réelle S 7 de dim<strong>en</strong>sion 7 et de rayon r cont<strong>en</strong>ue dans l’espace réelle R 8 ,|z 1 | 2 + |z 2 | 2 + |z 3 | 2 + |z 4 | 2 = r 2 . (2.20)qui s’écrit égalem<strong>en</strong>t sous la forme :∑ 4k=1 (x2 k + y2 k ) = r2 , avec z i = x i + iy i (2.21)les équations de l’intersection peuv<strong>en</strong>t être exprimées comme suit :⃗x 2 = r2 2et ⃗y 2 = r2 ,2⃗x.⃗y = 0. (2.22)30


Variétés de CY et Géométrie ToriqueLe premier terme de ces relations décrit la sphère S 3 d’équation ⃗x 2 = r2 2 . L’équation ⃗y2 = r2 2avec ⃗x.⃗y = x 1 y 1 + x 2 y 2 + x 3 y 3 + x 4 y 4 = 0 paramétrise la sphère S 2 fibrée sur la sphère S 3 .Ainsi, nous interprétons les variables x k comme des coordonnées de la sphère S 3 ∼ SU(2)et les y k comme celles de la sphère S 2 ∼ SU(2) de rayon r/ √ 2. En fixant les xU(1)k , les variablesy k doiv<strong>en</strong>t être orthogonales à ⃗x et se déplac<strong>en</strong>t sur la sphère S 2 . Ainsi, le conifold singulierest un cône de base S 3 × S 2 et la singularité correspond au point x = y = 0 comme lemontre la figure (2-3) suivanteFig. 2-3 – Conifold singulierLa base S 3 × S 2 <strong>du</strong> conifold singulier peut alors être vue comme la variété :B = SU(2) × SU(2)U(1) , (2.23)qui est isomorphe à SO(4) . Après avoir décrit la singularité <strong>du</strong> conifold, nous étudionsU(1)maint<strong>en</strong>ant la levée de la dégénéresc<strong>en</strong>ce qui peut être déformée de deux façons : (i) parremplacem<strong>en</strong>t <strong>du</strong> point singulier <strong>en</strong> utilisant une sphère S 2 , ce qui con<strong>du</strong>it au conifoldrésolu. (ii) <strong>en</strong> remplaçant le point singulier par une sphère S 3 , ce qui con<strong>du</strong>it au conifolddéformé [60, 97, 99, 111, 112].2.2.2 Conifold déformé et conifold résoluDans ce paragraphe, nous étudions les deux façons de résoudre la singularité <strong>du</strong> conifold,à savoir la déformation de la structure complexe et la résolution de la structure de Kahler.a) Conifold déforméDans le système de coordonnées locales (z i ), la déformation complexe de la singularité <strong>du</strong>conifold correspond tout simplem<strong>en</strong>t à la déformation de l’éq.(2.18) qui devi<strong>en</strong>t alors :z1 2 + z2 2 + z3 2 + z4 2 = µ, (2.24)31


2.2 Conifoldoù µ est un nombre complexe. Cette hypersurface algébrique de C 4 décrit précisém<strong>en</strong>t leconifold déformé T ∗ S 3 dont la base S 3 est donnée par la partie réelle de cette relation. Parune transformation linéaire des coordonnées w 1,3 = z 3 ∓ iz 4 et w 2,4 = iz 1 ∓ z 2 , on peutram<strong>en</strong>er l’equation ci-dessus à la forme usuelle suivante :w 1 w 2 − w 3 w 4 = µ. (2.25)En outre, le fait que la base S 3 est obt<strong>en</strong>ue <strong>en</strong> posant w 2 = ¯w 1 , w 4 = − ¯w 3 , cette relationest généralem<strong>en</strong>t invariante sous l’action(w 1 , w 2 , w 3 , w 4 ) → ( )e −iα w 1 , e iα w 2 , e −iβ w 3 , e iβ w 4 (2.26)où les paramètres α et β sont associés respectivem<strong>en</strong>t aux 1-cycles (1, 0) et (0, 1) de T 2 .Ces deux 1-cycles se dégénèr<strong>en</strong>t respectivem<strong>en</strong>t aux lieux w 1 = w 2 = 0 ou w 3 = w 4 = 0qui correspond<strong>en</strong>t aux deux géométries cylindriques suivantes :w 3 w 4 = −µ ,w 1 w 2 = +µ .(2.27)Ces géométries décriv<strong>en</strong>t des courbes complexes ; soit des surfaces de Riemann non compactes.La structure de la fibration <strong>du</strong> conifold déformée est donnée par la base R 3 paramétriséepar les axes de ces cylindres et Re (w 1 w 2 ) . Ainsi la fibre T 2 × R <strong>du</strong> conifold estreprés<strong>en</strong>tée par les cycles α et β et l’imaginaire de w 1 w 2 .Fig. 2-4 – Conifold déformé32


Variétés de CY et Géométrie Toriqueb) Conifold résoluLa singularité conique peut égalem<strong>en</strong>t être résolu par déformation de la structure de Kahler.Cette résolution consiste à remplacer le conifold singulier z1 2 + z2 2 + z3 2 + z4 2 = 0 par unevariété lisse appelée conifold résolu dont l’expression est défini par :− |z 1 | 2 − |z 2 | 2 + |z 3 | 2 + |z 4 | 2 = t, (2.28)<strong>en</strong>semble avec une action U(1)(z 1 , z 2 , z 3 , z 4 ) → ( )e −iα z 1 , e −iα z 2 , e iα z 3 , e iα z 4 . (2.29)Le conifold résolu est donné par O(−1) ⊕ O(−1) → P 1 où la droite projective P 1 estid<strong>en</strong>tifiée à S 2 . Les deux pièces O(−1) correspond<strong>en</strong>t aux ouverts U 123 = (z 1 , z 2 , z 3 ) avecz 4 ≠ 0 et U 124 = (z 1 , z 2 , z 4 ) avec z 3 ≠ 0.Fig. 2-5 – Conifold résoluDans ce qui suit nous fixons notre att<strong>en</strong>tion sur l’étude de la transition géométrique<strong>du</strong> conifold vue ses nombreuses applications <strong>en</strong> théorie des cordes topologiques et dans lacompréh<strong>en</strong>sion de certaines <strong>du</strong>alités <strong>en</strong> théorie des supercordes type II [107, 108].2.2.3 Transitions géométriquesUne transition géométrique de l’espace de compactification d’une théorie des supercordesest un changem<strong>en</strong>t de sa topologie pour lequel la physique est inchangée. On distingueess<strong>en</strong>tiellem<strong>en</strong>t deux types de transitions : la transition <strong>du</strong> conifold et la transitionde flop (flop transition).33


2.3 Variétés de CY toriquesFig. 2-6 – Transition géométrique <strong>en</strong>tre le conifold déformé T ∗ S 3 et le conifold résolu.Il est intéressant de remarquer que la théorie des cordes puisse donner des résultats équival<strong>en</strong>tsau niveau de la théorie effective à 4 dim<strong>en</strong>sions lorsqu’elle est compactifiée sur deuxvariétés de Calabi-Yau à trois dim<strong>en</strong>sions différ<strong>en</strong>tes reliés par une transition géométrique.C’est le passage <strong>du</strong> conifold résolu O(−1) ⊕ O(−1) → CP 1 à celui défomé T ∗ S 3 [110],[111]et vice versa [106] :O(−1) ⊕ O(−1) → CP 1 ↔ T ∗ S 3 .La dégénéresc<strong>en</strong>ce <strong>du</strong> 2-cycle S 2 <strong>du</strong> conifold résolu est accompagnée de l’apparition d’un3-cycle S 3 <strong>du</strong> conifold déformé et réciproquem<strong>en</strong>t comme le montre la figure (2-6). En effet,la transition géométrique permet de relier les versions <strong>du</strong> conifold et constitue une <strong>du</strong>alité<strong>en</strong>tre les théories des cordes vivant sur ces deux géométries non-singulières. Cette propriétéest à la base de la conjecture de Gopakumar et Vafa qui ont proposé que la théorie de cordetopologique ouverte sur le conifold résolu est <strong>du</strong>ale à la théorie de la corde topologiquefermée sur le conifold déformé. En language de théorie des champs, cette <strong>du</strong>alité relie lesthéories de jauge et celles gravitationnelles ; elle est communém<strong>en</strong>t connue sous le non de<strong>du</strong>alité jauge-gravité ou <strong>du</strong>alité corde ouverte-corde fermée. Cette <strong>du</strong>alité sera étudiée dansle chapitre 3 dans le cadre de la théorie de cordes topologiques.2.3 Variétés de CY toriquesLes variétés toriques constitu<strong>en</strong>t un sujet d’actualité et possèd<strong>en</strong>t un grand intérêt <strong>en</strong>théorie des cordes. Elles form<strong>en</strong>t une classe de variétés suffisamm<strong>en</strong>t riche pour permettreà la fois de tester et illustrer diverses conjectures <strong>en</strong> théorie de cordes. Elles ont ses applicationsdans de nombreuses branches <strong>en</strong> mathématiques, notamm<strong>en</strong>t <strong>en</strong> topologie et dans34


Variétés de CY et Géométrie Toriquela compactificatoion des théories des supercordes.L’un des objectifs de cette section est d’exhiber explicitem<strong>en</strong>t la condition de Calabi-Yaudans le cadre de la géométrie torique [114, 115]. Ainsi, après avoir intro<strong>du</strong>it quelques propriétésess<strong>en</strong>tielles, nous étudierons différ<strong>en</strong>tes formulations de la condition de Calabi-Yaupour les variétés toriques. Nous montrons <strong>en</strong> particulier que la variété de Calabi-Yau toriquedoit être non compacte. Nous donnons égalem<strong>en</strong>t une description intuitive de la géométrietorique <strong>en</strong> utilisant le ”quoti<strong>en</strong>t symplectique” [113]. Cette description diffère de laprés<strong>en</strong>tation mathématique habituelle de la géométrie torique, mais permet néanmoins decompr<strong>en</strong>dre les constructions géométriques des variétés toriques, leur sous-variétés lagrangi<strong>en</strong>nesspéciales et leurs réalisations <strong>en</strong> terme des modèles sigma linéaires supersymétriques[116].2.3.1 Variétés toriquesLes variétés toriques peuv<strong>en</strong>t être décrites soit à l’aide de Fans [117]- [118] et descoordonnées homogènes [119], ou être considérées comme des variétés symplectiques, ou<strong>en</strong>core comme une branche de Coulomb de l’espace des états supersymétriques <strong>du</strong> modèlesigma linéaire Jaugé (GLSM : Gauged Linear Sigma Model) [50]- [122].Rappelons tout d’abord la définition d’une variété torique. Pour cela, on considère l’espacecomplexe C m muni d’une action par un tore algébrique (C ∗ ) p , p < m ainsi que le sous<strong>en</strong>semble U ⊂ C m fixé par un sous-groupe de (C ∗ ) p . Le quoti<strong>en</strong>t géométriqueM = Cm \UG(2.30)dont les classes d’équival<strong>en</strong>ce suivantes :(w 1 , ..., w m ) ∼ (λ Q1 aw 1 , ..., λ Qm aw m ) a = 1, ..., m − 3. (2.31)est une variété torique de dim<strong>en</strong>sion trois complexe. De cette définition découle que, lesespaces complexes projectifs CP n sont des variétés toriques. A titre d’exemple le planprojectif CP 2 est une variété torique donnée par,CP 2 = (C 3 \{0})/C ∗ ,où le quoti<strong>en</strong>t est défini par la relation d’équival<strong>en</strong>ce suivante :(x, y, z) ∼ λ(x, y, z), où λ ∈ C ∗ . (2.32)35


2.3 Variétés de CY toriquesEn général, une variété torique de dim<strong>en</strong>sion complexe n est une fibration <strong>en</strong> tore T n (d’oùle nom torique) sur un espace de base B n réelle linéaire (pas nécessairem<strong>en</strong>t compact) avecdes bords où des 1-cycles de T n dégénèr<strong>en</strong>t. Dans le cas où la base B n est compacte, lavariété torique résultante sera aussi compacte. Cette variété torique est bi<strong>en</strong> adaptée àl’espace complexe C n qui est alors réalisé comme une fibration <strong>en</strong> tore T n sur une sousvariétéLagrangi<strong>en</strong>ne L de C n <strong>en</strong> utilisant l’action U(1) :z i → z i e iθ i, θ i ∈ [0, 2π] . (2.33)Localem<strong>en</strong>t, nous pouvons écrire C n comme suit :C n ≈ O + × T noù O + désigne le "octant positif" (|z i | 2 0). Cette réalisation signifie que chaque point deO + ∼ R n + est attaché au pro<strong>du</strong>it de n cercles obt<strong>en</strong>us <strong>en</strong> fixant |z i | et <strong>en</strong> laissant θ i variable.Il s’<strong>en</strong>suit que la sous variété lagrangi<strong>en</strong>ne pour θ i = cte est isomorphe au segm<strong>en</strong>t positifR n qui est représ<strong>en</strong>té par |z i | 2 ≥ 0, avec i = 1, · · · n, au lieu de |z i | afin de r<strong>en</strong>dre la baselinéaire 1 . L’action de T n avec l’action de U(1) n qui agit sur |z i | 2 recouvr<strong>en</strong>t la variété C n .Par la suite, on peut construire un grand nombre de variétés de Calabi-Yau qui possèd<strong>en</strong>t,<strong>en</strong> plus, la propriété d’être des variétés toriques.2.3.2 Variétés de Calabi-Yau toriquesLes variétés de Calabi-Yau toriques à n-dim<strong>en</strong>sions ; <strong>en</strong> particulier à trois dim<strong>en</strong>sionscomplexes, possèd<strong>en</strong>t une application mom<strong>en</strong>t particulière 2 . En effet considérons les coordonnéescomplexes z 1 , · · · , z k de la variété C k etµ a : C k → C (2.34)l’application mom<strong>en</strong>t avec a = 1, · · · , k − 3 est défini comme suit :k∑Q i a|z i | 2 = Re(t a ), (2.35)i=1où t a sont des mo<strong>du</strong>les complexes. La condition de Calabi-Yau impose quek∑Q i a = 0, (2.36)i=11 la projection de l’espase de base ∣ ∣z i∣ ∣ → ∣ ∣z i∣ ∣ 2 est appelée aussi le mom<strong>en</strong>t map2 l’action hamiltonni<strong>en</strong> d’un groupe de Lie sur une variété symplectique est associée à une applicationappelée l’application mom<strong>en</strong>t.36


Variétés de CY et Géométrie Toriqueet l’action <strong>du</strong> groupe G = U(1) k−3 sur les coordonnées est définie par :z j → exp(iQ j aα a )z j , a = 1, ..., k − 3. (2.37)Il s’avère qu’il existe une procé<strong>du</strong>re très générale, dite de quoti<strong>en</strong>t symplectique, qui permetde définir la variété de Calabi-Yau torique à trois dim<strong>en</strong>sions comme suit :M = ∩k−3 a=1µ −1 (Re(t a )), (2.38)Goù les k − 3 paramètres t a sont les mo<strong>du</strong>les de Kahler complexifiés.La description de la variété torique a égalem<strong>en</strong>t une réalisation physique <strong>en</strong> terme demodèles sigma linéaires supersymétriques. C’est une théorie de jauge à deux dim<strong>en</strong>sionsde groupe de jauge U(1) k−3 impliquant k − 3 superchamps vectoriels V a = V a(x, θ± , ¯θ ±)et k superchamps chiraux Φ iΦ i (x, θ ± ) = z i (x) + θ + ψ −i (x) + θ − ψ +i (x) + θ + θ − F −i (x) , (2.39)dont les élém<strong>en</strong>ts scalaires sont précisém<strong>en</strong>t les variables z i et ou les Q i a sont alors les chargesdes superchamps Φ i sous le groupe de jauge U(1) k−3 . Il s’<strong>en</strong>suit que dans la branche deCoulomb, les états supersymétriques de la théorie sont représ<strong>en</strong>tés par les équations deschamps auxiliaires D a = V a | θ 2 l’éq(2.35). En d’autres termes, la branche de Coulomb de±¯θ2 ±l’espace des états supersymétriques <strong>du</strong> GLSM est une variété torique M définie commedans éq(2.38). Une variété torique est de Calabi-Yau si et seulem<strong>en</strong>t si les charges Q i a∑satisfont la condition k Q i a = 0 quelque soit la valeur de a.i=12.3.3 Sous-variétés spéciales lagrangi<strong>en</strong>nes (SSL)Une sous variété spéciale lagrangi<strong>en</strong>ne L de dim<strong>en</strong>sion réelle n d’un espace de Calabi-Yau de dim<strong>en</strong>sion complexe n munie d’une (1, 1)-forme de Kahler ω et de (n, 0)- formeholomorphe Ω est définie par [77, 78] :ω| L ≡ 0, Im Ω = 0. (2.40)Dans l’espace Euclidi<strong>en</strong> R 2n = C n de coordonnées complexes (z 1 · · · z n ) avec z i = |z i | e iθ i,on distingue deux types (SSL) que nous noterons dans ce qui suit par L et M :(1) Sous variété spéciale lagrangi<strong>en</strong>ne LUne sous-variété spéciale lagrangi<strong>en</strong>ne L de dim<strong>en</strong>sion réelle n dans C n telle que la restrictionde ωω =n∑d |z i | 2 ∧ dθ i , (2.41)i=137


2.3 Variétés de CY toriquessur L est nulle. Cette condition correspond à fixer les angles θ iθ i = constante . (2.42)Ainsi, la sous variété lagrangi<strong>en</strong>ne spéciale L n’est autre que R n + comme on peut le voirsur la relation ci-dessus <strong>en</strong> pr<strong>en</strong>ant θ i = 0 mod 2π. A titre d’exemple, le plan complexe C 2de coordonnée (z 1 , z 2 ) = ( |z 1 | e iθ 1, |z 2 | e 2) iθ peut être vue comme une fibration T 2 sur R 2 +.Notons que la fibre T 2 dégénère <strong>en</strong> un cercle S 1 sur les bords |z i | = 0 ; soit sur les deuxdemi droites R + = ∂R 2 + et <strong>en</strong> un point à l’origine (0, 0). La sous-variété Lagrangi<strong>en</strong>ne Lde C 2 correspond tout simplem<strong>en</strong>t θ 1,2 = const qui est isomorphe au quart <strong>du</strong> plan réel.Fig. 2-7 – C 2 est une fibration T 2 sur (R + ) 2 .(2) Sous variété spéciale lagrangi<strong>en</strong>ne MCe type de variété spéciale lagrangi<strong>en</strong>ne M = D r × T n−r de l’espace complexe C n estréalisée comme une fibration <strong>en</strong> tore T n−r sur chaque point <strong>du</strong> sous-espace linéaire D r dedim<strong>en</strong>sion r ≤ n. Le sous espace D r ∼ R r + est caractérisé par les (n − r) contraintes :n∑i=1qui ré<strong>du</strong>is<strong>en</strong>t l’espace R n + <strong>en</strong> D rq α i∣ zi ∣ ∣ 2 = c α α = 1, · · · , n − r. (2.43)∼ R r + paraméterisé par r coordonnées réelles s β . Pourchaque valeur fixée α, nous définissons un hyperplan où les nombres <strong>en</strong>tiers q α i avec i =1..., n sont des vecteurs des charges normales et c α sont des constantes. Il s’<strong>en</strong>suit que |z i | 2sont exprimées <strong>en</strong> fonction des s β par le biais r vecteurs vβ i dont l’équation est donnée par :∣ zi ∣ ∣ 2 = v i βs β + d i β = 1, · · · r (2.44)où d i sont des constantes et s βsont des coordonnées qui paramétris<strong>en</strong>t la sous variété38


Variétés de CY et Géométrie Toriquelagrangi<strong>en</strong>ne D r avec les deux conditions ∑ ni=1 qα i vβ i = 0 et qα i d i = c α . La restriction de ωn∑ω = d ∣ ∣ zi 2 ∧ dθ ii=1= ∑ vβds i β ∧ dθ i (2.45)i,β= ∑ i,βds β ∧ d ( v i βθ i) .sur M est nulle et correspond à imposer la contrainte suivante :n∑vβθ i i = 0. (2.46)Cette condition est satisfaite <strong>en</strong> utilisant la condition q α i v i βcontrainte :i=1θ i = q i αϕ α ,= 0 et <strong>en</strong> choisissant laoù les ϕ α représ<strong>en</strong>t<strong>en</strong>t les angles qui paramétris<strong>en</strong>t le tore T n−r . En fait, l’informationsur la fibration torique d’une variété de Calabi-Yau torique est codée dans le diagrammetorique qui paramétris<strong>en</strong>t le lieu de dégénéresc<strong>en</strong>ce de la fibration [63, 70, 113].2.3.4 Diagrammes toriquesLes diagrammes toriques des variétés de Calabi-Yau sont des graphes qui représ<strong>en</strong>t<strong>en</strong>tle lieu de dégénéresc<strong>en</strong>ce de la fibration. Dans le cas des CY3, on trouve deux types degraphes selon la fibration <strong>en</strong> tore T 3 ou la fibration <strong>en</strong> R × T 2 [71, 113]. Dans cette soussectionnous illustrons quelques exemples de fibrations.Fibration <strong>en</strong> T 3Le diagramme torique de O(−3) → CP 2 avec une fibration T 3 est défini comme étantun graphe à trois dim<strong>en</strong>sions et correspond au bord de L où T 3 dégénère <strong>en</strong> T 2 jusqu’àce qu’il soit un point sur le sommet. Afin de trouver ce diagramme, nous comm<strong>en</strong>çons pardéfinir l’application mom<strong>en</strong>t donnée par :|z 1 | 2 + |z 2 | 2 + |z 3 | 2 − 3 |z 4 | 2 = t, (2.47)où t est le paramètre de Kahler. L’intersection des plans|z 1 | 2 = 0, |z 2 | 2 = 0, |z 3 | 2 = 0 et |z 1 | 2 + |z 2 | 2 + |z 3 | 2 = tf permet de représ<strong>en</strong>ter le diagramme torique de O(−3) → CP 2 .39


2.3 Variétés de CY toriquesz3z1z2Fig. 2-8 – La base torique de P 2 .Fibration T 2 × RDans cette fibration, un diagramme torique de la variété de Calabi-Yau torique noncompacte à 3 dim<strong>en</strong>sions est un diagramme à deux dim<strong>en</strong>sions qui représ<strong>en</strong>te les lieux dedégénéresc<strong>en</strong>ce de la fibration T 2 × R sur la base de R 3 .1) Le <strong>Vertex</strong> C 3La variété C 3 est un espace total de la fibration T 2 × R sur R 3 . Nous pouvons l’interprétercomme un espace de phase avec des coordonnées canoniques z i et le mom<strong>en</strong>t conjugué ¯z i∑où i = 1, 2, 3 où la forme symplectique donnée par ω = i 3 dz i ∧ d¯z i . Nous considéronsles trois Hamiltoni<strong>en</strong>s suivants :i=1r α = |z 1 | 2 − |z 3 | 2r β = |z 2 | 2 − |z 3 | 2r γ = Im(z 1 z 2 z 3 ),(2.48)qui définiss<strong>en</strong>t la base R 3 de la fibration. Le crochet de poisson {., .} associé à ω généretrois flux sur C 3 :∂ ɛ z i = {r ɛ , z i } .Alors que l’action de T 2 est donnée par :e iαr α+iβr β : (z1 , z 2 , z 3 ) → ( )e iα z 1 , e iβ z 2 , e −i(α+β) z 3avec α, β ∈ [0, 2π] tandis que la fibre R est générée par l’hamiltoni<strong>en</strong> r γ .40


Variétés de CY et Géométrie Torique(0,1)(1,0)(−1,−1)Fig. 2-9 – Le lieu de dégénéresc<strong>en</strong>ce de la fibration T 2 × R de C 3 dans la base, R 3 = (r α , r β , r γ )En effet, le vertex trival<strong>en</strong>t correspond au lieu de dégénéresc<strong>en</strong>ce <strong>du</strong> tore T 2 dans le sousespace R 2 de la base R 3 . Cette base est représ<strong>en</strong>tée par trois vecteurs v i = (p i , q i ) à deuxdim<strong>en</strong>sions : v 1 = (−1, −1) , v 2 = (0, 1) et v 3 = (1, 0) et satisfait les conditions suivantes :∑ 3i=1 v i = 0v i ∧ v j = p i q j − q i p j = 1.Le cycle paramétrisé par α = (0, 1) dégénère sur le sous-espace r α = r γ = 0, r β ≥ 0 oude manière équival<strong>en</strong>te sur z 1 = z 3 = 0. Le cycle indexé par l’indice β = (1, 0) dégénèresur le sous-espace r β = r γ = 0, r α ≥ 0 ou <strong>en</strong> terme de z i sur z 2 = z 3 = 0. Finalem<strong>en</strong>t, lecycle indexé par α + β dégénère sur le sous-espace r β − r α = r γ = 0, r α ≤ 0, de manièreéquival<strong>en</strong>t sur z 2 = z 1 = 0.2) La variété CP 1 Locale : O(−1) ⊕ O(−1) → CP 1Le conifold résolu est une variété de Calabi Yau non compacte qui admet une descriptiontorique paramétrisée par quatre coordonnées complexes :z 1 z 2 z 3 z 4Q i 1 1 −1 −1où les charges Q i sont les charges qui correspond<strong>en</strong>t à l’action torique sur les coordonnéeshomogènes z i avec la condition de Calabi-Yau ∑ Q i = 1 + 1 − 1 − 1 = 0. Cette variété estidéfinie comme un espace avec la condition suivante :|z 1 | 2 + |z 2 | 2 − |z 3 | 2 − |z 4 | 2 = t. (2.49)41


2.3 Variétés de CY toriquesFig. 2-10 – la base torique de la géométrie locale CP 1On peut représ<strong>en</strong>ter le conifold résolu de deux manières selon la fibration :i) la fibration <strong>en</strong> tore T 3 dont les équations de bords de la base sont données par leséquations suivantes :|z 1 | 2 = 0,|z 2 | 2 = 0,|z 3 | 2 = 0,|z 1 | 2 + |z 2 | 2 − |z 3 | 2 = t.Les intersections de ces plans donn<strong>en</strong>t le diagramme torique <strong>du</strong> conifold résolu.ii) la fibration <strong>en</strong> tore T 2 × R. Dans ce cas, le conifold résolu peut être représ<strong>en</strong>té d’unemanière systématique via la procé<strong>du</strong>re de collage de deux vertex C 3 . Le premier U 1 peutêtre défini par z 1 ≠ 0 ou représ<strong>en</strong>té <strong>en</strong> termes de trois coordonnées (z 2 , z 3 , z 4 ), dont leshamiltoni<strong>en</strong>sr α = |z 3 | 2 − |z 2 | 2(2.50)r β = |z 4 | 2 − |z 2 | 2caractéris<strong>en</strong>t le graphe trival<strong>en</strong>t de C 3 . Alors que pour le vertex U 2 défini par z 2 ≠ 0, lesfonctions (2.50) peuv<strong>en</strong>t être réécrites <strong>en</strong> termes de coordonnées (z 1 , z 3 , z 4 ) <strong>en</strong> utilisant lacondition (2.49), et les hamiltoni<strong>en</strong>s seront donnés par :r α = |z 1 | 2 − |z 4 | 2 − tr β = |z 1 | 2 − |z 3 | 2 − t.Ces fonctions génèr<strong>en</strong>t l’action de cercle :exp(iαr α + iβr β ) : (z 1 , z 3 , z 4 ) → (e i(α+β) z 1 , e −iβ z 3 , e −iα z 4 ).Dans ce vertex U 2 :Le cycle (0, 1) dégénère lorsque r α ≤ −t et r β = −t.42


Variétés de CY et Géométrie ToriqueLe cycle (−1, 0) dégénère lorsque r α = −t et r β ≤ −t.Le cycle (1, −1) dégénère lorsque r α − r β = 0 et r α ≥ −t.Les deux vertex sont reliés par le point commun donné par r α − r β = 0 avec t est leparamètre de Kahler associé au volume de CP 1 .Dans l’espace total O(−m 1 ) ⊕ O(−m 2 ) → P 1 qui s’est décrit comme le quoti<strong>en</strong>t de C 4 par(0,1)U1(1,0)(−1,−1)(−1,0)(1,1)U2(0,−1)Fig. 2-11 – Les vecteurs sont représ<strong>en</strong>tés par les cycles dégénérés sur les lignes.l’action suivante :(z 0 , z 1 , z 3 , z 4 ) → (sz 0 , sz 1 , s −m 1z 2 , s −m 2z 3 ), s ∈ C ∗ ,où la condition de Calabi-Yau est satisfaite lorsque m 1 + m 2 = 2. Les quatre diagrammestoriques sont représ<strong>en</strong>tés dans la figure (2-12).Fig. 2-12 – Quatre diagrammes toriques de l’espace total O(−m 1 ) ⊕ O(−m 2 ) → P 13) La variété CP 2 Locale : O(−3) → P 2Le diagramme torique de la variété de Calabi-Yau torique O(−3) → P 2est construit à43


<strong>du</strong> vertex U 3r α = |z 1 | 2 − |z 0 | 22.3 Variétés de CY toriquespartir la règle de collage de trois vertex C 3 . Dans ce cas, on considère quatre coordonnéesz 0 , z 1 , z 2 , z 3 , avec la contrainte de D-terme :|z 1 | 2 + |z 2 | 2 + |z 3 | 2 − 3|z 0 | 2 = t (2.51)Alors, on définit trois patchs U i avec z i ≠ 0, i = 1, 2, 3, puisque l’une de ces trois coordonnéesdoit être non nulle dans la variété X.(−1,2)U2(0,−1)(1,−1)U 3(0,1)(−1,1)U1(−1,−1)(1,0)(−1,0)(2,−1)Fig. 2-13 – Diagramme torique O(−3) → P 2 . Le collage de trois vertex trival<strong>en</strong>t C 3 .a) Dans le vertex U 3 = (z 0 , z 1 , z 2 ) où z 3 ≠ 0, les hamiltoni<strong>en</strong>s r α et r β sont donnés parr α = |z 1 | 2 − |z 0 | 2r β = |z 2 | 2 − |z 0 | 2 .b) Dans le cas de vertex U 2 = (z 0 , z 1 , z 3 ) où z 2 ≠ 0, l’hamiltoni<strong>en</strong> r α est le même que celuiEn revanche nous utilisons la contrainte éq(2.51), r β s’écrit comme :r β = t + 2|z 0 | 2 − |z 1 | 2 − |z 3 | 2avecexp(iαr α + iβr β ) : (z 0 , z 1 , z 3 ) → (e i(−α+2β) z 0 , e i(α−β) z 1 , e −iβ z 3 ).c) Le dernier vertex U 1 avec z 1 ≠ 0 est exprimé <strong>en</strong> termes de (z 0 , z 2 , z 3 ). En utilisant lacontrainte (2.51), les hamiltoni<strong>en</strong>s sont donnés par :r α = t + 2|z 0 | 2 − |z 2 | 2 − |z 3 | 2r β = |z 2 | 2 − |z 0 | 244


H.JehjouhLes diagrammes toriques sont fréquemm<strong>en</strong>t confon<strong>du</strong>s avec ceux de Feynman. Ces diagrammessembl<strong>en</strong>t être un premier pas vers une meilleure compréh<strong>en</strong>sion de vertex topologique.Les variétés toriques sont caractérisées par un graphe torique où les cycles d’unefibration T 2 se dégénèr<strong>en</strong>t. On montre dans les chapitres suivants comm<strong>en</strong>t calculer toutesles fonctions de partition des variétés de Calabi-Yau grâce aux diagrammes toriques. C’estdonc un outil utile et très puissant pour réaliser les calculs des amplitudes <strong>en</strong> théorie descordes topologiques.45


<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong> <strong>en</strong> Théorie des Cordes46


Chapitre 3Cordes et Invariants <strong>Topologique</strong>sSuite aux développem<strong>en</strong>ts réalisés dans le cadre de la théorie des supercordes et leurscompactifications à 4 dim<strong>en</strong>sions [9, 39, 45, 79, 90], la corde topologique quoique non physiques’est avérée un outil approprié pour approcher la description d’un grand nombre degrandeurs physiques. Elle a fourni plusieurs prédictions totalem<strong>en</strong>t imprévues <strong>en</strong> géométriecomplexe symplectique [37, 102]. C’est le cas par exemple <strong>du</strong> li<strong>en</strong> des cordes topologiquesavec les invariants de noeuds [18, 126, 127], les invariants de Gromov-Witt<strong>en</strong> [128, 129]et les diagrammes toriques des variétés de Calabi-Yau interprétées comme des graphes deFeynman de la théorie topologique [97, 153]. Ces aspects géométriques ont joué un rôleimportant dans le développem<strong>en</strong>t de la théorie des supercordes type II sur les variétés deCY à trois dim<strong>en</strong>sions complexes ; <strong>en</strong> particulier dans l’étude des trous noirs [17, 46, 47].Dans ce chapitre, nous intro<strong>du</strong>isons certains aspects de la théorie des cordes topologiques ;<strong>en</strong> particulier ses deux versions type A et type B. Ensuite nous étudions les branes topologiques<strong>du</strong> modèle A et <strong>du</strong> modèle B ; après quoi, nous considérons la dérivation de diversinvariants topologiques.3.1 Théorie des cordes topologiquesLa théorie des cordes topologiques est ess<strong>en</strong>tiellem<strong>en</strong>t une théorie conforme bidim<strong>en</strong>sionnelletwistée couplée à la gravité à 2D. C’est aussi un modèle sigma non linéairebidim<strong>en</strong>sionnel dont l’espace cible X est une variété de Calabi-Yau à trois dim<strong>en</strong>sionscomplexes [138, 140]. Il existe deux versions principales de corde topologique :(1) le modèle de corde topologique type A auquel nous nous référons souv<strong>en</strong>t par modèleA et qui con<strong>du</strong>it aux invariants de Gromov-Witt<strong>en</strong>.47


3.1 Théorie des cordes topologiques(2) le modèle topologique type B ou modèle B lié aux déformations de la structure complexede la variété de Calabi-Yau X.Les amplitudes dans la théorie de corde topologique inclu<strong>en</strong>t toutes les quantités holomorphesde la théorie de supercorde standard dont les valeurs sont protégées par la supersymétried’espace-temps. Les calculs de ces invariants topologiques sont étroitem<strong>en</strong>tliés à la théorie de Chern-Simons, aux invariants Gromov-Witt<strong>en</strong> et bi<strong>en</strong> d’autres modèles[39, 60, 136, 139, 140, 141, 142, 143].Dans cette section, nous étudions quelques aspects de la théorie supersymétrique N = (2, 2)que nous noterons généralem<strong>en</strong>t comme N = 2. Pour cela, nous comm<strong>en</strong>çons par intro<strong>du</strong>irel’algèbre superconforme N = 2 et deux types de modèles ayant cette superalgèbrecomme symétrie sous-jac<strong>en</strong>te <strong>en</strong> occurr<strong>en</strong>ce le modèle sigma non-linéaire N = 2 et le modèleGinzburg-Landau supersymétrique. Ensuite, nous examinons avec quelques détails lemodèle A et le modèle B qui sont liés <strong>en</strong>tre eux par la symétrie miroir.3.1.1 Modèle sigma non linéaire N = 2En théorie quantique des champs, un modèle sigma non linéaire désigne une théoriedans laquelle les champs fondam<strong>en</strong>taux représ<strong>en</strong>t<strong>en</strong>t des coordonnées dans une variétéRiemanni<strong>en</strong>ne souv<strong>en</strong>t désignée par espace-cible [35],[144]. En théorie des supercordes compactifiéessur des variétés de Calabi-Yau locale (correspondant à la limite à basse énergie),l’action classique de la supercorde se ré<strong>du</strong>it ess<strong>en</strong>tiellem<strong>en</strong>t à un modèle non linéaire surdes variétés toriques.L’intérêt de ce modèle provi<strong>en</strong>t <strong>du</strong> fait que la théorie des cordes topologiques peut êtreconstruite à partir des modèles sigma non linéaire N = 2 dont l’espace cible est une variétéde Calabi-Yau de trois dim<strong>en</strong>sions complexes. Dans cette optique, nous proposons de revoirbrièvem<strong>en</strong>t et dans un premier temps certains résultats utiles sur la supersymétrie N = 2et son twist. Ensuite nous abordons la corde topologique proprem<strong>en</strong>t dite.a) Supersymétrie N=(2,2)Dans les théories superconformes N = 2, le point de départ est la symétrie superconformeN = (2, 2). Cette symétrie se scinde <strong>en</strong> deux copies isomorphes chacune <strong>en</strong>g<strong>en</strong>drée parquatre courants conservés vivant sur la surface d’univers : Il s’agit de la super-algèbreconforme hétérotique (2, 0) et sa cousine la superalgèbre (0, 2) dont les courants conservés48


Théories de cordes topologiques, Invariants et D-branesont respectivem<strong>en</strong>t donnés par :super-algèbre (2, 0) : T L , J L , G + L , G− L,super-algèbre (0, 2) : T R , J R , G + R , G− R . (3.1)Ces courants, qui décriv<strong>en</strong>t le t<strong>en</strong>seur énergie impulsion (T L , T R ), la symétrie R (J L , J R )et les transformations supersymétriques conformes ( G ± L , G± R), correspond<strong>en</strong>t aux secteursholomorphe et anti-holomorphe de la surface d’univers ; c’est à dire :T L = T L (z) , J L = J L (z) , G + L = G+ L (z) , G− L = G− L (z)T R = T R (¯z) , J R = J R (¯z) , G + R = G+ R (¯z) , G− R = G− R (¯z) .Ces courants <strong>en</strong>g<strong>en</strong>dr<strong>en</strong>t respectivem<strong>en</strong>t la superalgèbre (2, 0) et (0, 2) qui sont des algèbresgra<strong>du</strong>ées de dim<strong>en</strong>sions infinies.Surperalgèbre (2, 0)Une façon de définir la super-algèbre conforme (2, 0) est <strong>en</strong> terme des pro<strong>du</strong>its des courantsà courtes distances. Ces pro<strong>du</strong>its d’opérateurs sont donnés par les relations suivantes :T (z) T (ω)∼T (z) G ± (ω) ∼ 3 2T (z)J(ω) ∼ J(ω)G + (z)G − (ω) ∼ 2c/3(z−ω) 3 + 2J(ω)J(z)G(ω) ∼ ± G± (ω)z−ωJ(z)J(ω) ∼ c/3c/2 2T (ω)+ + ∂ ωT (ω)+ termes reguliers(z−ω) 4 (z−ω) 2 (z−ω)G ± (ω)+ ∂ ωG ± (ω)+ termes reguliers(z−ω) 2 (z−ω)+ ∂ωJ(ω) + termes reguliers(z−ω) 2 (z−ω)(3.2)+ 2T (ω)+∂ ωJ(ω)+ termes reguliers(z−ω) 2 z−ω+ termes reguliers+ termes reguliers z−ωDans ces relations, les variables complexes z et w réfèr<strong>en</strong>t aux coordonnées de la surfaced’univers et le nombre c est la charge c<strong>en</strong>trale de la théorie dont la valeur dép<strong>en</strong>d <strong>du</strong> typede degrés de libertés dans la théorie. Un fermion libre a un c = 1/2 et un boson réel ac = 1. Les quatre courants interv<strong>en</strong>ants dans la superalgèbre conforme sont des opérateurslocaux exhibant les expansions de Laur<strong>en</strong>t suivantes :T (z) = ∑ n∈ZL n z −n−2 ,J (z) = ∑ n∈ZJ n z −n−1 ,(3.3)G ± (z) = ∑ n∈ZG n±a z −(n±a)− 3 2 .Le paramètre a (0 ≤ a ≤ 1) contrôle les conditions de bords des fermions. Des relationssimilaires sont égalem<strong>en</strong>t valables pour la superalgèbre (0, 2) . Il faudrait noter au passage49


3.1 Théorie des cordes topologiquesque l’étude des représ<strong>en</strong>tations de ces superalgèbres était à la base <strong>du</strong> développem<strong>en</strong>t dela théorie des supercordes et des phénomènes critiques de la mécanique statistique à 2D.b) Modèle sigma non linéaireLe modèle sigma non linéaire avec la supersymétrique N = 2 vivant sur une surface deRiemann Σ g (surface d’univers) est une théorie supersymétrque de champs bidim<strong>en</strong>sionnelle.Les champs de base dans cette théorie sont des champs scalaires complexes φ i (z, ¯z)qui peuv<strong>en</strong>t être vus comme des applications de la surface Σ g vers une variété Kahleri<strong>en</strong>neX de dim<strong>en</strong>sion n.φ i : Σ g → X ,(z, ¯z) → φ i (z, ¯z) .(3.4)Vu que la théorie est supersymétrique N = 2, il existe égalem<strong>en</strong>t des part<strong>en</strong>aires deschamps scalaires : ce sont des champs fermioniques ψ i ± et ψī± reliés au φ i par des transformationssupersymétriques.L’action <strong>du</strong> modèle sigma supersymétrique N = 2 décrivant le couplage de ces champss’écrit sous la formeS = 2t ∫ Σ d2 zg iī(∂z φ i ∂¯z φī + ∂¯z φ i ∂ z φī)+2t ∫ Σ d2 ziψī −D z ψ i −gīi + iψī +D¯z ψ i +gīi + R iīj¯jψ i +ψī +ψ j −ψ¯j− ,(3.5)où t est la constante de couplage qui jouera un rôle important <strong>en</strong> théorie de corde topologique.Le champg i¯j = ∂ i ∂ j K(φ, ¯φ) (3.6)est la métrique de Kahler alors que le t<strong>en</strong>seur R iīj¯j est le t<strong>en</strong>seur de courbure construit àpartir de la connectionΓ i jk = g i¯k∂ j g k¯k. (3.7)La dérivée covariante interv<strong>en</strong>ant dans l’eq(3.5) est définie comme :D¯z ψ i + = ∂ ∂¯z ψi + + ∂φj∂¯z Γi jkψ k +, (3.8)où les fermions ψ i ± et ψī± sont des sections des fibrés()()ψ i + ∈ Γψ i − ∈ ΓK 1 2 ⊗ φ ∗ (T 1,0 X)( )1 ¯K 2 ⊗ φ ∗ (T 1,0 X)où K, ¯K sont les fibrés canoniques et anticanonique de Σ., ψī+ ∈ Γ K 1 2 ⊗ φ ∗ (T 0,1 X)( )1, ψī− ∈ Γ ¯K 2 ⊗ φ ∗ (T 0,1 X) .(3.9)L’action S <strong>du</strong> modèle sigma non linéaire N = 2 ci-dessus est invariante sous les transformationssupersymétriquesδ = ɛ + Q − − ɛ − Q + − ¯ɛ + ¯Q− + ¯ɛ − ¯Q+ , (3.10)50


Théories de cordes topologiques, Invariants et D-braneoù Q ± et ¯Q ± sont les générateurs de l’algèbre supersymétrique N = 2 qui est une sousalgèbre de la superalgèbre conforme. Les paramètres fermioniques ɛ − , ¯ɛ − sont les paramètres<strong>du</strong> groupe de supersymétrie N = 2 sur la surface d’univers ; ce sont des sections de K 1 2tandis que ɛ + , ¯ɛ + des sections de ¯K 1 2 . Les transformations supersymétriques échang<strong>en</strong>tbosons et fermions et s’écriv<strong>en</strong>t explicitem<strong>en</strong>t comme suit :δφ i = ɛ − ψ i + − ɛ + ψ i − , δφī = −¯ɛ − ψī + + ¯ɛ + ψī−δψ i + = −i¯ɛ − ∂ z φ i + ɛ + ψ j −Γ i jmψ m + , δψī + = −iɛ − ∂ z φī − ¯ɛ + ψ¯j−Γī¯j ¯m ψ ¯m +δψ i − = −i¯ɛ + ∂¯z φ i − ɛ − ψ j +Γ i jmψ m − , δψī − = −iɛ − ∂¯z φī + ¯ɛ − ψ¯j+Γī¯j ¯m ψ +. ¯m(3.11)Notons au passage que l’action <strong>du</strong> modèle sigma non linéaire ne peut pas être globalem<strong>en</strong>tdéfinie sur une surface d’univers de g<strong>en</strong>re g ≠ 1. Dans le cas <strong>du</strong> tore de g<strong>en</strong>re g = 1, lefibré canonique est trivial et les fermions sont donc des scalaires.c) Modèle Landau GinzburgLe modèle Landau-Ginzburg N = 2 sur une variété Kahleri<strong>en</strong>ne de pot<strong>en</strong>tiel de Kahler Kest une théorie supersymetrique vivant sur une surface d’univers bidim<strong>en</strong>tionnelle (surfacede Riemann Σ ). En language des superchamps, l’action de ce modèle s’écrit <strong>en</strong> fonctiondes superchamps chiraux comme suit :∫S Σ = d 2 xd 4 θK ( Φ, ¯Φ ) ∫+ΣΣd 2 xd 2 θW (Φ) + cc, (3.12)où ( x, θ ÷ ±, θ − ±)est le superespace N = 2 à 2D, K = K(Φ, ¯Φ) est le superpot<strong>en</strong>tiel de Kahleret où W = W (Φ) est le superpot<strong>en</strong>tiel holomorphe <strong>en</strong> Φ.Avec ces rappels sur l’algèbre supersymétrique et les théories superconformes N = 2 à 2D,nous abordons la théorie de corde topologique.d) Algèbre twistée N=2Pour fixer les idées, nous comm<strong>en</strong>çons par rappeler les deux propriétés utiles suivantes : (1)La théorie des cordes topologiques est une théorie superconforme N = 2 à 2D twistée. Nousintro<strong>du</strong>isons les types de twistes possible juste après. (2) Il y a deux manières d’obt<strong>en</strong>ir unethéorie des champs topologiques à partir d’une théorie superconforme N = 2. Ce sont lemodèle-A et le modèle-B auxquels nous avons fait référ<strong>en</strong>ce auparavant. Ces deux modèlessont évidemm<strong>en</strong>t liés et la relation <strong>en</strong>tre eux est donnée par la symétrie miroir.L’algèbre définissant la symétrie sous jac<strong>en</strong>te à toute théorie topologique sur une variété deKahler X peut être formellem<strong>en</strong>t obt<strong>en</strong>ue à partir d’un changem<strong>en</strong>t de variables linéairesdes courants de l’algèbre superconforme. Ce changem<strong>en</strong>t, communém<strong>en</strong>t désigné par twist,est défini comme suit :T (z) → ˜T (z) = T (z) ± 1 ∂J(z) , (3.13)251


3.1 Théorie des cordes topologiquesoù le t<strong>en</strong>seur énergie impulsion T (z) est translaté par un terme 1 ∂J(z). Ce twist translate2les spins conformes s (valeurs propores de L 0 ) de tous les champs chargés Φ q s sous laR-symétrie de charge qs → s ± q 2 .Le t<strong>en</strong>seur énergie impulsion twisté satisfait les mêmes propriétés que celui de l’algèbrede Virasoro mais avec une charge c<strong>en</strong>trale c = 0 comme le montre les pro<strong>du</strong>its à courtesdistances suivants :T (z) T (ω) ∼2T (ω)+(z−ω) 2∂ωT (ω)(z−ω)(3.14)T (z) G (ω) ∼ 2G(ω) + ∂ ωG(ω)(z−ω) 2 (z−ω)T (z)Q(ω) ∼ Q(ω) + ∂ ωQ(ω)(z−ω) 2 (z−ω)T (z)J(ω) ∼ − ĉ + J(w) + ∂ωJ(ω)(z−w) 3 (z−w) 2 (z−ω)Q(z)G(ω) ∼ĉ+ J(w) + T (ω)(z−w) 3 (z−w) 2 z−ωJ(z)J(ω) ∼ĉ(z−w) 2J(z)G(ω) ∼ − G(ω)z−wJ(z)Q(ω) ∼ Q(ω)z−wavec ĉ = c et au lieu des générateurs supersymétriques 3G± , nous avons deux nouvellesquantités G (z) , Q (z) avec les poids conformes 2, 1 respectivem<strong>en</strong>tG (z) = ∑ nG n z n−2Q (z) = ∑ nQ n z n−1 . (3.15)Ainsi, étant donné une algèbre superconforme N = (2, 2), on distingue deux choix distinctspour le twist topologique. Ces deux choix qui sont associés aux signes dans l’éq(3.13)mèn<strong>en</strong>t à deux théories des cordes topologiques. La théorie avec un signe (−) dans l’éq(3.13)est une théorie de corde topologique de type A ou modèle A tandis que la théorie avec unsigne (+) est associée à la corde topologique de type B ou le modèle B.3.1.2 Modèles A et BIci, nous étudions les modèles topologiques A et B à partir des twists <strong>du</strong> modèle sigmanon linéaire N = 2 avec les deux courants gauche J L et droit J R dont les groupes decharges R associés seront notés U(1) V et U(1) A . Dans cette objectif, rappelons que laversion euclidi<strong>en</strong>ne <strong>du</strong> groupe de Lor<strong>en</strong>tz SO (1, 1) de la surface de l’univers vue commeune surface Riemann est SO (2) ∼ U(1) E . En utilisant ces groupes de charges, les théories52


Théories de cordes topologiques, Invariants et D-branede cordes topologiques A et B de groupe de symétrie U(1) tw E sont définies par les twistssuivantsmodèle A : U(1) E → U(1) tw E = U(1) E×U(1) V,U(1)modèle B : U(1) E → U(1) tw E = U(1) E×U(1) A.U(1)Les deux théories twistées sont appelées modèles A et B. Il s’avère que l’algèbre superconformetwistée N = (2, 2) reste inchangée sous la transformation Z 2 .J (z) → J (z) ¯J (¯z) → − ¯J (¯z)Par contre, les spineurs sous U(1) E devi<strong>en</strong>n<strong>en</strong>t après le twist des scalaires de spin conformes ′ = 0 et des vecteurs avec s ′ = 1 sous le groupe U(1) tw E .s = 1 ( 12 → s′ =2 2)− q = 0, 1 pour q = 1, −1. (3.16)Les supercharges scalaires dans les deux modèles A et B sont nilpotantes et sont vuescomme des opérateurs BRST dans les deux modèles :Modèle A : Q A = Q + + ¯Q − , Q 2 A = 0Modèle B : Q B = ¯Q + + ¯Q − , Q 2 B = 0 (3.17)Il est connu que le modèle A dép<strong>en</strong>d des mo<strong>du</strong>les de Kähler tandis que le modèle B dép<strong>en</strong>dde la structure complexe de la variété de Calabi-Yau.Modèle AAfin d’obt<strong>en</strong>ir le modèle A, nous redéfinissons les spins des champs comme suit :ψ i + ≡ χ i ∈ Γ (φ ∗ (T 1,0 X)) , ψī + ≡ ψī z ∈ Γ (φ ∗ (K ⊗ T 0,1 X))ψ i − ≡ ψ ī z ∈ (φ ∗ (K ⊗ T 1,0 X)) , ψī− ≡ χī ∈ Γ (φ ∗ (T 0,1 X)).Les champs fermioniques se combin<strong>en</strong>t <strong>en</strong> un scalaire χ ∈ Φ ∗ (T X) et des 1-formes holomorpheset anti-holomorphes. A la lumière de ces constructions, nous arrivons maint<strong>en</strong>antà écrire l’action <strong>en</strong> termes de nouvelles quantités :∫ ( (S = 2t d 2 z g iī ∂z φ i ∂¯z φī + ∂¯z φ i ∂ φī) )z + iψī z∂¯z χ i gīi + iψī ¯z∂ z χīgīi − R iīj¯jψ ī zψī zχ j χ¯jΣ(3.18)Maint<strong>en</strong>ant, nous posons ɛ + = ¯ɛ − = 0 et nous dénotons ɛ − = ¯ɛ + = −ɛ où ɛ est uneconstante fermionique. Nous posons égalem<strong>en</strong>t δ = ɛ(Q + + ¯Q − ). La supersymétrie de typeA est dans ce cas une symétrie scalaire et peut être définie sur des surfaces d’universarbitraires. Dans ce modèle, l’opérateur BRST quiQ A = Q + + ¯Q − , Q 2 A = Q + ¯Q− + ¯Q − Q + = 0,53


3.1 Théorie des cordes topologiquesagit sur les champs via l’anticommutateur et se transform<strong>en</strong>t comme suit :δφ i = ɛχ i δφī= ɛχīδχ i = 0 δχī = 0δψī z = −iɛ∂ z φī + ɛχ¯j Γī¯j ¯m ψ ¯m z δψ ī z = −iɛ∂¯z φ i + ɛχ j Γ i jmψ m¯z .(3.19)La remarque cruciale est que l’action peut être écrite comme suit :∫∫S = it d 2 z {Q, V } + t φ ∗ (K) (3.20)avec(V = g i¯j ψ īz∂¯z φ j + ∂ z∫Σ φ∗ (K) = ∫ Σ d2 zΣΣ)φīψj¯z(∂ z φ i ∂¯z φ¯j g i¯j − ∂¯z φ i ∂ z φ¯j g i¯j) (3.21)Ayant intro<strong>du</strong>it le modèle A, nous passons à la discussion des opérateurs physiques quisont des fonctions de φ et χ. A toute n-formesur X, nous pouvons lui associer l’opérateur local :W = W i1···i ndφ i1 · · · dφ i n(3.22)O W (P ) = W i1···i nχ i1 · · · χ i n(P ) , (3.23)où P est un point sur la surface d’univers. On peut alors montrer que {Q, O W } = −O dW .Modèle BDans le modèle B, les champs sont donnés par les relations suivantes :ψ i + ∈ Γ (K ⊗ φ ∗ (T 1,0 X)) , ψī± ∈ Γ (φ ∗ (T 0,1 X)) , ψ i − ∈ ( ¯K ⊗ φ ∗ (T 1,0 X) ) (3.24)Il est conv<strong>en</strong>able de faire les redefinitions suivantes( )ηī = ψī + + ψī − , θ i = g iī ψ ī+ + ψī−(3.25)En outre, nous combinons ψ i ± à une forme ρ ∈ Φ ∗ (T 1,0 X), où ψ i + = ρ i z et ψ i − = ρ ī z. L’action<strong>en</strong> termes de nouveaux champs se prés<strong>en</strong>te comme suit :⎛ (∫ g iī ∂z φ i ∂¯z φī + ∂¯z φ i ∂ φī) ⎞z + iηī (D z ρ ī z + D¯z ρ i z) g iīS = 2t d 2 z ⎜⎟⎝⎠ (3.26)Σ+iθ i (D¯z ρ i z − D z ρ ī z) + R iīj¯jρ i zρ ī zηīθ k g k¯j54


Théories de cordes topologiques, Invariants et D-braneSi on pose ɛ + = ɛ − = 0 et ¯ɛ − = −¯ɛ + = ɛ, on a δ = ɛ( ¯Q + + ¯Q − ). Comme dans le cas<strong>du</strong> modèle A, la supersymétrie dans le modèle B peut être globalem<strong>en</strong>t définie sur chaquesurface d’univers. Dans ce cas, l’opérateur BRST est :Q B = ¯Q + + ¯Q − . (3.27)La variation supersymétrique <strong>du</strong> modèle B se prés<strong>en</strong>te comme suit :δφ i = 0 ,δρ i = −iɛdφ i ,δ¯φī= −ɛηī ,δηī = δθ i = 0 .L’action ci dessus peut être donnée de la manière suivante :∫∫S = it {Q, V } + t W, (3.28)avecV = g i¯jΣ()ρ i z∂¯z φ¯j + ρ ī z∂ z φ¯jW = ∫ Σ(−θi Dρ i − i 2 R iīj¯jρ i ∧ ρ j ηīθ k g k¯j ) .Σ,(3.29)C’est une théorie topologique dans le s<strong>en</strong>s qu’elle est indép<strong>en</strong>dante de la structure complexeet de la métrique de Kahler sur X. Par la suite, nous discutons les observables locales <strong>du</strong>modèle B. Nous considérons les (0, p)-formes V dans X avec ses valeurs dans ∧ q T 1,0 XAlors, nous pouvons définir l’opérateur suivant :j 1···j qV = d¯zī1 · · · d¯zīp Vī 1···ī pψ j1···ψ jq(3.30)j 1···j qO V = ηī1 · · · ηīp Vī 1···ī pψ j1···ψ jq(3.31)Avant de conclure ce paragraphe, nous donnons les relations suivantes selon que les cordessont ouvertes et fermées.corde topologique ouverte ferméemodèle A Théorie de Chern-Simons Théorie de Gromov Witt<strong>en</strong>modèle B Théorie de Chern-Simons holomorphe Théorie de Kodaira-Sp<strong>en</strong>cerSymétrie miroirLa symétrie miroir est une propriété d’origine physique qui jusqu’à prés<strong>en</strong>t n’a pas une définitionmathématique définitive. En théorie des supercordes compactifiées sur des variétés55


3.1 Théorie des cordes topologiquesde Calabi-Yau, la symétrie miroir est une symétrie qui échange deux espaces de Calabi-Yau de même dim<strong>en</strong>sion. En d’autre terme, si deux espaces de Calabi-Yau différ<strong>en</strong>ts Xet son image Y (dimX = dimY ) par la symétrie miroir sont utilisés <strong>en</strong> tant que deuxespaces cibles, ils con<strong>du</strong>is<strong>en</strong>t alors à la même physique [101, 102, 104, 105]. Presque toutesles variétés de Calabi-Yau à trois dim<strong>en</strong>sions possèd<strong>en</strong>t une variété « miroir » où les h (1,1)mo<strong>du</strong>les de Kahler et h (2,1) mo<strong>du</strong>les complexes sont échangées. Cela veut dire que déformerla structure de Kahler de X serait équival<strong>en</strong>t à déformer la structure complexe de Y,h 1,1 (X) = h 2,1 (Y ), h 2,1 (X) = h 1,1 (Y ). (3.32)La symétrie miroir joue un rôle très important <strong>en</strong> théorie des cordes ; elle permet <strong>en</strong>tre autrede relier les deux modèles topologiques A et B. Le modèle A est s<strong>en</strong>sible à la géométriesymplectique via des D-branes lagrangi<strong>en</strong>nes ( lagrangian D-branes <strong>en</strong> anglais) alors que lemodèle B est s<strong>en</strong>sible à la géométrie complexe. Ainsi, étant donné une variété de Calabi-Yau X, il doit exister une variété de Calabi-Yau miroir Y tel que l’énergie libre Fg A (t; X)dans modèle A sur X et l’énergie libre Fg B (t; Y ) <strong>du</strong> modèle B sur Y coincid<strong>en</strong>t,Fg A (t; X) = Fg B (˜t; Y ), (3.33)où g, t et ˜t désign<strong>en</strong>t respectivem<strong>en</strong>t le g<strong>en</strong>re de la surface d’univers, les paramètres deKahler et les paramètres complexes. Dans le contexte de la symétrie miroir, les deuxversions topologiques A et B sont reliées comme le montre la figure suivante :Fig. 3-1 – Dualité N largeDans la sous section suivante, nous allons examiner brièvem<strong>en</strong>t la correspondance <strong>en</strong>treLandau-Ginzburg/Calabi-Yau, qui id<strong>en</strong>tifie le modèle sigma linéaire avec un certain superpot<strong>en</strong>tiel.56


Théories de cordes topologiques, Invariants et D-brane3.1.3 Modèle sigma linéaire jaugéLe modèle sigma linéaire jaugé à 2D avec une supersymétrique N = 2 peut être vucomme la réalisation des variétés complexes toriques <strong>en</strong> termes de superchamps chiraux Φ iet un certain nombre de superchamps de jauge V a portant la symétrie de jauge <strong>du</strong> modèle.Rappelons qu’un superchamp chiral Φ est une fonction sur le super-espace N = 2, D = 2 ;il conti<strong>en</strong>t un boson complexe φ et des fermions complexes gauche ψ + et droit ψ − ainsiqu’un champ complexe auxiliaire F .Cette théorie de jauge à 2 dim<strong>en</strong>sions avec un groupe de jauge abéli<strong>en</strong> U (1) s est représ<strong>en</strong>téepar des vecteurs superchamps V 1 · · · V s et n superchamps chiraux Φ 1 · · · Φ n dont la chargeest notée par Q a i . La forme <strong>du</strong> Lagrangi<strong>en</strong> <strong>du</strong> modèle sigma est :L = L cin + L W + L gauge + L F I,θ (3.34)où L cin et L gauge décriv<strong>en</strong>t respectivem<strong>en</strong>t les termes cinétiques des champs chiraux et dejauge, L W le superpot<strong>en</strong>tiel et L F I,θ le terme de Fayet IIliopolos (FI) avec le facteur thêtausuelle. En langage des superchamps, l’action <strong>du</strong> modèle supersymétrique s’écrit comme :avecS = S cin + S jauge + S W + S F I,θ , (3.35)S cin = ∫ d 2 x d 4 θ ∑ i Φ i(e2 P a Qa i V a)Φi ,∫S jauge = − 14e d 2 xd 4 θ ¯ΣΣ ,2S W = ∫ (3.36)d 2 y d 4 θW + cc ,∫S F I,θ = it2 √ 2 d 2 x dθ + d¯θ − Σ + c.c ,où e est le couplage de jauge et Σ est le superchamp chiral twisté qui vérifie la condition :¯D + Σ = D − ¯Σ = 0 (3.37)et où t = ir + θ avec r est le paramètre de Fayet-IIiopoulos.2πNous allons maint<strong>en</strong>ant examiner une théorie avec n superchamps chiraux S i de charge 1et un superchamp chiral P de charge −n dont les élém<strong>en</strong>ts bosoniques sont respectivem<strong>en</strong>ts i et p. Le superpot<strong>en</strong>tiel est donné par une fonction invariante de jaugeW = P.G (S 1 , · · · S n ) , (3.38)avec G un polynôme homogène de degré n et le pot<strong>en</strong>tiel bosonique s’écrit :U = |G (s i )| + |p| ∑ ∣ ( )2 ∂G ∣∣∣2 ∑∣ + 2 |σ| 2 |s i | 2 + n 2 |p| 2 , (3.39)∂si ii57


3.1 Théorie des cordes topologiquesavecD = −e 2 ( ∑i¯s i s i − n¯pp − r). (3.40)Dans le cas r ≫ 0 et p = 0, l’annulation <strong>du</strong> D-terme mène à :∑¯s i s i = r. (3.41)iPour trouver les états supersymétriques, nous pr<strong>en</strong>ons <strong>en</strong> considération la symétrie dejauge(s 1 , · · · , s n ) ∼ ( )e iϕ s 1 , · · · , e iϕ s n (3.42)qui tra<strong>du</strong>it le fait que s i vit sur l’espace projectif CP n−1 ∼ C n \0/C ∗ . Nous concluons quel’espace des vides supersymétriques classiques (D = 0) est isomorphe à une hypersurfacede degré n dans CP n−1 qui est une variété de Calabi-Yau. Dans ce cas, le modèle sigmasupersymétrique N = 2 à 2D décrit une réalisation physique des variétés de Calabi-Yau.Dans le régime non-géométrique (r ≪ 0), on peut réalisé égalem<strong>en</strong>t le modèle sigmalinéaire comme un modèle de Landau-Ginzburg. Les deux régions r > 0 et r < 0 sontséparées par une singularité <strong>en</strong> r = 0.3.1.4 Branes topologiquesSelon le type de modèle de cordes topologiques, on distingue des branes <strong>du</strong> type A etdes branes <strong>du</strong> type B pour désigner respectivem<strong>en</strong>t les D-branes des modèles topologiquesA et B [131], [132]. Il s’<strong>en</strong>suit que les conditions de bords dans la théorie des cordestopologiques doiv<strong>en</strong>t être compatibles avec le twist topologique. En particulier, les branesdans les modèles A et B doiv<strong>en</strong>t préserver respectivem<strong>en</strong>t la supersymétrie de type A etde type B.Nous représ<strong>en</strong>tons les branes de type A comme des sous variétés L a de l’espace de Calabi-Yau X à trois dim<strong>en</strong>sions,L a ⊂ X. (3.43)De plus, les applications φ i de la surface d’univers Σ g vers la variété de Calabi-Yau etqui définiss<strong>en</strong>t les champs φ i (z, ¯z) de la corde topologique, φ i : z ∈ Σ g → φ i (z, ¯z) ∈ X,obéiss<strong>en</strong>t à la condition :φ i (∂Σ g ) ⊂ L a ,⋃a58


Théories de cordes topologiques, Invariants et D-braneLes branes-A sont alors définies par des conditions de géométrie symplectique in<strong>du</strong>ites parla forme de Kahler et <strong>en</strong>roul<strong>en</strong>t les sous variétés lagrangi<strong>en</strong>nes L qui possèd<strong>en</strong>t la propriété :ω |L = 0, (3.44)où ω est la forme sympléctique de la variété de Calabi-Yau X.Les branes-B <strong>du</strong> modèle B sont définies par des conditions de géométrie complexe et port<strong>en</strong>tdes fibrés holomorphes. La structure complexe sur les branes B préserve leurs directionsnormales et tang<strong>en</strong>tes et par conséqu<strong>en</strong>t elle <strong>en</strong>roule les cycles holomorphiques dans lavariété X.Fig. 3-2 – La symétrie miroir <strong>en</strong>tre les deux modèles A et B.Les branes topologiques A et B vérifi<strong>en</strong>t égalem<strong>en</strong>t d’autres conditions de stabilité ; pourplus de détails voir [76, 130].3.2 Dualité corde ouverte / corde ferméeL’une des propriétés remarquables de la théorie des cordes topologiques est la correspondanceJauge/Gravité. Cette correspond<strong>en</strong>ce est considérée comme l’un des sujetsimportants dans le réc<strong>en</strong>t développem<strong>en</strong>t de la théorie des cordes non-perturbative. Dupoint de vue technique, la découverte de cette équival<strong>en</strong>ce est réalisée concrètem<strong>en</strong>t grâceà l’étude des D-branes. Dans la théorie des cordes topologiques, cette correspondance estréalisée par une équival<strong>en</strong>ce <strong>en</strong>tre les fonctions de partitions Z des modèles A et B. Cesfonctions de partition représ<strong>en</strong>t<strong>en</strong>t des fonctionnelles génératrices des invariants topologiquesdes invariants de Gromov-Witt<strong>en</strong> et Gopakumer-Vafa. Un exemple de cette <strong>du</strong>alitéest donné par la transition géométrique <strong>en</strong>tre le conifold résolu et celui déformé. Il s’agit, <strong>en</strong>effet, de l’équival<strong>en</strong>ce <strong>en</strong>tre les invariants de Gromov-Witt<strong>en</strong> définis sur le conifold résolud’une part et les invariants de la théorie de Chern-Simons définis sur S 3 d’autre part.59


3.2 Dualité corde ouverte / corde fermée3.2.1 Dualité jauge/ gravitéEn théorie de supercordes, les bosons de jauge A a µ, avec une constante de couplage dejauge g Y M et un groupe de jauge G que nous pr<strong>en</strong>ons dans la suite comme étant SU (N),correspond<strong>en</strong>t aux états non massifs de la corde ouverte, tandis que le graviton G µν estassocié à l’état non massif de la corde fermée. La <strong>du</strong>alité <strong>en</strong>tre cordes ouvertes et ferméesest interprétée alors <strong>en</strong> terme de correspondance Jauge/Gravité. Ce g<strong>en</strong>re de <strong>du</strong>alité a étéréalisé pour la première fois <strong>en</strong> 1974 par ’t Hooft dans la limite termodynamique N → ∞avec le paramètre gY 2 MN est fixé [137].Les amplitudes F de la théorie de jauge SU(N) ont un développem<strong>en</strong>t de typeF = ∑ gN 2−2g F g , (3.45)où la fonction F g représ<strong>en</strong>te la somme de tous les diagrammes de g<strong>en</strong>re g. Un tel développem<strong>en</strong>tressemble formellem<strong>en</strong>t à celui d’une théorie des cordes avec une constante decouplage g s = 1 N fig(3-3).Fig. 3-3 – Développem<strong>en</strong>t <strong>en</strong> termes de diagrammes de g<strong>en</strong>re g.Dans ce cadre, une découverte clé a été faite par Maldac<strong>en</strong>a qui a conjecturé la théorie deYang-Mills SU(N) superconformes à D = 4, N = 4 est <strong>du</strong>ale à la théorie des supercordesde type IIB définie sur une variété AdS 5 × S 5 . Ce résultat intéressant est connue souscorrespondance AdS/CF T [40, 87].3.2.2 Dualité de Gopakumar-VafaUn cas particulier de correspondance corde ouverte / corde fermée est connu sousle nom de la <strong>du</strong>alité de Gopakumar-Vafa. Cette correspond<strong>en</strong>ce peut être réalisée d’unefaçon précise dans le cadre de la théorie de corde topologique. En 1998, Gopakumar etVafa ont montré qu’il existe une équival<strong>en</strong>ce <strong>en</strong>tre les théories de jauge topologiques, quisont intrinsèquem<strong>en</strong>t plus simples que la théorie de super Yang-Mills, et les modèles dethéorie des cordes topologiques, qui à leurs tours sont aussi plus simples que les théories60


Théories de cordes topologiques, Invariants et D-branedes supercordes usuelles. La correspondance proposée par Gopakumar et vafa est l’un desexemples des <strong>du</strong>alités pour laquelle la théorie de Chern-Simons, avec un groupe de jaugeSU(N) formulée sur la sphère à trois dim<strong>en</strong>sions S 3 , est équival<strong>en</strong>te dans la limite N largeà la théorie des cordes topologiques fermées de type A sur le conifold résolu. Ceci se tra<strong>du</strong>itpar :Cordes topologiques fermées sur le conifold résolu ⇐⇒ Théorie de Chern-Simons sur S 3 .(3.46)La puissance de cette <strong>du</strong>alité réside dans le contexte de la théorie de Chern-Simons correspondanteà la théorie de jauge à 3 dim<strong>en</strong>sions qui vit sur le volume d’univers de D-brane.En effet, une action peut être décrite <strong>en</strong>tièrem<strong>en</strong>t <strong>en</strong> termes des formes différ<strong>en</strong>tielles :S CS = k ∫T r(A ∧ dA + 2 )4π3 A ∧ A ∧ A (3.47)Mavec A est une connection de jauge sur la variété M à trois dim<strong>en</strong>sions et k est uneconstante de couplage. De plus, l’action ne dép<strong>en</strong>d pas de la métrique et le résultat sera unethéorie topologique. Dans cette situation, la fonction de partition représ<strong>en</strong>te un invarianttopologique de la variété M∫Z (M) =[DA] e iS . (3.48)La classe des observables de cette théorie est donnée par la valeur moy<strong>en</strong>ne des boucles deWilson :WR K =〈T r R P e R 〉 ∫K A =DAe iSCS T r R P e R K A . (3.49)D’autre part, la comparaison <strong>en</strong>tre les amplitudes des deux théories de Chern-Simons et lathéorie de corde fermée est possible grâce à l’exist<strong>en</strong>ce d’un li<strong>en</strong> <strong>en</strong>tre les paramètres desthéories de jauge et de cordesλ =2πk + N ,t = 2πiNk + N , (3.50)où λ est la constante de la corde topologique et t est le mo<strong>du</strong>le de Kahler de la sphèreS 2 . Il est intéressant de voir qu’il existe un li<strong>en</strong> <strong>en</strong>tre la constante de couplage de la cordetopologique et le couplage de ’t Hooftλ = g 2 CS t = iNg 2 CS. (3.51)La correspondance proposée par Gopakumar et vafa a permis d’id<strong>en</strong>tifier les amplitudesde Chern-Simons FgCSle conifold résolu,(t) et celles F conifoldgdes cordes topologiques fermées <strong>du</strong> type A surFgCS (t) = Fg conifold (t). (3.52)61


3.3 Invariants topologiquesNotons au passage que la fonction de partition de la corde topologique est défnie par :Z top = e F top. F top = ∑ g=0F g g 2g−2s . (3.53)Signalons égalem<strong>en</strong>t que l’énergie libre F g de g<strong>en</strong>re g ≥ 1 de la théorie des cordes peut êtredirectem<strong>en</strong>t évaluée comme une somme sur le secteur des instantons à l’instar <strong>du</strong> modèlesigma topologiqueF g (t) = ∑ N g,β Q β , (3.54)βavec N g,β sont les invariants de Gromov-Witt<strong>en</strong> de la variété de Calabi-Yau. En effet,l’énergie libre de la théorie des cordes définie ici peut être directem<strong>en</strong>t obt<strong>en</strong>ue <strong>en</strong> calculantles invariants topologiques de la variété de Calabi-Yau.Il est intéressant de remarquer qu’il existe d’autres <strong>du</strong>alités. A trois dim<strong>en</strong>sions, Witt<strong>en</strong> amontré que les théories effectives de la corde topologique ouverte peuv<strong>en</strong>t être exactem<strong>en</strong>trésolues <strong>en</strong> les reliant à la théorie de Chern-Simons, via la correspondance :Cordes topologiques ouvertes sur T ∗ M ⇐⇒ Théorie de Chern-Simons sur M (3.55)où M est une sous variété lagrangi<strong>en</strong>ne de dim<strong>en</strong>sion 3 d’une variété de Calabi-Yau T ∗ M.Selon cette correspondance, le rang N <strong>du</strong> groupe de Jauge dans la théorie de Chern-Simonsdevi<strong>en</strong>t un <strong>en</strong>semble de N branes <strong>du</strong> type A <strong>en</strong>roulées sur la variété M dans la théorie descordes ouvertes.Il faut remarquer que la découverte de ces deux <strong>du</strong>alités éq(3.46, 3.55) a montré aussi unli<strong>en</strong> inatt<strong>en</strong><strong>du</strong> dans la limite N large <strong>en</strong>tre les cordes topologiques ouvertes <strong>du</strong> modèle Asur T ∗ S 3 et les cordes topologiques fermées sur le conifold résolu.Cordes topologiques ouvertes sur T ∗ S 3 ⇐⇒ Cordes topologiques fermées sur le conifold résolu(3.56)Dans la suite, nous nous limitons au modèle A et nous considérons l’étude des invariantsde Gromov-Witt<strong>en</strong> et Gopakumar-Vafa.3.3 Invariants topologiquesLes invariants topologiques constitu<strong>en</strong>t un outil fort pour étudier les amplitudes dela théorie des cordes topologiques. Parmi ces invariants topologiques, on distingue les62


Théories de cordes topologiques, Invariants et D-braneinvariants de Gromov-Witt<strong>en</strong>, ceux de Gopakumar-Vafa [9, 52] et les invariants des noeuds.Nous avons :corde fermée : Gromov-Witt<strong>en</strong> Gopakumar-Vafa/ Donaldson-Thomascorde ouverte : Gromov-Witt<strong>en</strong> ouverte invariants Ooguri-VafaDans la première sous section , nous étudions les invariants de Gromov-Witt<strong>en</strong>, <strong>en</strong>suit<strong>en</strong>ous considérons les invariants de Gopakumar-Vafa et nous terminons par les invariantsde noeuds.3.3.1 Invariants de Gromov-Witt<strong>en</strong>Connus comme une classe d’invariants, les invariants de Gromov -Witt<strong>en</strong> compt<strong>en</strong>tle nombre de courbes de g<strong>en</strong>re g dans une variété compacte symplectique ou projective.D’une manière générale, pour une famille de variétés de Calabi-Yau X de dim<strong>en</strong>sion 3,il est possible de déterminer les invariants de Gromov-Witt<strong>en</strong> <strong>en</strong> calculant la fonctionde partition de la corde topologique <strong>du</strong> modèle A. L’une des méthodes appliquée pourcalculer les invariants de Gromov Witt<strong>en</strong> consiste à considérer l’énergie libre F top de lathéorie de corde de type A qui <strong>en</strong>code les invariants énumératifs des applications φ stablesholomorphes des surfaces de Riemann Σ g de g<strong>en</strong>re g vers X :φ : Σ g → X avec φ (Σ g ) = [β] ∈ H 2 (X, Z) , (3.57)où β sont des classes d’homologie. La structure de l’énergie libre F top= F (t, X) où leprépot<strong>en</strong>tiel de Gromov-Witt<strong>en</strong> dép<strong>en</strong>d <strong>du</strong> paramètre de Kahler t = (t 1 , · · · t h 1,1) , de laforme de Kahler K et de la constante de couplage g s et elle peut se mettre sous la forme :F (t, X) = ∞ ∑g≥0g 2g−2s F g (t) avec F g (t) =∑[β]∈H 2 (X,Z)N g β e(K.β) , (3.58)avec(K.β) = (n 1 t 1 , · · · n h 1,1t h 1,1) (3.59)et où N g βsont les invariants de Gromov-Witt<strong>en</strong> de g<strong>en</strong>re g qui correspond<strong>en</strong>t à la classe β.On peut aussi écrire l’amplitude de la corde topologique de type A <strong>en</strong> terme de la fonctionde MacMahon :Z top = e F top= e P F g gs2g−263= M(q) χ(M)2 e F class+F(3.60)


3.3 Invariants topologiquesoù F est la contribution des instantons des surfaces d’univers qui est donnée par l’énergielibre F,F = ∑ g≥0∑βg 2g−2s N g β Qβ , (3.61)et où F class <strong>en</strong>code le nombre d’intersections a ijk des 2-cycles de g<strong>en</strong>re g = 0F class = ∑ 1a6gs2 ijk t i t j t k + ∑ 124 b it i . (3.62)Les termes pour le g<strong>en</strong>re g = 1 sont reliés à la seconde classe de Chern c 2 (M) comme∑∫bi t i = K ∧ c 2 (M) , (3.63)Mtandis que les applications constantes de g<strong>en</strong>re g > 1 contribue par une puissance à lafonction de MacMahon comme exhibé dans l’eq(3.60),M(q) = ∏ n=1(1 − q n ) −n avec q = e −gs . (3.64)Dans le cas où la variété de Calabi-Yau est M = C 3 , les invariants de Gromov-Witt<strong>en</strong> sontdonnés par [89, 135] :N g 0 = (−1)g |B 2g B 2g−2 |2g (2g − 2) (2g − 2)! , g ≥ 2 (3.65)avec B 2g est le nombre de Bernoulli [61]. Ce résultat donne l’amplitude de la corde topologiquesur la variété de Calabi-Yau C 3 dont la caractéristique d’Euler est χ(C 3 ) = 2. Cetteamplitude est exactem<strong>en</strong>t la fonction de MacMahon à 3 dim<strong>en</strong>sionsZ (q) C3top = M(q) = ∏ n=1(1 − q n ) −n . (3.66)Une telle fonction a de nombreuses interprétations : En physique, elle permet d’effectuer lescalculs des amplitudes de théorie des cordes et des modèles de la physique statistique. Enmathématique, elle permet de calculer explicitem<strong>en</strong>t le li<strong>en</strong> avec les invariants topologiques.3.3.2 Invariants de Gopakumar-VafaNous comm<strong>en</strong>çons par rappeler que la compactification de la théorie de supercordetype IIA sur une variété de Calabi-Yau M donne la théorie de supergravité N = 2, D = 4avec h 1,1 est un vecteur multiplet [1, 10]. Chaque multiplet conti<strong>en</strong>t un champ gravitonnel64


Théories de cordes topologiques, Invariants et D-branedont la courbure est R, un gravi-photon de courbure F ph et un champ scalaire complexeφ i qui correspond au mo<strong>du</strong>le de kahler de la variété M. Dans ce cadre, l’action effective dela théorie des supercordes devi<strong>en</strong>t une action de la théorie de supergravité de g<strong>en</strong>re g ≥ 1∫d 4 xF g(φi ) R 2 +F 2g−2ph,+ , (3.67)avec + représ<strong>en</strong>te la partie self <strong>du</strong>al. Ce terme décrit une diffusion de graviton et graviphotonsdont les expressions sont <strong>en</strong>codées dans la fonction F g . Les F g sont id<strong>en</strong>tifiées auxamplitudes de la corde topologique de type A de g<strong>en</strong>re g. Dans cette id<strong>en</strong>tification, leschamps scalaires sont interprétés <strong>en</strong> termes des mo<strong>du</strong>les t i de la variété M. Il s’avère quel’amplitude F 0(φi ) de g<strong>en</strong>re g = 0 permet de calculer le prépont<strong>en</strong>tiel de la théorie dejauge N = 2 dont la constante de couplage de jauge est donnée par :τ ij = ∂ i ∂ j F 0 (3.68)En suivant l’étude [89] et interprétant la courbure <strong>du</strong> graviphoton comme une constanteF (ph,+) = g s , l’éq(3.67) se ré<strong>du</strong>it à∫d 4 xR 2 +F top (t) . (3.69)L’analyse de la section précéd<strong>en</strong>te montre que la partie instantonique de l’énergie libre dela théorie de corde doit être nécessairem<strong>en</strong>t sous la forme [89] :F (t i ) =∑∞∑∞∑n g ββ∈H 2 (M) g=0 d=1Q dβd[d] 2−2g avec [d] = e d 2 − e − d 2 , et Q = e −t . (3.70)dont les nombres n g β, qui caractéris<strong>en</strong>t l’amplitude de manière élégante, ne sont autres queles invariants Gopakumar-Vafa.Ces invariants ont un li<strong>en</strong> avec les invariants de Gromov-Witt<strong>en</strong> ; ce sont des combinaisonslinéaires des invariants Gromov-Witt<strong>en</strong>. Par comparaison aux éqs (3.58) et (3.70), nousvoyons que les contributions des instantons F g , <strong>en</strong>codées dans les invariants de Gromov-Witt<strong>en</strong> N g β , sont déterminées par les invariants Gopakumar-Vafa nk β pour des g<strong>en</strong>res g ≥ k.A titre d’exemple, l’usage <strong>du</strong> développem<strong>en</strong>t de l’éq(3.70), nous obt<strong>en</strong>ons les contributionsdes instantons de g<strong>en</strong>re g = 0 de l’énergie libre de l’éq(3.54) :F 0 =∑∞∑β∈H 2 (M) d=165n 0 Q dββ(3.71)d 3


3.3 Invariants topologiquesDans le cas d’un seul invariant n 0 1 = 1 où l’indice β est id<strong>en</strong>tifié à une sphère S 2 , le termeinstantonique dans l’énergie libre <strong>du</strong> conifold résolu est donné par [60, 89] :−Q nF conifold = ∑ n [n] 2 = − ∑ n≥1g≥0g 2−2gs|B 2g |2g (2g − 2)! Li 3−2g (Q) (3.72)Q nn kavec Li k (Q) = ∑ et Q = e −t , t est la taille de P 1 . De plus, le term<strong>en</strong>≥1id<strong>en</strong>tifié aux invariants de Gromov-Witt<strong>en</strong> comme suit :|B 2g |2g(2g−2)!est|B 2g |2g (2g − 2)! × ( −n 2g−3) = N g n (3.73)Notons que toutes les informations nécessaires codées dans un nombre infini d’invariantsde Gromov-Witt<strong>en</strong> <strong>du</strong> conifold résolu sont maint<strong>en</strong>ant représ<strong>en</strong>tées par l’invariant deGopakumar-Vafa n 0 1.3.3.3 Boucle de Wilson, invariants de noeuds et d’<strong>en</strong>trelacsEnsemble avec les deux types d’invariants considérés ci-dessus, il existe d’autres invariantstopologiques ayant des interprétations physiques et mathématiques. Witt<strong>en</strong> a montréque certains invariants peuv<strong>en</strong>t être obt<strong>en</strong>us à partir de la théorie quantique des champspar l’intégral de Feynman. Ces invariants sont exprimés comme la fonction de partitionde système quantique [134]. L’exemple standard est donné par la théorie de Chern-Simonsoù l’interaction <strong>en</strong>tre les particules est décrite par des opérateurs de Wilson. La valeurmoy<strong>en</strong>ne dans le vide qui décrit la propagation le long <strong>du</strong> chemin C, donne l’invariant dejauge suivant :∮W R (C) = T r(P exp A), (3.74)Coù R est une représ<strong>en</strong>tation <strong>du</strong> groupe de jauge. Un autre aspect de ces observables estqu’elles fourniss<strong>en</strong>t des invariants de noeuds et de li<strong>en</strong>s. Ces derniers sont des objets mathématiquesins<strong>en</strong>sibles à la déformation de métrique. Et pour les illustrer, il est intéress<strong>en</strong>td’utiliser la <strong>du</strong>alité de Gopakumer-Vafa (GV). On comm<strong>en</strong>ce par considérer la théoriede Chern-Simons sur S 3 avec une boucle de Wilson W R (Γ) le long d’un noeud Γ qui estreprés<strong>en</strong>té par [126]q(s) ∈ S 3 (0 ≤ s ≤ 2π) , (3.75)Selon la <strong>du</strong>alité <strong>en</strong>tre la théorie de Chern-Simons sur S 3 et la corde topologique sur T ∗ S 3éq(3.55), la configuration correspondante W R (Γ) <strong>en</strong> théorie de corde ouverte sur T ∗ S 3 est66


Théories de cordes topologiques, Invariants et D-branedonnée par N branes type A <strong>en</strong>roulant la base S 3 . Comme conséqu<strong>en</strong>ce, il existe uneconstruction canonique de fibré conormale{}˜C Γ = (q(s), p)∣ p dq iids = 0, (0 ≤ s ≤ 2π)(3.76)pour laquelle on associe à chaque point q ∈ Γ qui est un noeud q(s) ∈ S 3 , une sous-variétéà 2 dim<strong>en</strong>sions de T ∗ q S 3 orthogonale à dqds . En effet, ˜C K est une sous variété à 3 dim<strong>en</strong>sionsde T ∗ S 3 d’une topologie R 2 × S 1 . La forme symplectiqueω =3∑dq i ∧ dp i ,i=1(q ∈ S 3 , p ∈ T q S 3) . (3.77)s’annule donc sur ˜C Γ . Ce qui r<strong>en</strong>d le 3-cycle ˜C Γ une sous variété Lagrangi<strong>en</strong>ne, qui intersectela base S 3 le long de la boucle q(s).pdqdsS 3q(s)Fig. 3-4 – Sous variété lagrangi<strong>en</strong>ne ˜C K intersecte S 3 dans T ∗ S 3 le long d’un noeud.En considérant M D-branes qui <strong>en</strong>roul<strong>en</strong>t la sous-variété ˜C Γ , on trouve la théorie de Chern-Simons sur ˜C Γ ainsi que la théorie de Chern-Simons sur S 3 , où les modes non massifs <strong>du</strong>système sont :- les champs de jauge A de la théorie de Chern-Simons SU(N) sur S 3 .- les champs de jauge à sur ˜C Γ ainsi qu’un champ scalaire φ dans une représ<strong>en</strong>tationfondam<strong>en</strong>tale de SU(N) × SU(M) qui existe sur l’intersection de S 3 ∩ ˜C Γ = Γ.Dans ce cas, les boucles de Wilson qui décriv<strong>en</strong>t les propagations des champs de jauge Aet à le long <strong>du</strong> chemin Γ sont données par :∮U = P exp A∮V = P expà (3.78)ΓΓ67


3.3 Invariants topologiquesL’action des champs des cordes devi<strong>en</strong>t une somme des actions de Chern-Simons pour lesdeux champs de jauge S CS CS(A) et S (Ã)∮ ( )T r¯φ d + A − Ã φ. (3.79)ΓLe premier terme correspond tout simplem<strong>en</strong>t à l’action de la théorie de jauge qui correspondau terme cinétique ¯φdφ. Il suffit d’effectuer une intégration sur ces champs pourobt<strong>en</strong>ir l’opérateur de Ooguri-Vafaexp∞∑n=11n (T rU n ) (T rV n ) = ∑ R(T r R U) (T r R V ) . (3.80)Alors, l’action de la théorie de Chern-Simons des champs de jauge A <strong>en</strong> prés<strong>en</strong>ce de probebrane 1 est sous la forme :∫DA e ∑ −S Cs(A)T r R UT r R V = ∑RR〈 ∮WRT Γ r R V, WR Γ = T r R P expΓA〉. (3.81)Ce qui prouve que la fonction de partition de la théorie de corde topologique est unefonction génératrice des invariants des noeuds pour toutes les représ<strong>en</strong>tations possibles.L’action effective des champs de jauge sur ˜C K est sous cette forme :) )S(Ãouverte = S(ÃCS + F ouverte (V ) (3.82))avec les corrections F(Ãouverte qui sont déterminées <strong>en</strong> intégrant sur les champs φ et A.Pour terminer cette analyse, nous allons voir l’action effective des champs de l’autre côtéde <strong>du</strong>alité. Sur le conifold résolu, l’action effective des champs s’écrit comme :) )S(Ã = S(ÃCS + F (V ) , (3.83))avec S(ÃCS est l’action de la théorie de Chern-Simons <strong>en</strong> prés<strong>en</strong>ce des corrections F (V ).Ces corrections sur la corde fermée provi<strong>en</strong>n<strong>en</strong>t des instantons de surface d’univers qui<strong>en</strong>roul<strong>en</strong>t les S 2 et qui finiss<strong>en</strong>t sur ˜C Γ .Dans le cas où Γ = ”unknot”, on trouve un accord <strong>en</strong>tre les corrections des deux côtés de<strong>du</strong>alité. Ces corrections sont analysées <strong>en</strong> détail dans [126]F op<strong>en</strong> (V ) | Γ=unknot = F (V ) | Γ=unknot .1 Probe brane désigne une brane qui ne va pas affecter le background.68


Théories de cordes topologiques, Invariants et D-braneUn autre résultat intéressant est l’étude de l’<strong>en</strong>semble N des noeuds où la valeur moy<strong>en</strong>nedans le vide <strong>du</strong> pro<strong>du</strong>it de boucles de Wilson s’écrit comme :〈γ W 1R 1(A) . . . W γ nR n(A) 〉 = 1 ∫ ( n)∏[DA] W γ n iRZ (M)nie iS , (3.84)les R i sont des représ<strong>en</strong>tations irré<strong>du</strong>ctibles <strong>du</strong> groupe de jauge SU(N). Cette expressionest extrêmem<strong>en</strong>t concise grâce à la prescription <strong>du</strong> pro<strong>du</strong>it de boucle de Wilson. En effet,on peut calculer l’invariant de deux représ<strong>en</strong>tations correspondantes à un même noeud.L’exemple intéressant dans ce cas est la construction <strong>du</strong> nombre d’<strong>en</strong>trelacs de Hopf quiconsiste à <strong>en</strong>trelacer deux noeuds :〈 ∮∮W R1 R 2= T r R1 P exp AT r R2 P expΓ 1i=1Γ 2A〉(3.85)Fig. 3-5 – Entrelac de HopfCette relation est importante car si on a une représ<strong>en</strong>tation triviale, cela permet de déterminerl’invariant ”unknot” W R• . Il faut souligner qu’il existe une id<strong>en</strong>tification <strong>en</strong>trel’invariant ”unknot” et la fonction de Schur|R| (−NW R = q 2 sR 1, q, q 2 , · · · q N−1) . (3.86)qui est exactem<strong>en</strong>t la dim<strong>en</strong>sion quantique d’une représ<strong>en</strong>tation R :dim q R ≡ W R (q) = W R (q) = s R (q ρ )De même pour les invariants des <strong>en</strong>trelacs de Hopf W λµ = W µλ <strong>en</strong> termes de fonctions deSchur est sous la forme :W µλ = W µ s λ (q µ+ρ ).69


3.4 Modèle B et espace twistorielIl est possible ainsi d’établir la fonction de partition Z d’un système de probe branes<strong>en</strong>roulant S 2 <strong>en</strong> appliquant la <strong>du</strong>alité de corde ouverte-fermée. Cep<strong>en</strong>dant, on peut calculerexactem<strong>en</strong>t la fonction de partition des configurations de plusieurs boucles de Wilson lelong de l noeuds Γ 1 , · · · Γ L dans S 3 qui est exactem<strong>en</strong>t la fonctionnelle génératrice desinvariants de li<strong>en</strong>sZ (V 1 , · · · V L ) =∑W Γ 1···Γ LR 1···R LR 1···R LL∏i=1T r Ri V i . (3.87)Notons finalem<strong>en</strong>t que ce résultat demande un bon <strong>en</strong>traînem<strong>en</strong>t avec les théories desnoeuds et <strong>en</strong>trelacs dans l’étude des amplitudes des théories des cordes topologiques. Onpourrait simplifier fortem<strong>en</strong>t le calcul des fonctions de partitions d’une manière analogue àcelle utilisée dans la technique de vertex topologique. Cela permet d’évaluer la correspondance<strong>en</strong>tre les fonctions de partitions de la corde topologique sur C 3 et la fonctionnellegénératrice des invariants d’<strong>en</strong>trelace de Hopf pour des représ<strong>en</strong>tations quelconques.3.4 Modèle B et espace twistorielDans cette section, nous prés<strong>en</strong>tons les résumés de nos deux contributions dans lagéométrie twistorielle. En s’intéressant premièrem<strong>en</strong>t aux notions de base de la géométriede l’espace twistoriel [162, 163, 164]. Par la suite nous prés<strong>en</strong>tons la contribution qui traitela <strong>du</strong>alité <strong>en</strong>tre l’espace twistoriel et la théorie de super-gravité N = 4 à 4 dim<strong>en</strong>sions [165].Cette <strong>du</strong>alité est motivée par la suggestion de Witt<strong>en</strong>, qui a conjecturé les contributions<strong>du</strong> modèle B dans le super-espace twistoriel CP 3|4 et les amplitudes de diffusion de lathéorie de super Yang-Mills N = 4, D = 4 [166]. Finalem<strong>en</strong>t, la dernière contribution estconsacrée à notre modèle twistoriel purem<strong>en</strong>t fermionique et sa relation avec l’espace cibleà 4 dim<strong>en</strong>sions.3.4.1 Espace TwistorielEspace des twisteurs de Roger P<strong>en</strong>rose intro<strong>du</strong>it <strong>en</strong> 1970 a retrouvé un grand intérêtrécemm<strong>en</strong>t dans le cadre de l’étude de la théorie de Yang-Mills et la théorie de supergravitévia la théorie des cordes. L’espace twisoriel est un espace de vecteurs à quatre (complexe)dim<strong>en</strong>sions, défini comme étant un espace de spineurs où les variables considérées sont desdoublets de spineurs de Weyl Z α = (λ a , µȧ), qui se transform<strong>en</strong>t par une représ<strong>en</strong>tation70


Théories de cordes topologiques, invariants et D-branede SU(2, 2) à quatre dim<strong>en</strong>sions comme suit :Z α = (Z 0 , Z 1 , Z 2 , Z 3 ) ←→ (λ a , µȧ) (3.88)Le complexe conjugué de Z est défini par :¯Z α = ( ¯Z 0 , ¯Z 1 , ¯Z 2 , ¯Z 3 ) ←→ (µ a , λȧ) (3.89)SpineursA quatre dim<strong>en</strong>sions, le groupe de Lor<strong>en</strong>tz complexifié est isomorphe à :SO(3, 1, C) ∼ = SL(2, C) × SL(2, C) (3.90)De plus, ses représ<strong>en</strong>tations <strong>en</strong> dim<strong>en</strong>sions finies sont classifiées comme (p, q) avec p et qsont des <strong>en</strong>tiers ou des demi <strong>en</strong>tiers. Les spineurs de chiralités positives et négatives setransform<strong>en</strong>t respectivem<strong>en</strong>t selon les représ<strong>en</strong>tations ( 1 2 , 0) et (0, 1 2 ).Nous écrivons généralem<strong>en</strong>t λ a , a = 1, 2 pour un spineur qui se transforme comme ( 1 2 , 0)et ˜λȧ, ȧ = 1, 2 pour un spineur qui se transforme comme (0, 1 ). Nous pouvons former2l’invariance de Lor<strong>en</strong>tz de deux spineurs de chiralité négative :〈λ, λ ′ 〉 = ɛ ab λ a λ ′b (3.91)avec 〈λ, λ ′ 〉 = − 〈λ ′ , λ〉. En particulier 〈λ, λ ′ 〉 = 0 si λ a = cλ ′a . Le vecteur mom<strong>en</strong>t p µ ,µ = 0, 1, 2, 3 peut être représ<strong>en</strong>té comme un bi-spineur : p µ = σ aȧµ p aȧ où le symbôleσ aȧµ apparaît comme le li<strong>en</strong> <strong>en</strong>tre les trois représ<strong>en</strong>tations fondam<strong>en</strong>tales de SO(3, 1). Endéfinissant, le vecteur g<strong>en</strong>re lumière par :p µ p µ = det (p aȧ ) = 0 (3.92)tel que, les bispineurs(peuv<strong>en</strong>t être représ<strong>en</strong>tés par p aȧ = λ a λ ′ ȧ. Les spineurs λ, ˜λ sonttransformés comme λ, ˜λ) (→ tλ, t −1˜λ).Variétés twistoriellesLa géométrie conforme de l’espace de Minkowski complexe CM est représ<strong>en</strong>tée dans l’espacetwistoriel T tel que :(λ a , µȧ) ∈ T ←→ {x aȧ : λ a = ix aȧ µȧ} ⊂ CM (3.93)où le point x aȧ ∈ CM est représ<strong>en</strong>té dans PT(CP 3 ) par un sous espace twistoriel projectifpassant par x aȧ . Ainsi, on peut représ<strong>en</strong>ter la géométrie conforme CM dont l’espace71


H. Jehjouhprojectif complexe -PT définit comme l’<strong>en</strong>semble des classes d’équival<strong>en</strong>ce des élém<strong>en</strong>ts T= C ∗ = {Z 1 , Z 2 , Z 3 , Z 4 }, sous l’action <strong>du</strong> groupe de multiplication C − {0}. L’expressionde l’espace twistoriel est la suivante :PT =TC − {0} = CP3 (3.94)CP 3 est une variété complexe équival<strong>en</strong>te à un espace projectif complexe de dim<strong>en</strong>sions 3,définit comme un espace des lignes complexes passant par l’origine de C ∗ .Super-variétés twistoriellesUne supervariété complexe (X, O) est une variété complexe X (variété bosonique ), <strong>en</strong>richiepar les directions fermioniques sur la variété X. Les plus importants exemples desupervariétés complexes sont les super-espaces projectifs CP (n|m) tel que :X = CP n et O = O(1) ⊕ . . . ⊕ O(1)} {{ }m3.4.2 Contribution : Pure fermionic twistor like modelRésuméhep-th/0605167. (3.95)Motivés par les résultats de l’espace twistoriel de P<strong>en</strong>rose et les propriétés spéciales del’espace de deux temps étudier par Bars, nous avons dérivé puis étudié le modèle twistorielpurem<strong>en</strong>t fermionique dont sa construction est fondée sur la résolution de la propriéténilpot<strong>en</strong>te (Υ m ) 2 de la ligne d’univers de fermions Υ m = Υ m (τ). Nous avons égalem<strong>en</strong>ttiré l’action des champs twistoriels qui décrit les champs fermioniques. Finalem<strong>en</strong>t, nousavons discuté le li<strong>en</strong> <strong>en</strong>tre le modèle twistoriel purem<strong>en</strong>t fermionique et l’espace cibleR (d,4−d) à 4 dim<strong>en</strong>sions.3.4.3 Contribution : Théorie des cordes twistoriellesAfrican Journal of Mathematical physics Volume 4 Number (2007) pages 65-97RésuméCe papier est consacré à l’étude de la théorie des cordes twistorielles. Cette théorie est unecombinaison de la théorie des cordes et les propriétés d’espace twistoriel. Plus particulièrem<strong>en</strong>t,nous sommes intéressés au modèle B dans la super-variété de Calabi-Yau. Cettedernière est un espace twistoriel <strong>en</strong> ajoutant des directions fermioniques, d’où l’espace72


H. JehjouhCP 3|4 . De plus nous avons étudié une autre version de la théorie de corde dans l’espacetwistoriel con<strong>du</strong>isant à la théorie de supergravité conforme N = 4 ; D = 4 et <strong>en</strong> particulierles cordes ouvertes dans l’espace twistoriel. Cette <strong>du</strong>alité est un caractère spécial del’espace de Twistor après utilisation de la correspondance de P<strong>en</strong>rose.73


¢¡¤£¦¥¨§©£¦©¡©¤ !©"¥¨§©#$%'&¥¨§(&*)++,.-/1012¨35464879:;=6:?A@CBDFEHGJI'DLKMDFNPOQERGSKMDFN ?UT IVNWXEHGJI'DZY'Y'DFN[D\W^]_MàDFGSbHGSc/deIfW BDgOQE1h.iEHGJjkDl monqpsr mtnuwv xzyZ{(|}(}¤|"~€ V ‚ƒ…„qv †5‡ˆ8‰Š5‹…ˆŒ~f 5ƒ…Fv „ Ž|’‘|¨~f ' ‚¡¤£¨¢¥¨¤&¦£¤Œ©6§(5 ¥ ¦6¤&\§§(¥¨¤©§¦¤&Z8\šS©6•©"£s£ª€§§8“06“J”Œ©•—–X˜¢š›8{(|, ‡'ß%, ‡}*‘‡=&;: ‡}@*»5àV‡#AŸß(‡#6*…}v&;:5‡‹54Y,‡}@*…+àV‡#AŸß(‡#6*…}\}¦Š+|½‡©"ß/‘‡}/Š Þ ‰‡Vßf}/‘Q, ‡'{(|1>P, ‡}v&;:|G"»à'$*»5àV|½ß%, ‡é‘‡\&;: '&G()2 ‡ Þ {¤‡é}0*|1"Š5{(|½‡©&G&‡ ‘‡ {%, ‡$&|’}¦‡{ÂàV‡VßeŠ Þ ‰‡à€ß(|]9R ‘Š'"%"-, ‡="…5|}(}¤'"…àV‡^2 }¦Š/"ˇ4©ß(‡#"…}¤|½Š/"uB:‡‡1'&G()2 ‡ Þ {¤‡Hß5‡`Z‹‡S&;: ‡}0*…5à'‡A>ߤ‡#6*…}*‹…|}Z‹‡S&½‡$*)*»‡©&,‹ß¤|1&½|’}¦‡!‘|{¤‡à€ß(‡#6 ‡#"ß/}¦Š+|Çß-*»Š+‹{/àVŠ/"…}ß({¤‹…|½{(‡FŠ5‹ Þ |‡#"E*ŒŠ5‹…{›{(‡9¨Š+{06R‹%&‡'{P*)&½‹…}¤|‡'‹{f}\5}0*»‡à€ß(}àVŠ/"…àV‡©*ß›‘‡}ß


‘B’ “Y”R•£–b——£–P˜#-š$›œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹8‹%4 {%, ‡à'‡#"ß(}*}¤‡M{%, ‡}¤‹%6 ‡.‘…$"…}E&½‡c9¨5|Çß[Z+‹Â: |1&}Š'"ß`*…5{@9¨5|Çß(‡#6 ‡#"ßÂ{%, ‡'‹…}(}¤|R2 àVŠ/63Au|1"‡'{ä‡#"ߤ{(‡&’*{(|àfˆ‡}¤}¤‡q‘‡}/ߤˆ-, ‡'Š+{¤|‡}F‘‡}Zà'Š5{f‘‡}F‡Vß &‡}ä}ß({¤‹»à€ß¤‹…{¤‡}e6*8ߤˆ-, ‡#6*8ß(|1Z‹‡}Z‘‡R& ß(ˆ-, ‡Š5{(|½‡!ß5‡5‹ß%, ‡!à'Š'"…}¦ß¤|½ß¤‹»&†:Q, ‡©"‡#6 ‡©"+ße*{(|1"…àV|1*…$&{(‡#6*{ŸZ‹…‡&½‡}›ß¤ˆ-, ‡'Š5{(|’àV|‡#"…}\àV‡}›ß¤{(Š5|’}/‘‡'{"|†2 ‡'{(‡}›$")"N, ‡'‡}'vZ‹|¢1Šà'àV‹)*N,‡'Š+{¤|‡›‘‡}Jà'Š5{f‘‡}ƒ$&’­àVŠ'"»}ß({¤‹…àVߤ|Š'" ‘‡FàV‡©&G&‡AªàV|)"N, ‡à'‡}(}¦|½ß¤‡N&;: |1"a>65{¤|’$"»àV‡/àVŠ/"59¨Š5{6 ‡yä'"…}ž&‡FàVŠ/"+ß(‡4©ß¤‡/‘‡e&’¦ß¤ˆ-,‡}@*…+àV‡Zà'| Þ &½‡+vBÇä{ƒ'&;: ‡}0*…5à'‡/ß©‹Z‹|…‡}ßN(+5{('"+ß(|½‡-*…5{P&’qàVŠ/"…‘|½ß¤|Š'"‘‡Æ/|’à'àV|GA³*%&ߤ‡Z‘‡&;:‡Vß%, ‡5v Ý g­"1‘‡›}¦‹…{06 Š'"ß(‡'{¡àV‡P*…{¤Š Þ &†2 ‡#6 ‡5ƒ8&;: ‡}@*…+àV‡wß5‡¦‡4ˆ| Þ $"ßj&’Rߤˆ-, ‡Š5{(|½‡!‘‡}FàVŠ+{(‘‡}\Š5‹%>+‡'{¤ß¤‡}!S{f$&1&†2‡‡Â}¦‹…{R&;: ‹ß(|G&|’}¤ß¤|Š'" ‘‡}>8{(|’ Þ &‡}!ß5‡{¦ß(‡ ߃)Z‹‡©&1Z‹‡­}¤Š5|½ßN&’Rߤˆ-, ‡Š5{(|½‡¦‘l: Š5{(|G(+|G"‡!àVŠ/"…}¤|‘N, ‡'{%, ‡'‡+ƒ'&‡}›{%, ‡}¤‹%&Çßf8ßf}B9¨Š+‹{0"…|}/}¤‡'{(Š'"ß\|‘‡©"ߤ|bZ+‹…‡}vBU‡‘'"…}-ÈVɇ#(+'&½‡©6 ‡#"ßZ‹%"Ì"Š5‹)>5‡5‹ X4‡1‘‡R{(‡àfˆ…‡'{fàfˆ‡*‹|’}Z+‹…‡h2?&†: ‡#"»àVŠ'"ß({¤‡1‘‡}!àVŠ+{(‘‡}FŠ+{(‘|G"…5|½{(‡}äŠ2 ‹ }¦‡‹%&½‡*…'*|½‡{äŠ'CŒ{(‡L,àVŠ+{(‘‡S9¨‡'{6K, ‡'‡`*…{¤Š‘‹|½ß&éߤˆ-, ‡Š5{(|½‡‘‡E(+{(8>©|½ß%, ‡+ƒX‘'"…} ¤ à$: ‡}¦ß&‡}¤‡àVߤ‡'‹…{1‘‹¯}¦|1"%(+‹%&½‡'ß1‘‡R‰¦‹)(5‡‘‡E&’ àVŠ+{(‘‡&’‡'ß(ß(}ž*âW}¤|1Z‹‡}S}¦‡©6 Þ &’ Þ &‡} 2­àV‡'‹%41‘‡e&’­}¦‹)*»‡{0(+{(8>©|½ß%, ‡›àVŠ'"%9¨Š5{61‡BÎØÏÕFE2Š5‹)>5‡'{¤ß¤‡›ß©|½ß%, ‡+ƒ5‡VßZ‹…8ß({¤‡¦‘|161‡©"…}¤|½Š/"…}'vSP"}¤‡!}¤‡'{>8$"ß›‘‡&’1‘‹…$&|½ß%,‡}@*»5àV‡Ù-_/|’}¦ß¤Š5{ƒ…à'‡Vߦß(‡H‘‹…'&½|½ß%, ‡¦‡}ßä‹%"‡H}0*-, ‡àV|Gg»à'ߤ|Š'".‘‡+&†: ‡}0*…+àV‡q‘‡RÙ-_/|’}¦ß¤Š5{‡#"h*…5{¦ß(|à'‹%&|½‡{j&½‡}FàVŠ+{(‘‡}F‘'"…}e&;:‡}w‹ß¤|1&½|’}¤ß¤|Š'"M‘‡&’1à'Š5{({¤‡}@*ŒŠ'"»‘$"…à'‡¦‘‡!¡‡#"{(Š+}¤‡5v$*…{©2‡Vß(‡¦à'Š'"…à'‡'{"…$"ßQ&’Hߤˆ-, ‡Š5{(|½‡Zß8{(|, ‡Vß%, ‡}X‘‹…$&‡}}¤‹%*»‡{0(ž,|G&1&½‡}w{f$*)*{¤Šàfˆ-, ‡'‡}O*»{P&’q}0Wa6K, ‡Vߤ{(|‡e61|{(Š5|{vBU‡#*Œ‡#"…‘…$"߃¡&àVŠ/"…}¦ß¤|½ß¤‹…'"+ß›|1"…}¦|…ߤŠ+‹ß¤‡}P&‡}‘‡'‹54‘‡}P*»|{¤‡}J‘‡j9¨'6‡'‡&½‡}P>8{(|, ‡Vß%, ‡}S‘‡+U›$&’ Þ |GA;Êw‹*{(|G(+|‘‡}B*{), ‡}¦‡©"+ß(‡/‹%"‡äŠ Þ }¦ß¤{(‹…à€ß(|½Š/"¢v¡u­{%, ‡}¤Š'&‹ß(|½Š/" ‘‡äàV‡'ߦß(‡c, ‡©"|G(/6 ‡à&’5}(}¤‡e"Š'66K,‡'‡Z‘…$"…} ~Q, ‡ Þ ‡‹»àVŠ5‹)*}¤‹%*»‡{0>85{¤|,8ߤߤ‡#"ß(|½Š/"úߦß({¤| Þ ‹-, ‡‡=2h&†:5Š/"…}X‰‹%(ž, ‡qàVŠ'"a>+‡#"… Þ &½‡­‘‡­{), ‡©*…{¤ß¤|{FàV‡+*…'*|‡'{F‡©"yä'"…}e&;:‡Â}¤‹|½ß©fH}ß(|G6¥, ‡‡39¨Š/"…‘'61‡©"ß($&‡*»Š+‹{&’.àVŠ'6*{), ‡ˆ‡#"…}¤|Š'"Z‹…8ß({¤‡*}¤‡àVߤ|Š'"…}6*8‰‡‹{(‡}+Z‹‡S"Š+‹…}¦ß¤{f|½ß¤Š/"…}­àVŠ/66'


¢ Ï ¢ £ ‡Vß £¢+Ï¢ »~‡‘B’ “Y”R•£–b——£–P˜#-š$›œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹8$"ߤ‡}'ƒk&’S*{(‡#6|G"ß({¤Š‘‹|{(‡&½‡}/ß5‡©61‡©"ßP&†: ‡4©ß(‡#"…}¤|½Š/"S9¨‡'{6 |½Š/"|bZ+‹…‡ä‘‡&R}@Wa6¥, ‡'ߤ{(|½‡j6 |{¤Š+|½{›‘‡}\}¤‹%*Œ‡'{>65{¤|#, ‡'ß%, ‡}S‘‡ÚÛ ‚0Á$¼$‚½ Z‹|}¤‡à€ß(|½Š/"*ߤ{f|½ß¤‡Þ &½‡}J‘‡e&’Zß(ˆ-, ‡Š5{(|½‡/‘‡}SàVŠ5{f‘‡}Xß5Š/"…}e*{),‡Vß(‹…‘‡‘‡?&’ }¤‹%*Œ‡'{(5{f8>©|Çß%, ‡ÂàVŠ/"59¨Š5{6 ‡3ÎÝÏÞFTßØÏàFa‘$"…}&‡e}¤‹%*Œ‡'{(‡}0*…5à'‡Â‡#"¯‘N, ‡'ߤ‡'{6 |G"»$"ß1}¤Š'" }0*Œ‡à€ß({¤‡2=&;:5‡{¦ß(‡Âß|½ß%, ‡\à'Š'"59¨Š+{06 ‡+viãX|1"…$&‡#6 ‡#"߃¡"…Š5‹…}wߤ‡{06 |1"Š'"…}8>5‡à-&½‡¦}@*Œ‡àVߤ{(‡ä‘‡}S,*…5{›‹)"‡qàVŠ/"…à&‹…}¤|½Š/"¢v£Âäeå­æVçR äéè+Š5Z+‹…Š'"…}e&‡}(5{f$"…‘‡}e&|G(/"‡}ƒ»à'Š'"…à'‡'{"…$"ßj&;: ‡}@*»5àV‡qß+‡#"ßH]üÂøý$M€ƒŠ2 ‹3üM‡'ßýH}¤Š'"ßF‘‡}e"…Š'6 Þ {¤‡}‡©"ߤ|‡'{f}\Š5‹.‘‡#6 |GA>‡©"+ß(|½‡{(}v}©:+‡àVߤŠ5{(|‡#&1&½‡/‘‡ õPö Hd÷%ø©Ô'øÈOMVƒŒ'&½Š+{(}O&‡j>+‡à€ß(‡'‹{!S{S|1&1&½‡‹{(}ƒž,Š'6 ‡#"ßeü+/Œƒ10Ï ÿ)ø#2#22tø0÷.‡}¦ßq‘N, ‡àV{(|Çß+*…5{¦‹)" Þ |]Aª}@*…|G"‡‹{eü ¢ ¨ ¢ ‘l: |1"…‘|’àV‡}H}@*|1"Š+{¤|‡#&’}*»Š+|G"ß%, ‡3 ¤é‡Vß"Š/"Ì*ŒŠ5|1"ß%, ‡4¤6‡+vXu‡*…+}¤}($(+‡Z‘‡Pü£/=2jü ¢ ¨ ¢ }¦‡ 9¨|½ß‡©"‹ß(|G&|’}¤'"+ßN&3*…5{¦ß(|½‡Zàfˆ…|½{f$&‡ä‘‡}P6*ߤ{(|à'‡}›‘‡äyä|½{f5à ~


;&Šà'} "Š'"\"‹)&}¦‘‡ /7 9¨Š5{6 ‡#"ßä{(‡}0*»‡à€ß(|G>+‡#6 ‡#"ß䑇‹54=6*8ß({¤|’àV‡} úû\úEZ+‹…|X}©:5‡{(}›‹%"…‡q‹ß({¤‡+v¢&†: |1"…}ßf{Z‘‡!à'‡Z‹|l*{%, ‡#&‹…‘‡+ƒXü / ‡}ß/{(‡#&|, ‡S2üB¢ ¨ ¢+>©|’3&’9¨Š5{6H‹)&½‡­}¤‹|G>8'"+ß(‡'fÝ/¨ ¢ ü+/ ¢¢ ¨ ¢ Ï": ü IC? :ž»D? ÏLü ‡'ßE? ü.}¤Š'"ße&‡}-*…5{¦ß(|½‡}›ß¤‡#6*ŒŠ5{(‡#&1&½‡­‡'ßF}@*…ߤ|’$&‡!{¤‡}@*Œ‡àVߤ|1>5‡#6 ‡©"+ßF‘‡eü / vߤ‡©&Z‹‡eü¨ ¢¦‡}¦ß\‹%"‡6*ߤ{(|à'‡úû`úq‘‡ä{f$"%(GFIHú5ø$&Š5{f}w}¤Š'"‡#45*{¤‡}¤}¤|Š'"*‡#"Âߤ‡'{6 ‡}›‘‡}\}0*|1"‡'‹{f}¡ ¢ƒB"Š5‹ß({¤‡+ƒ/*‹…|}Z‹‡QüB¢0 ‡}ßZ‘‡&’39¨Š5{61‡/fâÀ&¢}©: ‡©"…}¦‹…|ÇßZ'&½Š+{(}N&;: |’‘‡#"ß(|]g»à8ß(|½Š/"/ ü / ÏDÓ@LMH]ü£¢ ¨ ¢XM üH†úv ú'M‡3Z‹‡&½‡>5‡à€ß(‡'‹{Nü+/.‡}ß­‘‡(5‡#"…{¤‡#A;&‹%6 |†2 ‡'{(‡q}¦|¡‡Vß­}¤‡'‹%&‡#6 ‡#"ß!}¦|O&‡1‘N, ‡Vߤ‡{06 |1"…$"ß䑇3&’E6*ߤ{(|à'‡Z‹|i"Š5‹…}¦|1"59¨Š5{6¢ ¨ ¢ àVŠ+{¤{(‡}0*ŒŠ'"…‘'"ߤ‡¦‡}¦ß-"©‹%&>vJy¦$"…}Fà'‡­à'5}ƒ5&;:5‡à€ß¤‡‹{f}Z‹‡#&’àVŠ/")Z‹‡}v!S{Z}¤‹{fàV{(Š87P½ßƒŒ}¦|i* ‡'ßZ.}¤Š'"ߦ‘‡3(5‡©"{¤‡#A;&‹%6 |†2 ‡'{(‡5ƒŒ‹ß({¤‡©61‡©"ßߤ‡©&iZ‹‡3*¢ ¨ ¢ ÏQ ¢+¦©¨ ¢ ø)ý ¨‡à'{¤|1>5‡©"+ßžü0 ¨‘|½ßZ|1&}F}©:8$"ß(‡'f 9¨SR‡}¤‹%&½ß(ß(}S}¤Š'"ßJ‡4ˆ| Þ , ‡} Þ {¤|;2 ‡©>5‡#6 ‡©"+ßž*ŒŠ5‹{P&½‡}w‹ß¤{(‡}J}¤|G(/"…8ß(‹{(‡}S‘‡e&;: ‡}@*»5àV‡+vXP"ÂàV‡jZ‹|yä'"…}JàV‡jZ‹|Œ}¦‹…|Ç߃à'‡}S{%,õPö H†ú5øŸú'MiAª}($"…}O9¨5|½{(‡/{(‡àVŠ+‹{f}2­‹%"‡äàVŠ'6*%&‡4|Gg»à'ߤ|Š'"5AàVŠ/"…àV‡{0"…‡e&q}¤|1('"…ߤ‹{(‡H;ITI£J+JM£./&½‡(5{(Š5‹%*Œ‡F‘‡ZuŠ5{(‡#"ß0ô‡#"ß\|}¤Š'6 Š5{*ˆ‡S2 õžù H;ú5øTOM û õžù H;ú5øUTžMVvFyZ‡qà'‡ 9¨5|Ç߃5&½‡}/}@*…|G"‡‹{f}¡ ¢ ‡Vß4¦¨ ¢ }¤Š'"ßF‘‡}\Š Þ ‰‡Vßf}2 ‘‡‹54‡}¦ße&½Šà$&‡#6‡‡#&’}'ƒ©|1"…‘N, ‡©*»‡©"…‘'"+ßf}'vàVŠ/6*»Š}¤'"+ß(‡}ƒ©{%,‹…$"ß`2E&’e}¦|1('"…ߤ‹{(‡H‡‹…à&|’‘|½‡©"%"‡cH;IDIÕI+IMFŠ2 ‹\&‡(5{(Š5‹%*Œ‡R‘‡*uŠ5{(‡#"ß0ô1‡}¦ß&Šà''&½‡©61‡©"ߦ|’}¦Š/6 Š5{*ˆ‡c2ÖH;ú'MQû õWV H;ú'M£.5&‡}\{(‡#*…{%, ‡}¦‡©"ß(8ß(|½Š/"…}›‘‡}/}0*|1"‡'‹{f}/}¤Š'"ße*…}¦‡‹…‘Š'A>{%, ‡'‡©&G&‡}võWVXR$Y+îb¢Ÿ /ð­ñ ¢&R6[7 ‡#6 ‡j6*$"…|†2 ‡{¤‡?, ‡'ß('&, ‡‡e*…{N!¡‡#"{(Š+}¤‡F‘'"…} ' ¿ Z vSP"yä'"…}wà'‡Vߦß(‡*…5{¦ß(|½‡+ƒ/"…Š5‹…}|G"ß({¤Š‘‹|’}¤Š'"…}ž&‡}wß©|}¦ß¤‡+ƒ©ß¤Š5‹ßZ}@W}¦ß©2‡*…{/‹%"hF'A;>+‡à€ß(‡'‹{ƒ‡}¤‡#"ß%,[ ‹)"[6 Š/61‡©"ß/ߤŠßf$&]\ / {¤‡©*{%,¥^


;_q¢¨œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡à€ß(‡'‹{f}/‘‡+!S‹%&|GA—u‹ Þ $"»}@á©|õ /`f Ï J/,acedf\ a _ cd »ÔúBb9¨Š'"…‘…$6 ‡#"ß('&$Š2 ‹T&’?6 +}¤}¤‡H‡}ß­‹ {¤‡©*»Š}Hg\ a \ a Ïy¦$"…}&‡1à5}Þ $"»}@á©|$‘Š5|½ß7 ‡'ߤ{(‡R‹)"ÿ/M.u¢‡>5‡à€ß¤‡‹{+!S‹%&|GA—u‹6H‹)&Çß(|G*%&‡q‘‹hF$A³6 Š'6 ‡#"߃+ß(‡#&ÂZ‹‡_ cd \hd i\ cj õ /ÏJk\1/ H†úv M‡#&|à'|Çß%, ‡}.k.àVŠ'"»}ßf$"ߤ‡}'v Ö ‹…$"»‘L&½‡?6 Š©‘‹%&½‡.‘‹¯}@*|1"Ol kml$‡}¦ß"Š/"¾"©‹%&;.›'&½Š+{(}R&\*…{¤ß¤|’àV‹%&‡e‡}¦ß3"Š/"*ŒŠ5‹{*‘‡}Hˆ-,‡‡5v$u‡+>5‡à€ß¤‡‹{$\ / à'Š5{({¤‡}@*ŒŠ'"»‘T21‹%"M}@*…|G"‡‹{n©¨ ¢ ‡Vßj*…{¤‡©"…‘[&39¨Š5{6 ‡f&Š©à$&|}%,/4o ; nB¢pn ¨ ¢ H†úv M\/,a ‡}¦ß\{(‡#*…{%, ‡}¦‡©"ß%, ‡j*»{ß('"…‘|’}eZ‹‡G_/,a o 0 ¢# ¨ ¢ ¨ IK ¢; 0 ¨ ¢ ¨ _H†úv 5 M¢ ‡}ßF‹%"M}0*|1"‡'‹{F}0Wa6K, ‡Vß({¤|bZ‹‡5vO!S{/àVŠ/"…}5, ‡©Z‹‡©"+ߧ.a&†:u ¢ vnyZ‡¾*)&½‹…}ƒ*‹…|}Z‹‡ ߤŠ+‹ß }0*|G"…‡'‹{‡'ߤ{(|1Z‹‡`2EúÂ|G"»‘|à'‡}Z‡}¦ß&½‡3*{(Š©‘‹|Çߦ‡4©ß%, ‡'{(|½‡‹{F‘‡úÂ}@*|1"‡‹{(}ƒ»'&½Š+{(}j"Š5‹…}ä‘N, ‡‘‹|’}¦Š/"…}jZ‹‡&;: ‹)" ‘‡}9¨5àVߤ‡‹{(}}0W¡6¥,‡Vß({¤‡ ; n£¢a» Ý |G"»}¦|>ƒ‘‡z0`¢/‘Š5|½ß+7; uH†úv ~/M¨ ¢ ¨ Ï{s ; n ¢¢ u yx ¨ ¢ ¨ J|s}nwyx ¨ ¢ » ¨ ¢ w‡#"߃3&D*…|{(‡DHg\ / ø_ /,a Mé}¤‡ à'5{(+à€ß%, ‡'{(|}¤‡£*…{Ì&‡}a‘‡'‹54 }0*|1"‡'‹{f}KH*u ¢ øDn ¨ ¢$M€vÐU‡Vߤߤ‡K*»|{¤‡ ‡}¦ß ‹%"‡ãX|G"»$&‡#6‡}¤‡#"ßf8ߤ|Š'"‘‹ß


&;: ‡}@*…+àV‡z† €b¨¨¨¨¢¨¨b¨œb¨ ¢ œœ ¨¨¢ ¨»‡œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡à€ß¤Š+{¤|‡#&¡‘Š/"+ß&;: +‘‘|½ß¤|Š'" 5|G"»}¦|žZ‹‡&?6H‹%&½ß¤|1*%&|à8ß(|½Š/"\*…5{¦‹)"&|G"-,‡#g)"|‡}J‘‡&+6*'"|†2 ‡'{(‡/ˆ… Þ |½ß¤‹…‡#&1&½‡ Á vPUŠ'66 ‡j"Š+‹…}B&‡ZàVŠ/"…}¦ß(8ß(‡'{(Š'"…}w‘$"…} H†úvGÔ/Ô§M€v}(à'$&’|{(‡\‘‡ZàV‡}w}¤Š'&‹ß¤|Š'"…}}¦Š/"ß‘N,yä'"…}3"Šß({¤‡à5}1àV‡}R}¤Š'&‹ß¤|Š'"»}Au ¢ 9¨Š5{6 ‡#"ßR‹)" ‡}0*…5à'‡E>+‡à€ß(Š5{(|½‡©&‘‡?F ‘|16 ‡#"…}¤|Š'"…}1àVŠ/6*)&½‡#4©‡}H'*%*»‡©&,‡}@*…+àV‡­ß5‡à€ß(‡'‹{f}'v*{%,}ß({¤‹…àVߤ‹{(‡&½|1"-, ‡|{¤‡!}¤‹|G>8'"+ß(‡'fyä'"…}/àV‡'ßF‡}0*…5à'‡¦ßƒa&½‡}\ß


"©‹%&$}¦|`kÏÕÿ [‘{¤Š+|Çß(|½‡{F}¦|`k§^ÿ[(‹…àfˆ…‡'{/}¦|hkz¨^ÿ…v[{¤‡©*{%, ‡}¤‡#"ߤ‡*‹)"‡E&|G(/"‡‘l: ‹%"|1>5‡{(}R"©‹%&G&‡E*ŒŠ5‹…{4kcÏyä'"…}Hà'‡àVŠ/"ߤ‡4©ß(‡5ƒW€b¨Js¨ bnbÏœžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹©|’‘‡#"ßeZ+‹…‡!à'‡Vߤߤ‡9¨Š5{6 ‡ä‡}ß/{%, ‡'‡©&G&‡¦‘‡¦}¤|G(/"…8ß(‹{(‡H³I¥IÕJ+JMNZ+‹…Š5|bZ+‹Â: ‡©&G&‡ *Œ‡'‹ß7 ‡Vߤ{(‡*»Š}¦|½ß¤|1>5‡!Š5‹?"-, ‡©(+8ß(|G>+‡âÀ&¤‡}¦ß?,ˆ-, ‡©&½|’àV|½ß%, ‡Ak/»u¢‡!ß8'&½‡©"…àV‡‘‡}Rß


à''"ß-Û.€Ḧ Š+‹Ë€­MP*»{§fU‡}/à'Š'"…‘|Çß(|½Š/"…}/}¦Š/"ß/}(8ß(|}@9¨|½ß¤‡}\‡#".{(‡#6*%&’SRϵf € I-ÌÂÛ € €¿ J ½ ÿ“¾ È­ Èbqq ¨8¢ ¨¢¨Ô½ ÿ“¾ J­ Ô‡œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡à€ß(‡'‹{*ŒŠ+}¤|Çß(|½Š/"a‘‹£*ŒŠ5|1"ßA\ v yä‡ÂàV‡Vߤߤ‡*$"»$&1W©}¤‡ ‘Q,‹š`²E–„¡£¹$–d Ÿ˜£—€vWÑZŠß¤Š/"…}eZ‹‡&½‡¦ß+‡­àVŠ'6*%&‡4‡8M\‘$"…}-&;: ‡}0*…5à'‡!߃‹%"…‡!à'Š5‹{ Þ ‡Z‡}¦ßF‘|½ß¤‡­$&1(ž, ‡ Þ {(|1Z‹‡j&Š5{f}Z+‹Â: ‡©&G&‡*ŒŠ+}(}#2 ‡‘‡ä‹%"‡=, ‡©Z‹…ߤ|Š'"à{¤ß%, ‡}¦|‡#"%"…‡*ŒŠ'&1W¡"…Š'6 |]AB"?(ž,‡'‡©&}vZy¦$"…} &‡Rà'+}FŠ2 ‹é‡#&1&½‡1‡}¦ß "…Š'" '&G(ž, ‡ Þ {(|bZ+‹…‡5ƒ‡©&G&‡R‡}ß!‘|½ß¤‡qß({('"…}(àV‡#"»‘$"ߤ‡+väu‡R‘‡©(5{%, ‡`Ḧ Š+‹$&‡?2àVŠ©‡×ÂàV|‡#"ßf}Z{%,Š+{(‘{(‡§MÓ ‘l: ‹)"‡àVŠ+‹{ Þ ‡*$&1(ž, ‡ Þ {(|1Z‹‡*%&’$"‡‘l:


wÜÛ ¼¢~ x wÖÛ ¼Œ x ‘|16 |G"©‹-,~ » Ó » Ó3ÏDÓ|…àV‡Fß©|‘‡©"ßX‘l: |G6*ŒŠ+}¤‡'{S‘‡‹54RàVŠ/"…‘|½ß¤|Š'"…}÷Zà'Š'6*%&‡4‡}v‡§Z+‹…‡#"߃‹%"‡&|G(/"‡­‘$"…}N&;: ‡}@*»5àV‡äßP, ‡'{(|]g…‡©"ßߤ‡©&Z‹‡¤Â‡'ß$ã¦}¤Š'"ßF‘‡}-6*ߤ{(|à'‡}júûcúR‘Š'"ßj&½‡}E,Þ ÏéHg¤ ~øeã ~¨ ¢ M ‡Vß × Þ Ï Hg¤ ¢ øã ¨ ¢ M¢‡{¤‡-Z‹‡jÓ6LMr¤R‡VßPÓ6LMäã›}¦Š/"ßXß(Š5‹ßQ&½‡}S‘‡'‹54"Š/""©‹%&’}¡Š5‹*5‹6 Š+|G"…}B&†: ‹%"*‘l: ‡'‹54 ‡}¦ßB"©‹%&Ÿv¡yä‡RH†úvGÔ§÷/M€ƒ5|G&‘‡\ß(‡#&1&½‡e6*$"…|†25‹?6 Š+|G"…}-&;: ‹%"e‘‡}/‘‡‹54‘N, ‡'ߤ‡{06 |1"…$"ß(}\}¤Š5|½ße"Š'"["©‹%&¢Xg)".‘‡!{%, ‡‘‹…|½{(‡&;: ‡}@*»5àV‡S21‹%"‡&|G(/"‡5v*…5{(5|Çß/}¦‹f×*}($"ßeZ‹:‹cÓ@LM©ãGª ÏDÿ%.'&½Š+{(}ãZ‡}¦ßF|G"a>5‡{(}¤| Þ &½‡¦‡Vßj&’ }¦Š/&½‹ß¤|Š'".‘‡`H†úvGÔ©F¡M‡}¦ßZ‘Š'")"N, ‡'‡*…5{y¦$"…}-&‡qà'+}Š2¨ ¢ Ï JHvã ¼¢~ ¤5M ¢ ¨ ¢ ¢0Ø ¨ ¢#¢& ¢ » å‡#6 ‡*{(Š©àt, ‡‘‹{¤‡ä}#: ‡C¤‡àVߤ‹‡+*ŒŠ5‹{-&‡­à'5}-Ó6LMW¤§ª ÏDÿ%»B".àVŠ/"…à&‹…}¤|½Š/"¢ƒa*»Š+‹{/àfˆ…'Z‹‡¢Ø£¢ ¨ ¢q{%, ‡'‡©&†ø¡&’R*…5|½{(‡!‘‡}U‡Vߤߤ‡6[7‡©Z‹…ߤ|Š'"…}¨ ¢eÎ ØB¢ ¨ ¢& ¢ ÏDÿ%. 0¤Ï Ô'øŸúH;úv1Ô M‡#g)"|½ßH‹%"‡àVŠ5‹…{ Þ ‡$&1(ž, ‡ Þ {(|1Z‹‡ {%, ‡'‡©&G&‡EœA‘'"…}4TVÉ ‚ .$Š5‹ Þ |‡#" ‘…$"…}3ÈÂÉ ‚ }¤|P&‡}>85{¤|’ Þ &‡}H}¦Š/"+ß1àVŠ'6*%&‡4|Gge, ‡‡}v‘N,Þ ‡.‡}¦ß*‹%"‡M|1"ߤ‡'{f}¤‡à€ß(|½Š/" à'Š'6*%&†2 ‡Vß(‡[*…‹|}Z‹: ‡#&1&½‡h*…5{0>©|‡#"ß ‘: ‹%"…‡c*…|{¤‡.‘‡c*ŒŠ'&1Wa"V7 Š'6 ‡} ˆ…Š'6 Š'()2 ‡#"‡}U‡Vߤߤ‡MàVŠ+‹{‡|{(‡}›àfˆ…5à'‹%"e‘‡q‘‡#(+{%, ‡3Ồ Ḧ à¡25‘lfžÓ ~ ÏÕÓ Ï Ô§M€vSyä‡+*%&‹…}'ƒ)&½‡+(+‡#"{(‡­‘‡Rœz‡}¦ß-"©‹%&>v&|G"-,¢ ‡}¦ß ‹)"‡[9¨Š'"…àVߤ|Š'" ‘‡© ¢ » Ý |G"…}¤|>ƒQ&‡} ¢|1&G&‡'‹…{(}ƒB&‡}¾, ‡©Z‹…ߤ|Š'"…}hH;úv1Ô MR}¤‡e{), ‡}¦Š+‹…‘{(‡#"ß3Z‹…$"»‘-0¨!S{‡'‡}FˆŠ'6 Š'()2 ‡#"…‡}e*ŒŠ5‹{Rœ Z+‹…|¡‘‡#>©|½‡©"ß!'&½Š+{(}Z‹%"‡ àVŠ'*…|½‡ ‘‡4TVÉ ~ Š+‹TÈiÉ ~ }¦‡©&½Š/" }¤|i&‡} >8{(|]A}¤‡'{>5‡#"ß䑇1à'ŠŠ+{(‘Š'"%"-,&½‡}}¤Š'"ß\{%, ‡‡#&1&½‡}JŠ+‹àVŠ/6*%&½‡#4‡}ƒ5ßf$"…‘|’}-Z‹‡Ø ¢ ¨ ¢ ‘N, ‡Vß(‡'{61|1"‡©"+ßN&½‡*%&’5à'‡#6 ‡#"ß›‡Vße&†: Š5{(|‡#"ß(ߤ|Š'"‘‡&’R&|G(/"‡!‘$"…}Þ‡}@*…+àV‡äßv Ý |1"…}¤|Ÿƒ%&‡}$Ø£¢ ¨ ¢9¨Š5{6 ‡#"ße&½‡}e6 Š©‘‹%&½‡}/‘‡+&’3&|1('"‡+v&;:‡‡#&1&½‡¦‘‡&†: ‡}0*…+àV‡äß


ÏJïEH5‡à ‘|GC, ‡'{(‡#"ߤ‡}‡#4a*…{¤‡}¤}¤|½Š/"…}vu¢‡Ì6Çä{Mà': ‡}ßh&‡#g)"|½ß¤|Š'"Ë‘‡£Æ/Š/(5‡'{0AªyZ‡#A;Å |½ß¦ß?Z‹|"Š+‹…}|16*»Š+{¦ß(‡ ‘'"…}?"Šß({¤‡a$"…'&GW}¤‡5vUy¦$"»}à'‡ à'+‘{(‡5ƒ/‹%"‡ú}¤‹%*»‡{0>85{¤|, ‡Vß%,‘N,‡g­"|½‡¦àVŠ/66 ‡=, ‡Vßf$"ß\‹%"‡¦‡4©ß¤‡©"…}¦|Š'"?"…8ß(‹{(‡#&1&½‡+ƒ©‘…$"…}¡Trí Ë Û Š5‹[Èrí Ë Û ø©‘‡&’1‘Q, ‡g­"|Çß(|½Š/"‹…}¤‹‡#&1&‡!‘‡}N>8{(|, ‡'ß%, ‡}‡}¦ß/‘Q,Š5{f‘|1"…|{(‡}v‡#"-, ‡'{f$&>ƒ€‹%"‡}¤‹%*Œ‡'{>8{(|, ‡'ß%, ‡Sà'Š'6*%&‡4‡JŠ5‹­{), ‡‡#&1&½‡J‘‡‘|G6 ‡©"…}¦|Š'"?H*îˆl ÚÂMPḦ Š5‹q‘‡}¤‹%*»‡{(‘|G6 ‡#"»}¦|Š'"`HgîLJÚÂM@MB"+(ž,}¤‡!à'Š'"…}¦ß¤|½ß¤‹…‡!‘‡q‘‡'‹54[*…{¤ß¤|‡}f|„M-Òj"‡>85{¤|, ‡Vß%, ‡!àVŠ'6*%&‡4‡!Š5{f‘|1"…|{(‡¢ ‘‡H‘|G6 ‡#"»}¦|Š'">î *…5{('6K, ‡Vß({%, ‡‡j*…5{F‹%"‡+*…5{¦ß(|½‡+*…‹{¤‡©6 ‡#"ß Þ Š}¦Š/"|bZ+‹…‡Ḧ‡‡¦5|G"»}¦|l>8{(|, ‡Vß%, ‡ Þ Š+}¤Š'"|bZ‹‡!Š5‹.àVŠ+{0*…}/‘…$"…}-&‡+&'"%(+'(5‡6*8ߤˆ-, ‡#6*8ß(|1Z‹‡+v$*)*»‡©&,Ḧ ||bMQÒj"‡!à&’5}(}¤‡ä‘‡}P9¨Š/"…à€ß(|½Š/"…} õ }¤‹{&½Šà''&½‡©6 ‡#"ß›|’}¤Š'6 Š5{*ˆ‡}w‹%49¨Š/"…à€ß(|½Š/"…}›ˆŠ'&Š'6 Š+{0*ˆ…‡} 23>8$&‡'‹{f}›‘$"…}'&G()2 ‡ Þ {(‡ä‘‡ þ {(+}¤}06*$"%"ߤ‡#&VZ+‹…‡&;:‹ˆ” ‡}ßq‹%"£g Þ {%, ‡S>5‡àVߤŠ+{¤|‡#&SˆŠ/&½Š/6 Š5{*ˆ‡*‘‡Â{f$"%(©õú'*%*»‡©&, ‡&‡Sg Þ {%, ‡1ßf$"%(+‡#"ß+9¨‡'{6 |½Š/"|1Z‹‡‘Š'"ßG”I­*‡}¦ßR}¦Š/"Š2Þ {%, ‡ä‡©"h&½|1('"‡­x›‡'{(‡#ô|G"…|½‡©"[×C‘‡qà'‡}/}¤‹%*Œ‡'{>65{¤|#, ‡'ß%, ‡}J‡}¦ßZ‘N, ‡#g)"|‡*…{‘‹…'&ŸvSu‡g×éÏ ùËö ô ó ”éÏJï`HgîJ‘Ú[J^Ô§M»‡}¦ßq‹ß({¤‡Z+‹Â: ‹%"…‡(ž, ‡#"-, ‡'{f$&|}(8ß(|½Š/"e‘‹=g Þ {),B&1&‡R":‡G 2E&’e}¦|16*%&½‡‡'ß%,‡'{(‡#"…à'‡RZ‹‡Rà'‡Vߦß(‡3(ž, ‡#"-, ‡'{f$&|}(8ß(|½Š/"ß(|½‡©"ß!à'Š'6*ߤ‡1‘‡}!à'ŠŠ+{(‘Š'"%"-, ‡'‡}N9¨‡{06 |Š'"|bZ‹‡}v!yR: $*{§2 ‡} ~@Á ƒl&‡}¦‘|½{(‡àVߤ|Š'"…}‘|GC,|½Š/"|bZ+‹…‡}/àVŠ/"ߤ{(| Þ ‹…‡#"ßF8>5‡àä‹%"M}¤|G(/"‡!Š'*%*ŒŠ+}5, ‡S21à'‡#&‹|¢‘‡}FàVŠ/"+ß({¤| Þ ‹ß(|½Š/"…}/‘‡}/‘|½{(‡àVߤ|Š'"…} Þ Š+}¤Š'"…|1Z‹‡}'v9¨‡'{6‡4‡#6*%&‡£&½‡£*%&‹…}M|16*»Š+{¦ßf$"ß鑇}M}¤‹%*Œ‡'{>8{(|, ‡'ß%,uB:‡} ~@Ü ‘…$"…}cÈhí Ë Û‡}ßc&;: ‡}@*»5àV‡ }¤‹%*Œ‡'{*{¤Š5‰‡àVߤ|G9ÈVÉ í Ë ÛŠ Þ ß¤‡#"©‹[*…{§fí 2Iî I Ôqà'ŠŠ+{(‘Š'"%"-, ‡'‡} Þ Š+}¤Š'"|bZ‹‡}ÏÕÈiÉ2.Úaà'ŠŠ+{(‘`9¨‡'{6 |½Š/"|1Z‹‡}‡‡}ˆŠ'6 Š'()2 ‡#"…‡}›‘‡+ÈVÉ í Ë Ûu¢‡}/àVŠ©Š5{f‘Š/"%"-,‡'‡}N*…{ }¤Š'"ßF‘Š'")"N,~ ø#»©»#»øç í¡ž ~ lbHvç~ ø©»#»#»Âø‹>çm€ Œ (çm€[‡'ß Š2€ Œ€Â»Þ {), ‡‡¦x›‡'{(‡#ô|G"…|½‡©"%"‡­àVŠ+{¤{(‡}0*»Š/"…‘'"+ß(‡ä‡}ßj×éÏOï`HgÚhJ^Hgîy¦$"…}/à'‡!à5}ƒ5&&|1('"‡gIÕÔ§M@M»‡}*|1"ß%, ‡{¤‡}¤}($"ß ~@Ü ‡}¦ß?&‡ }¤‹%*»‡{¤‡}@*»5àV‡ËûòÈiÉHvõ ~ ø#22#2øeõ Û ž ~ l ü ~ ø2#2#2ü Û MSZ‹|¦‡}¦ßE>©‹Òj"Ë5‹ß¤{(‡é‡4‡©6*)&½‡Mß({©2‡#>©|‘‡#"ߤ‡H‘‹ à'+}j*{), ‡àt, ‡‘‡#"ßF8>5‡àç6€ Œóý çm€M‡'ßàVŠ/661‡H‹%"‡1‡4©ß¤‡©"…}¦|Š'"ó,€¢v¦yä'"…}&‡Rà'Š'"ߤ‡#4©ß¤‡R‘‹Š‘t2 ‡#&‡R× 6‡Š5{(|½‡R‘‡}¦àVŠ+{(‘‡}Fß(Š'*ŒŠ'&Š'(+|1Z‹‡}Z‹|X‡}¦ß&;: Š Þ ‰‡Vß­‘l: |G"ß%, ‡'{©7 ‡VßZ‘‡1à'‡R*…'*|½‡{ƒt&;: ‡}@*»5àV‡Hß+Š'&‹%6 ‡3(/&½Š Þ $&‡Hˆ…Š'&Š'6 Š5{*ˆ‡1‡Vß+"Š/"\"©‹%&1&½‡ ~


Ýqáq~ ‡'‡*…{‚©Ë Š4Œ ÈiÉ §Ë Ó … ¿§Ë H;úv1Ô ^ M*ŒŠ5|1"ß/‘…$"…}eÉœœžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡#6 ‡©"+ß\‹%"‡­‡#4©ß¤‡#"»}¦|Š'".}¦‹)*»‡{(}0Wa6K, ‡Vߤ{(|bZ+‹…‡ä‘‡}/{%, ‡}¤‹%&Çßf8ßf}wß({(5|Çß%, ‡}\‘'"…}-&‡}-*…5{¦ß(|½‡}-*{%, ‡àt, ‡‘‡©"+ß(‡}v*{%,Þ Š5{f‘sƒ}¤‡#&Š'"[& àVŠ+{¤{(‡}0*ŒŠ'"…‘'"…àV‡¦‘‡+!¡‡#"{(Š+}¤‡5ƒ5ÈiÉ ‚§Ë¿ ‡}¦ßF5}(}¦ŠàV|#, ‡23&’>5‡{(}¤|½Š/"}¤‹%*Œ‡'{f}@Wa6K, ‡Vß({¤|bZ‹‡!‘‡ õ ¿ vÙ$Š5‹ßF‘l:5‹ß¤{(‡*…{¤ß/‡VßZ‘‡+6*$"…|†2 ‡{¤‡!$"»$&Š'(5‹…‡ä‹Mà5} Þ Š+}¤Š'"|bZ‹‡5ƒ5*ŒŠ5‹{\ß(Š5‹ß$ØM‡'ߣ¢%.%&‡}?, ‡§Z‹…8ß(|½Š/"…}y3:¢ Î Ø ¢ ¨ ¢ ¢ ÏDÿ%. ¤Ï Ô'øŸú0¨I§¢ ¥ ¢¢ ÏDÿ%. ¤¦¥Ï Ô'ø#»©»#»lø0F‡'ߤ‡{06 |1"‡#"ßJ‹%"‡¦àVŠ+‹{ Þ ‡Z‘'"…}B&‡Z}¤‹%*Œ‡'{(‡}0*…5à'‡\ß6'&½‡‹{f}EÔ‡Vß`úvúyZ‡[*%&‹…}'ƒP&½‡E9¨|½ßZ‹‡?&½‡}1à'Š©‡×ÂàV|‡#"ß(}‡Vߤ{(|bZ+‹…‡}­‡©"3¤ ~ ø©»#»#»ÂøU¤ q 9¨Š5‹{"|½ß3F­H†ÓIéÔ§MHàVŠ©‡×Âà'|½‡©"+ßf} Þ Š+}¤Š'"…|1Z‹‡}H‡'ß39¨‡{06 |Š'"|bZ‹‡}vay3: ‹ß({¤‡[*…{¤ßƒ}¤Š'"ß1}0Wa6K,&;: |’‘‡#"ß(|]g»à8ß(|½Š/"|0 Œ ’&0¯‡Vß ¤ Œ ’ ¤ ߤ‡©&PZ‹‡4’IÅÈNJ ½ ´]¾EZ+‹…|J|G6*%&|bZ+‹…‡ Œ ’ ‡'ߢ Œ ’¢…ƒÂ"Š+‹…}­$*%*…{¤‡©"…‘TZ‹‡&;: ‡}@*…+àV‡!‘‡}j61Š‘‹)&½‡}/‘‡+& à'Š5‹{ Þ ‡": ‡}¦ßZ‹ß¤{(‡Z‹‡ÈiÉ ¿ž ‚§Ë¿ž ¿ »‡}¦‹)&Çßf8ß2RàV|½ß¤‡{\à'Š'"…à'‡'{"‡&’Hà'Š5{({¤‡}@*ŒŠ'"»‘$"…à'‡j(ž, ‡'Š'6¥, ‡'ߤ{(|1Z‹‡F}¤‹%*Œ‡'{¤ß85{¤|, ‡Vß%, ‡eÈiÉ ‚©Ë¿ ‡}¦ß‘O7 ‹5‹9¨|½ßNZ‹‡Zà'‡Vß›‡}0*…5à'‡/‡}¦ß›}¤|G6R‹%&Çßf$"-, ‡#6 ‡#"ßJ‹%"‡}@*»5àV‡ä}¦‹5AuB:‡Vß%, ‡J‘‡eU›'& Þ |GA†Êw5‹ ~À¿ vOÇä{§.5àV‡}$}¤‹%*Œ‡'{>8{(|, ‡'ߤ‡}l9¨Š5‹…{0"|’}(}¦‡©"+ßX‹%"‡\{©2 ‡#"…‡J|G"ß%, ‡'{(‡}(}¤'"ߤ‡*Œ‡'{¤ß85{¤|, ‡Vß%, ‡}S‘‡+U›'& Þ |]A;Êw‹*‡©"Âßf$"ßNZ‹‡Zߤˆ-, ‡Š5{(|½‡}w‘‡&‡}wߤ‡àfˆ)"|1Z‹‡}QZ‹|k*Œ‡'{61‡'ߦß(‡#"ß\‘:þ |1"%ô Þ ‹{().a*…{e&‡ Þ |’|’}/‘‹MàVŠ'"»àV‡#*ßZ‘‡&*‘‹»$&|Çß%, ‡AÀÙ *ŒŠ5‹{e&‡}Fàfˆ»$6*…}N9¨‡'{6 |½Š/"|bZ+‹…‡}vu'"…‘5‹5A‡}B*…5{N&;:


ûûbâbâœbbbb‹‘B’ “Y”R•£–b——£–P˜#-š$›œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5Š+|½{wߤŠ5‹ß-&‡}\‘‡‹54‹%"‡6*5}(}¤‡ – ÝävJB"[*»Š}¤'"+ß J Œ I4ƒ%&‡!}¤‹%*Œ‡'{*»Š5ߤ‡©"+ß(|'&1û *…{¤‡©"…‘[&39¨Š5{6 ‡HyÝ øD\øø74OMBÏòJ – ÝHgTMIKLØ/üVH=@? H85{¤|,ñ ÿ5î ñIëk ¢+¢ ð í $ñC ƒ ñ î#¨ ­¨ ¢ ëk £#XRsþî$"…'&GW}¤‡w‘N, ‡©>5‡©&½Š/*%*-, ‡‡ž*{%, ‡àt, ‡‘‡©66 ‡©"+ßX}¦‡\àVŠ/"…àV{%, ‡Vß(|}¤‡w‘'"…}X‘‡}X‡4‡©6*)&½‡}Xà'Š'"…à'‡'{"…$"ßÂ&;:85{¤|,ÜX¼Â¿f *…{6 |l&½‡}0Z‹‡©&}-"…Š5‹…}-*{), ‡}¦‡©"+ß(Š'"…}NZ‹‡#&bZ‹‡}F‹%"…}v$*)*…{(‹…}\‘$"»}/àV‡Vߤߤ‡q‘|{(‡à€ß(|½Š/"06“SÂ¥£ª¥’£/8 ìí¡îfïð$*{§2 ‡} ~@½ ƒa&½‡ 61|{(Š5|{›‘‹hU›$&’ Þ |GA†Êw5‹`ÈiÉ ‚§Ë¿ ‡}¦ß`, ‡©Z‹|1>6'&½‡©"ß2q‹)"EZ‹…5‘{(|bZ+‹…‡Z‘‡ÈVÉ ‚©Ë‚ ûSÈiÉ ‚§Ë‚ 2R&R&½|16 |Çß(‡ZŠ2yR:‡Vß({¤‡¦‘‡B­ˆ)&½‡{©M\‘‡ÈiÉ ‚§Ë¿ >P, ‡'{(|]g»‡MNJ Œ&‡*»{f$6[2N v‚§Ë¿ ‡}ß ‘N, ‡àV{(|½ß¤‡*…{ ‹%"B"A‡C¤‡V߃?ÈiÉŠ©‘l2 ‡#&‡Ë}¦|1('6*ó&|G"-, ‡|{(‡ à'Š'"ߤ‡©"…$"ßKZ‹…ߤ{(‡ àfˆ»$6*…} àfˆ|{f‹546Š+}¤Š'"|bZ‹‡}*‡'ß?Z‹…8ß({¤‡=9¨‡'{61|Š'"…|1Z‹‡}S"Šß%, ‡}Â{¤‡}@*Œ‡àVߤ|1>5‡#6 ‡©"+ß1)©Þ ‡VßD+¢Þ'ø(EDÏÞÿ%ø©»G»1»1ø0÷aߤŠ+‹…}‘‡ àfˆ…{(5‡}?I3Ô Á v5X


œFF)©ÞFFœGFF+¢ÞFF‚ò ~ â‚) Þ ò %F FÞò Þ9 ; ¼ S = ; x ÓkÛ wS‚FFI‚+ Þ ò F FÞ‚ò Þâ ‡'ß$Ù â 2>8$&‡'‹…{(}/‘'"…}eÈ›ƒ$‘N, ‡g)"|’}N*…{Ž©Š5|½ße&½‡}e"Š+‹%>5‡‹54àfˆ…'6*…}¡Ø¼ S = ; Ï"Ø â ƒ L ¼ SLŒ‚ò ~ â‚9 ; Ï"Ù â Ø â|1"…}¤|l&½Šà$&‡#6 ‡#"߃)&½‡6 |½{(Š5|{F‘‡+ÈiÉ ‚§Ë¿ ‡}¦ßj>©‹eà'Š'66 ‡¦‹)"‡­âW¡*Œ‡'{f}¤‹{@9¨+àV‡­‘‡­}¤‹%*»‡{@A@U›$&’ Þ |]A;Êw‹Ý‚ò ~ â÷/M|16*%&½|bZ‹‡‡©Z‹…ߤ|Š'"ÌHd÷v‚ò ~ âGFFœžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹9 J I-L ¼>= J IKL ¼>= J‡Vß({¤|‡-K‡'{(|]g»‡ Z+‹…|>P,‹ Š2L8Û Þ Ï‡'ßL Þ Ï‡'Š+{¤|‡ä‡}ßF|1"a>8{(|'"+ß(‡ *»{e&’1}0Wa6K,vQU‡Vߤߤ‡!ߤˆ-,b Þ ø.M ŒJMÞMLÛ Þ øb Þ Œ L ¼>= J 4 Þ ø54 Þ Œ L 9 Jg­"鑇q{(‡#*{(Š‘‹|{¤‡+&‡6 Š‘t2 ‡#&‡q}¤|G(/6 *à'Š5{({¤‡}@*ŒŠ'"»‘$"߃|1&9¨5‹ßä{(‡‘N, ‡g)"…|½{e&‡}äàfˆ…$6*…}e*ŒŠ5‹…{Z|ÑÏ3Ô5ƒ úƒ ÷1à'Š'66 ‡H}¤‹|ÇßÝâ Ï 7â IòÛ ‡'ß[Û â Ï 7â IòÛ »-!S{à'Š'"…}5, ‡©Z‹‡©"߃-&L9¨Š'"…àVߤ|Š'" H†÷%» ú'M*‘Š/"%"‡© ÏON ‚Û7 Û â J 7 Hâ ò â M-I J M£»»uB: ‹ß(|G&|}(8ß(|½Š/"鑇Hà'‡}F{), ‡}¦‹%&½ß(ß(}/5|G"…}¤|VZ‹‡R&†: |G"ß%, ‡#(+{(ߤ|Š'"}¤‹{ &½‡}Zàfˆ…'6*»}-9¨‡'{6 |½Š/"|1Z‹‡}Û‡#"‡[2&;: ‡#4a*…{¤‡}¤}¤|½Š/"}¤‹|1>6'"ߤ‡'fb ‡VßP4 6[2Hò ~ âL ¼:9Q¼:R ;â Ób â ÓI4 â û7 Û â Ó 7 Ó= ; ¼ S 9 ; x I-L ¼ S = ; wUSâ 4 â M})b‡#4a* % L ¼>9Q H= ; IÕÔPI3LT ž@R ;¼ S 9 ; I-L ¼ S HgL‹ ô Ï"L ¼:9Q ‰Š+‹‡+&‡¦{X7 Š'&‡­‘l: ‹%"h6H‹%&½ß¤|1*%&|à8ߤ‡‹{Z‘‡­u$$(5{f$")(5‡5vŠ2ߤ‡©&Z‹‡b âS =V;b â »…y¦$"…}Fà'‡­à'5}ƒ5&;:+ŠaZ‹-, ‡'‡ ‹%*»{f8>6'"ß=": ‡}ߣ*…+} $*)*…{(‡#"ߤ‡+vy¦$"…}aàV‡'ߦߤ‡¥>5‡{(}¤|½Š/"¢ƒ&’U‡©*»‡©"…‘$"߃H‡#&1&‡K*Œ‡'‹ß‡'*%*…{f87PÇß({¤‡i*…{X‹%" }¦|16*%&‡S, ‡àfˆ…$"%(+‡w‘‡}${X7 Š'&‡}X‘‡© ‡'ßOÛ ‡#"qßf$"ßžZ+‹…‡Q6R‹%&Çß(|G*)&½|’à'ߤ‡'‹…{¡‘‡›u$$(5{f$")(5‡5ƒ6|1"…}¤|¡Z‹‡{%,‡}vÂ!S5{¡à'Š'"…}%, ‡§Z+‹…‡#"ßž"Š5‹…}w8>5Š'"»}rØ â L Ù â ƒ@M L JM£»B"*…5{¦ß(|à'‹%&½|‡'{ƒ'&;: $*%*%&|’à'8ß(|½Š/"&‡}Jàfˆ…'6*…}BZ‹|)&‹|…}¤Š'"ß+}¤}¤ŠàV|,‘‡­àV‡qàfˆ»$"%(+‡#6 ‡#"ß2&†:


‚ò ~ â Ï Ô Ï{Ù 8>5‡àä‹%"‡+*»{¤ß¤|‡!‘‹hZ‹…+‘{(|1Z‹‡Z‹|¢àVŠ+|G"»àV|’‘‡*»Š+‹{$Ø‚¦Gž ~ ÛHò Þ‡#g)"|{F‘‡}/àfˆ…'6*…}¡Ù bÏL ¼:9 ; ‡'ß$Ø â Ï"L ¼>= ; ‡'ßj*»Š}¦‡{ ¦ Ù ~ ÏÙ ~ ‡'ß ¦âGž ~ ÛHÓ ¦ Ù â ÓI`tÓ[am5H+‡#6 ‡#"ßv\B"h*%&‹…}ƒ­&’*}0W¡6¥,‡Š5{(|½‡ Š+{¤|1(5|1"…$&‡ ‡}¦ßqߤŠ+‹ßR}¤|G6*%&‡#6 ‡©"+ß3&é}0Wa6K, ‡Vß({¤|‡SZ+‹…|}©:©|‘‡}!‘‡&†:5‡©6 ‡#"ßv Ý g­"a‘‡Hߤ{(Š5‹%>+‡'{!}¦Š/"‘N,|½{(Š5|{ƒ…|1&9¨‹ß *{(Šàt, ‡‘‡'{Z8>+‡à&’S6[7 ‡#6 ‡H}¦ß¤{f8ß%, ‡#(+|½‡+Z‹: ‹%*»{f8>6'"ßv\uB: |G"ß%, ‡#(5{f$&‡q‘‡Hàfˆ…‡#6 |G"\*»Š+‹{&‡6 Š‘t2 ‡#&‡H‘‡6þ ‘‹…'&¢‡}¦ßF‘Š'"%"-, ‡*…{uÓ­Û â Ó@ â ÓÛ â J b â Ó[4 â 5H â J|M@Meûâ ò ~b â 4 â M@M£»‡#4a*VHXL ¼>9 ; I


ò ~ âbaâ`~âbá Iò ~ €ò ~ âábâ / â I aò ~ €á Ma⢠â~ çâ í ¢ â~ < ÏÕÿ çbÏáwÜÛ Ë Û ž ~ x vwŽ©‹ Þ }5,‡œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹8{(|,$6*%&‡'‹{R‘‡ÂàV‡'ߦߤ‡S>85{¤|, ‡Vß%, ‡R‘…$"…}+&‡Âà'+}!Š5{f‘|1"…|{¤‡ ‹£"|1>5‡‹ú‘‡}HàVŠ'6*…+à€ß(|]g»à8ß(|½Š/"…}!‘‡}!ß(ˆ-, ‡Š5{(|½‡}tS‹£&†:‡àV‡}(}(|{¤‡ ‘l: 8ߦß({¤| Þ ‹‡{[2M}¤Š'" àVŠ+{¤{(‡}0*ŒŠ'"…‘'"ß\, ‡'ߤ‡©"…‘‹L&‡}+*{(Š'*…{¤|, ‡Vß%, ‡} &‡à'5{(+à€ß%, ‡'{(|}($"ßv*P"‘‡}qàVŠ+{(‘‡}'ƒ$|G&‡}¦ß"-,|1"…}0*|½{f$"ß/‘‹Mà'+}\‹…}¦‹…‡#&‘‹Mà'Š'"|G9¨Š'&’‘sƒ5&;: 5‹ß¤‡‹{F‘‹ ‚Á {%, ‡l, ‡àV{(|ÇßN&†:


÷§”§• ¤¡– ô(—“m¤¨˜6Šš ±ôz“†ò ~ âGH¢â~H~) â l I lò ~ €ââ †€€M»~û»‘B’ “Y”R•£–b——£–P˜#-š$›œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹©|‡#"ße‘l: ‹%"‡u¢‡ú‘‡'{"|‡'{.àVŠ/66~ ø Z ø(Z ‚ ø(Z ¿ M¡‘‡Fàfˆ…{(5‡} V H8'"+ß(‡ †È-É i ¼s~ H‡/ ~ø#»1»G»1ø7/ ~i M-ûÈ-É ˆ ¼s~ Hz/ø©»G»1»Gø‰/ Šˆ Mûe»1»1»)ûHd÷v ~/MÈ-É g ¼¢~ Hz/ ÛÛgø#»1»1»Gø‰/Þ Š+}¤Š'"|bZ‹‡}ZàVŠ5{({(‡}0*»Š/"…‘‡#"ßZ5‹54?>8{(|, ‡Vß%, ‡}(}\àVŠ'6*…+à€ß(‡}\ß(Š5{(|1Z‹‡}jZ‹|$}¤Š'"ß&;: ‡}@*»5àV‡qà'| Þ &½‡H‘‹‡Vߦ‘Š'"ßj&‡}j*»{¤ß¤|‡}Š‘t2 ‡#&‡A Ý à'Š'"…}¦ß¤{(‹|½ß\à'Š'66 ‡ä‹%"?6 Š©‘l2 ‡#&‡ä}¤|G(/6*&|G"‡{ V H


†â Ïˇ45*H©|½ß(}e*…{¤Š/*»Š}¦‡q‹)"‡R"Š+‹%>5‡©&G&‡R'&Çß(‡'{"…8ß(|G>+‡q‡#"éߤ‡'{6 ‡R‘‡3&‡'Š+{¤|‡›‘‡}wà'Š5{f‘‡}S‘…$"…}ž&;: ‡}0*…5à'‡\ß©|½ß%, ‡›àVŠ/"59¨Š5{6 ‡ß¤ˆ-,F¦{¤|‡#"Z‹: ‡#"1‹ß¤|1&½|’}($"ß¡‘‡}O>8{(| Þ &‡}$ß|½ß%, ‡RàVŠ/"59¨Š5{6 ‡+Î Ï F­øÂß Ï F…vàV‡'ߦß(‡S"Š5‹%>+‡5‹ß%,‡{¤‡}¤}¤Š'"»}h2 }¤Š'" }0*»‡à€ß¤{(‡c&|G"-, ‡|{(‡e‘'"…}S&‡é}¦‹)*»‡{¤‡}@*…+àV‡+ƒNZ‹‡c"Š+‹…}Âà'Š'6*…{(Š'"»}B"^*…5{¦ß(|à'‹%&½|‡'{E"Š5‹…}"…Š5‹…}*|1"ß%,‡}¦‹)&Çßf8ß(} 9¨Š+‹{0"…|}*…5{+&M‘‡}(àV{(|G*ß(|½Š/"a‘‡S&eß(ˆN, ‡'Š+{¤|‡ ‘‡ÂàVŠ+{(‘‡*‘'"…}+&;: ‡}@*»5àV‡1ß5‡à3&‡}q{%,‡Š5{(|½‡jZ‹|»‡}¦ßàVŠ/"…}ß(|Çß(‹-, ‡‡F‘‡ä‘‡‹54>5‡'{f}¤|½Š/"…}#fÂ&†: $&½ß¤‡{0"…ߤ|1>5‡F‘‡¦x›‡'{á5ŠX>©|Çßf}žZ‹|»à'Š'"…à'‡'{"‡e&HàVŠ5{f‘‡/Š+‹%>5‡{¦ß(‡U‡Vߤߤ‡/ß(ˆ-,‹ß(|G&|}($"ß‘‡}3>85{¤|’ Þ &‡}qß\ õWV HÑFBl FaM­‘Š/"+ß&=*…{¤ß¤|‡ Þ Š+}¤Š'"|bZ‹‡Â‡}ß õWV HÑFaMû õV HÑFaM€ƒJ}(àVŠ/"59¨Š5{6‡'ߤ{(|½‡!àVŠ+{¤{(‡}0*ŒŠ'"…‘'"ߤ‡Z‡}ßj&’ }@Wa6¥, ‡'ߤ{(|½‡}0W¡6¥,‡#"»}¦‡©6 Þ &‡!ߤŠ5ß($&¢‘‡}-(ž, ‡#"-, ‡{(ߤ‡‹{(}w‡}¦ßZàVŠ/"…}ß(|Çß(‹-, ‡!‘‡vJuB:½ \ í øe_ í Û ø0ßcøÿ í ¾Bø ½ – € ø õ € ¾ú‡Vß ½¾Qø‡©Z‹‡#"߃%&½‡RàVŠ'66R‹ß(ߤ‡'‹…{Z‘‡H‘‡‹54h(ž, ‡©"N, ‡'{f8ß(‡'‹{f}›‘‡H‘‡©(5{%, ‡-üé‡Vß ý"‡Z‹|$}¤Š'"ß/ß(Š5‹…}2>6'&½‡‹{f}F‘$"»}PK+»'!S5{FàVŠ/"…}5,‡©"-, ‡{(ߤ‡'‹…{(}䑇‘‡©(5{%, ‡ ǜ I ý©veSßf$"ßq‘Š'"%"-, ‡SZ+‹…‡S&½‡}q‘‡#(+{%, ‡}!‘‡}(ž, ‡©"N, ‡'{f8ß(‡'‹{f}Z}¤‡S*{%, ‡}¤‡#"ß(‡#"ßàVŠ/"ߤ|‡#"ßZ+‹…‡S&½‡}+(ž,àVŠ/661‡­}¤‹|½ß‡©"-, (ž,õ € _ í Û øßcø‡ Jú J+Ô ÿ Ô ú‘‡#(+{%,^ Ô


Îú5}\|G"»‘|à'‡}F}0*|1"Š5{(|‡#&’}[‹)"‡q‘|161‡©"…}¤|½Š/">kÏ ~ HÑF+JhÎÕM[IIG¢yx w¥:¨«ª ©¢¢¥¢ ¨¢ H´³ / ³m/5MY³ ¢ ¨ ¢ ;ô ¥¨¥IH‡³:­3M ¨w ¢yx ¥b¢ ¨”””G¨b¢¥¬,® ª ©I¥:¨ x Iµ°´¢ ½² w¥G¬§® ª w ¢#vx ©Iœ¥¨Mb¨w ¢yx ¥H‡³:­3M¬§® ª/ ¥© ¥:¨§ª w ¢#vx ³ ¢ ¨w ¢} q x œÚ¢ ;‘B’ “Y”R•£–b——£–P˜#-š$›œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡©&½Š/*%*»‡!àVŠ/661‡q}¤‹|½ß©fH*ØVø7¢/MF}¦‡­‘N,w ¥:¨ x¦ ÏDÚIA¢I¥ ¢ô ¥ ¢ I±°‡¢ ²”zw ¥:¨ x¢}}x wH´³ ®®¢}}x wI ° ¢ ²©¥:¨«ª w ¢yx I ° ¢ ‚ ² {¢¡¬¢¡¬ I ° ¢ ¿ ² ¥ w ¢yxI ° ¢ ‚ ²’j© ¥:¨«ªw ¢} q x© ¥:¨«ª° ¢ ¿²û‘¢} q I ° ¢ ¿² ©¥:¨§ªÓU©© ¥:¨«ª¢¡¬¢ Á² © ¥:¨§ª ¬ ³ ¢ ¨ ¢ ; °H´³£µaM’ ¨° ¢ Á² ¥ w ¢}}x© ¥:¨§ª¢ ³ ¨¥ w ¢}}xI"°´¢ ½² w ¥:¨ x ³6/Y³ / ;{ ©¥:¨§ª¨ ¢ ¨ M HI"°´¢ €²¢ I°‡¢ ܲ H´³6/Y³ / M ;HÑF…v1Ô§M‡#"-, ‡{('&Ÿƒ/&†: 5àVߤ|Š'"[&½|1"-, ‡|{¤‡!‘‡+&’Rߤˆ-, ‡Š5{(|½‡!‘‡q}¤‹%*Œ‡'{(5{f8>|½ß%, ‡ZàVŠ/"59¨Š5{6 ‡-ÎØÏÕF1‡}ßZ‘Š'"%"-, ‡'‡*…{§fB"[(ž,õ ÏÓ ¿ ØÜ ¢I¦HÑF…v ú'MÓ‡l, ‡àV{(|Çß‘…$"…}e"…Šß¤{(‡­à'+}#f‡VßF}¤‡­{%,õ ÏÓ ¿ ØVH†ÚH´³ / ³6/%M ;I ô ¢ ¥ H´³ / ³m/%M³ ¢ ¨ ¢ ;Úô ¥¨w ¨ ¢ ¨ yx {I¢”zw ¥:¨ x ³ / ³ / ;¥:¨ x I’{-© ¥:¨«ª w¢yx ³ ¢ ¨ ¢ ³ ¨ ; w© ¥:¨«ªI’ © ¥:¨«ª© ¥:¨§ª Ó © IKÓ³ ¢ ¨ ¢ ; ’ ¬¢¡¬¬§® ª© ¥:¨§ª© ¥:¨«ªw ¢} q xH´³>­3MIH‡³M ¥w ¢}}x H‡³£µaMM£»HÑF…v ÷/Mw ¢}}x^ ú


Î/ ¥¨w ¢yx ¥H´³:­Mb:óGϳ¥ ¢ ³\¢b¥ a¥¢¥¢¥óIb Ø õ¢¥ßM¢œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹85{¤|’$"ߤ‡}w‹»}¦‹‡©&G&‡}Z‘‹M}¦‹%*Œ‡'{(‡}0*…+àV‡!‘Š'"%"-, ‡'‡}Q*»{§f‡Vß‹hß ¢¥Š2³; ³» HÑF…v M¢¥ß; ¢ ¨ I¨ ¢ ; ß ¥¨ ¢ Ï ³“¢¢ ¨|t&½‡}/àVŠ'"ß({(5|G"ߤ‡}\‘Q, ‡‘‹|Çß(‡}/‘‡+&;:5‡©&½Š/*%*Œ‡#6 ‡#"ßH‘‹ }¤‹%*Œ‡'{fàfˆ…$6* àfˆ|{('&%¦P"L>5‡{¦ß(‹ú‘‡E&;:©|Çß%, ‡qàVŠ/"59¨Š5{6 ‡uB:‡} ‡#&|’àV|½ß%,F»v1P" ‡#CŒ‡'߃Â&;: $"…'&GW}¤‡1‘‡}­}¦Š/&½‹ß¤|Š'"…}!‘‡*}¤‡}T, ‡§Z+‹»8ߤ|Š'"»}䑇6 Š5‹)>5‡#6 ‡©"+ß+6h2 ‡©"‡15‹54 ‘|GC, ‡'{(‡#"ߤ‡}¦ˆN,Ï¿ŸÁ ƒ…àVŠ/66 ‡"Š+‹…}e&‡!‘N, ‡#>+‡#&Š'*%*Œ‡#6 ‡#"ßF‘'"…}/àV‡­}¤‹|Çߧf‡'Š5{(|‡‡àV{(|Çß(‡}N*…5{N&‡}Zàfˆ…$6*…}/‘‡&’1ߤˆ-,‘N,Ÿš$›GšX–„•0˜s¿uÀ^uB:


’¿¨)šX©·[—y·t–„¡k¤X•£–d˜#› É¢ ô¥ ô }0*|G"…Š5{(|½‡©&¢‡Vß ‘‡Rߤ‡©&G&‡1}¤Š5{¤ß¤‡Z‹‡In ¦ ¢ô ¥+}¦‡+ƒ5&½‡­àfˆ…'6* Þb¥f¥ôfGÏGÁôô¨~ x w¥³ ¢ ¨ ¢ ô Á Ø õ¨ô Ø õ‡#"ß‘‡'‹54*…5|½{(‡}‘‡ }@*…|G"‡‹{f}=H*nžø ¦ nÂM*‡'ßTHzÊ%ø ¦‡©6‡'ß Á ô//¨ b ôx w¥Hyõ%M@MÂI ¦ Ê ¢¥»—Ì š ÕŠÖ×Ø٠ͦΠm —Ò š ÒœžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡+vqyä'"…}!àV‡'ߦߤ‡Ê‡à'Š'6*»Š}¦‡!àVŠ/661‡}¤‡!‘N,¢¥~ x w¥I ¦ Ê ¢x w¥Ïn ¢‡§Z+‹»8ߤ|Š'"‘‡+6 Š5‹%>+‡#6 ‡#"ßHÑF»v ~¡M\‘‡#>©|‡#"ßF$&Š5{f}uB:@,ÏDÿÏÕÿ%»‡©"…}¦‹…|Çßj&’ }¦Š/&½‹ß(|½Š/"?(ž, ‡#"-, ‡'{f$&‡Z}¤‹|G>8'"+ß(‡'fâÀ&¢}©:¢¥HgØkMPÏÓ ¿ õ“5Hyõ MHdF…v1Ô©ÿ¡MHvõ%MVÎ sHdF…v1Ô'Ô8Mâ ó f 8fn ¢ H ô ¼ i L¥ ˆHyõ%M 5‡©"…$"ßZ‘‹aàfˆ…$6*é}0*|G"…Š5{(|½‡©& ô¢¥‡Vß ô ¼ iˆ ¥ˆ ¥Š'&‡!‘: ˆ-, ‡#&|à'|Çß%, ‡3J ~ ƒßf$"…‘|’}eZ‹‡ ô i*ŒŠ5‹{F‘Š/"%"‡{/‹%"Màfˆ…$6*.‘|G*i7‡à'{¤|½ß/‹%"Màfˆ…$6*eŠ5{f‘|G"»|{¤‡!‘l: ˆN, ‡#&|’àV|½ß%, ‡I ~ v‘N,ˆ ¥©’-œ ¥5†•˜£— ¨)šX©·­—‡§Z+‹»8ߤ|Š'"‘‡+6 Š5‹%>+‡#6 ‡#"ßF‘‹Màfˆ…$6*1{-© ¥:¨«ª w ¢yx Hvõ%M\}©:6'"ߤ‡}/³ ¢ w ¨ ³ ¨ ³:Ë/ ¢ ÏÕÿ).¢exg :¨ ¢ w ³ ³ ³q xg :x ¨ ³ ¢ w ¨ q¨ Lp/ ¢ ¨ ¢ ÏÕÿ)» HdF…v1Ô©÷¡M¹Ç¸¦¾fÄÀâ+³µ³¿ÄÀ*â5Ã\ÀÄ'´èq¸¹¿³¿Æ¸'³¿ÄÀÂâ5ÃPͦΠm ÃfÆ—JÑ+´³¿ÆªÃF¾fÄè­è­Ã›Æª·³µ ë,Ï Í ÷§ÐMÑð ÒÔÓÌ^ F


,n£¢R|1"…}¤|Z‹‡ ¦b¥G/¢ ¦ Ê yx Ê ¨ bI{n]wôb¢¥b¥ ¬§®/ª¢¨ ¢/ ¥ ;; b;;;;¥G/ÿˆ¥¼i¦ Ê ¢ ¦ Ê ¦ n ¨ b ÚˆMÂIœM Ø õLd ¨'L ¼Œ 'áá®/ ¥ÄÄ Ø b õœbMœžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡©61‡©"ß‘Š'"߇©Z‹…ߤ|Š'"…}›‹…}¤‹‡#&1&‡}}¤Š'"ße&½‡}E,”zw ¥:¨ x ÏÕÿ)øݳ ¢ ¨ ¢ ’ © ¥:¨§ª ¢¡¬ Ï ÿ)øÁ³ / ³ c Hg: /,a Mew ¢yx Hv: ced M w ¢#vx ­ ¥ÏÕÿ‡àV{(|G>+‡#"ß\{¤‡}@*Œ‡àVߤ|1>5‡#6 ‡©"+ße&‡}/àfˆ…'6*…}/‘l: ˆN, ‡#&|’àV|½ß%, ‡ÿ%ø ~ ‡'ßÔ5vZ‹|¢‘N,8>k2 ‡'{(‡|1"+ß%, ‡'{(‡}(}($"ß ‘‡M{(+}¤}¤‡#6 Þ &½‡{1ߤŠ5‹ßàV‡} {), ‡}¦‹%&½ß(ß(}*‘'"…} ‹%" }¤‡'‹%&!¡Š5‹{*à'Š'"…à#&½‹{(‡MàV‡'ߦߤ‡c*»{¤ß¤|‡5ƒ›|G&Z}©:Þ &‡‹H‡4ˆ| Þ $"ßO&½‡}$ˆ-, ‡#&|à'|Çß%, ‡}X|1"…}¤|5Z+‹…‡Q&‡}X{(‡#*{%, ‡}¤‡#"ßf8ߤ|Š'"»}s|½{({%, ‡‘‹»à€ß¤| Þ &‡}ß(õWV HÑF¡MPHÑ"Š5ß%, ‡ õWV HdF¡M7ÞOM‘‹R(+{¤Š+‹%*Œ‡‡'ß(ß(}V*âW©}¤|bZ+‹…‡}vOU‡ß( Þ &½‡‹Rà'Š'"ߤ|‡#"ß, ‡#($&‡#6 ‡#"ßV&’‘‡}+,V H


áœG€ó æ _ó Þ ‡'ßç_æ Ͼ` Þ ø a æ Ïqa Þ »`>8'"+ße&‡!ß8{(| Þ &½‡}$ß8{(|’$"ߤ‡}懇}N*…{ „‘Š/"%"-,æ Ïw³ ¥ Ï8³IÏ ; ³ ¥ Ï ; ³I Þù H@Ô'øUTžM­àfˆ»$6*M‘‡\‰¦‹%(+‡`Ḧ à'Š'"%"‡à€ß(|½Š/"­MN*»Š+‹{j&‡©Z‹‡©&PH;I3Ô8M›‡}¦ßj&*àfˆ…5{0(+‡­5}(}¦ŠàV|#, ‡‡2‹ÞàVŠ/"ߤ‡#"»$"ß/‹%"©à‡VßA$Þ$&Š5{f}jZ‹‡RÛæÞ ‡Vß Û Þ Þ *»Š+{¦ß(‡#"ß&’Âàfˆ…5{0(+‡`H5‡'{f}¤|½Š/"hZ‹…'"ߤ|bZ+‹…‡5ƒk&†: 5àVߤ|Š'"¾HÑF»vGÔ M/à'Š'"…‘‹|ÇßæŠ'6 ‡} ‘‡h6 Š5‹%>+‡#6 ‡#"ß*‘{(Š5|½ß1‡'ß(‹…àfˆ…‡5v u‡}*{(‡#6 |½‡{(}1}¤Š'"ß à'‡'‹54¯‘‡¯tZ|{(+}¦Š+{¤Š="…Šß%, ‡}[Hvã æ øL æ MR‡Vß2a‘‡}39¨Š'"ßX7Þ øL Þ Mø»à'‡#*Œ‡#"…‘'"ßj&½‡}F}¦‡àVŠ/"…‘}/}¤Š'"ßZ‘‡!ß


GÏÕÛ Þ ³] Þ IA{ ‡'ß ì=ÏÕÛ Þ Þ »{|1"…}¤|ž*»Š+‹{­8>5Š+|½{!‘‡}­'"Š'6*'&½|‡}!àVŠ/"59¨Š5{6 ‡} "©‹%&G&‡}SH – Ï ÿ¡M£øl&‡}­‘|GC, ‡'{(‡#"ß(}*…5{('6[2 ‡Vß({¤‡}Z‘Š+|G>+‡#"ß*ŒŠ5{¤ß¤‡{­‘‡}Ý¢,[¢‡¢¢Œó¢¢óó¢¢¢¢ó‡œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡{¦ß(‡J‹…}¤‹‡#&1&½‡Ḧ‹?&½‡}\|G"»‘|à'‡}/‘‡&;: '&G()2 ‡ Þ {¤‡Z‘‡!u|½‡*{(ŠX>©|½‡©"%"‡©"+ß/‘‡}N9¨5à€ß(‡'‹{f}\‘‡RUˆ…$"%A³!SߦߤŠ/"¢ƒà$: ‡}¦ße&†: $&1()2 ‡ Þ {(‡Z‘‡­àVŠ+‹{('"ß-Z‹|Š2*{(ŠàV‹{(‡¦à'‡}\|1"…‘|’àV‡}F‘$"»}/àV‡Vߤߤ‡q'&Çß(‡'{"…8ß(|G>+‡ 0 v‡'ß(ß(} *âW©}¤|bZ+‹…‡}¦‘‡*àV‡Vߤߤ‡ àVŠ+{(‘‡RŠ5‹%>+‡'{¤ß¤‡1}¦Š/"+ß­‘N, ‡àV{(|Çßf}F‡©" ß(‡'{61‡}䑇}¦Š'*-, ‡'{f8ߤ‡‹{f}e>5‡{¦ß(‡4\Z‹|S}¦Š/"ßu¢‡}c,‘‡}Jàfˆ…$6*…}O*…{¤|16*|{¤‡}'ƒ5‘‡F‘|G6 ‡©"…}¦|Š'"hÔ/‡'ßP"‡'‹ß¤{(‡}S}¤Š5‹…}ž&;: 5à€ß(|½Š/"*‘‡$à ù H@Ô§M ( vSy¦$"…}wàV‡Vߤߤ‡e>+‡'{f}¦|Š'"¢ƒ'&½‡}Jàfˆ…$6*…}*{(|G6*5|½{(‡}+&‡}R*%&‹…}R}¤|16*)&½‡}R}¤Š'"ßR&‡}Ràfˆ…$6*…}5ZHy M­‘‡‘|G6 ‡#"»}¦|Š'"Kÿé|G"a>8{(|'"ß(}q}¤Š5‹…}R&†: 5àVߤ|Š'" ‘‡‹à ù H@Ô'øUTžMÝ ‹ß¤{(‡#6 ‡©"+ßq‘|½ß Za‡}¦ß­|G"a>8{(|'"ß *ŒŠ5‹{ Œ MU 8>+‡àGM¢ÅˆTU‘l: Š2 ‹yZa‡}ß­‹%"‡39¨Š'"»à€ß¤|Š'"ÌZ‹‡#&’àVŠ/")Z‹‡1‘‹}¤‹{AÞ/»}¤‹%*»‡{­‡}0*…5à'‡HßàAàVŠ/"…‘‹|½ßä‹%4?*{(‡#6 |‡'{f}\ß5‡{¦ß(‡4 }¦‡ à'Š'"…}¦ß¤{(‹|’}¦‡©"ßT2L*…{¤ß¤|{‘‡}‡#4a*…{¤‡}¤}¤|½Š/"…}S&|G"-, ‡5|½{(‡} ‡©"yä‡=6*$"|†2‡©"…}¦|Š'"ÌÔ'»)U‡}›5‹ß¤{(‡}wß5‡©6 ‡#"ß©fŒ Þ H}M›‡}ßZ‘‡­àfˆ…5{0(+‡+I3Ô/.UŒ ŒM7ŒM ¼¢~Þ H}MN*ŒŠ5{¤ß¤‡&’1àfˆ»{(5‡J+Ô'.5|½{(‡}N*…5{/{('*%*ŒŠ5{¤ß/5‹54Þ(ž, ‡©"-, ‡{(ߤ‡'‹…{(}J‘‡à ù H85{¤|’$"ߤ‡}\‘‡\‰¦‹%(+‡¦ß¤‡#&VZ+‹…‡!S{\}¦‹{fàV{(ŠX7PÇ߃©à'‡}N9¨Š/"…à€ß(|½Š/"…}£ŒBÞ­‡V߇VßÊÔŒ Þ ÏO Þ ôÞ Ïw³ Þ 4P»‡#"߃sà'‡Vߦß(‡3*»{¤ß¤|‡R}¤‡Rà&À7 Šß(‹{(‡R*…5{ä‹%"…‡H|1"+ß%, ‡'{(‡}(}($"ߤ‡­|’‘N, ‡‡L, ‡©>5Š5Z‹N, ‡'‡­‘'"…} f~ vU‡©&G&‡AªàV|i"Š+‹…}Z|1"59¨Š+{06 ‡3Z‹‡ãX|G"»$&‡#6‘‡!¡‡©"{(Š+}¤‡5ƒlZ‹|J}¦‡{( Þ |‡#"ú‡45*%&|’àV|½ß%, ‡ ‘$"»}&e}¦‡à€ß(|½Š/"a}¤‹|1>6'"ߤ‡5ƒ@Z(5{X75à'‡[2[&’Âß({('"…}5‡©"+ßž&‡},Þ H}MS‘N,‡NÎÐÏ F)»kU‡'ߦߤ‡­|’‘N, ‡‡¦}¤‡'{f1‡45*%&|à'|Çß%, ‡'‡*…5{e& }¤‹|Çß(‡5vàVŠ/"59¨Š5{6^65


áóœóáõFóó¢á‡œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹6'&½‡‹{f}F‘$"»}e&½‡}e&|G(/"‡}jg Þ {%, ‡} ö ƒ ö H+‡#6 ‡#"ße&½‡}-*…{¤ß¤|’àV‹)&½‡}/8W+$"ß-&½‡}/ˆ-, ‡©&½|’àV|½ß%, ‡}H8$"߃FH‘l:‡Vßf8ß(}\‡}ßF‘l: ˆ-, ‡©&½|’àV|½ß%, ‡ ‚ vàfˆ…+àV‹%".‘‡­àV‡}E,|1&a": ‡}ßO*…+}X‘‹qߤŠ+‹ßJàVŠ+{¤{(‡àVß$‘l: |G(/"Š5{(‡'{i&½‡\}0*|G"3*»Š+{¦ß%, ‡P*»{VE…v¡B"R‡C¤‡V߃@¤!‡Vß4 ¤*{¤‡©"%"‡©"+ßÍé|’}X}¦|1('"…'&½Š/"…}iZ+‹Â:Þ &‡}F‡'ß &‡'‹{f}Z‘‡'‹54D, ‡Vßf8ßf}j*ŒŠ5{¤ß¤‡#"ß&†: ˆ‡#&|’àV|½ß%, ‡ ~ ‡VßJ ~ {(‡}0*»‡à€ß(|G>+‡#6 ‡#"ßv U‡©*»‡©"…‘'"+ßàfˆ…+àV‹%"鑇‹54h>6'&½‡‹{f}j*ŒŠ+}(}¦|&;: |1"…‘|’àV‡*ŒŠ5{¤ß¤‡!‹…à'‹%"‡äˆ‡#&|à'|Çß%, ‡6 5|}|1&¤}¤‡äߤ{f$"…}@9¨Š5{6 ‡äàVŠ/661‡ &R{(‡#*{%, ‡}¤‡#"ßf8ߤ|Š'"F ‘‹E(+{¤Š+‹%*Œ‡ õWV HdF¡Mw‘‡}"…‡‡'ߤ{(|½‡} }0W¡6¥,‡§Z‹‡#"ßF‡#".ß(‡#"…'"ßäàVŠ/6*ߤ‡q‘‹ }0*|1"鑇&;: |1"…‘|’àV‡.E…ƒ­&½‡}h, ‡'ß(8ßf}/{%, ‡}¤‹%&½ß($"ßf}/‘‡&’9¨Š'"…àVߤ|Š'"…}vN!S5{FàVŠ'"»}5,Œ Þ Hy øU0ÂM/‘‡#>©|‡#"%"…‡#"ßFàVŠ'66 ‡q}¤‹|Çßf‡'ß(ß(} Þ Š+}¤Š'"|bZ‹‡}/‘l: ˆ-, ‡©&½|’àV|½ß%, ‡úú£,‡'ß(ß(} Þ Š+}¤Š'"|bZ‹‡}/‘l: ˆ-, ‡©&½|’àV|½ß%, ‡3Ôú£,‡'ß(ß(}N9¨‡'{6 |½Š/"|bZ+‹…‡}/‘l: ˆ-, ‡©&½|’àV|½ß%, ‡ ‚ }¤‡¦ß¤{f$"…}@9¨Š5{6*$"ß›‡©"c®M‘‡ õWV HdF¡MÞ›vF¾,, ‡'ß(8ßf}/‘l: ˆN, ‡#&|’àV|½ß%, ‡} Ô­}¤Š'"ßE, ‡#&|G6 |1"-, ‡}‡©"MàVŠ'"»}¦|’‘N, ‡{('"+ßN&½‡}FàVŠ/"ߤ{f|1"+ß(‡}\}¤‹|G>8'"+ß(‡}©f!S{\|1&G&‡'‹…{(}ƒ5&½‡}F‘‡‹54H*s65{¤|’$"»àV‡­‘‡›‰¦‹%(+‡­‡VßF‘‡}FàVŠ/"+ß({(5|G"ß(‡}ƒ‡}ßF‘Š/"%"-, ‡*…5{ (Þ^¥^


¢¢,ø ùù¢¢¢¨¢ùù¢¢ ³r¨¢¢ááßáßá¢ßᢨ ¢ßáßáßá¢áøßááø¢¢á¨áᢢ¢‡œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹6'&½‡©"ߤ‡}¦‹%4=9¨Š'"…àVߤ|Š'"…}¦|1"a>65{¤|’$"ß(‡}¦‘‡*u¢Š5{(‡#"ßô‡*»Š+‹{+&½‡}!‘‡'‹54öŒ ¨ ¢ }¤Š'"ß\,‡#6Œ ¨ ¢ ƒ÷ ¢¨ ¢ »5!S{à'Š'"…}%,*ŒŠ5‹{Œ¢ Œ ¢ ø*÷*»Š+‹{¨Œ ¢ ‡Vß¡0$¨Þ ß(‡#"Š/"…}\‘‡'‹54à'Š5{({¤‡}@*ŒŠ'"»‘$"ß(}w‘‡­‘‡#(+{%, ‡ú‡§Z+‹…‡#"ßQ"Š+‹…}Š÷ú÷ / ø‘‡Z‘‡©(5{%, ‡Qôl, ‡'{(ŠZ+‹…|…}#: 5}(}¤Š©à'|½‡©"ß¡{(‡}0*Œ‡à€ß(|G>+‡#6 ‡#"߃+}¤‡#&Š'"&¦{©2 ‡#(/&½‡/‘‡ Þ 5}¤‡5ƒk2+ú\, ‡'ß(ß(} Þ Š+}¤Š'"|bZ‹‡}J‘: ˆ-, ‡#&|à'|Çß%,‡Vßw‘‡'‹54ˆ-, ‡©&½|’àV|½ß%, ‡3Ô X¼¡~¦‚ vúR‡VßZ‘‡‹545‹ß¤{(‡}/‘l:Mk»Xy¦$"…}RàV‡[Z+‹…|}¤‹|Ç߃ž"Š5‹»}+g%4Š'"»}&;: |1"ß%, ‡{©7 ‡'ßH}¦‹…{&U‡Vߤߤ‡'"…$&1W}¦‡S"…‡àVŠ/"…àV‡{0"‡?Z+‹…‡`&c9¨Š'"…àVߤ|Š'"éŒBÞBH9¨Š'"»à€ß¤|Š'"‡'{(‡¦}¤|G6 |1&5|½{(‡5v‡‘'"ß\‘‡+6*'"|†2Þ Hy M›‡#"[*{¤Šàt,ûÔü3ý,þ]ÿ  ¤ ¥ ÏÕÿ).X&½‡}J}(à'$&’|{(‡}¡‘‡Fu¢Š+{¤‡©"ß0ô+Hy ¢¢¡øe ¨¢ ø(³ ¨¢ M¡}¤Š'"ßJˆŠ/61Š/()2¢/ø‰³“¢‡Hdÿ)ø0ÿ%ø©Jú5ø©Jú'M‡©"‡}X‘‡Z‘‡#(5{),*‹|’}Z+‹…‡¢1‡'ߊ/()2 ‡©"‡}䑇1‘‡#(5{), ‡`J+Ô+vRÇä{ƒ¤‡#"aàVŠ/"…}¦|’‘N, ‡{('"ß&†: |1"¡>85{¤|’$"…à'‡R‘‡ä‰¦‹)(5‡¢Â}¦Š/"+ߦˆŠ'6|1"…}¤|ÂZ‹‡&’ÂàVŠ/"ߤ{f|1"+ß(‡I$Þ‡©&½|16 |G"-, ‡}v Ý |G"…}¤|>ƒ»|1&{(‡}¦ß¤‡H‘‡'‹54Màfˆ…$6*…}‡+ÿÂ}¤Š'"ßh,Þ ŒÞ I5‡'{f}¤|½Š/"e‘‡R&1ß(ˆN,‚§Ë¿ àVŠ/66 ‡!|1&¢‡}¦ßZ‘N, ‡àV{(|Çß/‘$"»} ( v}¤‹%*»‡{/‡}0*…5à'‡¦ß5‡{¦ß(‡­‘‹?6 Š‘l2 ‡©&½‡­xË}¤Š'"ß/‡#"D, ‡'ߤ{(Š5|½ß¤‡&½|’|’}¦Š/"yR:‡}0*…+àV‡A>ߤ‡©6*…}'vT!B&‹…}H‡©"…àVŠ+{¤‡+ƒž!¡‡#"{(Š+}¤‡Â}¦‹)('()2 ‡'{(‡Z‹: |1&}1àVŠ+{¤{(‡}0*»Š/"…‘‡©"+ß1‹548>5‡à&½‡}Ràfˆ»$6*…}H‘‡H‰¦5‹%(5‡‘…$"…}3&†:‡#(+|½Š/"‘‡ ÈiÉ ‚©Ë¿ ‘$"»}Q&’'Z‹‡©&G&‡&‡}¡ ¢‡#&#,HdÈiÉ ' ‚§Ë¿ øeï3M ~(~ Šl2 ‹E&‡!}¦‹%*Œ‡'{(‡}0*…+àV‡ ÈiÉ ‚§Ë¿ {(‡#*…{%, ‡}¦‡©"ߤ‡F‹%"…‡ä{%,~‡©61‡©"ß(}w‘‡, ‡#&, ‡#6 ‡©"+ߦ‘‡Rà'‡9¨|’}(àV‡5‹ ‘‡(5{(Š5‹%*Œ‡H‘‡ àVŠ+ˆŠ'6 Š/&½Š/(5|‡­‡}¦ß¦{(‡#*{), ‡}¦‡©"+ß%, ‡*…{"‡1}¦Š/"+ß*…5}ZߤŠ+‹ß&½‡}䑇‹54="©‹%&’}'vÒj"Š/"…‘‡+Z‹|}(8ߤ|’}@9¨|½ß-&‡}/ߤ{(Š5|’}-*{¤Š/*{(|, ‡'ß%, ‡}w}¤‹|1>6'"ߤ‡}‹%"‡9¨Š/"…à€ß(|½Š/".‘:ZhÏ Ó ; Þ u Ô5vI¦; M\5‹ß¤{(‡#6 ‡#"ßF‘|½ß˦ ZM‡}ßF‹%"…‡`Hdÿƒ1Ô§MÀA†9¨Š+{06 ‡H}eøÞ; ³z¦ Z[ÏÕÿ úvZ Œ ¦ Z-I ; ³U¥ ÷vI¦‡©Z‹|1>6'&½‡©"…àV‡\‘‡¡‰¦5‹%(5‡\‡}¦ßB>P, ‡'{(|Gg-, ‡'‡Q*»Š+‹{¡ß¤Š+‹ß¤‡-9¨Š/"…à€ß(|½Š/"¥é‘$"»}O&½‡F}¤‹%*Œ‡'{(‡}0*…5à'‡NÈiÉ ‚§Ë¿ vàV‡'ߦߤ‡[,^ ~


ÞÐZÞ‘B’ “Y”R•£–b——£–P˜#-š$›œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹5‡{¦ß(‡45‘8‰Š5|1"+ß(‡M‘‹‡}2Âà'‡}h, ‡Vß(ß(}vFuB: ‹%" ‘‡}Z{), ‡}¦‹%&½ß(ß(}F}¤‹{*{(‡#"…'"+ßf}F‘‡ ~‚ "Š5‹…}ä|1"59¨Š5{6 ‡RZ‹‡R‘'"…}j&;: ‡}0*…5à'‡­ß6'"ß©f-›G˜—¤·Iº ˜•0š$³˜#¥5•Ÿ—¼X˜#•£³˜Ï ¹¡º ˜§ #•£–‚¼8š$¡)P›1˜£—+ ¨)šX©·­—¹'˜+¢5šX¥[ó¡Ã¹'š$¡%—›Gš.¼8˜#•Ÿ—£–d¤X¡âÀ&¤}#:˜§¤X•£–d˜K¹/˜£—Ì Ÿ¤$•U¹'˜—̤X¥“¼X˜#•£³˜—\y¸ž–b—ŸÀ¤X•£–d˜#›„›G˜—T—#¤$¡)› –mº ˜—Ѩ¢ºŸ˜¥¥Ï¹$¥D©S¤¹äÐ ˜›G˜=“š¹'šX¡)—ÌÈiÉ ‚§Ë¿ ·)šX•Ì›G˜ É –dšX–b—ˆ¹'˜Ì›Gš;•0šX¡)—„¤$•£©š$†–d¤$¡ZcJ Œ ¦’l˜Ÿ À˜¥5•.¹'˜—£¥·)˜•´ó'•0š,¼©–„(º ˜` Ÿ¤X¡§¤X•£©S˜À^u¢‡6 Š‘‡R‘‡1àVŠ+{(‘‡R9¨‡'{6K, ‡'‡&‡R*)&½‹…}\, ‡©>©|‘‡©"ßä‘$"»}&‡R6 Š‘l2 ‡©&½‡1xz‡}ßɇ#9¨Š5{6 ߤ|Š'"ú‘‡E&’é}¦ß¤{(‹…à€ß(‹{(‡àVŠ/6*)&½‡#4©‡‘‡?ÈiÉ ‚§Ë¿ vÌU‡'ߦß(‡‘N, ‡#9¨Š5{6*8ߤ|Š'" ‘Š+|Çß3*{%, ‡}¤‡'{>5‡'{&’c9¨Š+{06 ‡`>+Š'&‹%6 ‡&’é‘N,‡'‡ Û v Ý |G"…}¤|>ƒ+|1&)9¨5‹ß›‘Š'"%"‡{P&‡/ß©|‘‹‡#&1&½‡©6 ‡#"ßž"Š'"1‘Q,à'5‹…}¤‡5ƒ8‘‡}X‡©"…}¦‡©6 Þ &‡}¡‘l: Š5‹)>5‡'{¤ß(} V‡} ‡9¨Š+{06¥,|1&’}à'Š5‹%>©{(‡#"ßQ&‡ä}¤‹%*Œ‡'{(‡}0*…5à'‡ÈiÉ ‚©Ë¿ vJyä'"…}Q&½‡¦à'5}›‘‡ú­‡#"»}¦‡©6 Þ &‡}Š5‹%>+‡'{¤ß(}©. V ~ }¤‡äà{f5à€ß%, ‡'{(|’}¦‡e*…5{© ~ ª ÏDÿ‡Vß-Z‹:$&Š5{f}eZ‹‡ V‡g)"|V*»Š+‹{¢ ª Ïòÿv Ý ‹LÔ\Š5{f‘{(‡5ƒ%&’‘N, ‡#9¨Š5{6*8ߤ|Š'"M}¤‡H‘N, ‡àV{(|Çßj>©|’S&‡H‘|GC, ‡Š'6 Š5{*ˆ|’}061‡q‘‡&‡}¦ßä‘N,9¨Š5{6 ‡ Þ ŒHdF…v ú'ú/MÞ I-¡Œ Þâ ñ Z‹|*Œ‡'{6 ‡Vß䑇­{f5}(}¤‡#6 Þ &½‡{e&½‡}/‡#"…}¤‡#6 Þ &‡} V‡Vߤ{(‡â ‘$"…}j&½‡‹{(}F|G"ߤ‡{(}¤‡àVߤ|Š'"…} Vâ ñ5vy¦$"…}HdF…v ú'ú/M€ƒ(Z‡}¦ßZ‹%"c*…5{('6[2‡}¦|16*$&/‡'ß1Œ â ñ¾Ï|1"5g)"|½ß%,JPŒñ â vò!J‹|’}0Z‹‡ ŒBÞJ ‡}ß‹)"Ëàfˆ…$6*>5‡à€ß¤Š+{¤|‡#&/ˆ…Š'&Š'6 Š5{*ˆ‡+ƒ›'&½Š+{(}&‡}}ß({¤‹…àVߤ‹{(‡}÷©â ñ ÷â ‡'ß V ñ›}¤Š'"ß,àVŠ/6*%&½‡#4‡}X‘‡ Vâ ñ›‘Š+|Çß-7 ‡'ߤ{(‡B&½‡N>5Š/&½‹)61‡N*{%, ‡}¤‡'{>8$"߇©(+'&½‡}¢‹3"|G>+‡5‹H‘‡N&½‡‹{¡|G"ߤ‡{(}¤‡àVߤ|Š'"¢v$B"R*)&½‹…}*Œ‡}¤‹{(‡ Û }¤Š5|½ßZ‘N, ‡#g)"|‡­}¦‹{&>8{(|, ‡Vß%, ‡¦‘N, ‡9¨Š+{06¥, ‡‡5vSyZ‡+6*'"|†2 ‡'{(‡¦‡#45*%&½|’àV|½ß¤‡+ƒ…àV‡àV|&‡qàfˆ…$6*M‘‡+>5‡à€ß(‡'‹{F$g)"cZ+‹…‡+&6|1"ߤ‡'{*{©2 ‡Vß(‡*…5{e&39¨Š+{06R‹%&½‡}©:³Þ ³]Þâ ñ ÏÕÿ)» Œ‡©Z‹‡©"ß &‡} Œ ‘Q, ‡à'{¤|1>5‡©"ߦ‹%" , ‡©&, ‡#6 ‡#"ß­‘‹Ì9¨|’}¤à'‡‹ú‘‹£(5{(Š5‹)*»‡ ~$HdÈiÉ ' ‚©Ë¿ ø7{ M¦Š2 ‹y{!S{­à'Š'"…}5,‡}¦ßq‹%"Ì9¨|’}(àV‡5‹‡}-*…{%, ‡}¦‡{0>8'"+ßN&½‡}/àfˆ…$6*…}->+‡àVߤŠ5{(|‡#&’}'v‘‡}->5Š'&‹%6Š'&Š'(5|‡ ; ³Xƒ‹%"Y, ‡©&, ‡#6 ‡#"ß/‘‡­àV‡(5{(Š5‹%*Œ‡­‘‡­àVŠ+ˆŠ'6 Š'&Š'(+|½‡¦‡}¦ßF‘N, ‡àV{(|Çß-*…{e&’39¨Š'"…àVߤ|Š'".‘l: Š'"…‘‡yä'"…}-&’1à'Š5ˆŠ/6ì=ÏÕÓ ; 7Þ _§Þ , ‡'|½ße&½‡}/àV{(|Çß©2 ‡'{(‡}\}¤‹|G>8'"+ßf}#fZ‹|sŠ; ³ 7 ìcÏDÿ15‹ß¤{(‡#6 ‡©"+ߣ³ Þ _! "Jö³ " _ÏÕÿ! ³ _"ÏÕÿ)øàV‡Z‹|¢}¤|G(/"|]g»‡Z+‹…‡ 7‡}¤‡'{>88ß(|½Š/"‘‡+>5Š/&½‹)61‡+vìa‡}¦ß/‹%"‡+*{%,!7ì Œ 7; ³\¥ ‡}¦ß-&;: |G"a>8{(|'"…àV‡­‘‡›‰¦5‹%(5‡!‹…}¤‹‡#&1&‡+>6'& Þ &‡*ŒŠ5‹…{\ߤŠ5‹ß¤‡q}¤‡à€ß(|½Š/"1¥ ‘‡-{ »ìEI‡}¦‹%&½ß(ß(}HÑF»v ú%Ô§M€ƒ©|1&¢}#: 8>k2 ‡'{(‡h,!S{'"…$&Š'(+|½‡¦‹54Â{),‘‡!àV‡6 Š‘l2 ‡©&½‡#A—x }¤Š5|½ßN&|, ‡'‡S2R&;: Š Þ ‰‡Vß쇩>©|‘‡©"ß-Z‹‡&R9¨Š'"»à€ß¤|Š'" 7&‹|¢àVŠ+{¤{(‡}0*»Š/"…‘'"+ß/‘‡+&’ àVŠ5{f‘‡!Š5‹)>5‡'{¤ß¤‡+vžU‡'ßFŠ Þ ‰‡'ßj": ‡}¦ß/5‹ß¤{(‡+Z‹‡ Œ‡àV{(|Ç߇Vßj&’1{¤‡©&ߤ|Š'".}#:@,; Þ _ Ó2 ; ³iH#žH‡E@îÌHvç¡M0M@M£»ÏŒHdF…v ú$÷¡MÞ~'ÿ


" GG+-, Ú/. 0+-, Ú1. 0G¢$Þã«GãÞ ŒbÏÞ㢢Þ;¨Þø; Þ œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹+Š'&‹%6 ‡j*{%, ‡}¤‡'{>8$"ßÒj"Â5‹ß¤{(‡/{%,‡©Z‹…ߤ|Š'"…}¡‘‡eÅa‡#Wa&‡9¨Š+{06*8ß(|½Š/"R‘‡-&’­}ß({¤‹»à€ß¤‹…{¤‡/àVŠ/6*%&½‡#4‡\‘…$"…}O&;: ‡}@*»5àV‡›ß©|; Þ&% Ó& HdF…v ú MÓÞ ‡©&½‡­x¦ƒ ( |16*ŒŠ+}¤‡`2 ã&‡}/à'Š'"…‘|Çß(|½Š/"…}/}¦‹…|G>8$"ß(‡}©f‹(' Š2‡}ßj&‡+>5Š/&½‹)61‡­‘l: ‹%"|1>5‡{(}/‘: ‹%"…‡qyRÔ#A Þ {f$"…‡5v¡yä'"…}FàV‡6 Š‘l2‡#"ß ; ³Bã-ÏDÿ‡©Z‹…ߤ|Š'"‘‡+61Š+‹%>5‡©6) &;:8{(|'"+ße*…5{e&Rß({('"…}ÑF| ‰‡#"ˆ©‹|’}vU‡ß¤‡#"»}¦‡‹{ |G"a>8{(|'"ßR‡}¦ß*àVŠ'"»}ß({¤‹|½ß=2\*…5{¦ß(|½{ ‘‡h&’ }¦ß¤{(‹…à€ß(‹{(‡" ‡{¤‡­‡}ßä‹%"Mߤ‡©"…}¤‡'‹{ ì ¥*…{¤‡}0Z‹‡Hà'Š'6*%&‡4‡ì%»lU‡Vߦß(‡R‘‡{0"…|†2à'Š'"ߤ{f|1"ßj*…{j&’à'Š'"…‘|Çß(|½Š/"öì Ï J+Ôq‡'ßä‘Š/"+ß"»&‡}/à'Š'6*ŒŠ+}($"ߤ‡}ëì Þ‹ÌÔ Ë ~ Š+{(‘{(‡!‘‡*»‡{¦ß(‹{ Þ 8ߤ|Š'"M‘‡5ìúàVŠp¦PG"…à'|‘‡#"ß\8>+‡à &½‡}%_ Þ$*%*{(Š84|G6*ߤ|Š'"E&|1"N, ‡5|½{(‡!‘‡A«y¦$"»}FàV‡­à'+}'ƒ5&;:‡}ßÞ "Ï8³"Jö³Þ _" _Z+‹…‡H‘l: $*{©2 ‡}j&½‡Hߤˆ-, ‡Š5{©2 ‡#6 ‡ % !¡‡©"{(Š+}¤‡5ƒ Ý ß(|GWˆ5Aª„F|½ß(àfˆ|1"5AŽ©|1"%(5‡{D)—ƒO›GšE—Ÿ;•£¥) #†¥5•0˜¡·k•SÐ ˜—#»¥)˜S Ÿ¤X©·k›1ŨÏa˜sìÑZŠß(Š'"…}˜‰ó'•0š É ›GẼ—£–˜#j—#˜¥5›1˜©S˜¡­j—£–z« Ï ÿk’y32 ‡}&½Š+{(}ƒ&;:ߤ‡©6*»}O": ‡}ß‘‹Rß(‡#"…}¤‡'‹…{w‘‡eÅa‡#Wa&ŸviâÀ&»}©:5‹ß¤{(‡+Z‹‡+&†: 5àVߤ|Š'"Ó ¿ Ø vû‘¢} qHdF…v ÷'ÿ¡MV ¢#y qŠ2‡'ߤ{(|1Z‹‡ä‡#"ߤŠ5‹ß\}¤‡}›|1"…‘|’àV‡}\‘‡!}0*|G"ÌH†ú%ø0ÿ/Mw‡VßeZ‹|¢àVŠ+{¤{(‡}0*ŒŠ'"…‘ ‡©"‡}ß-&½‡­àfˆ…'6*‘‡!u¢Š5{(‡#"ßôä}0W¡6¥,‹ V ¢#y q*»{¤ß¤|‡!5‹eàfˆ»$6*eß


ÍÞ Ï¢‹('f =‡}ßj&’ àVŠ+‹{ Þ ‡­‘‡+6 Š‘‹%&‡AØM‡Vߣ¢v¢GG¢G4¢Ô ¥:¨ ¬§® ¢[¥ ¢ ¢I¨ ¢ ¬h ¢ ® qFUGG$ãøœžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹|1"…}¦ß('"+ß(Š'"…}vSuB: 5àVߤ|Š'"e{%,Ôú V M HdF…v ÷%Ô8MÓ ¿ Ø vH V ¢} qû‘¢} q JZ‹|¢‘Š'"%"‡‡©Z‹|1>8$&‡#"ߤ‡S2*àV‡©&G&‡­‘‡+&’3(+{(8>©|½ß%, ‡¦àVŠ/"59¨Š5{6 ‡!'*{©2 ‡}S, ‡#&|161|1"…ߤ|Š'".‘‡ V‡}¦ß?,Ôú,û ¢} q M£»HdF…v ÷/ú/MÓ ¿ Ø vHvû¢#y q‡©Z‹…8ß(|½Š/"…}X}¦‹|1>8$"ߤ‡}#f‹%" ‘‡}${), ‡}¦‹%&½ß(ß(}V*{(|G"…à'|G*»‹54H‘N, ‡‘‹|½ßS‘‡wß(Š5‹ßJà'‡NZ‹|5*{), ‡à$2 ‡‘‡w}¤‡{%, ‡}¤‹%6 ‡‘…$"…}i&‡}O÷c,B"%g)"¢ƒX&;:ZhÏwZ ; ³OHžH´E¥î£HgçaM@MÔ5vI¦ÏŒ 2 ;Þ(_³iH#žH‡E@îÌHvç¡M0M@M£»; úvÂÓÞ÷vÂÓ ;Þã³iHžH´E¥î£HgçaM@M ; ‡'ߦߤ‡©"ßR‘‡Â{(‡#&|½‡{ƒ¡‘'"…}+&;: ‡}@*…+àV‡ ß©|Çß%, ‡äà'Š'"59¨Š+{06 ‡-Î ÏÕF…ƒ%ß ÏÕF*}#: 5‘'*ߤ‡©"ß{%,‡}¤‹%&½ß(ß(}q‘‡}L, ‡©Z‹…8ß(|½Š/"…}`HdF…v1Ô~/M­‡Vß[HdF…v ú$ÿ¡M=, ‡#"Š/"äR à''"ßR&½‡}Hàfˆ…'6*…}H‘‡?&†: ‡}0*…+àV‡A>ߤ‡©6*…}H‘N, ‡àV{(|½ß(}+*…5{R&‡}8>5‡à&½‡}q{%,}¤‹%*»‡{(àfˆ»$6*…}ß5‡RZ‹…+‘{(ߤ|bZ‹‡!‡#"Ëã6ƒ…‘Š'"%"…'"ßj&½‡­ß(‡'{6 ‡ Vú'MNâÀ"ß%,‡àV{(|Çß }©:


¢GÓ6¢@?¢‡©>6'&½‹-, ‡'‡¦‡#".‹ß(|G&|}($"ßߤ‡©&Z‹‡ÓpÞ­‡}ß[,¢ ø0Ó@0 ¨ ¢ ø0Ó ¤ ¥žMPÏ H†Ó6 ¢ øÓ6 £UØ ¨ ¢ ø0Óm £(¢ ¥žMHdÓmÚ‘Š/"+ß ô ¼ i¢;¢¨¢¢¢¢ ³ä¨¨ÏGGG¢¢¢¨¢¢v¢¢¢ Hv ÂøDØ£ Âø7¢6 ­M ¨¢¢¢ » HdF…v F/ÿ¡M¨M¢ 2 *»{¤ß¤|{U‘‡ HÑF»v F­Ô§M€ƒ |G& ‡}¦ßU|1"…‘|’}0*»‡©"…}¤ Þ &‡ ‘l: ‡#6*%&ŠXW5‡{é&’P‘‡}¤à'{¤|1*ߤ|Š'" ‘‡}¨¢¨ ¢ ‘N, ‹œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹©|’&½‡}[, ‡§Z+‹»8ߤ|Š'"»}|1"+ß(‡'{*{%,ß8{(| Þ &‡}/}¤Š'"ß/{%, ‡‡#&1&½‡}'vŠ2'"|†2 ‡'{(‡Z‡#4a*)&½|’àV|½ß¤‡+ƒU¦yä‡+6‡àV{(|½ß H*ØÂø‰¢/M/}©:


¢¢¢¨/ :ù!S{N&’*}¦‹|½ß¤‡+ƒ ¢¢¢¢¥v¢³¥ ¢ ³l¢GϳML ¦GG; ¢]¨ ¢ ¥ ¦ I¢: ¢¢¢bÏM¢¢ ³r¨¢¢ H ¦¢G¢¢ Ï{s£Hn ~ ¨~L¢ +¢aH´³ ¥ ¢ ÓmbJ¨Gˆ¢~ n~¢¨M ¦ n ¨ ¢¢M ¦ n ¨ ¢M¢¢¢7¢¢=7øDùMúMM¢I»¢¢¡M\‹Màfˆ…$6*.‘‹…'& ¦¢œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹8{(|'"+ß\‘‡qu¢Š+{¤‡©"ß0ô5v‡§Z‹‡#"ß !S{\àVŠ'"»}5,¢¡» HdF…v F¡ú/M¨¨ ¢ ³uB: |1"¡>85{¤|’$"…à'‡¦‘‡‰¦5‹%(5‡‡‘‹|½ßÞ Ï8³ Þ Û *Œ‡'{61‡'ßZ‘‡­àfˆŠ5|’}¤|½{e&’䉦‹%(+‡5v Ý |1"…}¤|¢|G&$}¦‡­‘N,{ © ¥:¨§ª 7 M¼¢~7 Ú' I ¤ ¥ 7 H¢ ¨õ ˆ ¥© ¥:¨§ªô ¼ i¢ Hy MPÏ JsIDH ¤I ¤ ¿ 7M£» HdF…v F/÷¡MI H ¤ ‚ M ¥ 7¥¼¨Úˆ¼¤¥ À âÀ&¢‡}¦ßFà&’|{jZ‹‡SHÑF…v ÷ M|G6*%&|1Z‹‡¹'’e‘A ˜Ï·k•0˜—Ÿ—Ÿ–d¤$¡‘¹/˜HdF…v F'FaMøù¢¡Hy øUØ£ Âø7¢@ kM ¨¢ Jq³r¨¥ Hy øUØ£ Âø7¢@ kM@M» HdF…v F M‡}¤‹%&Çßf8ß2cHÑF»vGÔ8Mw‡'ßRHÑF»vGÔ/Ô§M‡VßF‡©"‹ß(|G&|’}¤'"+ßj&‡9¨|½ßZ‹‡B".àVŠ/6*…{f$"ß/àV‡!{%,¥ Ï"sŸH¥ ø HdF…v F M&;: ‡#4a*…{¤‡}¤}¤|½Š/"‘‡{%, ‡}¦‹)&Çßf$"ߤ‡*{(‡#"…‘?&39¨Š+{06 ‡ ¢ ¥H}MBÏ 7¥ô iˆ ¥¤ ¨ 7 { © ¥:¨§ª I ¤ ¨ 7 ” w ¥:¨ x HdF…v F 5 MI’MN O>PKQIéH ¤IDH ¤© ¥:¨§ª 7©¥:¨§ªˆ ¥¼ i¼ iIH ¤ ‚ M ¥ 7­ ¨I ¤ ¿ 2;L ¼s~ IDH ¤ ‚ M ¨¼s~ ¥b ¼ Úˆ ¥˜#³˜#•£©–„¡kšX;–d¤X¡ˆ¹'˜˜©’-‘šι¡ºÀ âÀ&s{¤‡}ß(‡`21‘N, ‡Vß(‡'{61|1"‡{¢¢vB!¡Š+‹{/àV‡Vߤߤ‡!{f|’}¦Š/"¢ƒ©|G&l9¨5‹ßF‹ß(|G&|}¤‡'{/‡©"…àVŠ+{¤‡¦‹%"‡9¨Š+|}HÑF»v ÷ MQZ‹|¢{(‡#&|½‡¢[25¦ }¦‹|1>8$"ßj&’R{(‡#&’8ß(|½Š/"Ó6 ¢¢ IA¢ ¥ ¢¨¢Ïw¦¢@?Ø£¢ ¨ ¢ Ó6¥A » HdF…v F ^ M{-© ¥:¨«ª 7¼s~ˆ ¥© ¥:¨«ªMPÏ" £¢5H 7 Ú I ¤ ¥ 7¢¡Hyô ¼ iIÕH ¤H ¤ ‚ M ¥ 7bI¼¥Úˆ7 L ¼Œ ¼¤I ¤ ¿MÂI »1»1»F@~¡M HdF…v‹Ì»1»G»…‘N, ‡#*Œ‡#"…‘‡#"ßF‘‡}/àfˆ…'6*…}eZ‹|$*)*…{f|’}¤}¤‡#"ß/‘$"»}Š2¢­‡Vߥ »‡©Z‹…ߤ|Š'"…}+HdF…v F/÷/MVƒHXWYWaM›‡'ßRHÑF…v F@~¡MQ*»‡{06 ‡Vߤߤ‡©"+ߦ‘‡q‘N, ‡àV{(|½{(‡B".àVŠ'"»à&‹…}¦|Š'"ƒ5&½‡}E,J ‡©"ß(‡'{6 ‡q‘‡}Fà'Š'6*ŒŠ+}($"ߤ‡}‡©>5‡#&Š'*)*»‡©61‡©"ß/‘‡'¦‘‹.‘N,HgØÂø7¢¡M£»‡#"ß-&½‡­}¤‹%*Œ‡'{fàfˆ…$6*.$"ߤ|’àfˆ|{('& ¦ HgØI§¢ ; ¢%ø ;‡#($&‡#6; M Þ¡H‡#6 ‡*{¤Šàt, ‡‘‹…{¤‡¦{¤‡©&½|‡\,U‡Vߦß(‡+6[7 Þ ¦øù/ ù ò Sø ú ùø RD ùòf ùH©7 Øiø ; ¢/MBÏ ¦Ó ;H ; l ÞHdF…v ÿ¡MÓ ; ¨¢ H ;Vø ;0žø ;0žø ;¤ MVIé7¤ M HdF…v Ô8M¢aH ;ø ; ¨¨ ¢ ¦ ØB¢;ø ; 0žø ; ¤ M0M ¥-H‹ 7 Ø ¢ ¨ ¢ ÏØ ¢ ¨ ¢ I§¢ ¢ ¥ ; ¢ ¨Š2~$F


,¢¢ Œ ¢ H}MBÏ 7¢¦, Ú1. 0‡}¤‹%&½ß(ß+HdF…v ú/M-Z+‹…|‡45*{(|161‡R&½‡Hàfˆ»$6* Œ Þ ‡#".ß(‡'{6 ‡H‘‡ ¦IéH ¤ ‚ M ¨ùôMM¢WV«ÿ ÆMXÇ‡È ÂZY ÈM[§È ÿ¥¤ Ù\ˆ‡H‹ß¤ˆŠ+{(}-_›Š5‹%&’‘[&½|1á5‡qߤŠ ߤˆ»$"%áhÆ­v Ý ˆ%&$u$5$6*5{( $"»‘ Ý v»x›‡#&ˆ…‰Q9¨Š+{F‡5{0&|½‡{FàVŠ/&G&’ Þ AUߦß(‡'{vBÙ\ˆ‡#W'&}¤Š1ߤˆ…'"%áE!w{¤Š5ß({f}Qâ@â@â/y3Ô§ú§\$ú U©ÑÆY9¨Š5{F}¤‹%*%*ŒŠ5{¤ßvŠ5{f8ß(|½Š/"Š/"ß(ˆ|}j6—Mi¢M¢œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹8'"+ß(‡}IÕH ¤ ‚ M ¥ô2; i ˆ ¥ ¼T ¼Œ I ¤ ¥ 2;IÕH ¤¤ ¿ 2 ; Ú I; { 2b ¼ Úˆ ¥© ¥:¨«ª¼s~ © ¥:¨«ªI ¤ ¨ 7’ © ¨ ¬ ª” w ¥:¨ x 2;Œ>¥NH}MBÏ 7­¨¤ ¨7 L ~ IÕH ¤ IIDH ¤b Úˆ ¥~ ¥ˆ ¥© ¥:¨§ª2 ;¥:¨§ª2;b¨i ˆ ©IDH ¤ ‚ M ¥ˆ ¥ i2;{ © ¥:¨«ª I ¤ ¿ 2;¢ H}MBÏ8³OH©7 T I ¤¥ 2;bŒ]¨¥ˆÚ; { © ¥:¨§ª 2~IDH ¤ ‚ M ¥IDH ¤¤ ¿ 2 ; Ú MÂIÕ»#»#» Iô i2;¥ ˆ©¥:¨§ª‹¯³ ¢ Ï ÷ Š2¢ Ï ÷ ƒU³ä¨‡#*Œ‡#"…‘e‘‡}Zàfˆ…$6*…}eZ‹|¢'*%*…{f|’}(}¦‡©"+ß/‘'"…}£Œ ¢ ‡'ßMŒ ¥ v‡Vß»#»©»¤‘N,÷ / ø÷ú ù‡©"+ß/‘N, ‡©6 Š'"ߤ{%, ‡jZ‹‡&‡}\àfˆ…'6*…}‘‡&;: ‡}@*…+àV‡#AŸß(‡#6*…}›‘Q, ‡à'{¤|½ß(}Q*…5{Q&‡}›}¦‹)*»‡{(àfˆ…'6*…}Jß+Š'"…}Bg)"…$&‡#6‡Vßf8ß(} *âW}¦|bZ‹‡}!‘‡&’Âß(ˆ-, ‡Š5{(|½‡1‘‡1}¤‹%*Œ‡'{(5{f8>©|Çß%, ‡HàVŠ/"59¨Š5{6 ‡ Î Ï Fhß ÏéF…vRyä‡ àV‡9¨|½ßƒ{(|½‡©&}¦àVŠ/"+ß(|½‡©"%"‡©"+ßq‘‡}=,"Š+‹…}\{¤‡'ߤ{(Š5‹%>+Š'"…}‹%"‡q5‹ß¤{(‡­àVŠ/"5g…{6 ߤ|Š'"e‘‡+&1{(‡#&’8ß(|½Š/"‡©"+ß({¤‡&;: ‡}@*»5àV‡¦ß8{(|, ‡Vß%, ‡}¢‘‡-U›$&’ Þ |]A;Êw‹q‘‹»$&1&½‡}Â*…5{V&’Z}0Wa6K, ‡Vß({¤|‡ž6 |½{(Š5|{iZ‹|©àVŠ/"…}¦ß¤|½ß¤‹‡©"+ß8>5Š/"…}l*{),‡'{©7 ‡Vß*…{¤ß¤|’àV‹)&½|‡'{q‘'"…}&;:+‡'{¤ß¤‡\ßP,‡}¤‹%&½ß(ßàVŠ/"…}ß(|Çß(‹‡¦‹%"‡"Š5‹5AÏÕF…vPU‡Z{%,‡F|1"…8ߤߤ‡©"…‘‹‡*‹|’}Z+‹…‡¦à': ‡}ße&’Rà'Š5{f‘‡ZŠ+‹%>5‡{¦ß(‡Zß5‡‹ß%,}¦‹)*»‡{0(+{(8>©|½ß%, ‡¦àVŠ'"%9¨Š5{61‡-ÎÐÏF=2SZ‹…8ß({¤‡­‘|16 ‡#"…}¤|Š'"…}vàV‡‹54‘‡+&’W$"ߤ‰‹%(ž, ‡›|1"+ß%, ‡'{(‡}(}($"ßX‘l: $&1&‡'{J‹*‘‡©&³2!‘‡e&;: $"…'&GW}¤‡›‘‡}SŠ'*-, ‡'{f8ß(‡'‹{f}>+‡'{¤ß¤‡41‘$"…}O&‡F}¤‹%*»‡{¤‡}@*»5àV‡-ÈiÉ ‚§Ë¿ øÝÞ Š'"%"…‡S*…{¤ß¤|‡*‘‡*àV‡VßH5{¦ß(|à#&½‡*}©: ‡}ßq‘‡}ß(|G"-, ‡'‡=2e|G"ß(‡'{*{%, ‡Vߤ‡{¦à'‡Â}@*Œ‡àVߤ{(‡ Š Þ ß(‡#"©‹a‘…$"…}&;: ‡}@*»5àV‡#AŸß(‡#6*…}AT ~€ ‚ »‹%"‡|’‘Q, ‡'‡Z‹|¡`*Œ‡'{6 |}¦àV‡R*…5}(}¤'(5‡q}¤‡H{%, ‡}¤‹%6 ‡q‘'"…}&‡9¨5|ÇßZ‹: ‹)"‡9¨Š/"…à€ß(|½Š/" ‘‡Rà'Š©Š5{f‘Š'"%"-, ‡'‡}\ˆŠ'6 Š'()2 ‡#"…‡}ÞR‘‡uB:‡.õ ‘‡&†: ‡}0*…+àV‡1ß


—i ÖZ±©È±eM¸V´¦âRÃ(JÖF±Mcj]´€±§eÃf¹¿¹ÇÆfÏÊXæJ³¿Æ—Ä´ ÃfÄè­Ã(ª´ªÎRºwÀ©âHФ³ÇÃ(¹½â1Ê¡ÌÃfÄ'´ªÎ ——m Ê›±MlZ±ÅX¸³¿¹¿Ã(Î Ã(wÖF±K]+±Å¡¸'Æ—ÄÀ…ϩÀâ5Æf±ÇÏÊXæJ³¿Æ—Ä´Æ&g>ÀS ¸VÌÃfèq¸V³¿¾fÆwºÀ©âHØ¢ÌÎ5ƪ³¿¾fÆfÏ©Ò»ÄÀâ+ÄÀ —ȱ5ºF±Ù›·ÍÍÃ(ª¡Ã(po!±5ؤ±+ÊŒÄâ…Ï5ºwÀ@gŸÀª´Äâ5·¾(³¿ÄÀRÊ»Ä!ÊXæJ³¿Æ—ÄẃÊ¡Ì+ÃfÄ´ªÎ8Ï+Ò»ÄÀ©â5ÄÀ ¸'ÌÃ(èq¸'³¿¾€¸¹»ÈÄ8¾f³¿Ã(>ÎRÈ8·©â+Ã(ÀXÊŒÃ(êÆn—q ÖZ±Ø¤ÃfÀ5´ÄƪÃÏ©ÊXæJ³ÇÆ—Ä'´7r›·¸À8³ts€¸'³¿ÄÀ*ºÀâ ÁX·+´áÀâ ÈÑ©¸¾(Ã(³¿è­ÃÏg>À8€±K]+±©Ê¡Ì+ÃfÄ´€±©ØÌ8Î5Æf± ;_G` ^ šỒ ±—¸Æ—ä´›Ä'¼È5¾f³¿ÃfÀ+¾fÃÏ ;_WaG`´Ã€ç8·³µ´Ãfè­ÃfÀ8S¼Ä´JÌÃ/â5ÃfÍ´Ã(ÃZÄ'¼Â¸¾€Á¡¸¹¿¹¿·èPÃ(1ÖF±XؤÃfÀ+´ÄƪÃÏÊXæJ³ÇÆ—Ä'´*Ê¡Ì+ÃfÄ´ªÎ ë ºÀ ºwÑÑ+´Ä¸¾¦ÌúÊŒÄéÊ¡Ì+Ãwr›·©¸À8³ts€¸V³ÇÄÀxc¼ZФ³¿Ãf¹Çâ+ÆRºÀâôªôk±SºF±¡Ùä±O;_Wa f š f'ù ±—;_ ^ dš °°°h5ºF±'Ù›Äâ+ÍÃfÆfÏ8ÊXæJ³¿Æ—Ä'´—;_ ^©f šM b a Ï…ÊXæJ³ÇÆ—Ä'´6y\³Ç¸Í'´¦¸è­Æfϳ¿ÀÊ¡Ìà ÃfÄè­Ã(ª´³¿¾ZÕ›À+³ÇáÃ(´ÆªÃ ë Ⱦf³¿ÃfÀ¾fÃÏ ÃfÄè­Ã(ª´ªÎ8ÏÊ¡Ì×—m ÐX±SÁ¡¸'¾¤ÌĢsfÄMäHØs± Èá´¾(Ã-zÏSÒ»Ã(¾(·+´ÃfÆ­ÄÀúÊXæJ³¿Æ—Ä´ Ȫ´³¿À+ÍÆH¸Àâ ؤÃ(´ª·5´Ó©¸V³ÇáÃ1ý¡¸'ÀÍ'׳1³¿¹¿¹¿ÆHÊ¡ÌÃfÄ'´ªÎ8ÏJؤÄ8ÈéÖ$Ê~l‚f dGd bôf dGd b šƒdWd ù+ÏÌ+ÃfÑ+×¨Ì©Ô d b d ù ;_ ù—l䱤ÅXÃ(´{z6ÄVá³µÆfÏ¢ºÀeº¹µÃ(´À©¸V³¿á6ÃHȪ´³¿ÀÍ Ê¡ÌÃfÄ'´ªÎ.³¿ÀMÊXæJ³¿Æ—Ä´­È5Ñ©¸'¾fÃ!¼’Ä'´„l\÷›ùÂÈ5·+ÑÃ(´ª×¨ý¡¸'ÀÍ×;1³¿¹¿¹¿ÆfÏsØ¢ÌÎ5Æf± ÖwÃfá± Ò…Ã(ª€± _ °Ì³mf dGd ù šƒd[;`Gd[ ÏÌ+ÃfÑ+×¨Ì©Ô d ù d f d ù9b—Ì î lä± l›Ã}z´¦¸'ƪÄVáÏ5ÙZ±Mc›ÄÍ·5´³…äwÁ±5²X¸V¼"¸5Ï©Èןâ5·©¸¹¿³µ>έ¸À©âHÊŒÄÑĹ¿ÄͳǾf¸¹¢È8ª´³ÇÀ+ÍÆfÏK]Ù›Ú$Ø d ù ;d——————dWd ù špdWdW_ ÏÌ+ÃfÑ+×¨Ì©Ô d ù d ° ;`©a ±fo!±ÙÄ´³"ϺF±§o/¹¿Ãfè­è ÏÖF±Øs¸À©â5Ì©¸V´³ÇѸÀâ+ÃÏ+ÖZ±+Ê¡Ì+Äèq¸ÆfÏÁ±²$¸'¼’¸5ÏÖZ± ²$Ģz³¿¹½¸'À©âHÃ(ÚJ±M‡…¸'ƪ¹¿ÄVæ/Ï%1³¿´ª´Ä'´/È8Î5è­è­Ã(ª´ªÎ8Ï8á6Ũ±Ì†kļ$ÁX¹Ç¸fÎ ¸'ÌÃ(èq¸'³¿¾fÆQ1ÄÀ+ÄÍ´¦¸'ÑÌ+ÆfϺ-*ÈÏ©Øs´ÄVá³½â5ÃfÀ¾(ÃÏÖgpf dWd °5±Ì†q Á±©²$¸'¼’¸5ÏÊŒÄÑĹÇÄͳ¿¾€¸'¹Â1³¿´ª´Ä'´Æ›¸'À©âr›·©¸À8·+è Ö³¿À+ÍÆfÏ+³ÇÀ ȱ Ê›±ýX¸·ÌÃfÑ5×¨Ì©Ô d ù f'°5±d[;aÌ È± ³¿Äè¦Ó+³"Ï>c´Ó+³¿¼Ä¹Çâ+³¿À+ͦÌÃ\ŸæJ³ÇÆ—Ä'ẃÆ—ª´³ÇÀ+Í+ϧl›·¾(¹¨±Ø¢ÌÎ5Æf±+Å aÔ-_î°'× ˆad Ì+ÃfÑ+×¨Ì©Ô d ù d bGf° ` ±;`q Á± ןÙZ±ºÌÀ…Ïj1³µ´ª´Ä´Æ—Î5è­è­Ã(ª´ªÎaļqÁ¡¸'¹Ç¸Ó³µ×¨ýX¸·¯Æª·+ÑÃ(´èq¸À+³µ¼’ĹÇâ5ÆfÏ1Äâ…±›ØÌ8Î5Æf±wһ䪀±›º’f dî—————————œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹ÎRØs´ÃfƪÆfÏ -_ ^ `š ±²$Ĺ¿·è­Ã6f8ÏÈѳ¿ÀÄ'ẃºÀ©âRÊXæJ³¿Æ—Ä´-1Ã(ÌÄâ5Æ&gŸÀ*È5Ѹ¾fä³Çè­Ã ÃfÄè­Ã(ª´ªÎð ÖZ± ؤÃfÀ5´ÄƪÃWh©Ê¡ÌÃ\ŸæJ³ÇÆ—Ä'´›Ñ5´ÄÍ'´¦¸è Ï+ÖÃfÑ5€± ¸'Ì»± Ø¢Ì8Î+Æf± fÏ ` b;_WaWašk ұؤ±Ù·ÍÌ+Æ—ÄÀ…Ï©ÊXæJ³¿Æ—ĴƸÀâ Øs¸'´ª³¿¾f¹¿ÃfÆfÏ©Ò»Ãf¾¤·+´Ã7l›Ä'ÃfÆJ³¿À Ø¢ÌÎ5ƪ³¿¾fÆ _©aÁ¡¸'è¦Ó+´³Çâ5ÍÃ/Õ›À+³¿á6Ã(´Æª³µ>ÎHØs´ÃfƪÆfÏ -_W_ÔŠš ±ÈÑ+´³¿ÀÍÃ(´ª×¨²$ä´¹½¸'Í+ÏÅXÃ(´¹¿³¿À»Ï ;_ ^ _š ±ù+± ¸'Ì+Ãfèq¸'³¿¾€¸'¹¢È5Ä8¾(³Çä>ÎRÒ»Ãf¾¤·+´Ã7l›Ä'ÃfÆÈ5Ã(´³¿ÃfÆ b `;_G_Wdš ±ô m ÖZ±u]ÄWsfƸϡºwÑÑ+¹Ç³¿¾€¸V³¿ÄÀÆäļwƪÌÃf¸'¼w¾fÄÌ+Äè­Ä¹ÇÄÍÎM³¿ÀM>æJ³¿Æ—Ä'´€ÏXº â5³¿ÆªÆªÃ(´ª¦¸'³¿ÄÀ ƪ·Ó+è­³µªÃ€âe³¿ÀMÑ©¸V´ª³½¸'¹¡¼’·+¹tv¹¿è­ÃfÀ8äÄ'¼JÌÃÈ5Ѹ¾fÃ(ןʡ³¿è­ÃÏØÌ8Î5Æf±+ÖÃfÑ5€± ` Áô Ì ºZ±Ø¤±ÙÄâ+ÍÃfÆ¢Ã(Xȱ'Ù›·+ÍÍ䪀Ï8ÊXæJ³¿Æ—ÄŹy\³Ç¸Í'´¦¸è­ÆfÏ6È5·+´áÃ(Î5Æ¥g>ÀäÙ›³¿Í̦ÚÀ+Ã(´ÍΦØÌ8Î5ƪ³¿¾fÆy\³Ç¸Í'´¦¸è­ÆfÏ»ØÌ8Î5ƪ³¿¾€¸ ù6ºe.Ä'´{z|c¼ŒÖÄÍÃ(´Ø¤ÃfÀ5´ÄƪÃÏÀâ+Æf±©È±ºZ±5Ù›·+ÍÍÃ(ªwÃ(€±©¸'¹"±cê¼’Ä'´¦âRÕ›À+³¿á6Ã(´Æª³µ>ÎHØs´ÃfƪÆfÏ -_W_ ^ š ±ô¨î ±'ÐX±Vº¡³µÎ¸Ì…Ï ÃfÄè­Ã(ª´ªÎ6c¼ ¸'·ÍáФ³¿Ãf¹Çâ+ÆfÏÒ»Ã}sf³¿ÄÀ³Ð+Ã(´è­³Ç¸Àú¾f¸â+Ã(è­³½¸~l›Ģsf³¿ÄÀ©¸¹¿ÃXâ+Ãf³8Ò…³¿À¾fÃf³¸'À©âZÈ5¾(·Ä¹Ç¸pl›Ä'´èq¸¹¿ÃÈ5·+ÑÃ(´³¿Ä´Ãϩس¿Æ¸5Ï ;_WaG_š ±ô¨ð ºZ±ÐÃ(´ÓÃ(´€Ï©È5·+ÑÃ(´ª>æJ³¿Æ—Ä'´ÆºwÀ©â1ÁXÄÀ+¼Ä´èq¸'¹¤È·ÑÃ(´Æ—Î5è­è­Ã(ª´ªÎ8Ï9l·¾f¹"±+Ø¢ÌÎ5Æf±Å °Wf;_Wa ^ š bGb±ô i Á±©ÖF±+Ò»ÃfÓ5´·Àh ÃfÄè­Ã(ª´ªÎHļÊXæJ³¿Æ—Ĵƪѩ¸'¾fÃfÆfÏÈ5³¿è­ÄÀ+Æ¡æ¡Ä´{zƪÌ+ÄÑ ¹¿Ãf¾(·5´ÃWh€—·¹µÎ|f dGd ùô k ÚJ±e³µªÃfÀ…Ïؤä´ª·+´Ó©¸V³¿á6à ¸·+ÍÃFÊ¡Ì+ÃfÄ´ªÎR¸Æw¸¦Èª´³¿ÀͦʡÌÃfÄ'´ªÎR³¿À1ÊXæJ³ÇÆ—Ä'´/ÈÑ©¸'¾fÃÏ©ÁXÄè­è¦·+À»± ¸'Ì…± ØÌ8Î5Æf±KfWbGff dGd ù š ^ _ × fWb^5ÏÌ+ÃfÑ+רÌÔ d ° f ;aÔ ±ô n ȱ'ÈÃ(̳"ÏÈ·ÑÃ(´èq¸'À³µ¼’Ĺ½â5ÆfÏVÖ³¿Í³½âj ¸'À³µ¼’ĹÇâ+Æ¢¸Àâj1³µ´ª´Ä´¢È8Î5è­è­Ã(ª´ªÎ8Ï}l·¾f¹"± Ø¢ÌÎ5Æf±fÅXù6° d-_W_ ù š ° × b d Ï'Ì+ÃfÑ+רÌÔ _ ù d ù ^ `ô q ±¥e.ŵ¼>Ï¢ÊXæJ³¿Æ—Ä´ ÃfÄè­Ã(ª´ªÎ¸'À©â ¸'·ÍÃqÌÃ(Ä´ªÎ8Ϥ¹¿Ãf¾¤·+´ÃqÀÄ'ÃfÆ ë gŸÀ8ª´Äâ+·¾¤Ä´ªÎ¾fÄ·5´ÆªÃRÄÀe>æJ³¿Æ—Ä'´¦ÍÃfÄè­Ã(ª´ªÎe¸ÀâÍ6¸'·ÍÃ/Ì+ÃfÄ´ªÎHͳ¿á6Ã(À*¸'JÌÃe1Ä⸀á6ÃZÈ5·+è­è­Ã(´JȾ¤Ì+Ä8Ĺ…ÄÀ èq¸VÌÃfèq¸V³¿¾€¸¹ÑÌ8Î5ƪ³¿¾fÆS³¿À]·+ÀÃf dGd bf dWd ù špd b ` ÏÌ+ÃfÑ+×¨Ì©Ô d ù d ° ^ aÌ ô…lä±ÅXÃ(´{z6ÄVá³µÆJÃ(wÒ±a1Ĺ"ÏÁX·+Ó³¿¾FÊXæJ³¿Æ—Ä´³Ç¸¹È8ª´³¿ÀÍ!Ф³¿Ãf¹Çâ Ê¡Ì+ÃfÄ´ªÎ8ϧ]Ù›Ú$Ø d ù d ùÌïÌ lä± ÅXÃ(´{zÄ'á³µÆeäMÚS± eú³¿ªÃ(À»ÏHÁXÄÀ+¼’Ä'´èq¸¹HÈ·ÑÃ(´Í'´¦¸Vá³µ>ί³¿À >æJ³¿Æ—Ä´ª×¨Æ—ª´³¿ÀÍ ÌÃfÄ'´ªÎ8ÏC]Ù›Ú$Ød ù d ^f dWd ù š(dWdG_ ÏqÌ+ÃfÑ+×ÌÔ d ù dGẀd b ±ð g¤±Å¡¸'´ÆfÏ©ÊXæJ³¿Æ—Ä´Æ\¸'À©âfÊXרÑÌ8Î5ƪ³¿¾fÆ-hÌ+ÃfÑ+×¨Ì©Ô d b d f dW` b8±Ì±ºwÍ6¸À¸Í³Ç¾FÃ(Á±²X¸V¼"¸Ï51³µ´ª´Ä´\ÈÎ5è­è­Ã(ª´ªÎ!¸Àâ*È5·+ÑÃ(´èq¸À+³¿¼Ä¹Çâ+ÆfÏ5Ì+ÃfÑ+רÌÔ d ù d ° -_ f8±Ìim o!±ÙÄ´³…Ã(›Á±²$¸'¼’¸5Ïa1³µ´ª´Ä´Æ—Î+è­è­Ã¤ª´ªÎ8Ï8ÌÃfÑ5רÌÔ dGdWd fWfWfGf±ÌºZ± l›Ã(³¿{s}z6Ã/Ã(Á± ²X¸V¼"¸ÏMl/÷7fZÆ—ª´³ÇÀ+ÍÆw¸ÀâRÌÃ>æJ³¿Æ—Ä´³Ç¸¹sÁ¡¸¹Ç¸Ó+³¿×¨ýX¸·…ÏÌ+ÃfÑ+רÌÔ d ù d f fG^̆nÀⅱ š ë 1³µ´ª´Ä´Æ—Î5è­è­Ã(ª´ªÎg¤Ï f Ì+ÃfÑ+רÌÔ _[;dÔˆa ±î m ±ºwÍ6¸À¸Í³Ç¾FÃ(Á±²X¸V¼"¸Ï©Ø¤Ã(´ª·5´Ó©¸V³ÇáÃ5y\Ã(´³¿á¸'³¿ÄÀ Ä'¼V1³µ´ª´Ä´›ÈÎ5è­è­Ã(ª´ªÎ8Ï8Ì+ÃfÑ+×¨Ì©Ô d f dW_Ô °W^±î(ô]+±Øs¸'´{z­Ã(wÈ8׉]+±5ÖÃ(Î8Ï+È·ÑÃ(´ªŸæJ³¿Æ—ÄŚÄ'´Ó³µ¼’Ĺ½â5Æ ëž ¸'·ÍÛÌÃfÄ'´ªÎH¸è­Ñ+¹Ç³µ·â+ÃfÆS¸'À©â­ÄÑĹÇÄͳ¿¾€¸'¹ŒÆ—ª´³¿ÀÍÆfÏM]Ù›Ú$Ø ff dGd ù šf dGd b š f°'ù× fWbGf¦ÌÃfÑ5ÌÔ d ù ;aÔ ±Œ èq¸'ÀÀ»Ï©Ã(±Še.Ĺ¿¼ŸÏ¤Ê¡Ì+ÃäÄÑĹ¿ÄͳǾf¸¹¡Åןè­Äâ5Ãf¹¤ÄÀeḨè­³ÇÀ+³µ×ŸÆª·+ÑÃ(´ª>æJ³¿Æ—Ä'´\ƪѩ¸'¾fí¸'À©âÂƪ·ÑÃ(´Æ—Î5è!×îªî¢ºZ±Šy¦±…ؤÄÑÄVáÏŒÁ±ŒÈ‹è­Ã(ª´³¿¾FÅXÄÍÄè­Ä¹ÇÀ8ÎRè­ÄÀ+ÄÑĹ¿Ã\Ãfç·¸'³¿ÄÀ+ÆfÏ>]Ù›Ú$Ø ;df dWd b šƒd bG^Ï©Ì+ÃfÑ+רÌÔ d b d b -`[ ±îªð ±5ؤĹ¿³¿¾€¸Æ—ª´ÄhÊXæJ³¿Æ—Ä´ÈÑ©¸'¾fÃfÆfÏ/1³µ´ª´ÄẃÈ8Î5è­è­Ã(ª´ªÎä¸À©âHÈÃf¹µ¼×Ÿâ+·¸¹MoF¸'̹¿Ã(Ṕ ¸'À³µ¼’Ĺ½â5ÆfÏÁXÄÀ8ª´³ÇÓ+·+³¿ÄÀ­Ä/ÌÃwÑ+´Ä8¾(ÃfÀâ5×ÈÎ5è­è­Ã(ª´³¿ÃfÆfÏKf a ן° ]·¹µÎf dWd bÏ>]Gg l›ÖFÏy\·Ó+À©¸Ï+Öw·ÆªÆª³Ç¸±³¿ÀÍÆwļŒÌÃ5e.Ä'´{zƪÌÄÑŽªÈ·ÑÃ(´Æ—Î5è­è­Ã(ª´³¿ÃfÆS¸'À©âr›·©¸À8·+èi ÖZ±+Öw³Ç¾(¾f³"ÏÈ·ÑÃ(´Á¡¸'¹½¸'Ó³µ×¨ý¡¸'·»ñ ƸÀ©âHƪÑÃf¾f³Ç¸¹»¹Ç¸Í´¦¸'Àͳ½¸'ÀÆfÏÌ+ÃfÑ+×¨Ì©Ô d b fG^ù5±îk ±©ÖľfÃ}zHä&lZ± eM¸â5Ì8æS¸5ÏMc›À*Á¡¸'¹Ç¸Ó³µ×¨ýX¸· ƪ·+ÑÃ(´èq¸À+³¿¼Ä¹Çâ+ƃg¦ÏÌ+ÃfÑ+רÌÔ d ù d ^ ^W^±îm ±©ÖľfÃ}zHä&lZ± eM¸â5Ì8æS¸5ÏMc›À*Á¡¸'¹Ç¸Ó³µ×¨ýX¸· ƪ·+ÑÃ(´èq¸À+³¿¼Ä¹Çâ+Æf±Kgg¤Ï+ÌÃ(Ñ+רÌÔ d ù -dWd ^ ±îî n ȱ¢Ø¤±‘o/·èq¸'ŹÃ( ±sؤĹ¿³Ç¾f¸Æ—ª´Ä+ÏSȪ´³¿ÀÍÆ䳿À.ŸæJ³ÇÆ—Ä'´!ƪ·+ÑÃ(´ÆªÑ©¸'¾fÃ1¸'À©â.è­³µ´ª´Ä´äÆ—Î5è­è­Ã(ª´ªÎ8ÏŒØÌ8Î5Æf± Ò»Ã(ª€±¤Å Ồ;_f dGd b šf dGd b š ù d©a רù ^aÌ+ÃfÑ+×d ù dWaGdGdW_ ± ÌÔm Á±È‹ ðŒ èq¸ÀÀ…ϧc›ÀRÌÛ譳ÇÀ+³µ×ŸÆª·+ÑÃ(´¦¸èäÓ³µ>æJ³¿Æ—Ä'´Jƪѩ¸¾(ÃF¸À©â@l ÷j^¦Æª·+ÑÃ(´Jý¡¸'ÀÍ×V1³¿¹¿¹¿ÆÌ+ÃfÄ´ªÎ8Ï5ÌÃ(Ñ+רÌÔ d b d ^ ° a ±ð(ô…ÖZ±5ºw̹һ¸¸'èq¸'´¦¸5Ï+ºF±ÅXÃf¹¿Ì©¸;€(ÏÒ±5Åw±Ky´³Çƪƪ³"Ï+ÚS±Ùä±+È+¸'³Çâ+³"ÏKc›ÀHÒ»Ä8¾€¸'¹ŒÁ¡¸'¹Ç¸Ó³µ×¨ýX¸·1È5·Ñä´èq¸À³µ¼Ä¹Çâ+ÆS¸'À©âqÊ¡ÌÃ(³¿´B1³¿´ª´Ä'´ÆfÏ]+± ØÌ8Î5Æf±º° _f dWdG`š b _W` bV× b _Wa ^­ÌÃ(Ñ+רÌÔ dWẀdÔ f b~


——ý5g>ÊSØsÏ+È8ÄÀ8έŴÄ8ÄWz š ÏÁXÄÀ8ª´³¿Ó·+³¿ÄÀqÄb d ý$À¸'´Æ$ļ—i ȱÐÃ(´ª´¦¸V´¦¸!Ã(ű>‡·è­³¿À+Ä+ÏȪ´·+¾(·+´Ã\Ä'¼Ò»³¿ÀÃf¸'´³tsfÀâÂÈ·ÑÃ(´Í'´¦¸Vá³µ>ÎR¸'À©â ÁXÄÀ+¼Ä´èq¸¹¤È5·+ÑÃ(´Í´¦¸€á³µ>Î8Ïl›·+¾f¹"±©Ø¢Ì8Î+Æf±+Å °ùð;_Wa ^ š ° dÔ ±—f—œžŸ•£–d ŸšX¡£¢5¤X¥5•£¡kš$›ž¦l§=š$„¨)˜©SšX;–d ŸšX›ªBä«8—£–d —=¬­¤X›­®T¯R¤?°c±³²¡´/´5µ@'·'¸X¹c›À+×>È5Ì+Ãf¹¿¹cl š È5·+ÑÃ(´Í´¦¸€á³µ>Îq³¿ÀÂÈ·ÑÃ(´ÃfƪѸ¾fÃ7l·¾f¹"± Ø¢ÌÎ5Æf± Å ;aWa ë °WfWbÏ -_ ^ ±ð k ]+±/±/ ¸¹Çâ+¸¾fÃfÀ¸5Ï+Ê¡ÌÃw¹Ç¸'´ÍÃ5l¹¿³Çè­³µXļ»Æª·ÑÃ(´¾fÄÀ+¼Ä´èq¸¹Kv©Ãf¹Çâ!ÌÃfÄ'´³¿ÃfÆS¸Àâ!ƪ·ÑÃ(´Í'´¦¸€á5³µŸÎ8Ϻ›â5á± Ê¡Ì+ÃfÄ´€±/ ¸VÌ»±Ø¢Ì8Î+Æf±;_G_ ^ š f'° × fGbWf!ÌÃfÑ5רÌÔ _WaÔ f dGd ±ð m Ò±£y\ĹǸÀ…Ï›Á± ÖF±Zl\¸Ñ+ѳ"ÏwÚS±£e ³µªÃfÀ…ÏSý¡¸'ÀͳǸ'À ÈÎ5è­è­Ã(ª´ªÎM³¿À ü ÷ ù È5·+ÑÃ(´¾fÄÀ5¼’Ä'´èq¸¹Jý¡¸'ÀÍ×;1³¿¹¿¹¿Æ1Ê¡Ì+ÃfÄ´ªÎ8ÏJÌ+ÃfÑ+×ÌÔ d ù dÔ fVù6°5±~ 5


Chapitre 4Modèle Cristallin <strong>du</strong> <strong>Vertex</strong> C 3Depuis la découverte de la théorie des cordes topologiques, plusieurs voies ont étéexplorées pour calculer leurs amplitudes. La méthode <strong>du</strong> vertex topologique, qui sera traitéedans les chapitres suivants et qui consiste à évaluer l’expression de ces amplitudes surles variétés de CY3 toriques, est sans doute une méthode sophostiquée vue son analogieformelle avec la méthode des graphes de Feynmann des théories quantiques des champs.Dans ce chapitre, qui peut être vu comme étude préliminaire au vertex topologique, nousétudions la fonction de partition de la corde topologique type A pour le cas d’une variété deCY3 simple à savoir l’espace complexe C 3 . Ainsi, nous montrerons à travers cette étude quela fonction de partition de la corde toplogique <strong>du</strong> modèle A sur C 3 est fortem<strong>en</strong>t liée à unmodèle classique de la mécanique statistique [11], [149, 150] à savoir le cristal fon<strong>du</strong> et d’oùle nom de variété de CY cristalline [91], [63]-[68]. Les premiers calculs exacts de ce modèledans le cadre de la théorie des cordes topologiques sont <strong>du</strong>s à Okounkov, Reshetikhin etVafa [91, 71, 151] qui proposèr<strong>en</strong>t une <strong>du</strong>alité <strong>en</strong>tre les cordes topologiques <strong>du</strong> modèle Asur une variété de CY3 et le cristal fon<strong>du</strong> sur lequel on impose des contraintes aux bords.L’origine de cette équival<strong>en</strong>ce découle de l’id<strong>en</strong>tification de la constante de couplage de lacorde topologique g S avec l’inverse de la température absolue T,g S = 1 T . (3.1)Avant d’étudier cette <strong>du</strong>alité, rappelons qu’un cristal à géométrie tri-dim<strong>en</strong>sionnelle peutêtre naïvem<strong>en</strong>t imaginé comme fait de cubes dont les atomes sont répartis de façon régulièredans l’espace 3D. Les excitations correspond<strong>en</strong>t à la suppression des cubes élém<strong>en</strong>taires<strong>du</strong> cristal et par suite nous permett<strong>en</strong>t d’associer à chaque atome A un poids q donné par,q = e µ/T , (3.2)105


4.1 Variétés de CY toriques et cristal fon<strong>du</strong>interprété comme le poids de Boltzmann d’une configuration physique donnée. C’est aussil’énergie de la suppression de l’atome A avec un pot<strong>en</strong>tiel chimique µ et une températureT .Dans la première section de ce chapitre, on prés<strong>en</strong>te avec quelques détails la construction<strong>du</strong> modèle de mécanique statistique décrivant le cristal <strong>en</strong> utilisant la variété de Calabi-Yautorique. Après avoir donné la définition de Calabi-Yau quantique ainsi que celle <strong>du</strong> cristalde Calabi-Yau, nous discutons la méthode de matrice de transfert pour laquelle la fonctionde partition de la théorie de corde topologique sur la variété de Calabi-Yau torique C 3s’exprime <strong>en</strong> termes des opérateurs de création et d’annihilation. Dans les sections 4.2 et4.3, nous exhibons <strong>en</strong> détails la formulation de la fonction de partition perp<strong>en</strong>diculaire <strong>du</strong>cristal fon<strong>du</strong> dans les deux cas standard et raffiné. Dans la section 4.4, nous analysonsla fonction de partition <strong>du</strong> conifold résolu. Dans les sections 4.5 et 4.6, nous étudions lesinvariants topologiques dans le modèle cristallin. Nous terminons ce chapitre par notrecontribution sur la fonction de MacMahon généralisée interprétée comme une fonction decorrelation <strong>en</strong> théorie de champs conformes c = 1.4.1 Variétés de CY toriques et cristal fon<strong>du</strong>L’étude de la <strong>du</strong>alité <strong>en</strong>tre les amplitudes de la théorie de corde topologique et les fonctionsgénératrices des configurations <strong>du</strong> cristal a été une source d’inspiration importantepour interpréter la variété de Calabi-Yau quantique C 3 comme une fonte de cristal.4.1.1 Variété de Calabi-Yau quantiqueL’exemple le plus important des variétés de Calabi-Yau quantiques est la variété C 3ayant une structure de fibration T 2 × R sur la base R 3 et une forme symplectique quis’écrit :3∑3∑3∑ω = dz j ∧ d¯z j = d|z j | 2 ∧ dθ j = dp j ∧ dθ j (3.3)j=1j=1j=1Il se trouve que (x, y, z) = (p 1 , p 2 , p 3 ) = (|z 1 | 2 , |z 2 | 2 , |z 3 | 2 ) sont des coordonnées qui paramétris<strong>en</strong>tl’octant positif de R 3 qui devi<strong>en</strong>t une base de la fibration T 2 × R et θ i sont descoordonnées dans la fibre T 3 . En effet, nous pouvons considérer θ i comme des positions etp i comme des mom<strong>en</strong>ts dont la condition de quantification canonique est donnée par :[θ j , p k ] = ig s δ jk106


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>où g s joue le rôle de la constante de Planck . Il s’<strong>en</strong>suit que la base de l’espace de Hilbertest définie comme :3∑ψ n1 ,n 2 ,n 3(θ k ) = exp(i n j θ j ), (3.4)avec 0 ≤ n j ∈ Z, dont les états peuv<strong>en</strong>t être id<strong>en</strong>tifiés avec les boîtes dans la position(n 1 , n 2 , n 3 ) ∈ Z 3 . L’unité de distance est déterminée par le paramètre quantique g s . Cettevariété de Calabi-Yau quantique est gelée à la température T = 0, dans ce cas, l’octantpositif Z 3 est complètem<strong>en</strong>t rempli. Par contre dans le cas de la température non nulle,le cristal comm<strong>en</strong>ce à fondre. C’est sur ce modèle cristallin qu’il existe un li<strong>en</strong> avec lesdiagrammes de Young généralisés.j=14.1.2 Cristal de CYLa relation <strong>en</strong>tre le modèle topologique C 3 et le modèle de la physique statistique estréalisée concrètem<strong>en</strong>t grâce à l’équation (3.1) qui relie la constante de couplage avec latempérature. En particulier, lorsque g s → 0, le paramètre q → 1, <strong>en</strong> fait la limite où lecristal comm<strong>en</strong>ce à fondre. Par ailleurs, une configuration d’énergie zéro décrit un octantFig. 3-1 – Représ<strong>en</strong>tation de la fusion <strong>du</strong> cristalpositif dans un espace R 3 rempli par des atomes qui représ<strong>en</strong>t<strong>en</strong>t des boîtes localisées à despoints <strong>en</strong>tiers dans un réseau à 3 dim<strong>en</strong>sions. Cette analyse est <strong>du</strong>e uniquem<strong>en</strong>t à la formesymplectique dans la limite de grand paramètre de kahler (3.3). Un atome à la position(x 0 , y 0 , z 0 ) représ<strong>en</strong>te une boîte remplie ayant la forme suivante :{(x + s x , y + s y , z + s z ) | 0 ≤ s x , s y , s z ≤ 1}107


4.1 Variétés de CY toriques et cristal fon<strong>du</strong>dont l’excitation peut être interprétée comme l’arrachem<strong>en</strong>t des atomes d’un coin <strong>du</strong> cristalavec la condition qu’aucun atome ne sera dans la région{(x, y, z) | x < x 0 , y < y 0 , z < z 0 } .Chaque arrachem<strong>en</strong>t contribue un facteur de l’éq(3.2) et donne lieu à une autre configurationappelée les diagrammes de Young à 3D et la fonction de partition aura la formesuivante :Z = ∑ q |π| (3.5)πFig. 3-2 – Cristal fon<strong>du</strong> et la partition plane 3d.avec |π| est le nombre d’atomes arrachés. Il est important de savoir que la somme surtoutes les configurations fon<strong>du</strong>es repro<strong>du</strong>it la fonction de partition de la corde topologiquede C 3 . Un résultat intéressant de cette analyse est que la fonction de partition de la cordetopologique <strong>du</strong> modèle A sur C 3 est liée à la fonction génératrice éq(3.5) qui est connuecomme la fonction de Macmahon à 3 dim<strong>en</strong>sions :Z C3top = ∑ πq |π| = M(q) =∞∏n=11(1 − q n ) n . (3.6)Avant d’<strong>en</strong>tamer la méthode de matrice de transfer, il faut m<strong>en</strong>tionner que les atomes dansla région x ≥ N ne peuv<strong>en</strong>t pas être arrachés, l’<strong>en</strong>tier N doit être relié au mo<strong>du</strong>le deKahler par t = g s N. Cette condition est interprétée comme un “mur”, qui avec les deuxmurs originaux constitu<strong>en</strong>t le diagramme torique <strong>du</strong> conifold résolu.4.1.3 Matrice de transfertDans ce paragraphe, nous allons étudier un formalisme particulièrem<strong>en</strong>t utile pourconstruire des amplitudes de la théorie des cordes topologiques. Okounkov et Reshetikhin108


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>[152] ont proposé la méthode de matrice de transfert. Cette méthode peut être définiecomme l’opérateur d’évolution qui porte l’information d’une tranche à autre. Conformém<strong>en</strong>tà certains résultats <strong>en</strong> théorie combinatoire, notamm<strong>en</strong>t ceux de Percy MacMahon[92], Okounkov et Reshetikhin ont pu retirer le résultat classique de MacMahon sur lafonction génératrice des partitions planairesZ fermée (q) = M(q) :=n=1(1 ∏ − q n ) −n . (3.7)En effet, cette fonction génératrice n’est autre que la fonction de partition <strong>du</strong> cristal fon<strong>du</strong>,définie comme une somme sur toutes les configurations de partition possibles et l’exposantde q n’est autre que le nombre d’atomes fon<strong>du</strong>s.Z cristal (q) =∑q #boxes . (3.8)3d partitionLa méthode d’Okounkov et Reshetikhin est généralisée pour faire face au vertex topologique<strong>du</strong> modèle A sur les variétés de Calabi-Yau à trois dim<strong>en</strong>sions. Or, la connaissance de laméthode de matrice de transfert permet de déterminer la fonction de partition de l’<strong>en</strong>sembledes cristaux, <strong>en</strong> termes des opérateurs de la théorie des champs conforme à deux dim<strong>en</strong>sions(fermion libre à 2d). La fonction de partition <strong>du</strong> cristal s’exprime comme une séqu<strong>en</strong>ced’opérateurs qui agiss<strong>en</strong>t sur les états associés aux asymptôtes des diagrammes de Young.Le principe de cette méthode consiste à associer à chaque partition 3d une séqu<strong>en</strong>ce departitions 2d. On pourra trouver ce résultat à partir de la coupure diagonale de π par desplans x = y − a, (voir Figure (3-3)) .Fig. 3-3 – Coupure <strong>en</strong> tranche diagonale de la partition 3D π par des plans x = y − a,Ces partitions 2d obéiss<strong>en</strong>t à la condition d’<strong>en</strong>trelacem<strong>en</strong>t :µ(t) < µ(t + 1), t < 0, µ(t + 1) < µ(t) t ≥ 0 (3.9)109


4.1 Variétés de CY toriques et cristal fon<strong>du</strong>Si ces deux partitions 2D µ et ν s’<strong>en</strong>trelac<strong>en</strong>t, alorsµ 1 ≥ ν 1 ≥ µ 2 ≥ ν 2 ...Dans ce qui suit, µ k désigne le nombre de boites dans la k-ème ligne de µ. Dans ce formalisme,on utilise la correspondance <strong>en</strong>tre les partitions 2d des tranches diagonales et lesétats de Fock dans le secteur NS d’un fermion complexe ou un boson (via la bosonisation).d∏|µ〉 = ψ ∗ −a iψ −bi |0〉 (3.10)avecψ (z) = ∑ ψ n+1 z −n−1 ,2n∈Zi=1ψ ∗ (z) = ∑ n∈Zψ ∗ n+ 1 2z −n−1 ,{ψ n+12}, ψ ∗ −m−= δ 1 m,n (3.11)2et |0〉 est l’état qui est annihilé par tous les ˜ψ r , r ∈ Z + 1/2. Pour un diagramme de Youngν, on peut définir ses coordonnées de Frob<strong>en</strong>ius par :{ }boite noire : µi − i + 1 | i = 1, 2, · · ·2boite blanche :{j − µtj − 1 2| j = 1, 2, · · ·}(3.12)où µ t est le transposé de diagramme de Young. Il s’avère que dans le formalisme de matricede transfert, les élém<strong>en</strong>ts matriciels <strong>en</strong>tre les partitions sont donnés par les opérateursvertex qui sont définis comme des opérateurs de création et d’annihilation. En termes demodes J ±k , sont les modes de courant de Noether U(1) 1 , ces opérateurs sont donnés parΓ ± (z) = exp( ∑ n>0z ∓nn J ±n) J(z) =: ψ(z) ∗ ψ(z) : = ∑ m∈Z z−m−1 J m ,et satisfont les relations de commutation suivantes :(3.13)[J n , J m ] = −nδ n+m,0 , [J n , ψ k ] = ψ k+n , [J n , ψ ∗ k] = −ψ ∗ k−n . (3.14)Etant donné deux partitions λ et µ, les opérateurs de création et d’annihilation Γ ± vérifi<strong>en</strong>tles équations suivantes :1 courant fermioniqueΓ + (x)Γ − (x) =〈λ| Γ + (x) |µ〉 =〈µ| Γ − (x) |λ〉 =( )1 − x Γy − (x)Γ + (x),{Sµ/λ (x) λ ≺ µ0 sinon{Sµ/λ (x) λ ≺ µ0 sinon,.(3.15)110


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>En comparant ces formules avec les relations d’<strong>en</strong>trelacem<strong>en</strong>t dans l’éq.(3.9), nous remarquonsque Γ ± décrit l’évolution des partitions. Plus précisém<strong>en</strong>t, l’évolution négative à unmom<strong>en</strong>t x ≤ −1 est donnée par Γ + tandis que l’évolution à un mom<strong>en</strong>t positif x ≥ 0 estnotée par Γ − . La fonction de MacMahon, nommée aussi la fonction génératrice de nombrede partition plane Z 3D (q) := Z ∅∅∅ (q), sera repro<strong>du</strong>ite <strong>en</strong> utilisant la notion de matrice detransfert(∏ ∞)Z cristal (q) = Z ∅∅∅ (q) = 〈∅| q L 0Γ + (1)avec l’opérateur q L 0t=0q L 0( −1 ∏t=−∞Γ − (1) q L 0)|∅〉 (3.16)qui déplace la tranche diagonale par une unité. Le mode zéro deVirasoro L 0 est l’hamiltoni<strong>en</strong> qui compte le nombre de boîtes |µ| dans le diagramme deYoung µL 0 |ν >= |ν||ν > q L 0|ν >= q |ν| |ν > . (3.17)Nous allons décrire <strong>en</strong> détail l’analyse intuitive pour compr<strong>en</strong>dre la forme de la fonctionde partition <strong>en</strong> termes des opérateurs de création et d’annihilation :Tout d’abord, on comm<strong>en</strong>ce par l’état de vide qui décrit une tranche ou une partition 2dà la position a = −∞.Lorsqu’on agit sur cet état par l’opérateur de création, on obti<strong>en</strong>t une somme sur toutesles partitions possibles. En passant de a = −∞ à 0 <strong>en</strong> faisant agir l’opérateur de créationΓ − sur cette somme de façon à ce qu’il <strong>en</strong>trelace ces partitions dans la tranche précéd<strong>en</strong>te.Ainsi de suite jusqu’à ce qu’on atteigne la tranche principale a = 0.On procède de la même manière pour la tranche principale. Lorsqu’on agit sur cet étatavec l’opérateur d’annihilation Γ + , on détruit les partitions précédemm<strong>en</strong>t créées dans latranche ”a” et qui sont <strong>en</strong>trelacées sur la tranche précéd<strong>en</strong>te a − 1 avec a est positive.Cette analyse, avec les opérateurs q L 0, donne la somme sur toutes les partitions 3D q |π| quisatisfont à la condition d’<strong>en</strong>trelacs. Notons qu’on définit l’état <strong>du</strong> vide comme l’état quiest annulé par l’opérateur Γ + . Alors, nous pouvons déplacer l’opérateur d’annihilation surla droite et utiliser les relations de commutation <strong>en</strong>tre les deux opérateurs Γ ± éq(3.15). Lafonction génératrice des partitions 3D est la fonction de MacMahon à 3 dim<strong>en</strong>sionsZ 3D (q) = M(q) := ∏ n=1(1 − q n ) −n (3.18)Nous remarquons que cette expression est égale à la fonction de partition <strong>du</strong> modèle Afermé sur C 3Z C3top = M (q) = ∑ πq |π| avec q = e −gs . (3.19)111


Cela a fait apparaître d’autres équival<strong>en</strong>ces qui se reli<strong>en</strong>t à d’autres variétés de Calabi-Yaubeaucoup plus complexes ainsi que des <strong>en</strong>sembles de partitions planes plus compliquées[69, 70, 133]. En particulier, les amplitudes de vertex topologiques C λµν (q) seront reliéesaux comptages d’un <strong>en</strong>semble particulier de partitions planes avec des conditions de bordsdonnées par les partitions (λ, µ, ν) sur les trois axes de C 3 [91]∑C λµν (q) ∼ M (q) −1 q |π| (3.20)π→{R 1 R 2 R 3 }Dans la suite, nous nous référons donc régulièrem<strong>en</strong>t à la fonction de MacMahon pourtrouver les amplitudes <strong>du</strong> modèle de type A.4.2 Fonction de partition perp<strong>en</strong>diculaireSuivant les deux derniers paragraphes, de nombreux résultats ont été obt<strong>en</strong>us à partir<strong>du</strong> formalisme de matrice de transfert. Le cristal Calabi-Yau qui est défini comme unesomme statistique sur les partitions 3d, a été construit à partir des tranches diagonalesqui satisfont la condition d’<strong>en</strong>trelacem<strong>en</strong>t. Dans ce cas, la fonction de partition de la cordetopologique de C 3 est égale à la fonction de MacMahon à trois dim<strong>en</strong>sions, qui est unefonction génératrice des partitions. Dans ce paragraphe, nous allons montrer que la fonctionde partition asymptotiques aux diagrammes de Young λ, µ, ν est égale à l’amplitude devertex topologique C λµν (q) multipliée par la fonction de Macmahon M (q)M(q)C λµν (q) = q 1 2 (‖λ‖2 +‖µ‖ 2 +‖ν‖ 2) P λµν(q cristal ) (3.21)avec q cristal = q −1 . En fait, il existe une relation qui transforme le q vertex au q cristalq vertex → 1 q = q cristal, (3.22)une telle opération est connue dans la théorie de cordes topologiques pour laquelle onéchange les branes dans le cristal aux anti-branes dans le vertex. On <strong>en</strong> dé<strong>du</strong>it alors unerelation <strong>en</strong>tre les paramètres <strong>du</strong> vertex topologique et les modèles cristaux.4.2.1 Formule P λ,µ,ν (q)Notre objectif est la construction de la fonction de partition perp<strong>en</strong>diculaire P (λ, µ, ν)par l’utilisation de la technique de matrice de transfert. Cette partition perp<strong>en</strong>diculaire112


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>est la fonction génératrice dont les asymptotes <strong>en</strong> direction des trois axes de coordonnéessont donnés par trois partitions λ, µ et ν. Avant de calculer cette fonction, il convi<strong>en</strong>t derappeler qu’une partition 3D est définie à l’intérieur <strong>du</strong> boite π et s’écrit sous cette forme :[1, N 1 ] × [1, N 2 ] × [1, N 3 ] , (3.23)avec les conditions de bords données par trois partitions dans les plans x i = N i , pour plusde détails sur les partitions planes voir l’annexe. En effet, lorsque la face de π se trouvedans le plan x i = N i , on parle d’un diagramme de Young ori<strong>en</strong>té de sorte que sa longueurλ 1 soit dans la direction de x 2 . De même pour les deux autres partitions qui sont définiesd’une manière cycliquem<strong>en</strong>t symétrique. Alors, on peut aisém<strong>en</strong>t prés<strong>en</strong>ter les différ<strong>en</strong>tespartitions à l’intérieur de la boite π. Dans cet exemple, on prés<strong>en</strong>te les partitions λ = (3, 2),µ = (3, 1) et ν = (3, 1, 1) dans la figure.3-4Fig. 3-4 – Une partition 3D où les conditions de bords sont les partitions 2D.Cep<strong>en</strong>dant, on peut calculer la fonction P (λ, µ, ν) qui existe comme une série de puissanceformelle <strong>en</strong> q <strong>en</strong> terme de la fonction de partition P N1 ,N 2 ,N 3(λ, µ, ν) :P (λ, µ, ν) = lim N1 ,N 2 ,N 3 →∞q −N 1|λ|−N 2 |µ|−N 3 |ν| P N1 ,N 2 ,N 3(λ, µ, ν). (3.24)dont chaque boite π est associée à un poids q vol(π)4.2.2 Calcul de la fonction de partition perp<strong>en</strong>diculaireIl est maint<strong>en</strong>ant possible de calculer la fonction perp<strong>en</strong>diculaire <strong>en</strong> utilisant la techniquede matrice de transfert. Il faut m<strong>en</strong>tionner que cette technique consiste à diviser la113


4.2 Fonction de partition perp<strong>en</strong>diculairepartition 3d <strong>en</strong> tranches diagonales. En comparant ces dernières avec les tranches perp<strong>en</strong>diculaires,la fonction P λµν(q) peut être obt<strong>en</strong>ue <strong>en</strong> ajoutant des boites à la partition et parconséqu<strong>en</strong>t, un nouveau facteur de la forme q −n(λt )−n(µ) sera aussi ajouté à l’expression dela fonction de partitionP λµν(q) = q −n(λt )−n(µ) P diag (λ, µ, ν) (3.25)avec n(µ) = ∑ (i − 1) µ i . Et la fonction de partition diagonale se prés<strong>en</strong>te sous la formeisuivante :P diag (λ, µ, ν) = ∑ π\π 0∏a∈Zq |π a|(3.26)où la somme est sur toutes les partitions 3D dont la fonction de partition P diag compte lespatitions 3D où les conditions de bords sont définies au long des trois axes et π 0 est unepartition 3D avec le minimum de boîtes.Fig. 3-5 – Une partition π (λ, µ, ν) avec extra boites sur une partition π • (λ, µ, ν).En appliquant le formalisme de matrice de transfert et <strong>en</strong> intro<strong>du</strong>isant des opérateursde création et d’annihilation, la fonction diagonale peut s’écrire comme suit :⎛⎞ ⎛⎞P diag (λ, µ, ν) = 〈 λ t∣ ∣ ⎜ ∏⎝q L 0⎟Γ + (1) ⎠ q L 0⎜ ∏⎝ Γ − (1)q L 0⎟⎠ |µ〉 (3.27)N 1 −1termesN 2 −1termesoù λ, µ, ν sont des partitions 2d et λ t est une partition transposée. Dans cette expression,nous avons un nombre infini d’opérateurs Γ ± (1) qui agiss<strong>en</strong>t sur les deux états. Leurs ordressont déterminés par la forme de la partition ν. Donc, il s’agit par la suite de déterminerl’expression générale de la fonction de partition perp<strong>en</strong>diculaire <strong>en</strong> termes des fonctions deSchur.114


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>Formule d’opérateurMaint<strong>en</strong>ant, nous appliquons à la formule (3.27) les trois transformations suivantes : Commutons les opérateurs q L 0à l’extérieur. Les opérateurs de création et d’annihilationseront conjugués aux Γ ± (q ... ) Commutons les opérateurs de création à gauche et les opérateurs d’annihilation àdroite <strong>en</strong> utilisant la formule (3.15). Ecrivons l’expression résultante comme une somme sur tous les états intermédiaires|η〉 〈η| .Résultat :L’expression la plus générale de la fonction diagonale est la suivante :avec q −ν−ρP diag (λ, µ, ν) = Z(ν)q ‖ − ‖λ‖2 µ t ‖ 2− 2 2∑s λ t /η(q −ν−ρ )s µ/η (q −νt−ρ ) (3.28)η={q −ν 1+ 1 2 , q −ν 2+ 3 2 , q −ν 3+ 5 2 · · · } et Z(ν) est un facteur multiplicatif obt<strong>en</strong>u àpartir de la relation de commutation des opérateurs de création et d’annihilation. Il dép<strong>en</strong>dseulem<strong>en</strong>t de la forme de ν et peut être dé<strong>du</strong>it <strong>en</strong> utilisant la symétrie cyclique commesuit :P (∅, ∅, ν) = P (ν, ∅, ∅). (3.29)En appliquant cette <strong>du</strong>alité sur les deux côtés d’équation (3.28)Z(ν) = q − ‖ν‖22 s ν t(q −ρ )M(q). (3.30)Finalem<strong>en</strong>t, <strong>en</strong> t<strong>en</strong>ant compte des équations (3.21,3.28,3.30), nous retrouvons l’amplitudede vertex topologique <strong>en</strong> termes des fonctions de Schur,C λ,µ,ν = q k(µ) s ν t(q −ρ ) ∑ ηs λ t /η(q −ν−ρ )s µ/η (q −νt −ρ ). (3.31)Dans le chapitre suivant, nous utiliserons le vertex topologique pour le calcul des fonctionsde partition des cordes topologiques ouvertes et fermées.4.3 La fonction perp<strong>en</strong>diculaire raffinéeDans la section précéd<strong>en</strong>te, nous avons exposé de façon systématique la méthode decalcul de la fonction de partition perp<strong>en</strong>diculaire Z 3D (q) où nous avons limité notre étude115


4.3 Version raffinée de la fonction perp<strong>en</strong>diculaireà un seul paramètre q. Dans cette section, on va suivre la même démarche de calcul afinde déterminer la fonction de partition raffinée <strong>en</strong> termes d’autres paramètres (q, t), ce quiva motiver l’intro<strong>du</strong>ction des invariants Gopakumar-Vafa [55].4.3.1 Fonction de partition avec un nombre infini de paramètresL’analyse effectuée dans ce qui suit, concerne le calcul de la fonction de partition <strong>en</strong>t<strong>en</strong>ant compte d’un nombre infini de paramètres. En liaison avec ce qui est intro<strong>du</strong>it auparavant,l’<strong>en</strong>tier ”a” est utilisé pour décrire la position de chaque tranche sur le diagrammede Young. Pour toute partition 2D, nous pouvons répartir les coins de la représ<strong>en</strong>tationd’une partition correspondante <strong>en</strong> deux groupes : coin intérieur et extérieur.Nous prés<strong>en</strong>tons les coins intérieur et extérieur par leurs coordonnées u i et v i respectivem<strong>en</strong>t.Leur projection sur la ligne réelle est représ<strong>en</strong>tée dans la Fig 3-6−5 −4 −3 −2 −1 0 1 2 3 4 5Fig. 3-6 – Coins intérieur et extérieurM∑i=0v i = M−1 ∑i=0u i M = # de coins exérieur (3.32)Il est commode de prés<strong>en</strong>ter un autre <strong>en</strong>semble de paramètres { x ± m|m ∈ Z + 1 2}et lesid<strong>en</strong>tifier sous la forme q a :x + m+1x +m= q m+12m > v M or u i − 1 > m > v i ,x + x − u i − 1 u 2 i + 1 2x − x + v i + 1 v 2 i − 1 2= q −1u i,= q vi ,(3.33)x − mx − m= q m+12m < v 1 or v i+1 − 1 > m > u i .116


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>En utilisant la méthode de matrice de transfert, la fonction de MacMahon raffinée estcalculée de la même manière que dans le cas standard. En effet la fonction de MacMahon<strong>en</strong> termes d’opérateurs de création et d’annihilation est sous la forme suivante :Z 3D (q) = ∑ ∏0∏ ( )q |η aa |∏ ∞ ( )= 〈0| Γ − x + Γ−k+ 1 + x − |0〉−k+ 1 π a∈Zk=−∞2 k=12∏= ∞ ∞∏ () −1 (3.34)1 − x + x − k 1 − 1 −k 2 2 + 1 2k 1 =1 k 2 =1Les variables x + et x − sont exprimées <strong>en</strong> termes des variables q comme suit :x + k+ 1 2x − − 1 2= k−1 ∏i=0q i , k ≥ 1= 1 , x − −k+ 1 2= k−1 ∏i=0q −i k ≥ 2(3.35)Nous dé<strong>du</strong>isons alors l’expression de la fonction de MacMahon :()∞∏ ∞∏k∏1 −1 k 2 −1 −1∏Z 3D (q) = 1 − q i q −j . (3.36)k 1 =1 k 2 =1Quand q i = q où i ∈ Z, nous obt<strong>en</strong>ons la fonction de MacMahon standard à 3 dim<strong>en</strong>sions :∞∏ (Z 3D (q) =) 1 − qk −k(3.37)k=1Après avoir démontré que l’égalité pour laquelle tous les paramètres sont ègaux l’un àl’autre est <strong>en</strong> accord avec celle initialem<strong>en</strong>t obt<strong>en</strong>ue, il est naturel d’ét<strong>en</strong>dre l’étude <strong>en</strong> intro<strong>du</strong>isantdes partitions λ et µ non-triviales. La fonction de partition, où ν est la partition2D le long de la direction préférée et λ = µ = ∅, est donnée par :νi=0j=1q tλ µFig. 3-7 – ν est une partition 2D représ<strong>en</strong>tée le long de la direction préférée.Z ν (q) = P ∅∅ν =∏(i,j)∈ν c (1 − q (i,j) ) −1 , q (i,j) =117∏(a,b)∈H(i,j)q b−a , (3.38)


4.3 Version raffinée de la fonction perp<strong>en</strong>diculaireoù H(i, j) est la longueur d’équerre de l’<strong>en</strong>semble des boites (i, j). La fonction de partitionZ 3D (q) peut être exprimée comme un pro<strong>du</strong>it sur toutes les boîtes. Chaque boîte contribuepar un facteur de (1 − x) −1 où x est le pro<strong>du</strong>it de paramètres q i indexé par la longueurd’équerre.Nous considérons maint<strong>en</strong>ant la fonction de partition perp<strong>en</strong>diculaire raffinée sur laquell<strong>en</strong>ous imposons des conditions de bords :P λ,µ,ν (q) = ∑ π\π 0∏k∈Zq |λ(k)|k(3.39)où la somme est sur toutes les partitions 3D. Chaque partition π peut être tranchée detelle sorte que nous obt<strong>en</strong>ons des partition 2D π(a) le long de la diagonale. La fonction departition perp<strong>en</strong>diculaire dans le cas où λ et µ sont des partitions non triviales, pr<strong>en</strong>d laforme :P λµν (q) = Z ν (q) ∑ ηs λ t /η(x + )s µ/η (x − ) (3.40)avec x ± = {x ± m|m ∈ Z + 1 }. Dans la sous-section suivante, nous allons limiter notre analyse2à l’étude de l’expression de la fonction de partition <strong>en</strong> termes de deux paramètres q et t.4.3.2 P λµν (q, t)Comme nous avons vu dans la section précéd<strong>en</strong>te, nous pouvons calculer la fonction departition <strong>en</strong> choisissant les paramètres q t . Si ces derniers se ré<strong>du</strong>is<strong>en</strong>t à :{q , α ≥ 0q t =t , α < 0alors la fonction de partition sera comme :Z 3D (t, q) =(3.41)∞∏(1 − t i q j−1 ). (3.42)i,j=1Par la suite on va considérer le cas où ν est non trivial. Il existe une transformation <strong>en</strong>trex ± = {x ± m|m ∈ Z + 1 } et {t, q} qui est représ<strong>en</strong>tée par :2{x + m|m ∈ D + } = {t i q −ν i| i = 1, 2, 3, · · · }{x − m|m ∈ D − } = {q j−1 q −νt j | j = 1, 2, 3, · · · }(3.43)avec D + et D − représ<strong>en</strong>t<strong>en</strong>t respectivem<strong>en</strong>t l’<strong>en</strong>semble des boîtes noires et blanches dansle diagramme de Maya correspondant à la partition ν 2 . Si nous considérons la i-ème boîte2 Voir l’annexe.118


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>blanche <strong>du</strong> côté gauche, le nombre de boîtes noires à droite sera donné par ν i . Ce quiimplique qu’il y’a une correspondance une à une{(m 1 , m 2 )| m 1 ∈ D − , m 2 ∈ D + , m 1 ≥ m 2 } → {(i, j) ∈ ν}{(m 1 , m 2 )| m 1 ∈ D − , m 2 ∈ D + , m 1 < m 2 } → {(i, j) /∈ ν}(3.44)et la fonction Z ν est donnée comme suit :∏Z ν =(1 − x + m 2x − m 1) −1 =m 1


4.4 Modèle <strong>du</strong> cristal fon<strong>du</strong> et conifold résoluEt <strong>en</strong> utilisant la relation <strong>en</strong>tre les variables <strong>du</strong> diagramme de Maya et celles des diagrammesde Young nécessite une expression de la fonction diagonale <strong>en</strong> termes des paramètresq et tP diag (λ, µ, ν) = t − |λ|2 q −|µ|2 Z ν (t, q) ∑ η( qt) |η|/2sλ t /η(t −ρ q −ν )s µ/η (t −νt q −ρ ) .En multipliant par q −n(λt) t −n(µ) , la fonction perp<strong>en</strong>diculaire s’écrit <strong>en</strong> termes de P diag (λ, µ, ν)comme suit :P λ,µ,ν (t, q) = = q −n(λt )− |µ||µ|−n(µ)− 2 t 2 Z ν (t, q) ∑ η( qt) |η|/2sλ t /η(t −ρ q −ν )s µ/η (t −νt q −ρ )= q − |λ||22 t − ||µt || 22 Z ν (t, q) ∑ η( qt) |η|+|λ|−|µ|2s λ t /η(t −ρ q −ν )s µ/η (t −νt q −ρ ).(3.52)La connaissance de la fonction de partition perp<strong>en</strong>diculaire (3.52) aussi bi<strong>en</strong> que l’expressionde la fonction de MacMahon à 3 dim<strong>en</strong>sions (3.46) , est suffisante pour donnerl’expression explicite de vertex topologique raffiné <strong>en</strong> termes des fonctions de schur :C λµν (t, q) = q f(ν) t g(ν) q ‖λ‖2 + ‖µ‖2 P λµν (t,q)2 2M(t,q)= q f(ν) t ( ) g(ν) ‖µ‖ 2q 2t k(µ) Z2 ν(t,q)tM(t,q)∑= q f(ν) t ( ) g(ν) ‖µ‖ 2q 2t k(µ)2 ˜Zν (t, q) ∑ tηη( qt) |η|+|λ|−|µ|2s λ t /η (t −ρ q −ν ) s µ/η(t−ν t q −ρ)( qt) |η|+|λ|−|µ|2s λ t /η (t −ρ q −ν ) s µ/η(t−ν t q −ρ)(3.53)les fonctions f(ν) et g(ν) sont données par f(ν) + g(ν) = ||ν||2 . Il s’avère que cette relation2n’est pas suffisante pour fixer les deux fonctions f(ν) et g(ν). Or, le choix le plus courantest de pr<strong>en</strong>dre g(ν) = 0 et par conséqu<strong>en</strong>t le vertex topologique raffiné s’exprime <strong>en</strong> termesde la généralisation de la fonction de Schur :⎧⎪⎨C ∅∅ν (t, q) = q ||ν||22 ˜Zν (t, q)⎪⎩= q k(ν)2 Q ν (q −ρ ; t, q),= ( ) ||ν|| 2q 2Pt ν t(t −ρ ; q, t)(3.54)où P ν t(t −ρ ; q, t) et Q ν (q −ρ ; t, q) sont respectivem<strong>en</strong>t des fonctions de Macdonald et leur<strong>du</strong>ales. Finalem<strong>en</strong>t, après la considération des éq(3.54) et (3.53), la forme générale <strong>du</strong>vertex topologique raffiné est :C λµν (t, q) = ( ) ‖µ‖ 2 +‖ν‖ 2q 2t k(µ)2 Ptν t(t −ρ ; q, t) ∑ η( qt) |η|+|λ|−|µ|2s λ t /η (t −ρ q −ν ) s µ/η(t−ν t q −ρ)Notons que pour le cas où t = q, le vertex topologique raffiné n’est autre que le vertextopologique standard <strong>en</strong> pr<strong>en</strong>ant P ν (q −ρ ; q, q) = s ν (q −ρ ).120


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>4.4 Modèle <strong>du</strong> cristal fon<strong>du</strong> et conifold résoluDans le contexte <strong>du</strong> modèle cristallin, il est démontré que l’analyse <strong>du</strong> cristal fon<strong>du</strong>permet de repro<strong>du</strong>ire la fonction de partition sur le conifold résolu. Nous allons analyser<strong>en</strong> détails le modèle C 3 avec un nombre arbitraire de "murs". A partir de ce modèle, onétablit une relation <strong>en</strong>tre l’amplitude <strong>du</strong> modèle cristallin et les invariants des noeuds. Icinous n’allons évoquer que trois exemples parmi bi<strong>en</strong> d’autres des modèle cristallins <strong>du</strong>vertex.4.4.1 Modèle cristallin avec un seul murEn se basant sur les résultats fournis dans la section précéd<strong>en</strong>te, nous calculons l’amplituderésultante <strong>du</strong> modèle cristallin fon<strong>du</strong> d’un conifold résolu. Ce dernier est une simplemodification d’un modèle de cristal C 3 dont les atomes sont dans la région x ≥ N.Fig. 3-8 – Géométrie torique <strong>du</strong> conifold résolu se termine sur un mur sur l’axe y à une distancequi correspond au paramètre de kahler.Un ”mur” à la position x = N sur l’axe x est un axe interne ou compact au delà<strong>du</strong>quel le cristal ne peut pas être fon<strong>du</strong>, l’axe y est considéré comme un axe externe ouune direction non compacte et N est lié au mo<strong>du</strong>le de Kähler t par l’équation suivantet = g s N. Le fait de considérer le mur sur la N me tranche est représ<strong>en</strong>té <strong>en</strong> insérant lesN opérateurs Γ − qui agiss<strong>en</strong>t sur le vide. Alors, la fonction de partition <strong>du</strong> modèle crisal121


4.4 Modèle <strong>du</strong> cristal fon<strong>du</strong> et conifold résolufon<strong>du</strong> sur le conifold résolu est de la forme suivante :( ∞) (∏N)Zcristal P1 (q, t = g s N) = 〈0| q L 0∏Γ + (1) q L 0Γ − (1) q L 0|0〉n=1m=1∏Z cristal (q, t = g s N) = 〈0| ∞ ) ∏ NΓ +(q n− 1 2 Γ −(q − (m− 1 2) ) |0〉n=1= 〈0| e − P n>0= M(q)e − P nm=1αnn[n] e− P n>0Q nn[n] 2 .1−q Nnn[n] α −n|0〉(3.55)où Q = e −t = q N . Lorsque N → ∞, on trouve un accord parfait avec la fonction de partition<strong>du</strong> modèle cristallin C 3 . Ce modèle peut être obt<strong>en</strong>u par la réécriture de l’expression de lafonction de partition de la théorie de Chern-Simons sur S 3 .4.4.2 Modèle cubiqueLe modèle cubique est une généralisation naturelle des modèles avec des parois prés<strong>en</strong>tésdans la section précéd<strong>en</strong>te. On verra donc comm<strong>en</strong>t calculer la fonction de partition detoutes les partitions 3d placées dans un cube de taille N × M × L. Il est donc nécessaired’intro<strong>du</strong>ire un nombre fini de termes dans Z cube = 1+· · ·+q LMN , avec LMN est le volume<strong>du</strong> cube. En d’autres termes, nous mettons trois murs, dans les positions suivantes : x = M,y = L et z = N comme le montre la figure(3-9) ci-dessous :Fig. 3-9 – Modèle cristallin cubique de la taille M × L × N.Dans ce cas, la fonction de partition, grâce au formalisme de matrice de transfert,est représ<strong>en</strong>tée par un nombre fini de Γ ± alors que le troisième mur est représ<strong>en</strong>té par122


Modèle Cristallin <strong>du</strong> <strong>Vertex</strong>l’opérateur de projection 1 d t ≤N(∏ L )Z cube = 〈0| Γ +(q ) (∏ M n− 12 1 d t ≤N Γ −(q − (m− 1 2) )) |0〉 . (3.56)n=1Cette fonction génératrice est égale au pro<strong>du</strong>it <strong>du</strong> la fonction de partition de la cordefermée <strong>en</strong> utilisant le formalisme de vertex topologique par la fonction de MacMahonm=1Z cube = M (q) Z C .De manière simple, l’évaluation de la fonction de partition est non triviale, puisque la forme<strong>du</strong> projecteur est plus compliquée. Or, c’est exactem<strong>en</strong>t le type de fonctions génératricesdes partitions planes que l’on peut obt<strong>en</strong>ir à partir des outils combinatoires démontréspar MacMahon [92]. Ainsi, <strong>en</strong> utilisant ce résultat, on obti<strong>en</strong>t finalem<strong>en</strong>t l’expression desfonctions de partition :Z cube = Z 1 Z 2 =∏ 1 − q N+i+j−11 − q h(i,j)(i,j)∈M Loù M L désigne la partition 2d avec L lignes et h(i, j) est la longueur d’équerre avec lesélém<strong>en</strong>ts (i, j) (voir l’annexe). Il est important de souligner que le résultat de l’équation(3.47) se ré<strong>du</strong>it <strong>en</strong> fonction de MacMahon standard à la limite M, N, L → ∞.Le modèle cristallin dans un cube est très prometteur pour étudier autres modèles pluscompliqués. Nous avons jugé utile que le collage de tels cubes correspond à la constructiondes modèles cristallins plus complexes liés aux espaces de CY toriques. Ces modèlesaurai<strong>en</strong>t deux caractéristiques : le paramètre g s lié à la taille d’une boîte élém<strong>en</strong>taire, etles paramètres de Kahler des variétés de CY donnés par q M , q N , et q L . L’étude <strong>du</strong> cristalCY reste <strong>en</strong>core un problème ouvert. Il s’agit, dans ce cas, d’un nouveau terrain <strong>en</strong>core àexplorer.4.5 Invariants topologiques dans le modèle cristallinLa question qui se pose : comm<strong>en</strong>t peut-on concilier les invariants des noeuds et <strong>en</strong>trelacssous le modèle cristallin <strong>du</strong> conifold. En fait, pour dévoiler les invariants dans lemodèle cristallin, il faut insérer des branes qui correspond<strong>en</strong>t aux invariants de la théoriede Chern-Simons. Dans cette section, nous allons prés<strong>en</strong>ter une étude directe des invariantsde la théorie de Chern-Simons dans le modèle cristallin.123


4.5 Invariants topologiques dans le modèle cristallinFig. 3-10 – Branes et anti-branes dans les slices positives et négatives de cristal C 3 .4.5.1 Invariant unknotConsidérons l’exemple d’une seule brane sur l’axe y <strong>du</strong> vertex qui est décrite par lemo<strong>du</strong>le a dans le cristal. La fonction de partition s’écrit donc <strong>en</strong> termes de la fonction deMacMahon comme :Z(a, q) = M(q)L(a, q), (3.57)où L(a, q) est la fonction génératrice des invariants unknot 3 . On la développe de la manièresuivante avec [n] = q n 2 − q −n2 :L(a, q) = e P ∞n=1 ann[n] a= 1 + „ « + a 2 q2+ · · ·q 1 2 −q −1 (q 2 −1)(q−1)2= ∑ (3.58)R W R•a |R|de sorte que cette fonction peut être exprimée par une somme sur toutes les représ<strong>en</strong>tationsR associées aux lignes dans les diagrammes de Young, par exemple 1 = □, 2 = □□, etc.D’où W R• correspond à un invariant unknot. Une telle fonction de partition des invariantsdénoués (unknot) qui est <strong>en</strong> accord avec les résultats des théories de Chern-Simons étudiésdans le cadre de la transition géométrique, ne peut exister que si nous avons une seulebrane insérée dans le cristal C 3 .3 dénoué <strong>en</strong> français124


H. Jehjouh G<strong>en</strong>eralized MacMahon G(q) as q-deformed CFT Correlation Function4.5.2 Invariant Entrelacs de HopfCherchons à déterminer la fonction de partition ˜Z(a, b, q) qui régit la fonction génératricedes invariants d’<strong>en</strong>trelacs de Hopf. Pour cela, insérons une antibrane sur l’axe des xet une brane sur l’axe y. La fonction de partition sera sous cette forme :˜Z(a, b, q) =Z(a, b, q)M(q)(3.59)avecZ(a, b, q) = ∑ Rq k R +k R2 W R t P ta|R| b |R| (3.60)où k R = |R| + ∑ i R i (R i − 2i) et M(q) est la fonction de MacMahon à trois dim<strong>en</strong>sions. Ilest intéressant de remarquer que le vertex avec une représ<strong>en</strong>tation triviale est fortem<strong>en</strong>t liéaux termes des invariants de Hopf d’<strong>en</strong>trelacs W P R , qui peuv<strong>en</strong>t être exprimés <strong>en</strong> termesdes fonctions de schur comme :W P R = q k R/2 C P R t •= s P (q ρ ) s R(qρ+R )= q 1 2 (k R+k P ) ∑ η s R t /η (q ρ ) s P t /η (q ρ ) .(3.61)En appliquant la même procé<strong>du</strong>re des deux cas précéd<strong>en</strong>ts, on peut effectuer une fonctiongénératrice des invariants d’<strong>en</strong>trelacs de Hopf avec n branes sur l’axe des y et m antibranessur l’axe des x. Alors, la fonction de partition aura la forme suivante :˜Z(a 1 , · · · a n ; b 1 , · · · b m ; q) = ∑ ∑ q k R +k R2 W P t R ta|R 1|1 · · · a |R n|n b |P 1|1 · · · b |P m|m (3.62)où R = (R n , ..., R 1 ) et P = (P m , ...P 1 ) sont des partitions 2D. Enfin, l’un des aspects lesplus intéressants est que les résultats obt<strong>en</strong>us puiss<strong>en</strong>t donner des informations utiles surla fonction de partition des invariants de la théorie des cordes topologiques <strong>en</strong> prés<strong>en</strong>ce desbranes.Pour conclure ce chapitre, remarquons que la découverte de la <strong>du</strong>alité <strong>en</strong>tre le modèlede la physique statistique et la théorie de corde topologique <strong>du</strong> modèle-A sur une variétéde Calabi-Yau de dim<strong>en</strong>sion 3 a montré un li<strong>en</strong> inatt<strong>en</strong><strong>du</strong> <strong>en</strong>tre les efforts fournis pourconstruire les amplitudes de la théorie des cordes topologiques type A et B sur le conifoldrésolu. Tout cela suggère que la variété de Calabi-Yau cristal pourrait être un bon candidatpour dé<strong>du</strong>ire la relation <strong>en</strong>tre la théorie des cordes topologiques et les modèles de lamécanique statistique.125


4.6 Contribution : G<strong>en</strong>eralized MacMahon4.6 Contribution : G<strong>en</strong>eralized MacMahonNuclear Physics B, Volume 801, Issue 3, 1 October 2008, Pages 316-345, arXiv :0801.2661.RésuméDans ce papier, nous nous sommes basés sur deux outils très importants : la méthode dematrice de transfert et la théorie de champs quantique q-deformée afin de construire lathéorie de champs conforme à 2d de la fonction MacMahon généraliséeG d (q) =∞∏k=1[ (1 ) ]− qk − (k+d−3)!(k−1)!(d−2)!, d ≥ 2, (3.63)Différ<strong>en</strong>ts résultats fur<strong>en</strong>t obt<strong>en</strong>us. En effet, nous avons montré que les opérateurs de vertexΓ ± (z) de la théorie des champs conformes à 2 dim<strong>en</strong>sions, apparaiss<strong>en</strong>t comme le premierniveau de l’hiérarchieΓ (p)− (z) |0〉 = exp( ∞∑n=1)iz n J −nn (1 − q n ) p−1 |0〉 , p ≥ 1 (3.64)Nous avons dérivé l’expression explicite de la fonction de MacMahon généralisée G d (q) <strong>en</strong>termes de ces opérateurs de vertex Γ ± (z) q-déformés. Suite à sa forme générale, nous avonsinterprété cette fonction G d (q) comme une fonction de corrélation de (d + 1)-points dansla théorie de champs q-déformée c = 1.126


Nuclear Physics B 801 [FS] (2008) 316–345www.elsevier.com/locate/nuclphysbG<strong>en</strong>eralized MacMahon G d (q) as q-deformed CFT 2correlation functionLalla Btissam Drissi a,b , Houda Jehjouh a,b , El Hassan Saidi a,b,c,∗a Lab/UFR-Physique des Hautes Energies, Faculté des Sci<strong>en</strong>ces, Rabat, Moroccob GNPHE, Groupem<strong>en</strong>t National de Physique des Hautes Energies, Siège Focal: FS, Rabat, Moroccoc Académie Hassan II des Sci<strong>en</strong>ces et Techniques, Collège des Sci<strong>en</strong>ces Physiques et Chimiques, Rabat, MoroccoReceived 18 January 2008; received in revised form 27 February 2008; accepted 4 March 2008Available online 16 March 2008AbstractUsing Γ ± (z) vertex operators of the c = 1 two-dim<strong>en</strong>sional conformal field theory, we give a 2d-quantumfield theoretical derivation of the conjectured d-dim<strong>en</strong>sional MacMahon function G d (q). We interpretthis function G d (q) as a (d + 1)-point correlation function G d+1 (z 0 ,...,z d ) of some local vertex operatorsO j (z j ). We determine these operators and show that they are particular composites of q-deformedhierarchical vertex operators Γ (p)± , with a positive integer p. In agreem<strong>en</strong>t with literature’s results, we findthat G d (q), d 4, cannot be the g<strong>en</strong>erating functional of all d-dim<strong>en</strong>sional g<strong>en</strong>eralized Young diagrams.© 2008 Elsevier B.V. All rights reserved.Keywords: Topological string; <strong>Vertex</strong> operators; Young diagrams and solid partitions; c = 1 2d conformal field model;G<strong>en</strong>eralized MacMahon function; q-deformed QFT 21. Intro<strong>du</strong>ctionThe study of two-dim<strong>en</strong>sional (2d) MacMahon function G 2 , and its 3d-g<strong>en</strong>eralization G 3 appearin many areas of statistical physics, such as crystals growth, crystals melting, Bose–Einsteinstatistics and dimer model [1–7]. Rec<strong>en</strong>tly, these functions have known a revival of interest inconnection with topological string theory [8–10]; in particular in the study of BPS black holes,giv<strong>en</strong> by branes wrapping collapsed cycles in Calabi–Yau orbifolds, and in the infinite n limit* Corresponding author at: Lab/UFR-Physique des Hautes Energies, Faculté des Sci<strong>en</strong>ces, Rabat, Morocco.E-mail addresses: drissilb@gmail.com (L.B. Drissi), jehjouh@gmail.com (H. Jehjouh), h-saidi@fsr.ac.ma(E.H. Saidi).0550-3213/$ – see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.nuclphysb.2008.03.006


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 317of quiver gauge theories[11–15]. MacMahon functions G 2 (q) and G 3 (q) arealsousedintheexplicit computation of the amplitudes of A-model topological string on local Calabi–Yau manifolds[5,16–18].In [5], it has be<strong>en</strong> shown that the topological A-model partition function Z 3d on the complexspace C 3 , which coincides exactly with the 3d-crystal melting partition function Z crystal ,isgiv<strong>en</strong> by 3d-g<strong>en</strong>eralized MacMahon function G 3 . This is an important result since topologicalamplitudes for the full class of toric Calabi–Yau threefolds X 3 with a planar toric geometry arerecovered just by gluing the C 3 -vertices [16]. Amplitudes involving op<strong>en</strong> strings are also recovere<strong>du</strong>p on inserting special Lagrangian D-branes captured by boundary conditions on the edgesof the 3-vertices [18]. An evid<strong>en</strong>ce for a “topological 4-vertex”, in the case of toric Calabi–Yauthreefold with non planar toric geometry, has be<strong>en</strong> also studied in [19] and would, roughly, bedescribed by a 4d-ext<strong>en</strong>sion of the g<strong>en</strong>eralized MacMahon function G 3 .MacMahon functions G 2 and G 3 appear as well in repres<strong>en</strong>tation theory of infinite dim<strong>en</strong>sionalLie algebras and in topologically twisted U(1) gauge theories [20–23]. They are respectivelythe (specialized) character ch R of the basic repres<strong>en</strong>tations of the sl(∞) and the affinealgebra ŝl(∞) at large c<strong>en</strong>tral charge c [22,24–26]. In the limit c →∞, it has be<strong>en</strong> moreoverobserved in [27] that the basic repres<strong>en</strong>tation of ŝl(∞) is closely related to the partition functionof a three-dim<strong>en</strong>sional free field theory. MacMahon G 2 and G 3 are also the partition functionsof the 4d- and 6d-topologically twisted U(1) gauge theory giv<strong>en</strong> by a D4- and D6-branes fillingrespectively C 2 and C 3 [11,15].For higher dim<strong>en</strong>sions, it has be<strong>en</strong> checked in [28] that G 4 (q) cannot be the g<strong>en</strong>erating functionalof the 4d-g<strong>en</strong>eralized Young diagrams leading th<strong>en</strong> to the two basic questions:(i) What is the right g<strong>en</strong>erating functional of g<strong>en</strong>eralized d-dim<strong>en</strong>sional Young diagrams ford 4?(ii) What is the exact interpretation of G 4 (q) and G d (q) in g<strong>en</strong>eral?These questions are not trivial and their answer needs developing more involved mathematicalmachinery. Nevertheless, a first step towards the answer of these questions is to start by deep<strong>en</strong>ingthe study of the conjectured g<strong>en</strong>eralized d-dim<strong>en</strong>sional MacMahon function G d . In particularthe issues regarding its interpretation in 2d-quantum field theory and its explicit derivation usingcorrelation functions of local vertex operators.A natural way to reach this goal is to use the “transfer matrix” approach [5,18,29] and borrowideas from q-deformed QFT 2 [30–32]. This method has be<strong>en</strong> successfully used for the particularcase d = 3 and could, à priori, ext<strong>en</strong>ded to higher d-dim<strong>en</strong>sions. In topological string onlocal Calabi–Yau threefolds, the key idea in getting topological closed string partition functionrelies on expressing Z3dclosed as a particular vev 〈0|T |0〉 of some hermitian transfer matrix operatorT . This operator can be factorized as A + A − with A + and A − being composite local vertexoperators of the two-dim<strong>en</strong>sional c = 1 conformal field theory. Implem<strong>en</strong>tation of op<strong>en</strong> stringsleads to Z op<strong>en</strong>3d∼ C νμλ and is achieved as 〈ν t |A + (λ)A − (λ t )|μ〉 by inserting boundary states |σ 〉,σ = ν, λ, μ, described by asymptotic 2d-Young diagrams. If we let string interpretation aside,this construction could be applied as well for higher d-dim<strong>en</strong>sions where G d is expected to playa c<strong>en</strong>tral role.This paper has two main objectives:


318 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345(1) Give a conformal field theoretical derivation of the conjectured d-dim<strong>en</strong>sional g<strong>en</strong>eralizedMacMahon function G d expressed by the following formula,G d (q) =∞∏ [(1 − qk ) − (k+d−3)! ] (k−1)!(d−2)!, d 2,k=1together with the two special ones filling the hierarchy,G 1 (q) = 11 − q , G 0(q) = 1.Recall that in combinatorial analysis, the function G 3 (q) can be defined as the g<strong>en</strong>erating functionalof 3-dim<strong>en</strong>sional partitions Π (3) ext<strong>en</strong>ding the usual 2d-partitions μ = Π (2) to higher3-dim<strong>en</strong>sions. Refined studies regarding G 4 function have revealed that it is not the g<strong>en</strong>eratingfunctional of 4d partitions [18,28,33].To fix the ideas, it is interesting to recall that expanding G 2 (q) as a q n power series like,∞∏( ) 1∞∑G 2 (q) =1 − q k = p 2 (n)q n ,k=1n=0one gets the number p 2 (n) of 2d-partitions (Young diagrams) containing n boxes. From thisview, G 2 can be physically interpreted as the exact partition function Z 2 = Tr(q H ) of a twodim<strong>en</strong>sionalstatistical physics system with,(q = exp− 1KT),and <strong>en</strong>ergy spectrum E k = k. HereT is the absolute temperature and the constant K is theBoltzmann one. For instance, G 2 (q) is the partition function of the c = 1 free Bose gas. There,the Hamiltonian is giv<strong>en</strong> by H = ¯hω ∑ kN k with N k = a + k a k being the operator number ofparticles and <strong>en</strong>ergy spectrum E k = ¯hωk.Similar expansions can be also made for G d (q) which th<strong>en</strong> read as followsG d (q) =∞∑p d (n)q n , d 3.n=0For the case d = 3, the number p 3 (n) is precisely the number of 3d-partitions; but for d = 4, th<strong>en</strong>umber p 4 (n) is not the total number of 4d-partitions as it has be<strong>en</strong> explicitly checked in [28].(2) The second objective of the pres<strong>en</strong>t study is to show that G d (q) can be remarkably interpretedas a (d +1)-point correlation function G d+1 of some q-deformed vertex operators O j (x j ),i.e.(1.1)G d (q) = G d+1 (x 0 ,x 1 ,x 2 ,...,x d )with x j = q j , j = 0,...,d; and(1.2)G d+1 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 ) ···O d (x d )|0〉.The O j (x j )’s will be determined in terms of the usual vertex operators Γ ± = exp(Φ ± ) of thec = 1 two-dim<strong>en</strong>sional bosonic conformal field theory [34]; but also others, d<strong>en</strong>oted like Γ (p)± ,involving q-deformed QFT 2 . This result gives:(1.3)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 319vertex operators. In Section 4, we derive the g<strong>en</strong>eralized MacMahonfunction G n for 4d and 5d using transfer matrix method and q-deformed vertex operators Γ ±(3)and Γ (4) . In Section 5, we give the result for g<strong>en</strong>eric d-dim<strong>en</strong>sions. In Section 6, we derive(i) A q-deformed 2d quantum field theoretical proof of the conjectured MacMahon functionG d .(ii) An interpretation of G d using q-deformed c = 1 conformal field theory rather than CFT 2free field theory with c<strong>en</strong>tral charge c →∞.(iii) For d 4 G d cannot be the g<strong>en</strong>erating function of d-g<strong>en</strong>eralized partitions; but rather ofa subclass of d-partitions with very specific boundary conditions.The organization of this paper is as follows:In Section 2, we intro<strong>du</strong>ce the usual vertex operators Γ ± of the c = 1 2d conformal modeland give some of their properties ess<strong>en</strong>tial for the next steps. In Section 3, we revisit the CFT 2derivation of 3d-g<strong>en</strong>eralized MacMahon function G 3 using transfer matrix method. We also intro<strong>du</strong>cethe q-deformed Γ (2)±±O j (x j ) vertex operators involved in G d (q) re-interpreted as (d + 1)-point correlation functionG d+1 (z 0 ,...,z d ) in q-deformed c = 1CFT 2 . In the conclusion section, we summarize the mainresults of the paper accompanied with a discussion. In App<strong>en</strong>dices A and B, we give more detailson the proofs of id<strong>en</strong>tities used in the pres<strong>en</strong>t study.2. <strong>Vertex</strong> operators: Useful propertiesIn this section, we explore some basic properties of the vertex operators Γ ± (z) in c = 1 2dconformalfield theory. We study their commutation relations algebra in connection with thecounting of the Hilbert space states and the 2d-partitions (Young diagrams). We also give specialfeatures of Γ ± (z) which has motivated us to look for the relations (1.2)–(1.3).2.1. <strong>Vertex</strong> operators in c = 1 CFT 2As the c = 1 field vertex operators Γ ± (z), z ∈ C, have be<strong>en</strong> well studied and are quite knownin 2d conformal field theory [29,35,36], we shall come directly to the main points by consideringthe three following materials needed for the study of Γ ± (z) and their ext<strong>en</strong>sions to be consideredin this study:(1) U(1) Kac–Moody algebraIn CFT 2 on the complex line C parameterized by the coordinate z, theU(1) Kac–Moodyalgebra is g<strong>en</strong>erated by the holomorphic curr<strong>en</strong>t J(z) obeying the following operator pro<strong>du</strong>ctexpansion (OPE)1J(z 1 )J (z 2 ) = + regular terms.(z 1 − z 2 ) 2(2.1)Using the Laur<strong>en</strong>t expansion,J(z)= ∑ z −n−1 J n ,n∈Z∮ dzJ n =2iπ zn J(z),(2.2)the above OPE algebra reads as follows[J n ,J m ]=nδ n+m,0 .We have, amongst others, (J n ) † = J −n and J n |0〉=0forn 1.(2.3)


320 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345(2) c = 1 conformal modelIn the 2D conformal field theoretic realization of Eqs. (2.1)–(2.3), one distinguishes two freefield theoretic realizations of the c = 1 conformal repres<strong>en</strong>tation:(i) The free bosonic realization using a single real (chiral) boson Φ(z) with the usual two-pointcorrelation function,Φ(z 1 )Φ(z 2 ) =−ln(z 1 − z 2 ) + regular terms.(ii) The free fermionic realization using a complex one compon<strong>en</strong>t fermion ψ(z). In this case,the two-point correlation function that have a singular term is,ψ ∗ (z 1 )ψ(z 2 ) = 1z 1 − z 2+ regular terms.The two-point functions ψ(z 1 )ψ(z 2 ) and ψ ∗ (z 1 )ψ ∗ (z 2 ) are regular.The U(1) Kac–Moody curr<strong>en</strong>t J(z)is giv<strong>en</strong>, in the bosonic repres<strong>en</strong>tation, by:J(z)= ∂Φ(z)i∂z ,while it has the following form J(z) =:iψ ∗ (z)ψ(z): in terms of fermions. Below, we shallmainly focus on the bosonic case; the link with fermionic repres<strong>en</strong>tation can be done by usingbosonization ideas.Expanding the 2d chiral scalar field as(2.4)(2.5)(2.6)Φ(z) = ∑ n∈Zz −n Φ nand rearranging it as Φ(z) = Φ − (z) + Φ 0 + Φ + (z), we can write the above expansion as,Φ − (z) = i ∑ n>01n zn J −n , Φ + (z) =−i ∑ 1n z−n J nn>0(2.7)(2.8)wherewehaveusedΦ n = 1(2.9)in J n, n∈ Z ∗ .This id<strong>en</strong>tity follows directly by comparing Eqs. (2.6)–(2.7) and (2.2). Notice also that the zeromode Φ 0 acts trivially; it will be ignored in follows.(3) <strong>Vertex</strong> operators: Level 1There are various local field vertex operators that we will <strong>en</strong>counter in this pres<strong>en</strong>t study. Thesimplest ones, named as level 1,aregiv<strong>en</strong>byΓ ± (z) = exp ( Φ ± (z) ) , z∈ C.(2.10)The other vertex operators Γ (p)± (z), to be intro<strong>du</strong>ced later on, will be named as level p vertexoperators. Substituting Φ ± (z) by their expression (2.8), thelevel 1 vertex operators (Γ ± (z) ≡Γ ± (1) (z)) read also as follows(Γ − (z) = exp i ∑ )1n zn J −n ,n>0


(Γ + (z) = exp −i ∑ )1n z−n J n .n>0L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 321(2.11)These objects may be interpreted as the g<strong>en</strong>erating functionals of monomials of the J m operators.For instance, we have for the leading terms,J −1 = ∂Γ −(0),i∂z(J2−1 − iJ −2)=∂ 2 Γ − (0)(i∂z) 2(2.12)and similar relations for their adjoints. Notice that since the states of the Hilbert space of c = 1conformal theory repres<strong>en</strong>tation are giv<strong>en</strong> by∏(J −ni ) λ i|0〉, J ni |0〉=0, n i ,λ i ∈ N,(2.13)i1it follows that the stateΓ − (z)|0〉,is the g<strong>en</strong>erating functional of the basis states (2.13) of the c = 1CFT 2 Hilbert space.(2.14)2.2. Algebra of the Γ ± vertex operatorsThe action of the local operators Γ ± (z) on the Hilbert space states of the c = 1 2d-conformalfield theory exhibits a set of special properties inherited from the algebra of the J ±n modes (2.3).Some of these properties are revisited in what follows:(1) The Γ ± (z) operators obey the algebra,Γ ± (x)Γ ± (y) = Γ ± (y)Γ ± (x), x, y ∈ C,(Γ + (x)Γ − (y) = 1 − y ) −1Γ − (y)Γ + (x),x(2.15)which can be easily established by using Eqs. (2.4)–(2.3).(2) The operator q L 0 acts on Γ + (1) and Γ − (1) as a translation operator as shown below,q L 0Γ ± (1)q −L 0= Γ ± (q).This relation will play a crucial role later on, in particular wh<strong>en</strong> using transfer matrix method.(3) Using the properties J n |0〉=0 and 〈0|J −n = 0forn>0, we have moreover:(i) For all positions z, the operators Γ ± (z) act on the vacuum as the id<strong>en</strong>tity operator:(2.16)So we haveΓ + (z)|0〉=|0〉, 〈0|Γ − (z) =〈0|.(2.17)and〈0|Γ − (z)|0〉=1,〈0|Γ + (z)|0〉=1,(2.18)〈0|Γ − (z)Γ + (w)|0〉=1,〈0|Γ ± (z)Γ ± (w)|0〉=1.(2.19)Notice in passing that, viewing Γ ± (z) as local operator fields, Eqs. (2.18) and (2.19) may beinterpreted respectively as 1-point and 2-point Gre<strong>en</strong> functions. Notice moreover that since Γ + (z)


322 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345and Γ − (w) are non-commuting operators, we have,〈0|Γ + (z)Γ − (w)|0〉 ≠〈0|Γ − (z)Γ + (w)|0〉.(2.20)We will develop this issue much more later wh<strong>en</strong> we come to the derivation of Eqs. (1.2)–(1.3)by using correlation functions.(ii) As Γ − (z) involves all monomials in J −ni ,J λ −n ≡ ∏ i1(J −ni ) λ i,(2.21)where λ = (λ 1 ,λ 2 ,...) is a 2d-partition, the state Γ − (z)|0〉 is re<strong>du</strong>cible and is giv<strong>en</strong> by a sumover all possible 2d-partitions λ. In particular we have for z = 1,∑Γ − (1)|0〉= |λ〉.(2.22)2d partitions λA similar relation is also valid for 〈0|Γ + (1). More g<strong>en</strong>erally, this relation ext<strong>en</strong>ds as Γ − (1)|μ〉and involves the Schur function SμSchur (q) [35]. With these tools, we are in position to proceedfor higher dim<strong>en</strong>sional g<strong>en</strong>eralizations.3. The 3d-MacMahon function revisitedOur main objectives here are:(i) revisit the derivation of Z 3d ,(ii) use CFT 2 explicit computations to give argum<strong>en</strong>ts which support the exist<strong>en</strong>ce of a hierarchyof level p vertex operators Γ (p)± .To reach this goal, we first give some details on 3d-partitions (known also as plane partitions)and its g<strong>en</strong>erating functional Z 3d . Th<strong>en</strong> we pres<strong>en</strong>t the explicit computation of the function Z 3<strong>du</strong>sing transfer matrix method. As m<strong>en</strong>tioned in the intro<strong>du</strong>ction, Z 3d is precisely the amplitudeof the topological 3-vertex of closed strings on C 3 . There, the q-parameter is giv<strong>en</strong> byq = exp(−g s ),(3.1)with g s being the topological string coupling constant. Z 3d is also the partition function of cornermelting 3d-crystals.3.1. Plane partitions and 3d-Hilbert statesTo begin notice that, from the view of combinatory analysis, the 3d-MacMahon function G 3dcan be defined by the following partition function∑Z 3d =q |Π (3)| ,(3.2)3d partitions Π (3)where |Π (3) | is the number of boxes of the 3d-g<strong>en</strong>eralized Young diagram. This relation may bealso writt<strong>en</strong> as∑ 〈Z 3d =Π(3) ∣ qH ∣ Π(3) 〉 ,(3.3)3d partitions Π (3)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 323withH ∣ ∣Π (3)〉 = E ∣ ∣Π (3)〉 , E = ∣ ∣Π (3)∣ ∣.The Hilbert space states |Π (3) 〉, to which we shall refer as “3d-Hilbert states”, are the quantumstates associated with Π (3) . The relation (3.2) has a remarkable combinatorial interpretation; itis the g<strong>en</strong>erating function of the p 3 (n) number of 3d-partitions Π (3) with n boxes. The p 3 (n)number can be determined by expanding Z 3d like,Z 3d (q) =∞∑n=0p 3 (n)q n , p 3 (n) = ∂n Z 3d (0)n!∂q n .Notice also that 3d-partitions Π (3) are 3d-g<strong>en</strong>eralizations of Young diagrams and can be decomposedas a sequ<strong>en</strong>ce 1 of 2d-partitions Π t(2) like,Π (3) = ∑ t∈ZΠ (2)t ,where t parameterizes the slices. For fixed integer t, the 2d-partition Π t(2) = (Π a,a+t ) a∈N ∗ liveson the diagonal plane b = a + t of the cubic lattice N ∗ × N ∗ × N ∗ parameterized by the positiveintegers (a,b,c). The diagonal decomposition (3.6) is useful here in the s<strong>en</strong>se it is used in thetransfer matrix method for computing Z 3d . There exist an other decomposition of Π (3) namelythe so-called perp<strong>en</strong>dicular decomposition relevant for the study of the topological vertex.Expressing the number |Π (3) | of boxes of 3d-partition in terms of 2d ones, namely∣ Π(3) ∣ ∑= ∣ (2)Π ∣ ,(3.7)we can put Eq. (3.2) in the form∑( ∏Z 3d =3d partitions Π (3) ttt(2)|Πq t |).To get “3d-g<strong>en</strong>eralized Hilbert states” |Π (3) 〉, it is interesting to first recall 2d-g<strong>en</strong>eralized Hilbertspace states |Π (2) 〉≡|λ〉. In the language of the U(1) Kac–Moody algebra repres<strong>en</strong>tations, theHilbert space states of the c = 1CFT 2 have the structure|λ〉=|λ 1 ,...,λ i ,...〉,and are completely characterized by 2d-partitions,λ = (λ 1 ,...,λ i ,...), λ 1 λ 2 ···,λ i ∈ N.(3.4)(3.5)(3.6)(3.8)(3.9)(3.10)The g<strong>en</strong>erating functional of these states is giv<strong>en</strong> by Γ − (1)|0〉= ∑ λ |λ〉,Eq.(2.22). G<strong>en</strong>eralizedHilbert space states |Π (3) 〉, associated to 3d-partitions Π (3) may be built out 2d-partitions withinterlacing relations [5].1 3d-partitions Π (3) have integer <strong>en</strong>tries (Π a,b 0) such that Π a,b Π a+i,b+j i, j 0. These are 3d g<strong>en</strong>eralizationsof the usual Young diagrams described by the 2d-partitions λ = (λ 1 ,λ 2 ,...) with λ a λ a+1 . The partitions Π (3) haveseveral properties; in particular the diagonal slicing in terms of 2d partitions Π t(2) used in the transfer matrix method.The diagonal slicing of Π (3) = (Π a,b ) is obtained by setting b = a + t where t ∈ Z parameterizes the sequ<strong>en</strong>ce Π t (2) .For fixed t, Π t (2) may be thought of as λ with parts λ a = Π a,a+t .


324 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345The g<strong>en</strong>erating functional of 3d partitions requires, in the framework of transfer method, thefollowing 2 ( ∏−1(Ψ − (1)|0〉= Γ− (1)q L )) 0|0〉,(3.11)t=−∞together with(∏ ∞(〈0|Ψ + (1) =〈0| qL 0Γ + (1) )) .t=0(3.12)Notice in passing that in Eqs. (3.11)–(3.12), the pro<strong>du</strong>cts ∏ −1t=−∞ (···) and ∏ ∞t=0 (···) are tak<strong>en</strong>over diagonal slices of the 3d partitions. These pro<strong>du</strong>cts are typical ones in the transfer matrixmethod where a 3d partition is thought of as a bound state from the slice at t =−∞(in-state) tothe slice at t =+∞(out-state). The action by the operator Γ − (1) allows to g<strong>en</strong>erate all possible2d-partitions on a giv<strong>en</strong> diagonal slice as shown on Eq. (2.22). The relation (2.16) permits tomove from a slice to an other by creating all possible partitions interlacing with the partitions inthe previous slice.Therefore, using Eq. (2.16) and q L 0|0〉=|0〉, the states (3.11)–(3.12) can be rewritt<strong>en</strong> as(∏ ∞(Ψ − (1)|0〉= Γ − qk )) |0〉,(3.13)k=0and similar relation for 〈0|Ψ + (1). We de<strong>du</strong>ce from this relation the two following:(i) 3d partitions can be realized in terms of an infinite 2d ones,(ii) the g<strong>en</strong>erating functional of 3d partitions are captured by the local vertex operators 3( s∏(Ψ − (1) = lim Γ − qk )) q sL 0,(3.14)s→∞k=0and its <strong>du</strong>al Ψ + (1). These Ψ ± operators will be d<strong>en</strong>oted later asΨ ± (z) = Γ ± (2) (z),(3.15)but to keep the notations simpler, we will mom<strong>en</strong>tary use Ψ ± (z) and come later to the Γ ±(2)wh<strong>en</strong> we consider p-dim<strong>en</strong>sional g<strong>en</strong>eralization.3.2. More on 3d g<strong>en</strong>erating functionThe partition function Z 3d g<strong>en</strong>erating 3d-g<strong>en</strong>eralized Young diagrams is giv<strong>en</strong>, in the transfermatrix language, as follows( ∞) (∏−1)∏Z 3d =〈0| q L 0Γ + (1) q L 0Γ − (1)q L 0|0〉.(3.16)t=0t=−∞2 Note that Ψ ± (1) corresponds to Ψ ± (z) with z = 1. Note also that Ψ ± (1) dep<strong>en</strong>d on the q-parameter; it has be<strong>en</strong>dropped out for simplicity of notations.3 It should be noted that a 3d partition is a collection of Young diagrams. However, an arbitrary collection of Youngdiagrams do not correspond to a 3d partition.


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 325Splitting q L 0 as q L 02 q L 02 and commuting each of the operators q L 02 to the left and the other tothe right by using Eq. (2.15), we get∞∏ (Z 3d =〈0| Γ + q−t− 21 ) ∏∞ (Γ − ql− 1 )2 |0〉.t=0l=1Th<strong>en</strong> commuting the Γ − ’s to the left of Γ + , we obtain( ∞[∏ ∏ ∞ ( ) ]) 1Z 3d =(1 − q j+l .)l=0j=1By setting j + l = k, we can bring this relation to(∏ ∞ ( ) ]) 1Z 3d =(1 − q k ,)and th<strong>en</strong> toZ 3d =k=1[ k∏j=1∞∏( )1(1 − q k ) k ,k=1(3.17)(3.18)(3.19)(3.20)which is precisely the usual form of the 3d-MacMahon function. Before proceeding ahead noticethe four following:(1) Z 3d asafreeCFT 2 with c<strong>en</strong>tral charge c →∞The expression (3.18) of Z 3d is very suggestive. Setting∞Z (l)2d = ∏( )1(1 − Q l q j , Q l = q l ,)j=1which, roughly, describes a partition function Z 2d , we could th<strong>en</strong> rewrite Eq. (3.18) like∞∏Z 3d ∼ Z (l)2d .l=0(3.21)(3.22)Se<strong>en</strong> that each Z (l)2d is associated with a c = 1 free CFT 2 repres<strong>en</strong>tation, it follows that Z 3d couldbe interpreted as the partition function of a free CFT 2 repres<strong>en</strong>tation with c →∞. In Section 6,we develop an alternative interpretation of Z 3d using correlation of c = 1 q-deformed vertexoperators.(2) <strong>Vertex</strong> operators Ψ ± (z): Level 2Using Eqs. (2.11)–(3.11), it is not difficult to check that Ψ ± (z) is also a local vertex operatorwhose explicit expression in terms of the J ±n modes, reads as,(Ψ − (z) = exp i ∑ 1nn>0(Ψ + (z) = exp −i ∑ n>0( zn1 − q n )J −n),1n( z−n1 − q n )J n).The explicit derivation of these relations is giv<strong>en</strong> in App<strong>en</strong>dix A, Eq.(A.2). Notice that:(3.23)


326 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345(i) Γ ± (z) and Ψ ± (z) are related by the mappingz ±n →z±n1 − q n .(3.24)Since for q → 0, Γ ± (z) and Ψ ± (z) coincide, it follows that the operators Ψ ± (z) can be interpretedas a q-deformation Γ ± (z).(ii) Γ ± (z) and Ψ ± (z) share most of the basic quantum properties since both of them involvethe same Kac–Moody mode operators J ±n ,(3) TranslationsThe operator q L 0 acts also as a translation operator on Ψ ± (z) in the same manner like forΓ ± (z)q L 0Ψ ± (z)q −L 0= Ψ ± (qz).(3.25)This property allows us to define 4d-g<strong>en</strong>eralization from 3d one in quite similar manner as wehave done in going from 2d to 3d. We will come back to this feature later.(4) Z 3d as a “2-point correlation” functionUsing Ψ ± (1) vertex operators, the partition function Z 3d can be put in the simplest formZ 3d =〈0|Ψ + (1)q L 0Ψ − (1)|0〉.By help of the id<strong>en</strong>tity q L 0Ψ − (1)q −L 0 = Ψ − (q) Eq. (3.25),wealsohave(3.26)Z 3d =〈0|Ψ + (1)Ψ − (q)|0〉,(3.27)where Z 3d appears as just the 2-point correlation function of the level 2 vertex operators Ψ + (1)and Ψ − (q). It happ<strong>en</strong>s that Eq. (3.27) is not the unique way to define Z 3d . Let us comm<strong>en</strong>t brieflyaspects of this issue; g<strong>en</strong>eral results will be giv<strong>en</strong> in Sections 5 and 6.(i) Eq. (3.27) can be also expressed as follows( −1)∏Z 3d =〈0|Γ + (1)q L 0Ψ − (1)q L 0|0〉.(3.28)t=−∞This expression will allow us to get the definition of higher dim<strong>en</strong>sional g<strong>en</strong>eralizations ofMacMahon function; see Eq. (5.2). This relation can be put in the simple formZ 3d =〈0|Γ + (1)q L 0Ω − (1)|0〉,or equival<strong>en</strong>tlyZ 3d =〈0|Γ + (1)Ω − (q)|0〉,where we have set( s∏(Ω − (1) = lim Ψ − qk )) q sL 0.s→∞t=0(3.29)(3.30)(3.31)This local vertex operator should be thought of as the level 3 of the hierarchy we have refereedto earlier; i.e.Ω ± (1) = Γ (3)± .(3.32)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 327(ii) Along with the two repres<strong>en</strong>tations (3.27) and (3.30) the partition function Z 3d can beexpressed as well like( ) 1Z 3d =〈0|Ω + Γ − (1)|0〉,(3.33)qwhere we have used the correlation of Ω + (x) and Γ − (y) rather than Γ + (x) and Ω − (y).(iii) The diversity in expressing Z 3d as a 2-point correlation function, let us suspect that Z 3dcould be expressed as a more basic objects. In exploring this idea, we have found that the adequateinterpretation of Z 3d (q) is as a special 4-point correlation functionZ 3d (q) = G 4 (x 0 ,x 1 ,x 2 ,x 3 ),(3.34)of vertex operators O j (x j ) involving differ<strong>en</strong>t Γ (p)± levels,G 4 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 )O 3 (x 3 )|0〉.(3.35)To fix the ideas keep in mind the two following:(α) the vertex operator O 0 (x 0 ) stands for Γ + (1) and the other operators will be explicitlygiv<strong>en</strong> in Section 6;seeEqs.(6.34).(β) the observed diversity in defining Z 3d (q) corresponds just to decomposing (3.35) by usingWick theorem combined with Eqs. (2.18)–(2.19).4. Ext<strong>en</strong>sion to 4d and 5dWe first show that the leading terms of the g<strong>en</strong>eralized MacMahon function can be realized as2-point functions of some vertex operators of c = 1 2d conformal field theory.Th<strong>en</strong>, we use this feature to derive the g<strong>en</strong>eral formula for G d (q). In Section 5, we considerthe interpretation of G d (q) as (d + 1)-points correlation function G 4 (x 0 ,x 1 ,...,x d ) involvingvertex operators O j (x j ) as in Eq. (1.2).4.1. Z 1d and Z 2d as 2-point functionsBefore studying 4d and 5d g<strong>en</strong>eralizations, it is interesting to start by revisiting the 1d and2d cases. This is an important thing for getting the full picture on the conjectured MacMahonfunction G d .We start by noting the two following:(1) Recall that the 1d-MacMahon function corresponds to,Z 1d = 1(4.1)1 − q .This function can be exactly interpreted as the two-point correlation 4Z 1d = G 2 = G 2 (z 0 ,z 1 ),(4.2)of the vertex operators Γ + (1) and Γ − (q) as shown below,Z 1d =〈0|Γ + (1)Γ − (q)|0〉.(4.3)4 Z pd is the MacMahon function G pd ; it should not be confused with its interpretation as (p + 1)-points correlationfunction G p+1 = G(x 0 ,x 1 ,...,x p ) to be studied in Section 6; see also Eqs. (1.1).


328 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345This relation describes just the bosonization of Eq. (2.5) and can be rewritt<strong>en</strong>, by using theHamiltonian L 0 , as follows,Z 1d =〈0|Γ + (1)q L 0Γ − (1)|0〉.(2) A quite similar interpretation can be also giv<strong>en</strong> for 2d-MacMahon function,Z 2d = ∏ k1(1 − qk ) −1 .(4.4)(4.5)This relation can be expressed in terms of operators vertex as follows,( ∏)Z 2d =〈0|Γ + (1)q L 0Γ − (1)q L 0|0〉.k1(4.6)Indeed using (2.16), we can bring it to( ∏ (Z 2d =〈0|Γ + (1) Γ − qk )) |0〉.k1(4.7)Th<strong>en</strong> moving the operators Γ − (q k ) to the left and Γ + (1) to the right by using Eqs. (2.15), we getthe desired result.Notice that using Eq. (3.11), we learn that Z 2d can be also defined as two-point correlationfunction as followsZ 2d =〈0|Γ + (1)Ψ − (q)|0〉.(4.8)This relation involves the correlation of two vertex operators of differ<strong>en</strong>t levels namely Γ +(level 1) and Ψ − (q) (level 2). Remark also that though Γ + and Γ − do not appear on equalfooting in Eqs. (4.6)–(4.8), positivity of Z 2d is <strong>en</strong>sured because Γ + and Γ − are positive definedoperators. Notice moreover that we also haveZ 2d = G 3 = G 3 (z 0 ,z 1 ,z 3 ),but this feature will be discussed later on once we give the derivation proof of the conjecturedMacMahon function G d .4.2. Z pd derivation for p = 4, 5The property that Z 1d (4.3), Z 2d (4.8) and Z 3d (3.27) can be all of them interpreted as 2-pointcorrelation functions of some giv<strong>en</strong> vertex operators is very remarkable. It happ<strong>en</strong>s in fact thatthis feature is a more g<strong>en</strong>eral property valid also for higher-dim<strong>en</strong>sional g<strong>en</strong>eralizations. Let usdescribe this feature here for the 4d and 5d cases.Motivated by the above analysis, 4d- and 5d-g<strong>en</strong>eralizations of the MacMahon function canbe th<strong>en</strong> defined as well as 2-point correlation functions of some local operators as followsZ 4d =〈0|Ψ + (1)Ω − (q)|0〉,Z 5d =〈0|Ω + (1)Ω − (q)|0〉,(4.9)(4.10)where Ω ± (q) are vertex operators of some hierarchy level (level 3) which remain to be specified.These relations have be<strong>en</strong> motivated by the following,


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 329and alsoZ 2d =〈0|Γ + (1)Ψ − (q)|0〉,Z 3d =〈0|Ψ + (1)Ψ − (q)|0〉,(4.11)Z 0d =〈0|I id (1)Γ − (q)|0〉,Z 1d =〈0|Γ + (1)Γ − (q)|0〉,(4.12)where I id stands for the id<strong>en</strong>tity operator (of level zero). To get the Ω ± (z) operators, we require:(i) The Ω + (z) and Ω − (z) are local CFT 2 vertex operators that should obeyΩ − (x)Ω − (y) = Ω − (y)Ω − (y),Ω − (x)Ψ − (y) = Ψ − (y)Ω − (y),Ω − (x)Γ − (y) = Γ − (y)Ω − (y),Ω − (0) = 1,and similar relations for Ω + (x).(ii) We should also haveso thatΩ − (q) = q L 0Ω − (1)q −L 0Z 4d =〈0|Ψ + (1)q L 0Ω − (1)|0〉,Z 5d =〈0|Ω + (1)q L 0Ω − (1)|0〉,in analogy with the transfer matrix method used previously.(iii) We impose the commutation relations(4.13)(4.14)(4.15)Ψ + (1)Ω − (q) = G 4 (q)Ω − (q)Ψ + (1),Ω + (1)Ω − (q) = G 5 (q)Ω − (q)Ω + (1),(4.16)where G 4 (q) and G 5 (q) stand for the 4d- and 5d-g<strong>en</strong>eralized MacMahon functions giv<strong>en</strong> by,G 4 (q) =G 5 (q) =∞∏[( ) (k+1)! ] 1 (k−1)!2!1 − q k ,k=1∞∏[( ) (k+2)! ] 1 (k−1)!3!1 − q k .k=1A solution of these constraint relations is giv<strong>en</strong> by( −1( ∏ ∏−1(Ω − (1) =Γ− (1)q L )) )0q L 0,Ω + (1) =t 2 =−∞( ∞∏t 2 =0q L 0t 1 =−∞(∏ ∞(qL 0Γ + (1) ))) ,t 1 =0(4.17)(4.18)


330 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345or equival<strong>en</strong>tly likeΩ − (1) =( −1 ∏t=−∞(Ψ− (1) ) q L 0( ∞)∏Ω + (1) = q L 0Ψ + (1) .t=0),(4.19)To check that these relations solve indeed the above constraint equations, let us give some explicitdetails.4d caseFirst consider the 4d-partition function Z 4d expressed in (4.15) which we rewrite by substituting(4.19) as follows,( −1)∏Z 4d =〈0|Ψ + (1)q L 0Ψ − (1)q L 0|0〉.(4.20)t=−∞By help of Eq. (4.14), it reads also like(∏ ∞(Z 4d =〈0|Ψ + (1) Ψ − ql )) |0〉.l=1Th<strong>en</strong> commuting Ψ − (q l ) to the left by using the id<strong>en</strong>tity[ ∞()]∏ 1Ψ + (1)Ψ − (x) =(1 − xq k−1 ) k Ψ − (x)Ψ + (1), x < 1,k=1 l=1k=1see also App<strong>en</strong>dix A Eqs. (A.6) for g<strong>en</strong>eral case, we get∞∏ ∞∏()1∞∏ s∏(Z 4d =(1 − q l+k−1 ) k =s=1 k=1)1(1 − q s ) kwhich, up on using ∑ sk=1 k = s(s+1)2, can be also put in the formZ 4d =∞∏(s=11(1 − q s ) s(s+1)2),(4.21)(4.22)(4.23)(4.24)that should be compared with Eq. (4.17). Notice that like for Z 3d , the 4d partition function canbe expressed in differ<strong>en</strong>t, but equival<strong>en</strong>t, ways: We have the results⎧〈0|Ψ + (1)Ω − (q)|0〉,⎪⎨ 〈0|Ω + (q 1 Z 4d =)Ψ −(1)|0〉,(4.25)〈0|Γ + (1)Υ − (q)|0〉,⎪⎩〈0|Υ + (q 1 )Γ −(1)|0〉,where we have set


(∏ ∞(Υ − (q) = Ω − qk ))k=1which should be thought of as Υ − = Γ (4)− .L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 3315d caseSimilarly, we have for the 5d-g<strong>en</strong>eralization (4.15),(∏ ∞)Z 5d =〈0| q L 0Ψ + (1) q L 0Ω − (1)|0〉,t=0(4.26)(4.27)where we have substituted Ω + in terms of pro<strong>du</strong>ct of Ψ + (4.19). Next using the fact that q L 0 actsas a translation operator, we can put Z 5d as follows(∏ ∞(Z 5d =〈0| Ψ + q−l )) Ω − (1)|0〉.(4.28)l=1Th<strong>en</strong> using the id<strong>en</strong>tity( ) [1 ∏ ∞ ( ) s(s+1) ] ( )1 21Ψ + Ω − (1) =x1 − xq s Ω − (1)Ψ + ,xwe obtainZ 5d =∞∏ ∞∏[(s=1 l=1s=1) s(s+1)121 − q l+sThe next step is to put it in the form∞∏ k∏[( ) s(s+1)1 2Z 5d =1 − q kwhich givesZ 5d =k=1 s=1k=1],].∞∏[( ) k(k+1)(k+2) ]161 − q k .In getting this relation, we have used the id<strong>en</strong>tityk∑ s(s + 1) k(k + 1)(k + 2)= ,26s=1(4.29)(4.30)(4.31)(4.32)(4.33)proved in App<strong>en</strong>dix B. Here also we have differ<strong>en</strong>t, but equival<strong>en</strong>t, ways to define Z 5d . Later on,we will give the exact numbers of ways for g<strong>en</strong>eric Z pd .5. Result by in<strong>du</strong>ctionFirst notice that the expression (3.27) of the partition function Z 3d can be also put in the formZ 3d =〈0|Γ + (1)q L 0( −1 ∏t 2 =−∞(Ψ− (1)q L 0 )) |0〉(5.1)


332 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345or equival<strong>en</strong>tly by using Eq. (3.11), like( −1[( ∏ ∏−1Z 3d =〈0|Γ + (1)q L 0t 2 =−∞t 1 =−∞(Γ− (q)q L 0 )) q L 0This relation as well as Eqs. (4.4)–(4.6) suggest us the structure of the p-dim<strong>en</strong>sional partitionfunction Z pd in terms of CFT 2 ’s vertex operators Γ ± . For doing so, we need to intro<strong>du</strong>ce thefollowing hierarchy of local vertex operators( −1[( ∏Γ − (n+1)∏−1∏−1((1) = ···Γ− (1)q L )) ] )0q L 0···q L 0(5.3)for n 1, together withΓ (0)t n =−∞t 2 =−∞− = I id, Γ −(1) (z) = Γ −(z).Eq. (5.3) can be also defined as follows,Γ (n+1)− (1) =−1∏t=−∞t 1 =−∞( (n) ) Γ − (1)qL 0, n 1.A similar relation can be writt<strong>en</strong> down for Γ + (n+1) (1). TheΓ (p)− , referred to as the level p vertexoperator, obey quite similar relations that the ones associated to Γ − (1), in particularΓ (p)− (q) = qL 0Γ (p)− (1)q−L 0, p 0.])|0〉.(5.2)(5.4)(5.5)More details, concerning these high level operators, are pres<strong>en</strong>ted in App<strong>en</strong>dix A.Based on the preceding results realized for lower dim<strong>en</strong>sions, it follows that the p-dim<strong>en</strong>sional partition functions Z pd can be defined as,Z pd =〈0|Γ + (1)Γ (p)− (q)|0〉, p 0.This relation, which has be<strong>en</strong> explicitly checked for p = 0, 1, 2, 3, 4 and 5, reads also asZ pd =〈0|Γ + (1)q L 0Γ (p)− (1)|0〉.Commuting Γ (p)− (q) to the left of Γ +(1), we can show by in<strong>du</strong>ction that for p 2Z pd =∞∏[( ) (k+p−3)! ] 1 (k−1)!(p−2)!1 − q k .k=1Proof by in<strong>du</strong>ctionWe suppose that Eq. (5.9) holds for level p; th<strong>en</strong> prove that it holds as well for level (p + 1);that is,Z (p+1)d =〈0|Γ + (1)Γ (p+1)− (q)|0〉,and find that it is giv<strong>en</strong> by∞∏[( ) (k+p−2)! ] 1 (k−1)!(p−1)!Z (p+1)d =1 − q k .k=1(5.6)(5.7)(5.8)(5.9)(5.10)(5.11)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 333Indeed, we start from the definition of Z (p+1)d ,Z (p+1)d =〈0|Γ + (1)Γ (p+1)− (q)|0〉and express it, by using (5.5), as( ∏−1Z (p+1)d =〈0|Γ + (1)q L 0or equival<strong>en</strong>tly likeZ (p+1)d =〈0|Γ + (1)( ∞∏l=1t=−∞Γ (p)− (1)qL 0Γ (p) (− ql )) |0〉.)|0〉(5.12)(5.13)(5.14)Th<strong>en</strong> we commute Γ (p)− (ql ) to the left of Γ + (1) in Eq. (5.10), we get after some computations,∞∏ ∞∏[( ) (k+p−3)!1] (k−1)!(p−2)!Z (p+1)d =(5.15)1 − q l+k .l=1 k=1Setting s = (l + k), we can rewrite this relation as followsZ (p+1)d =∞∏s=1 k=1s∏[( ) (k+p−3)! ] 1 (k−1)!(p−2)!1 − q s .(5.16)At first sight, this expression seems differ<strong>en</strong>t from the desired result; however explicit computationleads exactly to the right result; thanks to the combinatorial id<strong>en</strong>tity,s∑ (k + p − 3)! (s + p − 2)!= , p 2,(5.17)(k − 1)!(p − 2)! (s − 1)!(p − 1)!k=1which is showed in App<strong>en</strong>dix B.These computations give an explicit proof for the derivation of the expression of g<strong>en</strong>eralizedMacMahon function. Thanks to the “transfer matrix method” and to the hierarchy of level pvertex operators Γ (p)± Eq. (5.5).6. G n (q) as (n + 1)-point correlation functionSo far we have se<strong>en</strong> that n-dim<strong>en</strong>sional g<strong>en</strong>eralization of MacMahon function G n (q) withn 2, can be interpreted as 2-point correlation functions of some composite vertex operators.We have also se<strong>en</strong> that there are differ<strong>en</strong>t, but equival<strong>en</strong>t ways, to express G n (q) as 2-pointcorrelation functions. Using Γ ± (r) (s)and Γ ± vertex operators, one can check that for any positivedefinite integers r and s such that r + s − 1 = n,wehave,G n (q) =〈0|Γ (n−s+1)+ (1)Γ − (s)(q)|0〉, 1 s n, n 1.The 2( r+s−22) + 1 possibilities are all of them equal to each other. This diversity in definingG n (q) suggests us to look for a more refined definition of it. We have found that the adequateway to define G n (q) is like a (n + 1)-point correlation function as giv<strong>en</strong> below,(6.1)G n (q) = G n+1 (x 0 ,x 1 ,x 2 ,...,x n )(6.2)


334 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345withG n+1 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 ) ···O n (x n )|0〉,where the x j = x j (q) and O j (x j ) are some vertex operators that have to be specified. In thisway, the diversity (6.1) appears just as a manifestation of applying Wick theorem to (6.3) for itsdecomposition in terms of two-points correlation functions. To fix the ideas, think about O 0 (x 0 )as giv<strong>en</strong> byO 0 (x 0 ) = Γ + (1)and all remaining others as giv<strong>en</strong> by vertex operators involving pro<strong>du</strong>cts of Γ − (y) only, that is:(6.3)(6.4)O j (x j ) ∼ ∏ Γ − (y), j = 1,...,n.Since Γ − (y) and any pro<strong>du</strong>ct of Γ − (y) has vacuum expectation values equal to one,)〈0|Γ − (y)|0〉=1 =〈0|(∏Γ− (y) |0〉,it follows thatn∏〈0| O l (x l )|0〉=1.l=1Th<strong>en</strong>, by using Wick theorem G n+1 re<strong>du</strong>ces to(6.5)(6.6)(6.7)G n+1 = ∏ k〈0|O 0 (x 0 )O k (x k )|0〉,k= 1,...,n.(6.8)Let us build the Gre<strong>en</strong> function G n+1 step by step, starting from Eq. (5.7) and using the resultsobtained above:6.1. Leading termsBelow, we give the explicit computation of G n for n = 2, 3, 4, 5.(1) G 1 (q) as 2-point propagatorComparingG 1 (q) =〈0|Γ + (1)Γ − (q)|0〉withG 2 =〈0|O 0 (x 0 )O 1 (x 1 )|0〉we getO 0 (x 0 ) = Γ + (1),O 1 (x 1 ) = Γ − (q).(2) G 2 (q) as a 3-point functionStarting from the expression Eq. (5.7) for G 2 ,G 2 (q) =〈0|Γ + (1)Γ −(2) (q)|0〉,and using the special property established in App<strong>en</strong>dix A;seeEq.(A.26),(6.9)(6.10)(6.11)(6.12)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 335Γ (2)−(q) = Γ(1)−(2)( (q)Γ − q2 )we can bring G 2 (q) into the formG 2 (q) =〈0|Γ + (1)Γ (1)−Th<strong>en</strong>, comparing with(q)Γ(2)−G 3 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 )|0〉,(q2 ) |0〉.we get, in addition to Eqs. (6.11), the following:O 2 (x 2 ) = Γ +(2) (q2 ) .(6.13)(6.14)(6.15)(6.16)(3) G 3 (q) as a 4-point functionWe start from the expression of G 3 (q),G 3 (q) =〈0|Γ + (1)Γ −(3) (q)|0〉th<strong>en</strong> use the id<strong>en</strong>tity,Γ (3)−(q) = Γ(2)−and substitute Γ (2)(3)( (q)Γ − q2 ) ,(6.17)(6.18)(q2 )) Γ (3) (q3 ) .(6.19)− (q) by Eq. (6.13), we get( 2∏Γ − (3) (1)(q) = Γ − (q) l=1Comparing withΓ (2)−−we obtainG 3 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 )O 3 (x 3 )|0〉,(6.20)O 0 (x 0 ) = Γ + (1),O 1 (x 1 ) = Γ −(1)( (q),2∏O 2 (x 2 ) = Γ (2) (− q2 )) ,l=1O 3 (x 3 ) = Γ −(3) (q3 ) .(6.21)(4) G 4 (q) as a 5-point functionStarting fromG 4 (q) =〈0|Γ + (1)Γ −(4) (q)|0〉th<strong>en</strong> using the id<strong>en</strong>tities,Γ − (4) (3) (4)( (q) = Γ − (q)Γ − q2 ) ,Γ −(4) (q2 ) = Γ −(3) (q2 ) Γ −(4) (q3 ) ,Γ −(4) (q3 ) = Γ −(3) (q3 ) Γ −(4) (q4 ) ,(6.22)(6.23)


336 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345we obtain at a first stageΓ (4)−(q) = Γ(3)−(q)Γ(3)−(q2 ) Γ (3)−(q3 ) Γ (4) (q4 ) .At a second stage, we substitute Γ − (3)(3)(q) and Γ − (q2 ) by Eq. (6.18), we getΓ − (3) (2) (3)( (q) = Γ − (q)Γ − q2 ) ,Γ −(3) (q2 ) = Γ −(2) (q2 ) Γ −(3) (q3 ) .Putting back into Eq. (6.24), we obtainΓ − (4) (2) (2)( (q) = Γ − (q)Γ − q2 ) Γ −(2) (q2 ) Γ −(3)Next replacing Γ (2)−(q3 ) Γ (3)−(q3 ) Γ (3)−(q3 ) Γ (4) (q4 ) .−(6.24)(6.25)− (q) by Eq. (6.13), we <strong>en</strong>d with the following, (q2 ))( 3∏ (q3 )) Γ (4) (q4 ) .(6.26)( 3∏Γ − (4) (1)(q) = Γ − (q) l=1Comparing withwe obtainΓ (2)−l=1Γ (3)−G 5 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 )O 3 (x 3 )O 4 (x 4 )|0〉,O 0 (x 0 ) = Γ + (1),O 1 (x 1 ) = Γ −(1)( (q),3∏O 2 (x 2 ) = Γ (2) (− q2 )) ,l=1−(6.27)( 3∏O 3 (x 3 ) = Γ (3) (− q3 )) ,l=1O 4 (x 4 ) = Γ −(4) (q4 ) .With these results on lower values of p, it is straightforward to derive the g<strong>en</strong>eric picture.6.2. G<strong>en</strong>eric result(6.28)We start from the expression ofG p (q) =〈0|Γ + (1)Γ (p)− (q)|0〉, p 1.Th<strong>en</strong> we use the id<strong>en</strong>tity,withΓ (p)−p k =p−1(q) = ∏k=0( pk∏l k =1Γ −(k+1) (qk+1 )) ,(p − 1)!k!(p − k − 1)! , 0 k p − 1,whichisprovedinApp<strong>en</strong>dix A,Eq.(A.37), we can bring G p (q) into the form,(6.29)(6.30)(6.31)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 337∏G p (q) =〈0|Γ + (1)By comparing with,we obtain(p−1 pkk=0∏l k =1Γ −(k+1) (qk+1 )) |0〉, p 1.G p+1 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 ) ···O p (x p )|0〉,(6.32)(6.33)O 0 (x 0 ) = Γ + (1), O 1 (x 1 ) = Γ − (1)( (q), ∏ pj[ (j)( O j (x j ) = Γ qj )]) , j 2.(6.34)l j =1−For more details on the derivation of this relation, see App<strong>en</strong>dix A: Eqs. (A.27)–(A.37).Eqs. (6.34) complete the interpretation of G d as a (d + 1)-points Gre<strong>en</strong> function.7. Discussion and conclusionIn this paper, we have giv<strong>en</strong> a 2d-conformal field theoretical derivation of the g<strong>en</strong>eralizedMacMahon function by using ideas from “transfer matrix method” and q-deformed QFT 2 .Among our results, we m<strong>en</strong>tion:(1) The usual vertex operators Γ ± (z) of the bosonic c = 1 conformal field theory appear asthe level one of the following hierarchy,( ∞Γ (p)− (z)|0〉=exp ∑n=1)iz n J −nn(1 − q n ) p−1 |0〉, p 1where q = exp(−g s ). These local operators, which coincide in the limit q → 0; that is wh<strong>en</strong> g sgoes to ∞, can be obtained from Γ ± (z) by making the substitutionz n →z n(1 − q n , p 2.) p−1The Γ (p)− s form th<strong>en</strong> an infinite hierarchy of q-deformed vertex operators and obey commutationrelations quite similar to those satisfied by the level one Γ −(1) (z) = Γ −(z). In particular we have,Γ (1)+(7.1)(7.2)(1)Γ(p)−(q) = G p(q)Γ (p) (1)(q)Γ + (1),−where G p (q) is precisely the g<strong>en</strong>eralized p-dim<strong>en</strong>sion MacMahon function. We also have thefollowing g<strong>en</strong>eral relation,〈0|Γ + (1) (z 1)Γ − (l+1) (z l )|0〉= ∏ ··· ∏ [ ∏() ] 1(7.4)(1 − q k 0+k 1 +···+k l z l.k l =0 k 1 =0 k 0 =0z 1)This relation can be giv<strong>en</strong> an interpretation as l copies of c =∞free CFT 2 repres<strong>en</strong>tations.Indeed, setting z lz 1= q, q k 0+k 1 +···+k l= Q k q k 0 with Q k = q k 1+···+k landZ 2 (Q k ,q)=∞∏(k 0 =01(1 − Q k q k 0)we can put the right-hand side of (7.4) like,), k = (k 1 ,...,k l ),(7.3)(7.5)


338 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345∞∏(k 1 ,...,k l )=0Z 2 (Q k ,q)∼ ( [Z 2 ] ∞) l .This factorization suggests that, roughly, G l+1 (q) could be interpreted as giv<strong>en</strong> by the pro<strong>du</strong>ctof l copies of infinite pro<strong>du</strong>cts of Z 2 . Since from 2d conformal free field theory view, each Z ∞ 2copy should be described by a free field CFT 2 repres<strong>en</strong>tation with c =∞,theG l+1 (q) partitionfunction would th<strong>en</strong> correspond to a c<strong>en</strong>tral chargec = k l ,with k →∞.(2) Using the above level p vertex operators, we have shown that the p-dim<strong>en</strong>sional g<strong>en</strong>eralizedMacMahon function is giv<strong>en</strong> by the following two-point correlation functionG p (q) =〈0|Γ (1)+(p)(1)Γ − (q)|0〉.(3) The level p vertex operators Γ (p)− satisfy several remarkable properties, in particular theycan be realized as cond<strong>en</strong>sates of vertex operators of lower levels as shown below,Γ (p)−(z) = Γ(p−1)−(z)Γ (p)− (qz),so that G p (q) can be defined as a particular (p + 1)-point correlation function as giv<strong>en</strong> below,G p+1 =〈0|O 0 (x 0 )O 1 (x 1 )O 2 (x 2 ) ···O p (x p )|0〉,(7.6)(7.7)(7.8)(7.9)(7.10)where the O j (x j ) are giv<strong>en</strong> by Eqs. (6.32)–(6.33). This correlation function can be expressed indiffer<strong>en</strong>t forms by using Wick theorem and the property (6.7).(4) Based on the field theoretical derivation giv<strong>en</strong> in the pres<strong>en</strong>t study, we learn that the functionG p (q) with p 4 cannot be the g<strong>en</strong>erating functional of the p-dim<strong>en</strong>sional g<strong>en</strong>eralizedYoung diagrams.Recall that for the case p = 3, solid partitions Π (3) ext<strong>en</strong>ding Young diagrams have g<strong>en</strong>erallythree boundaries giv<strong>en</strong> by 2d partitions λ, μ and ν. The typical g<strong>en</strong>erating functional ofall possible 3d partitions Ψ (3) with boundaries ∂(Ψ (3) ) = (λ,μ,ν) is giv<strong>en</strong> by the correlationfunction C λμνC λμν =〈ν t (|A + (λ)A − λt ) |μ〉,(7.11)where A + (λ)A − (λ t ) is the transfer matrix operator described previously. For the simplest casewhere ∂(Ψ (3) ) = (∅, ∅, ∅), the correlation function C ∅∅∅ is precisely the g<strong>en</strong>erating functional of3d partitions.For higher values of p;sayp = 4, one has 4d g<strong>en</strong>eralized Young diagrams Ψ (4) . This 4d partitionshave g<strong>en</strong>erally four 3-dim<strong>en</strong>sional boundaries captured by 3d partitions Λ (3) , Σ (3) , Υ (3)and Π (3) . The typical g<strong>en</strong>erating functional of all possible 4d partitions Π (4) with boundaries∂(Ψ (4) ) = (Λ (3) ,Σ 3 ,Υ (3) ,Π (3) ) is giv<strong>en</strong> by the correlation function C ΛΣΥ Ψ . This functionalext<strong>en</strong>ds (7.11) and can be defined as〈〈Λ(3) ∥ ( A′− Σ 3 ,Υ (3)) A ′ (+ Σ 3 ,Υ (3))∥ ∥ Π(3) 〉〉 ,(7.12)where A ′ − (Σ3 ,Υ (3) )A ′ + (Σ3 ,Υ (3) ) is some g<strong>en</strong>eralized transfer matrix operator acting on 3dpartition states ‖Π (3) 〉〉. It is this function that would g<strong>en</strong>erate the 4d g<strong>en</strong>eralized Young diagramswith boundaries Λ, Σ, Υ and Π.


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 339Moreover, using the fact that 3d partitions Π (3) may themselves be sliced in terms of 2dpartitions, one can usually bring the correlation function C ΛΣΥ Ψ to the form,〈〈ς,τ,υ‖A ′ +[(λ,μ,ν); (ζ,η,θ)]A′−[(λ t ,μ t ,ν t) ; ( ζ t ,η t ,θ t)] ‖α, β, γ 〉〉,(7.13)where ‖ϑ, σ, ϱ〉〉 is a 3d partition boundary state expressed in terms of 2d partitions |ϑ〉⊗|σ 〉⊗|ϱ〉. In the particular case α = β = γ =∅and ς = τ = υ =∅, the correlation function becomes〈〈∅, ∅, ∅‖A ′ +[(λ,μ,ν); (ζ,η,θ)]A′−[(λ t ,μ t ,ν t) ; ( ζ t ,η t ,θ t)] ‖∅, ∅, ∅〉〉.In the special case ζ = η = θ = λ = μ = ν =∅, the above quantity simplifies as(7.14)〈〈∅, ∅, ∅‖A ′ + [∅]A′ − [∅]‖∅, ∅, ∅〉〉.(7.15)From this g<strong>en</strong>eral relation, we see that the MacMahon function G 4 (q) =〈0|Ψ + (1)Ω − (q)|0〉Eq. (4.10) appears as a very particular correlation function and th<strong>en</strong> cannot be the g<strong>en</strong>eratingfunctional of all possible 4d partitions.Acknowledgem<strong>en</strong>tsThis research work is supported by the program Protars III D12/25. H.J. would like to thankICTP for kind hospitality where part of this work has be<strong>en</strong> done. The authors thank B. Sz<strong>en</strong>droifor helpful suggestion.App<strong>en</strong>dix A. <strong>Vertex</strong> operators Γ (n)± (x)In this app<strong>en</strong>dix we describe the vertex operators Γ ± (n) (x) and their commutation relationsalgebra.We first study the level n vertex operators Γ ± (n) (x) and their main properties starting byΓ ± (2) = Ψ ±. Th<strong>en</strong>, we give their algebra.A.1. Level 2 vertex operatorTo begin notice that the operators Ψ ± (1) Eq. (3.11), d<strong>en</strong>oted also as Γ ± (2) (1), can be put in theform,[( s∏(Ψ − (1) = lim Γ − qt )) ]q sL 0,s→∞t=0[ ( s∏Ψ + (1) = lim q sL (0Γ + q−t ))] .(A.1)s→∞t=0Using the expression of Γ ± (z) Eq. (2.11), we can rewrite Ψ ± (1) as follows:( −1)∏∞∏Ψ − (y) = Γ − (y)q L (0= Γ − q k y ) ,t=−∞k=0( ∞)∏∞∏Ψ + (x) = q L (0Γ + (x) = Γ + q −k x ) ,t=0t=0(A.2)


340 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345or equival<strong>en</strong>tly like( ∑ i y n )Ψ − (y) = expn (1 − q n ) J −n ,n1(Ψ + (x) = exp − ∑ i x −n )n (1 − q n ) J n .n1(A.3)Let us compute the algebra of these vertex operators.First, we have,q L 0Ψ ± (z)q −L 0= Ψ ± (qz),showing that q L 0 acts as a translation operator. We also have(A.4)Ψ ± (x)Ψ ± (y) = Ψ ± (y)Ψ ± (x).(A.5)To get the commutator betwe<strong>en</strong> Ψ + (x) and Ψ − (y), we can do it in two ways which, by theircomparison, allow us to get a new id<strong>en</strong>tity:(i) Computation by using pro<strong>du</strong>cts of Γ ± .Wehave,Ψ + (x)Ψ − (y) = ∏ (Γ + q l x ) ∏ (Γ − q k y )l0 k0∞∏(= 1 − q s y ) −(s+1)Ψ − (y)Ψ + (x),(A.6)xs=0in particular(∏ ∞(Ψ + (1)Ψ − (q) = 1 − qt ) )−tΨ − (q)Ψ + (1).Noticet=1Γ + (1)Ψ − (y) = Γ + (1) ∏ k0Γ −(q k y )(A.7)= ∏ k0(1 − q k y ) −1 Ψ− (y)Γ + (1)(A.8)and alsoΓ + (x)Ψ − (y) = Γ + (x) ∏ k0Γ −(q k y )= ∏ (1 − q k y ) −1Ψ − (y)Γ + (x).xk0(A.9)We also haveΨ + (x)Γ − (1) = ∏ k0Γ +(q −k x ) Γ − (1)= ∏ k0(1 − q k x −1) −1 Γ− (1)Ψ + (x).(A.10)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 341(ii) Computation using directly Eqs. (A.3). We get,( ∑ 1 y n x −n )Ψ + (x)Ψ − (y) = expn (1 − q n ) 2 Ψ − (y)Ψ + (x).n1By comparing the two expressions (A.6) and (A.11), we get the following id<strong>en</strong>tity( ∑ 1 y n x −n ) ∞∏(expn (1 − q n ) 2 = 1 − q s y ) −(s+1),xn1s=0or equival<strong>en</strong>tly like,∑ 1 y n x −n ∞n (1 − q n ) 2 =− ∑[ ((s + 1) ln 1 − q s y )].xn1s=0(A.11)(A.12)A.2. G<strong>en</strong>eric q-deformed operatorsHere we give the expressions of the g<strong>en</strong>eric q-deformed operators and some useful propertiesof their algebra.The starting point is the vertex operators(Γ − (z) = exp i ∑ )(1n zn J −n , Γ + (z) = expn>0−i ∑ n>0)1n z−n J n(A.13)and the aim is:(1) compute for n 1, the following hierarchy of composite vertex operators( ∞[Γ − (n+1) ∏ ∞(∏ ∞) ] )∏(z) = ··· Γ − (z)q L 0q L 0···q L 0.t n =1t 2 =1t 1 =1(A.14)For n = 0, we have just Γ −(1) (z) = Γ −(z). Similar quantities can be writt<strong>en</strong> down for Γ + (n+1) (z);we shall not report them here.(2) Derive the id<strong>en</strong>tity (6.30).For these purposes, we proceed by using in<strong>du</strong>ctive method:q-deformed vertex operators Γ (2)−(a) Case Γ −(2) (z)In this case, we have(3)(z) and Γ − (z):Γ (2)−Notice thatΓ (2)−∞ (z) = ∏( Γ− (z)q L ) 0.t=1(z) = lims→∞t=1s∏ (Γ− (z)q L ) 0.(A.15)(A.16)


342 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345By using the fact that q L 0 acts as translation operator on Γ − (z), we get([ s∏Γ −(2)( ((z) = lim Γ− q k z ))] )q sL 0.s→∞k=0(A.17)For simplicity, we consider the action of Γ − (2) (z), on the vacuum, that reads as( (z)|0〉= ∏ ∞(Γ − q k z )) |0〉.(A.18)Γ (2)−k=0Substituting Γ − (q k z) by its expression (A.13), we find(Γ − (2) (z)|0〉=exp i ∑ )∞∑ q knn zn J −n |0〉n>0 k=0(= exp i ∑ 1 z n )n (1 − q n ) J −n |0〉.n>0(A.19)Notice that Γ − (2) (z)|0〉 can be decomposed as follows, (Γ − (2) (z)|0〉=exp i ∑ ) (1(A.20)n zn J −n exp i ∑ ∞ )∑q n q knn zn J −n |0〉n>0n>0 k=0or equival<strong>en</strong>tly like(Γ − (2) (z)|0〉=exp i ∑ ) (1n zn J −n exp i ∑ 1 (qz) n )n (1 − q n ) J −n |0〉n>0n>0(A.21)showing that we have:Γ (2)−(1) (2)(z)|0〉=Γ (z)Γ (qz)|0〉, z∈ C.−t 2 =1−(b) Case Γ −(3) (z)Here, we have∞Γ −(3) (z)|0〉= ∏ [ (2) ] ∞∏ (Γ − (z)qL 0 (2)( |0〉= Γ − q k z )) |0〉.k=0Substituting Γ −(2) (qk z) by its expression (A.2), we find(Γ − (3) (z)|0〉=exp i ∑ )1∞∑q kn z nn (1 − q n ) J −n |0〉n>0 k=0(= exp i ∑ 1 z n )n (1 − q n ) 2 J −n |0〉.n>0Finally, if we rewrite the above relation as follows( ∑Γ − (3) (z)|0〉=exp iz n ) (J −n ∑ i(qz) n )J −nn(1 − q n exp) n(1 − q n ) 2 |0〉n>0n>0(A.22)(A.23)(A.24)(A.25)


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 343we obtain the relationΓ (3)−(z)|0〉=Γ(2)−(3)(z)Γ − (qz)|0〉as claimed in Section 6 Eq. (6.18).(A.26)Higher levelsFrom the above analysis, it is not difficult to check that the explicit expression of the vertexoperators Γ (p)− (z) acting on the vacuum is giv<strong>en</strong> by, (A.27)( ∞Γ (p)− (z)|0〉=exp ∑n=1Moreover using the id<strong>en</strong>tity,z n p−1(1 − q n ) p−1 = ∑ (qz) n(1 − q n ) k−1 ,k=1)iz n J −nn(1 − q n ) p−1 |0〉, p 1.we can decompose the above relation as followsΓ (p)−(z)|0〉=Γ(1)−p−1(z) ∏k=2Γ −(k) (qz)|0〉.Th<strong>en</strong>extstepistousetherelationz n(1 − q n ) p−1 = z n(1 − q n ) p−2 + (qz) n(1 − q n ) p−1 ,that imply the equalityΓ (p)−(p−1)(z)|0〉=Γ (z)Γ (p) (qz)|0〉.−−Doing the same for the first term for the right-hand sidez n(1 − q n ) p−2 = z n(1 − q n ) p−3 + (qz) n(1 − q n ) p−2 ,Eq. (A.30) can be brought to the formz n(1 − q n ) p−1 = z n(1 − q n ) p−3 + (qz) n(1 − q n ) p−2 + (qz) n(1 − q n ) p−1 ,leading th<strong>en</strong> toΓ (p)−(z)|0〉=Γ(p−2)−(z)Γ (p−1)−(qz)Γ (p)− (qz)|0〉.We can repeat this operation successively to <strong>en</strong>d with the two following:(i) the expression of level p vertex operator as we have used it in Section 6 Eq. (6.32),(A.28)(A.29)(A.30)(A.31)(A.32)(A.33)(A.34)withZ pd =〈0|T |0〉,∏p jT = Γ + (1) (z) Γ (j+1) (− q j z ) ,i=1(A.35)(A.36)


344 L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345where p j are giv<strong>en</strong> byp j =(p − 1)!, j = 0,...,p− 1.j!(p − j − 1)!(A.37)(ii) Using the decomposition for Γ (p)− (z),Γ (p)− (z) = O 0(x 0 )O 1 (x 1 )O 2 (x 2 ) ···O p−1 (x p−1 )(A.38)we getO j+1 (x j+1 ) =∏p ji=1Γ (j+1) (− q j z ) , j = 0,...,p− 1.(A.39)App<strong>en</strong>dix B. Combinatorial Eq. (5.17)Here we want to derive the id<strong>en</strong>tity (5.17) namely,s∑k=1C p−2k+p−3 = Cp−1 s+p−2, p 2.This is a standard combinatorial id<strong>en</strong>tity; its proof follows from basic property [37],C k n+1 = Ck−1 n + C k n .(B.1)(B.2)Applying this id<strong>en</strong>tity to Cn k and putting it back into the above relation, we get,(B.3)Cn+1 k = Ck−1 n + C k−1n−1 + Ck n−1 .By in<strong>du</strong>ction, it results,C k n+1 =n∑j=k−1C k−1j.Setting k = p − 1 and n = s + p − 3, we recover the id<strong>en</strong>tity (B.1).(B.4)Refer<strong>en</strong>ces[1] H.N. Temperley, Statistical mechanics and the partition of numbers I: The transition to liquid helium, Proc. R. Soc.London A 199 (1949) 361;H.N. Temperley, Statistical mechanics and the partition of numbers II: The form of crystal surfaces, Proc. CambridgePhilos. Soc. 48 (1952) 683.[2] E.J. van R<strong>en</strong>sburg, The Statistical Mechanics of Interacting Walks, Polygons, Animals and Vesicles, Oxford Univ.Press, Oxford, 2000.[3] C. Weiss, M. Holthaus, From number theory to statistical mechanics: Bose–Einstein cond<strong>en</strong>sation in isolated traps,Chaos Solitons Fractals 10 (1999) 795.[4] V. Elser, Solution of the dimer problem on a hexagonal lattice with boundary, J. Phys. A 17 (1984) 1509.[5] A. Okounkov, N. Reshetikhin, C. Vafa, Quantum Calabi–Yau and classical crystals, hep-th/0309208.[6] R. K<strong>en</strong>yon, An intro<strong>du</strong>ction to the dimer model, math.CO/0310326.[7] R. K<strong>en</strong>yon, A. Okounkov, S. Sheffield, Dimers and amoebae, math-ph/0311005.[8] D. Ghoshal, C. Vafa, c = 1 string as the topological theory of the conifold, Nucl. Phys. B 453 (1995) 121, hepth/9506122.[9] E. Witt<strong>en</strong>, Ground ring of two-dim<strong>en</strong>sional string theory, Nucl. Phys. B 373 (1992) 187, hep-th/9108004.


L.B. Drissi et al. / Nuclear Physics B 801 [FS] (2008) 316–345 345[10] E.H. Saidi, M.B. Sedra, Topological string in harmonic space and correlation functions in S 3 stringy cosmology,Nucl. Phys. B 748 (2006) 380–457, hep-th/0604204;E.H. Saidi, Topological SL(2) gauge theory on conifold, hep-th/0601020.[11] M.R. Douglas, G. Moore, D-branes, Quivers, and ALE instantons, hep-th/9603167.[12] R. Ahl Laamara, M. Ait B<strong>en</strong> Haddou, A. Belhaj, L.B. Drissi, E.H. Saidi, RG Cascades in hyperbolic quiver gaugetheories, Nucl. Phys. B 702 (2004) 163–188, hep-th/0405222.[13] M. Ait B<strong>en</strong> Haddou, A. Belhaj, E.H. Saidi, Geometric <strong>en</strong>gineering of N = 2 CFT_{4}s based on indefinite singularities:Hyperbolic case, Nucl. Phys. B 674 (2003) 593–614, hep-th/0307244.[14] M.A. B<strong>en</strong>haddou, E.H. Saidi, Explicit analysis of Kähler deformations in 4D N = 1 supersymmetric quiver theories,Phys. Lett. B 575 (2003) 100–110, hep-th/0307103.[15] J.J. Heckman, C. Vafa, Crystal melting and black holes, hep-th/0610005.[16] M. Aganagic, A. Klemm, M. Marino, C. Vafa, The topological vertex, Commun. Math. Phys. 254 (2005) 425–478,hep-th/0305132.[17] A. Iqbal, N. Nekrasov, A. Okounkov, C. Vafa, Quantum foam and topological strings, hep-th/0312022.[18] A. Iqbal, C. Kozcaz, C. Vafa, The refined topological vertex, hep-th/0701156.[19] L.B. Drissi, H. Jehjouh, E.H. Saidi, Topological strings on local elliptic curve and non-planar 3-vertex formalism,arXiv: 0712.4249 [hep-th], Lab/UFR-PHE, 0701.[20] V.G. Kac, A.K. Raina, Repres<strong>en</strong>tations of Infinite Dim<strong>en</strong>sional Lie Algebras, Bombay Lectures on Highest Weight,World Sci<strong>en</strong>tific, 1987.[21] V.G. Kac, Infinite Dim<strong>en</strong>sional Lie Algebras, third ed., Cambridge Univ. Press, 1990.[22] E.H. Saidi, M. Zakkari, The G<strong>en</strong>erating functional of higher conformal spin ext<strong>en</strong>sions of the Virasoro algebra,Phys. Lett. B 281 (1992) 67–71.[23] H. Awata, M. Fukuma, Y. Matsuo, S. Odake, Repres<strong>en</strong>tation theory of the W 1+∞ algebra, Prog. Theor. Phys.Suppl. 118 (1995) 343–374, hep-th/9408158.[24] E. Fr<strong>en</strong>kel, V. Kac, A. Ra<strong>du</strong>l, W.-Q. Wang, W 1+∞ and W(gl N ) with c<strong>en</strong>tral charge N, Commun. Math. Phys. 170(1995) 337–358, hep-th/9405121.[25] V.G. Kac, A.K. Raina, Bombay Lectures on Highest Weight Repres<strong>en</strong>tations of Infinite Dim<strong>en</strong>sional Lie Algebras,World Sci<strong>en</strong>tific, 1987.[26] M. Ait B<strong>en</strong> Haddou, Thèse de doctorat de 3ème cycle, Sur Les Repres<strong>en</strong>atations des Algèbres de Lie de dim<strong>en</strong>sionsInfinies, Faculté des Sci<strong>en</strong>ces, Rabat, 1994.[27] H. Awata, M. Fukuma, Y. Matsuo, S. Odake, Determinant formulae of quasi-finite repres<strong>en</strong>tation of W 1+∞ algebraat lower levels, Phys. Lett. B 332 (1994) 336–344, hep-th/9402001.[28] V. Muston<strong>en</strong>, R. Rajesh, Numerical estimation of the asymptotic behaviour of solid partitions of an integer, J. Phys.A 36 (2003) 6651, cond-mat/0303607.[29] A. Okounkov, N. Reshetikhin, Random skew plane partitions and the Pearcey process, math/0503508.[30] C. Vafa, Two-dim<strong>en</strong>sional Yang–Mills, black holes and topological strings, hep-th/0406058.[31] M. Aganagic, H. Ooguri, N. Saulina, C. Vafa, Black holes, q-deformed 2d Yang–Mills, and non-perturbative topologicalstrings, Nucl. Phys. B 715 (2005) 304–348, hep-th/0411280.[32] R. Ahl Laamara, A. Belhaj, L.B. Drissi, E.H. Saidi, Black holes in type IIA string on Calabi–Yau threefoldswith affine ADE geometries and q-deformed 2d quiver gauge theories, Nucl. Phys. B 776 (2007) 287–326, hepth/0611289.[33] O. Foda, M. Wheeler, BKP plane partitions, JHEP 0701 (2007) 075, math-ph/0612018.[34] P. Ginsparg, G. Moore, Lectures on 2D gravity and 2D string theory (TASI 1992), hep-th/9304011.[35] A. Okounkov, N. Reshetikhin, Correlation function of Schur process with application to local geometry of a random3-dim<strong>en</strong>sional Young diagram, math.CO/0107056.[36] N. Saulina, C. Vafa, D-branes as defects in the Calabi–Yau crystal, hep-th/0404246.[37] I.S. Gradshteyn, I.M. Rhyzik, Tables of Integrals, Series and Pro<strong>du</strong>cts, sev<strong>en</strong>th ed., Academic Press, 1980.


Chapitre 5<strong>Vertex</strong> <strong>Topologique</strong> UsuelDans le cadre de la théorie des cordes topologiques sur les variétés de CY3 toriques,le calcul des amplitudes A top correspondantes a constitué un défi important pour la communautédes physici<strong>en</strong>s des théories des supercordes. Les recherches <strong>en</strong>treprises dans cedomaine ont permis d’explorer plusieurs voies comme nous l’avons déjà évoqué dans cemanuscrit. Dans leur travail original [71], Aganagic et al ont élaboré un formalisme basésur l’intro<strong>du</strong>ction <strong>du</strong> vertex topologique C λµν qui s’est avéré un outil sophitiqué pour lecalcul des expressions explicites des l’amplitudes topologiques. L’objectif principal de cechapitre consiste à contribuer à l’étude <strong>du</strong> vertex topologique usuel 1 et à l’utiliser pourcalculer la fonction de partition Z top ainsi que les amplitudes A top de la théorie des cordestopologiques. Après avoir intro<strong>du</strong>it le vertex topologique C λµν et les ingrédi<strong>en</strong>ts qui lesaccompagn<strong>en</strong>t comme le framing et les règles de collage, nous illustrons le formalisme <strong>du</strong>vertex topologique sur quelques exemples concrets. Nous établirons égalem<strong>en</strong>t la relation<strong>en</strong>tre des invariants topologiques et le vertex topologique. Nous clôturons ce chapitre parla pés<strong>en</strong>tation de nos travaux dans cette matière.5.1 <strong>Vertex</strong> topologique I : formalismeDans cette première partie, nous étalons les grandes lignes prés<strong>en</strong>tées dans [71, 93] quipermett<strong>en</strong>t le calcul des amplitudes <strong>du</strong> modèle A sur les variétés de Calabi-Yau toriqueslocales. L’idée est de placer les paires lagrangi<strong>en</strong>nes D-brane/anti-D-brane sur les axes <strong>du</strong>diagramme torique afin de découper l’<strong>en</strong>semble des variétés de Calabi-Yau <strong>en</strong> C 3 . Notons1 Dans le chapitre suivant, nous considérons une version raffinée de ce vertex.157


5.1 Formalisme <strong>du</strong> <strong>Vertex</strong> topologique I : (Ingrédi<strong>en</strong>ts de base)que chaque vertex correspond à un patch C 3 , sa topologie ainsi que sa structure complexesont obt<strong>en</strong>ues <strong>en</strong> précisant la manière dont les C n doiv<strong>en</strong>t être collées <strong>en</strong>semble. Or, cetteprocé<strong>du</strong>re de collage est obt<strong>en</strong>ue <strong>en</strong> sommant sur les diagrammes de Young. Ainsi, on seretrouve avec des amplitudes de la corde topologique ouverte sur C 3 <strong>en</strong> prés<strong>en</strong>ce de troisLagrangi<strong>en</strong>nes D-branesZ (V i ) = ∑ C λµν T r λ V 1 T r µ V 2 T r ν V 3 . (3.1)λ,µ,νL’amplitude <strong>du</strong> vertex peut être prés<strong>en</strong>tée sous différ<strong>en</strong>tes formes et peut être obt<strong>en</strong>ue <strong>en</strong>termes de fonctions de Schur commeC λ,µ,ν = q k(µ) s ν t(q −ρ ) ∑ ηs λ t /η(q −ν−ρ )s µ/η (q −νt −ρ ). (3.2)Dans ce qui suit, nous explicitons l’expression des amplitudes de la théorie de corde fermée<strong>en</strong> termes des fonctions de partitions des cordes ouvertes.5.1.1 <strong>Vertex</strong> et amplitude de la corde ouverteLes Lagrangi<strong>en</strong>nes de D-branes qui ont été découvert par Harvey et Lawson [123] jou<strong>en</strong>tun rôle très important dans le contexte <strong>du</strong> vertex topologique. L’<strong>en</strong>semble des lagrangi<strong>en</strong>nesD-branes L 1,2,3 avec une topologie C × S 1 qui <strong>en</strong>roule la fibre T 2 et ses projections sur labase R 3 qui sont des lignes, données par :L 1 : r α = 0, r β = r ∗ 1, r γ ≥ 0L 2 : r β = 0, r α = r ∗ 2, r γ ≥ 0L 3 : r α − r β = 0, r α = r ∗ 3, r γ ≥ 0.(3.3)avec ri∗ correspond au mo<strong>du</strong>le de L i . L’amplitude de la corde topologique ouverte correspondanteaux nombres de N i D-branes sur la i me Lagrangi<strong>en</strong>ne L i sur C 3 est de laforme :∑3∏Z =⃗ k (1) , ⃗ k (2) , ⃗ k (3) C ⃗k (1) , ⃗ k (2) , ⃗ k (3)1z ⃗k (i)i=1T r ⃗k (i)V i (3.4)où V i est la source de ligne de Wilson sur la i me D-brane, V i = P exp[ ∮ A 1 ] le long de S 1 ,∞∏T r ⃗k V = (trV j ) k j, (3.5)j=1158


Amplitudes des cordes topologiquesL2L3L1Fig. 3-1 – C 3 avec trois piles de D-branes lagrangi<strong>en</strong>ne.etz ⃗k = ∏ jk j ! j k j.Notons qu’il existe k j trous d’indice j de sorte que la somme h = | ⃗ k| = ∑ j k j est le nombretotal des trous sur une D-brane, et l = ∑ j jk j est le nombre d’<strong>en</strong>roulem<strong>en</strong>t. L’amplitude <strong>du</strong>vertex C ⃗k (1) , ⃗ k (2) , ⃗ ks’exprime <strong>en</strong> terme de constante de couplage g (3) s de la théorie des cordes.Il se trouve que le vertex C est l’objet de base qui est utilisé pour calculer l’amplitude <strong>du</strong>modèle A avec bords. En utilisant la règle de collage, nous pouvons dé<strong>du</strong>ire les amplitudesde la corde fermée sur les géométries toriques.5.1.2 Symétrie de vertexConsidérons une transformation SL(2, Z) qui agit sur la fibre T 2 de C 3 <strong>en</strong> prés<strong>en</strong>ce deD-branes. Le vertex dép<strong>en</strong>d directem<strong>en</strong>t de trois paires (f i , v i ) avecf i ∧ v i = 1 (3.6)où (f i , v i ) forme une base ori<strong>en</strong>tée de H 1 (T 2 ). En outre, si nous ori<strong>en</strong>tons les axes à l’intérieurvers le vertex, nous écrivons :∑v i = 0 v i ∧ v j = ± 1 pour i = j (3.7)iNous choisissons alors un ordre cyclique des vecteurs v i dans R 2 , dans ce cas nous avons :v 2 ∧ v 1 = v 1 ∧ v 3 = v 3 ∧ v 2 = 1. (3.8)159


5.1 Formalisme <strong>du</strong> <strong>Vertex</strong> topologique I : (Ingrédi<strong>en</strong>ts de base)Un choix naturel de framing f i basé sur cette symétrie est la suivante :avec la propriété f i ∧ v i = 1.(f 1 , f 2 , f 3 ) = (v 2 , v 3 , v 1 )TSv 2v3ST −1v 3’v1 = v 1’TSST −1v 2’Fig. 3-2 – Les transformations SL(2, Z) agiss<strong>en</strong>t sur les branes situer aux lignes externes dansle diagramme torique de conifold résolu.D’où le vertex a une symétrie cyclique dans le framing canoniqueC R1 ,R 2 ,R 3= C R3 ,R 1 ,R 2= C R2 ,R 3 ,R 1.5.1.3 Collage des vertexDans cette sous section, nous cherchons à exprimer le pro<strong>du</strong>it de deux fonctions departitions de la théorie des cordes ouvertes <strong>en</strong> termes de l’amplitude de la corde fermée. Ils’agit <strong>en</strong> effet de découper le diagramme torique <strong>en</strong> deux graphes, dont chacun d’eux peutêtre associé aux amplitudes de la théorie des cordes ouvertes.Z(Γ) ∼ Z(Γ L ) × Z(Γ R )En effet, le diagramme Γ est décomposé <strong>en</strong> Γ L et Γ R qui peuv<strong>en</strong>t être id<strong>en</strong>tifiés commedes surfaces de Riemann de la corde ouverte <strong>en</strong> termes des g<strong>en</strong>res g et les nombres d’<strong>en</strong>roulem<strong>en</strong>t.L’expression de l’amplitude de la corde fermée consiste alors à coller les deuxdiagrammes <strong>en</strong>tre eux par leurs pattes∑ exp(−l(Z(Γ L ) ⃗ k)t)⃗k ∏ Z(Γk j !j k j R ) ⃗k .⃗ kj160


Amplitudes des cordes topologiquesoù t est la taille <strong>du</strong> P 1 . Dans la procé<strong>du</strong>re de collage, nous devons être att<strong>en</strong>tifs à ce queles deux branes soi<strong>en</strong>t définies sur le même framing. En outre, nous ne devons pas oublierque les variétés X L et X R qui correspond<strong>en</strong>t aux graphes Γ L et Γ R , sont s’équipées parle choix des structures complexes, ce qui in<strong>du</strong>it l’ori<strong>en</strong>tation naturelle des bords des deuxdisques dans X L,R . Afin de coller les deux disques dans P 1 , il faut que les deux bords soi<strong>en</strong>tori<strong>en</strong>tés de façon opposés, ce qui mène à remplacer les branes par des anti-branes. Or, celaest équivaut à multiplier les amplitudes par un facteur (−1) h , où h est le nombre de bordsde la surface de Riemann. Il se trouve que le collage a une interprétation physique de tellesorte d’avoir les N D-branes sur la patte de X L et les anti D-branes sur X R . A partir deces résultats, l’opération de collage se tra<strong>du</strong>it tout simplem<strong>en</strong>t par la fonction suivante :Z(X) = ∑ QZ(X L ) Q (−1) l Qe −l(Q)t Z(X R ) Q t5.1.4 Règles de collagePour calculer les amplitudes de la corde fermée à partir <strong>du</strong> formalisme de vertex topologique,nous utilisons les diagrammes toriques des variétés de CY qui respect<strong>en</strong>t touteune série de règles bi<strong>en</strong>s définies, que l’on appelle : les règles de collage :i) Les axes <strong>du</strong> graphe sont marqués par des vecteurs v i qui <strong>en</strong>cod<strong>en</strong>t un des cycles dela fibre T 2 et se dégénère sur la i-ème ligne. Pour chaque axe de diagramme torique, onassocie une représ<strong>en</strong>tation R i .ii) Pour une variété de Calabi-Yau, le graphe peut être divisé <strong>en</strong> vertex trival<strong>en</strong>ts tel quechaque C 3 correspond à un patch U a avec a = 1, 2.iii) On associe à chaque vertex un triplet de vecteurs (v i , v j , v k ) qui se réuniss<strong>en</strong>t au sommet.Dans le s<strong>en</strong>s inverse des aiguilles d’une montre, le vertex −(v i , v j , v k ) est équival<strong>en</strong>t à(v j , v k , v i ).iv) Si les pattes <strong>du</strong> vertex sont des lignes <strong>en</strong>trantes, alors à chaque patch U a est associé lefacteur C RI ,R J ,R K. Sinon, nous remplaçons la représ<strong>en</strong>tation correspondante par son transposéque multiplie le facteur (−1) l(R) . Par exemple, on considère le vertex U a qui partageavec le vertex U b la i me axe donné par un triplet de vecteurs (v i , v′ j , v′ k ). En supposant icique le vecteur v i est représ<strong>en</strong>té par une ligne sortante de U a et <strong>en</strong>trante à U b . L’amplitudecorrespondante est obt<strong>en</strong>ue <strong>en</strong> faisant la somme sur les représ<strong>en</strong>tations le long de l’i-èmeaxe qui connecte les deux vertex sous cette forme :∑R iC Rj R k R ie −l(R i)t i(−1) (n i+1) l (R i) q −n iκ Ri /2 C R ti R′ j R′ k(3.9)161


5.2 Formalisme <strong>du</strong> <strong>Vertex</strong> topologique II : (Calcul des amplitudes)où l’<strong>en</strong>tier n i est défini comme suit :n i = |v k ′ ∧ v k |avec |v k ′ ∧ v k | est égal à v k ′ ∧ v k si les deux vecteurs v k ′ et v k sont des lignes <strong>en</strong>tranteset −v k ′ ∧ v k dans le cas contraire . Ce facteur reflète le fait que le framing sur l’ième axedevrait être le même sur les deux côtés de collage. Le facteur (−1) (n i+1) l (R i) dans (3.9) estle signe associé au framing.v) Notons que les pattes <strong>du</strong> graphe sont des lignes droites dans le plan. Nous associons unparamètre de Kahler t i =x i .x i √p 2 +t 2 pour la i-ème axe dans la direction (p i, q i ) de longueurvi) Pour une patte non-compacte <strong>du</strong> graphe, la représ<strong>en</strong>tation correspondante est nécessairem<strong>en</strong>ttriviale R = 0 (notée aussi R = •).Naturellem<strong>en</strong>t, le vertex peut être utilisé pour calculer les amplitudes de la corde ouvertesur les variétés de Calabi-Yau toriques aussi bi<strong>en</strong> pour les cordes fermées. Lorsque on placeles D-branes sur les axes non-compacts <strong>du</strong> graphe, nous modifions simplem<strong>en</strong>t la (vi)èmerègle, <strong>en</strong> sommant sur toutes les représ<strong>en</strong>tations arbitraires R. Cette règle est représ<strong>en</strong>téepar l’intro<strong>du</strong>ction de T r R V où V est la matrice d’holonomie sur la D-brane. Supposonsque la D-brane est sur l’i-ème axe de diagramme torique, l’amplitude correspondante peutêtre obt<strong>en</strong>ue <strong>en</strong> modifiant la (v)ème règle de collage :v′) Pour le cas d’une seule D-brane sur l’i-ème axe qui connecte les deux vertex U a et U b ,l’amplitude est donnée <strong>en</strong> sommant sur toutes les représ<strong>en</strong>tations gauche et droite Q L i , Q R idéfinies sur les deux cotés <strong>du</strong> D-brane par :∑R i Q L i ,QR iC Rj ,R k ,R i ⊗Q L i (−1)s(i) q f(i) e −L(i) C R ti ⊗Q R i ,R′ j,R′ kT r Q LiV i T r Q RiV −1i (3.10)avec L(i), f(i) et s(i) sont respectivem<strong>en</strong>t la longueur, le framing et les signes des facteurssur cet axe :L(i) = l(R i )t i + l(Q L i )r i + l(Q R i )(t i − r i )f (i) = p i k Ri ⊗Q L/2 + (n + p i) ki R ti ⊗Q R/2is (i) = l (R i ) + p i l ( )R i ⊗ Q L i + (n + pi ) l ( Ri t ⊗ Q R i).(3.11)Par la suite, nous allons formaliser ce que nous avons évoqué dans des exemples intéressants.Ces applications vont nous permettre de donner une démonstration rigoureuse des règlesde collage.162


Amplitudes des cordes topologiques5.2 <strong>Vertex</strong> topologique II : Calcul des amplitudesDans la section précéd<strong>en</strong>te, nous avons prés<strong>en</strong>té tout les ingrédi<strong>en</strong>ts nécessaires dansl’usage <strong>du</strong> formalisme <strong>du</strong> vertex topologique pour le calcul des amplitudes de la cordetopologique. Pour ce formalisme, l’idée c<strong>en</strong>trale est d’associer une amplitude à chaquediagramme torique. Afin de mieux illustrer ce fait et dans l’objectif de mieux compr<strong>en</strong>dreles différ<strong>en</strong>tes régles de la première partie concernant le vertex topologique, nous consacronsla partie qui suit à des exemples concrets très pratiques et utilisables dans le cadre descordes topologiques.5.2.1 L’espace complexe C 3On va calculer la fonction de partition de la corde topologique <strong>en</strong> utilisant le vertextopologique lorsqu’une pile de lagrangi<strong>en</strong>nes D-branes se termin<strong>en</strong>t sur l’un des axes <strong>du</strong>C 3Z(q; V ) = ∑ C ∅ ∅ ν (q −1 ) T r ν V (3.12)νavec T r ν V =s ν (x) et dont x = {x 1 , x 2 , · · · } sont les valeurs propres de la matrice d’holonomieVZ(q; V ) = ∑ ν C ∅ ∅ ν(q −1 ) s ν (x)= ∑ ν s ν(q ρ ) s ν (x)∏= ∞ (1 + q −i+ 1 2 x j )i,j(3.13)Dans le cas d’une seule lagrangi<strong>en</strong>ne D-brane, x = (−Q, 0, 0, 0, ), la fonction de partitionest définie comme suit :Z(q; Q) = ∏ (1 − Qq −i+ 1 2 ) (3.14)i5.2.2 Variété locale P 2Nous allons maint<strong>en</strong>ant examiner un exemple plus complexe, à savoir la variété deCalabi-Yau non-compacte O(−3) → P 2 connue sous le nom locale P 2 . Il s’agit, <strong>en</strong> effet, dedéterminer l’amplitude correspondante <strong>en</strong> utilisant le formalisme <strong>du</strong> vertex topologique. Ilest commode de formuler cette amplitude <strong>en</strong> utilisant le diagramme torique décrit dans lechapitre 2. La variété de Calabi-Yau X = O(−3) → P 2 peut être décrite dans le language<strong>du</strong> modèle sigma linéaire N = 2 à 2d par 4 champs chiraux z µ (µ = 0, 1, 2, 3),|z 1 | 2 + |z 2 | 2 + |z 3 | 2 − 3|z 0 | 2 = t (3.15)163


5.2 Formalisme <strong>du</strong> <strong>Vertex</strong> topologique II : (Calcul des amplitudes)avec l’action U (1) qui agit sur z i sous cette forme:(z 0 , z 1 , z 2 , z 3 ) → ( e −3iα z 0 , e iα z 1 , e iα z 2 , e iα z 3). (3.16)Cette équation est le D-terme de GLSM qui définit une sous variété de C 4 . Nous <strong>en</strong>roulonsN D4-branes respectivem<strong>en</strong>t suivant les diviseurs D i qui sont défini par (3.15) et par leséquations suivantes :D i : X i = 0, i = 1, 2, 3. (3.17)Ces sous variétés ne sont que l’espace total d’une ligne fibrée de degrés −3 sur P 1 .Di ∼ O(−3) → P 1 .Chaque deux diviseurs D i et D j s’intersect<strong>en</strong>t le long d’un plan complexe donné par z i =z j = 0. La base de ⋃ D i est une chaîne de trois plans projectifs P 1 qui se connecte <strong>en</strong> uniseul point p i . Pour trouver une relation <strong>en</strong>tre l’amplitude de la corde ouverte topologique etla théorie de Yang-Mills, il est utile d’ajouter les M D4-branes <strong>en</strong>roulées les sous variétésLagrangi<strong>en</strong>nes.L 1 : |X 1 | 2 − |X 0 | 2 = 0 , |X 2 | 2 − |X 0 | 2 = c 1L 2 : |X 2 | 2 − |X 0 | 2 = 0 , |X 3 | 2 − |X 0 | 2 = c 2(3.18)L 3 : |X 3 | 2 − |X 0 | 2 = 0 , |X 1 | 2 − |X 0 | 2 = c 3avec c i sont des constantes et satisfont la condition 0 < c i < t. On <strong>en</strong>roule les piles MD4-branes sur les trois sous variétés Lagrangi<strong>en</strong>nes.Grâce aux régles de collage, l’amplitude de P 2 local est donnée par :Z P 2 =∑R 1 ,R 2 ,R 3(−1) P i l(R i) e − P i l(R i)t q P i k R iC•R2R t 3 C •R 1 R t 2 C •R 3 R t 1(3.19)où t est le paramètre de kahler de O(−3) → P 2 . L’intérêt de cette amplitude (3.19) réside<strong>du</strong> fait qu’elle est compatible avec celle obt<strong>en</strong>ue dans [170]. Le vertex peut être écrit <strong>en</strong>termes des invariants des noeuds comme suit :C •R2 R t 3 = W R 2 R 3q −k R 3 /2 . (3.20)5.2.3 Les branes non-compactesOn utilise le vertex topologique, l’amplitude des branes non compactes <strong>du</strong> conifold résolupeut être calculer au différ<strong>en</strong>tes positions. Dans cette sous-section, nous allons décrire164


Amplitudes des cordes topologiquesQ 1(−2)R 2R 1(−2)QR 3(−2)Q3 2Fig. 3-3 – Diagramme torique de O(−3) → P 2 est obt<strong>en</strong>u à partir de collage de trois vertextrival<strong>en</strong>t.les branes non-compactes dans le contexte de vertex topologique, basé sur l’exemple de conifoldrésolu. On considère l’insertion des branes sur les lignes sortantes dans le diagrammetorique de conifold résolu [172]. L’amplitude de branes non-compactes sur la patte externepeut être obt<strong>en</strong>ue par la succession suivante :C ν∅∅ − → ∑ µ(e −r ) |µ| (−1) p|µ| q 1 2 pk µC ν∅µ T r µ V (3.21)où p désigne le framing. Par conséqu<strong>en</strong>t, l’amplitude des A-branes ”f 3 ”, avec le framing pà la position ”r” sur l’axe externe v 3 <strong>du</strong> diagramme torique de conifold résolu est donnée<strong>en</strong> appliquant le formalisme de vertex topologiqueZ f 3A (V, r, p) = 1Z tv (g s , t)[∑ ∑µ ν(e −r ) |µ| (−1) p|µ| q 1 2 pk µC ∅∅ν t(−e −t ) |ν| C ν∅µ]T r µ V (3.22)avec Z tv (g s , t) correspond à la fonction de partition de la corde topologique <strong>du</strong> modèle Asur le conifold résolu éq(2.18).En termes des fonctions de schur, la fonction de partition est donnée par :[Z f 1 ∑3A(V, r, p) = (e −r ) |µ| (−1) p|µ| q 1Z tv 2 (p+1)k µµ165]∑s ν t(−e −t q ρ )s µ t /λ(q ρ )s ν/λ (q ρ ) T r µ V.ν,λ(3.23)


5.2 Formalisme <strong>du</strong> <strong>Vertex</strong> topologique II : (Calcul des amplitudes)v 3’v 2v3f3v1 = v 1’v 2’Fig. 3-4 – A-brane f 3 sur la ligne externe v 3 .Il est clair que l’amplitude au cas des branes non-compactes sur la patte externe v′ 3 estexprimée par :v 2v3v 3’f 3’v1 = v 1’v 2’Fig. 3-5 – A-brane f ′ 3 sur la ligne externe v′ 3.Z f ′ 3A (V, r, p) = 1Z tv (g s , t)[∑ ∑µ ν(e −r ) |µ| (−1) p|µ| q 1 2 pkµ C ∅µν t(−e −t ) |ν| C ν∅∅]T r µ V (3.24)Cette fonction est id<strong>en</strong>tique à la fonction de partition Z f 3. Ceci est possible grâce à lasymétrie <strong>du</strong> vertex topologique.Dans le cas des anti-branes, ces fonctions de partitions peuv<strong>en</strong>t être obt<strong>en</strong>ues par l’inversion166


Amplitudes des cordes topologiquesde l’ori<strong>en</strong>tation de branes. Par conséqu<strong>en</strong>t, l’amplitude des anti-branes à la position ”r”sur la ligne externe v 3 est donnée par :[ ]Z f 1 ∑ ∑3Ā(V, r, p) =(e −r ) |µ| (−1) p|µt| q 1Z tv 2 pk µ t C ∅∅ν t(−e −t ) |ν| C ν∅µ t T r µ V (3.25)(g s , t)µ νIl paraît que les amplitudes dans les deux cas branes et anti-branes sont complètem<strong>en</strong>tdiffér<strong>en</strong>tes dans le cas de conifold résolu. D’autre part, étant donné, les (anti) A-branes surles pattes externes v 2 et v ′ 2, il existe deux expressions des amplitudes <strong>en</strong> utilisant le vertextopologique :v 23v 2v3vf 2v 3’v1 = v 1’v 3’v1 = v 1’f ’ 2v 2’v 2’Fig. 3-6 – A-brane f 2 sur la ligne externe v 2 et A-brane f ′ 2 sur la ligne externe v ′ 2.Z f 21A(V, r, p) == ∑ µ[ ] ∑(e −r ) |µ| (−1) p|µt| q 1 2 pkµ C ∅∅ν t(−e −t ) |ν| C νµ∅ T r µ Vν)]T r µ V∑Z tv µ([(e −r ) |µ| (−1) p|µt| q 1 2 pkµ q 1 2 N|µ| s µ q N 2 −i+ 1 2(3.26)etZ f 2Ā(V, r, p) = ∑ µ(−1) |µ| [ (e −r ) |µt| (−1) p|µt| q 1 2 pk µ t q − 1 2 N |µ t | sµ(q N 2 −i+ 1 2)]T r µ V .(3.27)Ceci nous con<strong>du</strong>it à trouver la relation <strong>en</strong>tre les deux amplitudes des (anti) A-branes surles lignes externes v 2 , v 3 de la manière suivanteZ f 2Ā= Z f 3A (V, r, −p − 1) , Zf 2A= Z f 3Ā(V, r, −p − 1)5.2.4 La ”casquette” et le ”pantalon”L’amplitude de la théorie de corde topologique de modèle A est déterminée <strong>en</strong>tièrem<strong>en</strong>tà l’ordre g de la théorie des perturbations <strong>en</strong> termes de la théorie des champs topologiques167


5.2 Formalisme <strong>du</strong> <strong>Vertex</strong> topologique II : (Calcul des amplitudes)v 23v 2v3vf 2v 3’v1 = v 1’v 3’v1 = v 1’f ’ 2v 2’v 2’Fig. 3-7 – A-brane f 2 sur la ligne externe v 2 et A-brane f 2 ′ sur la ligne externe v′ 2.quantiques sur la surface de Riemann [167, 168].X = L 1 ⊕ L 2 → Σ. (3.28)L’espace total X est une variété de Calabi-Yau définie comme les fibres holomorphes derang 2 sur la surface de Riemann dont la classe de Chern de la fibre normale annule laclasse canonique d’une surface de Riemanndeg(L 1 ) ⊕ deg(L 2 ) = −χ (g) (3.29)où (χ = 2 − 2g) est la caractéristique d’Euler. La surface de Riemann est obt<strong>en</strong>ue à partir<strong>du</strong> collage de pantalon (P), annuli (A) et casquettes (C).(−1,0)(1,0)Fig. 3-8 – Les surfaces de Riemann correspondant à l’amplitudes de cap C (−1,0) de la premièreclasse Chern (−1, 0) et pantalon P (1,0) avec (1, 0) .Les transformations holomorphiques de la surface d’univers au variété de Calabi-Yau Xsont des transformations holomorphiques g et h. Afin d’obt<strong>en</strong>ir une telle transformation,on ajoute des lagrangi<strong>en</strong>nes D-branes. Ajouter et annuler des D-branes dans une variété deCalabi-Yau correspond à couper ou coller les bases des surfaces de Riemann. L’opérationde collage permet la composition de l’amplitude <strong>du</strong> modèle topologique.Pour calculer les amplitudes de la corde topologique ouverte <strong>du</strong> modèle A des variétésde Calabi-Yau, nous comm<strong>en</strong>çons tout d’abord par calculer la fonction de partition de la168


Amplitudes des cordes topologiquesFig. 3-9 – La surface de Riemann est obt<strong>en</strong>u à partir le collage de pantalon (P), annuli (A) etcasquettes (C)variété de Calabi-Yau <strong>du</strong> casquette et pantalon et par la suite nous appliquons le formalismede collage. Alors, avant de procéder à ce formalisme, nous rappelons que l’expression de lafonction de partition de la variété de ”casquette” et de ”pantalon” est donnée par :Z top (C (−1,0) ) = ∑ Rd q (R)q −k R/4 T r R U. (3.30)où R est une représ<strong>en</strong>tation de groupe de Lie SU(∞) et le paramètre q est lié à la constantede couplage g s . La matrice d’holonomie <strong>du</strong> champ de Jauge sur les D-branes autour d’uncercle qui provi<strong>en</strong>t de l’intersection <strong>en</strong>tre une D-brane et Σ est donnée par :U = P e i H A .Le coeffici<strong>en</strong>t d q (R) est la dim<strong>en</strong>sion quantique de la représ<strong>en</strong>tation des groupes symétriquescorrespondant aux tableaux de Youngd q (R) = ∏ □∈R1[h(□)] q. (3.31)Dans cette expression, le pro<strong>du</strong>it porte sur toutes les cases dans les tableaux de Young deR, h est la longueur d’équerre et d q (R) est le même que l’amplitude de vertex topologiqueC R,0,0 :d q (R) q k(R)/4 = C R,0,0 = W R0 . (3.32)169


5.2 Formalisme <strong>du</strong> <strong>Vertex</strong> topologique II : (Calcul des amplitudes)Dans ce cas, nous examinons le modèle A avec une pile des D-branes de Calabi-Yau X = C 3 .De la même manière, nous pouvons calculer la fonction de partition de ”pantalon” Calabi-Yau P (1,0) . Maint<strong>en</strong>ant, pour 3 séries de représ<strong>en</strong>tations correspond<strong>en</strong>t à trois piles deD-branes avec holonomies U i , i = 1, 2, 3. La fonction de partition est donnée par :Z top ( P (1,0)) = ∑ R1d q (R) qk R/4 T r R U 1 T r R U 2 T r R U 3 . (3.33)Il suffit de considérer les deux équations (3.30, 3.33) et recoller l’<strong>en</strong>semble de Σ L et Σ R auniveau de leur bord commun pour former l’amplitude de Σ L∪R . La formule d’orthogonalitédes caractères :∫dU T r R1 U T r R2 U −1 = δ R1 R 2(3.34)permet de définir les cercles de bord des deux surfaces qui sont ori<strong>en</strong>tées dans des s<strong>en</strong>sopposés. Elle est munie d’une opération d’inversion qui agit sur l’holonomie U → U −1 . Lafonction de partition de Σ L∪R est formellem<strong>en</strong>t définie par l’intégrale de chemin suivant :∫Z top (Σ L∪R ) = dU Z top (Σ L ) (U) Z top (Σ R ) ( U −1) . (3.35)On calcule cette fonction de partition à l’aide de la propriété suivante :Z top (Σ) (U) = ∑ RZ topR (Σ) T r RU (3.36)Le résultat sera sous la forme suivante :Z top (Σ L∪R ) = ∑ PZ topP(Σ L) Z topP (Σ R) . (3.37)Par exemple, on considère l’amplitude <strong>du</strong> modèle-A correspondant à la variété de Calabi-Yau dont la première classe de Chern est donnée par (2g − 2 + p, −p). L’espace totalcomporte une fibre sur une surface de Riemann. Une manière de réaliser une surface deRiemann fermée de g<strong>en</strong>re g et (g − 1) poignées est de pr<strong>en</strong>dre (2g − 2) pantalons de typeP (1,0) avec l’insertion de p annuli de type A (−1,1) . Ce dernier est obt<strong>en</strong>u à partir <strong>du</strong> collage<strong>en</strong>tre le pantalon P (1,0) et la casquette C (0,−1) . La fonction de partition est donnée par :Z top (Σ g ) = ∑ R( ) 2g−2 1q kR/2(p+g−1) e −t|R| . (3.38)d q (R)Notons que dans la limite N → ∞, d q (R) est la dim<strong>en</strong>sion quantique dim q (R) de lareprés<strong>en</strong>tation de U(N). Dans la suite, nous allons considérer comme autre exemple, une170


Amplitudes des cordes topologiquessurface de Riemann de g<strong>en</strong>re g mais avec h trous. Cela correspond à l’insertion de hpantalons. On Choisit (h − r) de type P (1,0)et r de type P (0,1) . Nous obt<strong>en</strong>ons un<strong>en</strong>ouvelle réalisation de Calabi-Yau et la condition de CY est satisfaite quand((deg (L 1 ) , deg (L 2 )) = 2g − 2 + h + p ′ , −p ′) , (3.39)avec p′ = p − r. L’amplitude de la corde topologique sera de la forme suivante :Z top (Σ g,h ) = ∑ R( ) 2g−2 1q k R/4(deg(L 1 )−deg(L 2 )) e −t|R| T r R U 1 · · · T r R U h . (3.40)d q (R)Alors que, l’amplitude associée à une surface de Riemann de g<strong>en</strong>re g et h trous <strong>en</strong> prés<strong>en</strong>cedes D-branes dans la fibre est exprimée sous cette forme :Z top (Σ g,h ) = ∑ RW RR1 · · · W RRhq k R/4(deg(L 1 )−deg(L 2 )) T rW 2g−2+hR V 1 · · · T r Rh V h . (3.41)R0Nous avons utilisé le fait que W R0 = d R q k R/4 . Cette étude a ouvert de nouveaux perspectivespour une compréh<strong>en</strong>sion plus profonde des invariants de la théorie de Chern-Simon et leformalisme de vertex topologique.5.3 <strong>Vertex</strong> <strong>Topologique</strong> et Théorie de Chern-SimonsLa correspondance clé <strong>en</strong>tre la théorie de Chern-Simons et la théorie de corde topologiqueest considérée comme un des exemples importants [155, 156, 157, 158]. Du pointde vue mathématique, son application a permis de donner une description rigoureusepour les invariants de Gromov-Witt<strong>en</strong> [64],[128], des noeuds et des invariants d’<strong>en</strong>trelacs[125, 126, 159]. D’autre part, l’une des applications à la physique les plus intéressantes oùcette correspondance est réalisée concrètem<strong>en</strong>t a été possible grâce au formalisme <strong>du</strong> vertextopologique qui a permis d’obt<strong>en</strong>ir une solution complète de la théorie de corde topologiquesur la variété de Calabi-Yau. Dans cette section, nous démontrons des résultats analoguesà ceux qui précèd<strong>en</strong>t, mais dans le contexte de la théorie de Chern-Simons, dans laquelleon dérive le vertex trival<strong>en</strong>t <strong>en</strong> termes des observables de la théorie de Chern-Simons.5.3.1 Ingrédi<strong>en</strong>ts de Chern-SimonsDans le but de trouver les amplitudes et extraire la formule explicite de vertex, nousaurons besoin tout d’abord de définir certains invariants de Chern-Simons des <strong>en</strong>trelacs de171


5.3 <strong>Vertex</strong> <strong>Topologique</strong> et Théorie de Chern-SimonsHopf W λµ = W µλ dans les représ<strong>en</strong>tations de SU(N). Dans un premier temps, nous allonsbrièvem<strong>en</strong>t rappeler les différ<strong>en</strong>ts ingrédi<strong>en</strong>ts importants pour notre étude qui sont donnéspar les relations suivantes :q = exp(g s ) = exp( ) 2πi, λ = q N .k + NDans la <strong>du</strong>alité avec la théorie des cordes topologiques sur une classe de variété de Calabi-Yau [155], nous avons t = Ng s et λ = e t . Considérons W R ≡ W R0 qui est liée aux invariantsde Chern-Simons dénoués dans une représ<strong>en</strong>tation arbitraire R. Pour le cas de l’invariantd’<strong>en</strong>trelacs de Hopf, les invariants W R1 R 2pour les représ<strong>en</strong>tations R 1 , R 2 sont donnés par :(W R1 R 2) U(N)= q l 1 l 2N (WR1 R 2) SU(N)(3.42)où l i est le nombre total de boîtes dans le tableau de Young R i , i = 1, 2. Le préfacteurq l 1 l 2N apparaît dans l’éq(3.42) est une correction qui a été établie dans [171].5.3.2 Expression de l’amplitudeEnergie libreNous rappelons certaines propriétés des amplitudes <strong>du</strong> modèle A topologique. On définitalors l’énergie libre <strong>du</strong> modèle A topologique F (X) comme la série formelle :F (X) =∞∑g=0gs2g−2 F g (t)où F g (t) est l’énergie libre de g<strong>en</strong>re g. Elle peut être définie comme une somme sur les deuxclasses d’homologies des instantons de surface d’univers de g<strong>en</strong>re gF g (t) = ∑ QN g,Q e −Q.t .où le vecteur Q ∈ H 2 (X, Z) est nommé par la classe d’homologie, t est le paramètre deKahler et N g,Qsont des invariants de Gromov-Witt<strong>en</strong>. L’énergie libre F (X) peut êtreutilisée pour compter les états BP S sur les variétés de Calabi-Yau associées aux D2-branes <strong>en</strong>roulant les courbes holomorphes dans X. Cela implique que l’énergie libre est dela forme :F (X) =:∞∑∑∞∑n=1 Q∈H 2 (X) g=0n g Q (2sinh(ng 2g−2 e−nQ.ts/2))n .172


Amplitudes des cordes topologiquesDans cette formule, q = e g sest liée au nombre quantique SU(2) L ⊂ SO(4) indiquant lareprés<strong>en</strong>tation de spin de la particule à 4 + 1 dim<strong>en</strong>sions, Q.t est la masse des états BP Set n g Qest un <strong>en</strong>tier qui compte le nombre des états BPS avec les nombres quantiques Q etg.Expression de l’amplitudeL’expression de vertex trival<strong>en</strong>t dans le modèle A est dé<strong>du</strong>it grâce à la <strong>du</strong>alité topologiquede N large [155] qui relie les amplitudes de Chern-Simons avec celles des cordestopologiques.Dans sa formulation originale, cette conjecture qui a été annoncée dans [169], et démontréedans l’article [67], stupile que l’équival<strong>en</strong>ce <strong>du</strong> modèle A de la théorie de corde topologiqueouverte de N D-brane sur S 3 de T ∗ S 3 et le modèle A de la théorie de corde fermée surY = O(−1) ⊕ O(−1) → P 1 est maint<strong>en</strong>ant bi<strong>en</strong> établie [155][126, 160, 161].On considère la variété T ∗ S 3 avec N D-branes sur S 3 , avec l’<strong>en</strong>roulem<strong>en</strong>t de N 2 D-branesautour de lagrangi<strong>en</strong>ne L 2 sur la ligne de S 3 , ainsi que les N 1 et N 3 D-branes autour L 1 etL 3 respectivem<strong>en</strong>t sur les deux autres pattes.Nous intro<strong>du</strong>isons les holonomies des champs de Jauge sur les branes L i notées respectivem<strong>en</strong>tcomme V i et calculées dans la représ<strong>en</strong>tation R i . La fonction de partition est donnéepar :Z (V 1 , V 2 , V 3 ) =∑1Z S3 R 1 ,R 2 ,R 3(−1) |R 1|+|R 2 |+|R 3 | 〈 〉T r R t1U 1 T r R2 U 2 T r R t3U 3× T r R t1V 1 T r R2 V 2 T r R⊗R3 V 3 T r R1 V1−1(3.43)où le facteur Z S3est une fonction de partition de S 3 . La valeur moy<strong>en</strong>ne〈T rR t1U 1 T r R2 U 2 T r R t3U 3〉est considérée comme un observable dans la théorie de Chern-Simons qui correspond auxinvariants des <strong>en</strong>trelacs de Hopf. Pour la théorie <strong>du</strong>ale, nous nous retrouvons avec unevariété Y = O(−1) ⊕ O(−1) → P 1 . Les D-branes qui <strong>en</strong>roul<strong>en</strong>t les cycles compacts disparaiss<strong>en</strong>t,alors que les D-branes sur les cycles non-compacts surviv<strong>en</strong>t après la transition.Après avoir effectuer la limite N large où t = Ng s t<strong>en</strong>d vers l’infini, Y devi<strong>en</strong>t une variétéC 3 . On peut cep<strong>en</strong>dant dé<strong>du</strong>ire l’amplitude de la théorie de corde topologique. Lesquantités ainsi obt<strong>en</strong>ues sont données comme suit :〈T rR t1U 1 T r R2 U 2 T r R t3U 3〉= ZS 3 W R t1 R 1W R3 R 1W R1 •173


H. JehjouhL’amplitude des pro<strong>du</strong>its de deux boucles de Wilson correspond au même noeud mais avecdes représ<strong>en</strong>tations différ<strong>en</strong>tes〈T r R1 UT r R2 U〉 = ∑ RN R R 1 R 2〈T r R U〉 .Et le vertex topologique peut être dérivé <strong>en</strong> utilisant la transition de Gopakumar VafaC R1 R 2 R 3= q k R +k 2 R32∑Wc R 1P 3 P 1c Rt 3 R t1 R 1W R3 R 1P 3 P 2W R1 •P 1 ,P 2 ,P 3Pour conclure, nous avons développer un méthode de calcul très puissant permet, d’unepart de calculer les amplitudes de la théorie des cordes topologiques sur des variétés deCalabi-Yau toriques et d’autre part de trouver les invariants topologiques. Il est importantde pouvoir insérer ces résultats dans un cadre plus général afin d’avoir une fonction departition plus raffinée.Pour conclure, on vi<strong>en</strong>t de développer des méthodes de calcul trèspuissantes, d’une part pour leur relations avec les invariants des théories topologiques,et d’autre part pour le calcul des fonctions de partitions d’un <strong>en</strong>semble des variétés deCalabi-Yau. Il est important de pouvoir insérer ces résultats dans un cadre plus généralafin d’avoir une fonction de partition plus raffinée.5.4 <strong>Contributions</strong>Dans cette section, nous étalons deux articles de recherche, le premier traite la dérivationde la fonction de partition de la théorie de corde topologique sur les variétés deCalabi-Yau toriques dans la limite large de structure complexe. Le deuxième propose lemodèle sigma supersymétrique jaugé N = 2 à 2D pour une nouvelle classe de variétéde Calabi-Yau torique à 3 dim<strong>en</strong>sions ainsi que le formalisme <strong>du</strong> vertex topologique nonplanaire.5.4.1 Topological String on Toric CY3s in Large Complex StructureLimitNuclear Physics B, Volume 813, Issue 3, 1 June 2009, Pages 315-348,arXiv :0812.0526.RésuméDans cette contribution, une étude explicite de la théorie de corde topologique sur une174


H. Jehjouhclasse importante des variétés de Calabi-Yau toriques fut étalée. Plus précisém<strong>en</strong>t, nousnous sommes intéressés aux variétés qui décriv<strong>en</strong>t le vide supersymétrique <strong>du</strong> modèle sigmalinéaire jaugé N = 2 à 2D avec le superpot<strong>en</strong>tiel W (Φ) qui est non nul. Pour ce fait, nousavons prés<strong>en</strong>té, tout d’abord, la méthode de vertex topologique planaire qui est associéeau D-terme dans le modèle sigma supersymetrique dans le cas où le superpot<strong>en</strong>tiel W (Φ)est nul. Puis, nous avons donné la liaison <strong>en</strong>tre la fonction génératrice des variétés de CYtoriques et les invariants de Gromov-Witt<strong>en</strong>. Par la suite, nous avons dérivé la structuregénérale <strong>du</strong> vertex topologique C (np) tout <strong>en</strong> montrant que ce vertex est associé avec lesF et D-termes dans le modèle sigma supersymetrique. Nous avons égalem<strong>en</strong>t calculé lafonction de partition Z top de la courbe elliptique locale <strong>en</strong> se basant sur le formalisme nonplanaire déjà intro<strong>du</strong>it dans la publication précéd<strong>en</strong>te.5.4.2 Non Planar Topological 3-<strong>Vertex</strong> FormalismNuclear Physics B, Volume 804, Issue 3, 1 December 2008, Pages 307-341,arXiv : 0712.4249.RésuméDans cette publication, nous avons proposé un nouveau formalisme décrivant un vertextopologique que nous avons baptisé « non planaire » puisqu’il va au delà <strong>du</strong> 3- vertexusuel. Ce formalisme permet de calculer les amplitudes des cordes topologiques sur lafamille de Calabi-Yau locale donnée par :X (m,m−,0) = O(m) ⊕ O(−m) → E (t,µ) ,dans la limite où sa structure complexe est large, ce qui revi<strong>en</strong>t à poser |µ| → ∞. Notonsque la base E (t,µ) correspond à la courbe elliptique avec t est le paramètre de Kahlerdégénéré. Dans la limite où la structure complexe est large |µ| → ∞, la courbe elliptiqueE (t,∞) est réalisée comme le bord de la variété torique locale P 2 . Dans cette contribution,la réalisation développée <strong>du</strong> modèle sigma supersymmetrique N = 2 à 2D pour la nouvelleclasse de variété de Calabi-Yau torique précitée. Nous proposons une représ<strong>en</strong>tation toriquepour cette famille de classe de CY. Avec tout ce matériel <strong>en</strong> main, nous avons appliqué notreanalyse pour le cas des surfaces de Riemann locales de g<strong>en</strong>re g O(m)⊕O(2−2g−m) → Σ g .175


176H. Jehjouh


Nuclear Physics B 804 [PM] (2008) 307–341www.elsevier.com/locate/nuclphysbNon-planar topological 3-vertex formalismLalla Btissam Drissi a,b , Houda Jehjouh a,b , El Hassan Saidi a,b,∗a Lab/UFR – Physique des Hautes Energies, Faculté des Sci<strong>en</strong>ces, Rabat, Moroccob GNPHE, Groupem<strong>en</strong>t National de Physique des Hautes Energies, Siège focal: Faculté des Sci<strong>en</strong>ces, Rabat, MoroccoReceived 4 January 2008; accepted 15 April 2008Available online 30 April 2008AbstractUsing embeddings of complex curves in the complex projective plane P 2 ,wedevelopanon-planar topological3-vertex formalism for topological strings on the family of local Calabi–Yau threefolds X (m,−m,0) =O(m) ⊕ O(−m) → E (t,∞) . The base E (t,∞) stands for the deg<strong>en</strong>erate elliptic curve with Kähler parametert; but a large complex structure μ;i.e.|μ|→∞. We also give first results regarding A-model topologicalstring amplitudes on X (m,−m,0) . The 2D U(1) gauged N = 2 supersymmetric sigma models of the deg<strong>en</strong>erateelliptic curve E (t,∞) as well as for the family X (m,−m,0) are studied and the role of D- and F-termsis explicitly exhibited.© 2008 Elsevier B.V. All rights reserved.Keywords: Topological string; Topological vertex; Hypersurfaces in the local projective plane; Supersymmetric linearsigma models with D- and F-terms1. Intro<strong>du</strong>ctionTopological string theory [1–3] is a powerful method to deal with the 4D N = 2 supergravityPlanck limit of the compactification of type II superstring on Calabi–Yau (CY) threefolds X 3[4,5]. The study of local Gromov–Witt<strong>en</strong> theory of curves in non-compact Calabi–Yau threefolds[6–10] and the OSV conjecture [11], relating microscopic 4D black holes to 2D q-deformedYang–Mills theory [12–18] and [41–43], have giv<strong>en</strong> an additional impulse to the revival interestin the study of the topological field and string theories [19–26] and [27]. Several important resultshave be<strong>en</strong> obtained in the few last years; in particular the developm<strong>en</strong>t of the topological tri-* Corresponding author at: Lab/UFR – Physique des Hautes Energies, Faculté des Sci<strong>en</strong>ces, Rabat, Morocco.E-mail addresses: drissilb@gmail.com (L.B. Drissi), jehjouh@gmail.com (H. Jehjouh), h-saidi@fsr.ac.ma(E.H. Saidi).0550-3213/$ – see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.nuclphysb.2008.04.025


308 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341vertex (to which we refer below as planar 3-vertex) method for computing the A-model partitionfunction for non-compact toric CY3s [28,29] and the interpretation of this vertex in terms of 3dpartitions of melting crystals g<strong>en</strong>eralizing U(∞) Young tableau [30,31].Moreover the planar 3-vertex and its refined version [32,33] have be<strong>en</strong> shown particularlypowerful. They agree with the Nekrasov’s partition function of N = 2 SU(N) gauge theory[34–40] and provide more insights into non-perturbative dynamics of string field theory. Thepower of the topological 3-vertex method may be compared with the power of Feynman graphstechnique in perturbative φ 3 quantum field theory (QFT). This formal similarity betwe<strong>en</strong> thetoric web-diagrams and the Feynman graphs op<strong>en</strong>s a window on the following issues:First, the use of perturbative QFT results to motivate topological stringy analogues, in particulartoric web-diagrams with higher-dim<strong>en</strong>sional vertices such as the typical φ 4 to be consideredin this study;Second, the developm<strong>en</strong>t of new techniques to <strong>en</strong>large the class of toric Calabi–Yau threefoldsto which the topological vertex formalism applies.Recall that for non-compact toric Calabi–Yau threefolds X 3 with toric web-diagram Δ(X 3 ),the planar 3-vertex method allows to compute explicitly the A-model topological string amplitudes.The topological vertex method is a Feynman-rules like technique where the Feynmangraphs, the vertices of these graphs, the mom<strong>en</strong>ta, and the propagators correspond respectivelyto the toric web-diagrams Δ(X 3 ), the 3-val<strong>en</strong>t vertices C λμν , Young diagrams λ, and the weights(−) (n+1)|λ| e −t|λ| q − n 2 κ(λ) where n <strong>en</strong>codes the framing.Motivated by:(1) the formal correspond<strong>en</strong>ce betwe<strong>en</strong> toric web-diagrams of local Calabi–Yau threefolds andQFT Feynman graphs,(2) the two classes of toric Calabi–Yau threefolds describing the vacua of supersymmetric sigmamodel with (W(Φ i ) ≠ 0) and without superpot<strong>en</strong>tial (W(Φ i ) = 0), and(3) a special feature 1 of the local 2-torus O(m) ⊕ O(−m) → E (t,∞) where the elliptic curve 2E (t,μ)(1.1)is in the base of the local Calabi–Yau threefold X (m,−m,0) rather than in the fiber.We address in this paper, the two following points:(a) We propose in this study a toric repres<strong>en</strong>tation for the family of the local 2-torii with fixedfinite Kähler parameter t and a large complex structure μ; say |μ|→∞,O(m) ⊕ O(−m) → E (t,∞) ,(1.2)with integer m. Thedeg<strong>en</strong>erate elliptic curveE (t,∞) ,(1.3)1 In toric Calabi–Yau 3-folds X 3 with typical fibration B × F , the torii appear g<strong>en</strong>erally in the fiber F . In the localelliptic curve we are considering in this paper, the 2-torus is in the base B.2 The 2-torus has one Kähler parameter t and one complex parameter μ. As these parameters play an important rolehere, we will exhibit them below by referring to the elliptic curve as E (t,μ) . Further details are giv<strong>en</strong> in App<strong>en</strong>dix A.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 309describing the base of the above local Calabi–Yau threefolds (1.2), will be realized as the(toric) boundary of the complex toric projective plane P 2 ; see Sections 3, 5 and 6 as wellas App<strong>en</strong>dix A for more details.With this repres<strong>en</strong>tation at hand, we can th<strong>en</strong>:(α) circumv<strong>en</strong>t, at least for the particular case of the deg<strong>en</strong>erate E (t,∞) , the usual difficultyregarding the lack of a toric diagram for the 2-torus.In addition to the large complex structure limit constraint |μ|→∞, the other price to payin this set up is to consider a non-planar 3-vertex formalism rather than the standard planar3-vertex one based on the R × T 2 special Lagrangian fibration of C 3 . The reason behindthe emerg<strong>en</strong>ce of the non-planar 3-vertex is that the toric Calabi–Yau 3-fold X (m,−m,0) isrealized as a non-compact toric hypersurface of the complex four-dim<strong>en</strong>sional toric manifoldO(m) ⊕ O(−m) → P 2 .(1.4)For later use, we will refer to the local geometries (1.2) and (1.4) as H 3 and Y 4 respectively.The geometry Y 4 will be also promoted to the Calabi–Yau 4-fold X 4 = O(−m − 3) →W P 3 1,1,1,m .(β) Draw the lines for computing the topological amplitudes by using a non-planar 3-vertexformalism.In the pres<strong>en</strong>t study, we will mainly set up the key idea by:(i) building the toric web-diagram Δ(X 3 ) of the local deg<strong>en</strong>erate elliptic curve X (m,−m,0) ,(ii) give first results regarding the structure of the topological non-planar 3-vertex and the partitionfunction of X (m,−m,0) as well as their relation to the 4-vertex of the ambi<strong>en</strong>t C 4 localpatch and to the g<strong>en</strong>eralized Young diagrams.(b) We develop the supersymmetric linear sigma model field theory setting of the local deg<strong>en</strong>erateelliptic curve X (m,−m,0) .More precisely, we show the two main following things:(α) theplanar 3-vertex method is associated with the auxiliary D-terms in supersymmetricsigma models.The non-planar 3-vertex formalism we will be considering here corresponds to the case wherewe have both D-terms and F-terms.(β) We use the sigma model for local P 2 to in<strong>du</strong>ce the N = 2 supersymmetric gauged model forthe local elliptic curve X (m,−m,0) . The underlying complex geometry of a such constructionwas noticed in the Witt<strong>en</strong>’s original work [47]. Here, we give explicit details regarding theimplem<strong>en</strong>tation of F-terms.


310 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341The organization of this paper is as follows: In Section 2, we give an overview of the topologicalplanar 3-vertex method. In Section 3, we exhibit briefly first results concerning thetopological non-planar 3-vertex by considering the example of a Calabi–Yau threefold H 3 .Thetoric 3-fold H 3 is realized as a hypersurface of a four-dim<strong>en</strong>sional complex Kähler manifold. InSection 4, we review the main points of the U(1) gauged supersymmetric sigma model realizationof the local P 2 . In Section 5, we consider the sigma model for the deg<strong>en</strong>erate local ellipticcurve X (m,−m,0) . As the question of the toric realization of T 2 is a crucial point, we divide thissection in three parts: We first study the realization of the deg<strong>en</strong>erate elliptic curve E (t,∞) byusing the compact divisor of P 2 . Th<strong>en</strong> we give explicit details regarding the U(1) gauged supersymmetricsigma model realization of the local deg<strong>en</strong>erate elliptic curve X (m,−m,0) .Next,westudy the mo<strong>du</strong>li space of the supersymmetric vacua associated with X (3,−3,0) . In Section 6, weext<strong>en</strong>d the construction to the case of local elliptic curve X (m,−m,0) . In Section 7, wegivetheconclusion and in App<strong>en</strong>dix A where we show that ∂(P 2 ) is precisely E (t,∞) .2. Topological vertex methodIn this section, we consider the topological 3-vertex method used for the computation of theA-model topological string amplitudes. We illustrate this method through some examples of noncompacttoric Calabi–Yau threefolds namely:(1) the complex space C 3 , with special Lagrangian fibration as R × T 2 , playing the role of theplanar 3-vertex,(2) the resolved conifold obtained by gluing two planar 3-vertices,(3) local P 2 made of three planar 3-vertices.Th<strong>en</strong>, we consider an example of toric Calabi–Yau threefold (1.2) where one needs intro<strong>du</strong>cingnon-planar 3-vertices (and 4-vertices). This local Calabi–Yau threefold is precisely the onegiv<strong>en</strong> by the local deg<strong>en</strong>erate elliptic curve X (m,−m,0) = O(m) ⊕ O(−m) → E (t,∞) realized asa hypersurface in (1.4).2.1. Tri-vertex method: A brief reviewThe topological 3-vertex formalism computes the partition functionZ X3 (q)of the local toric Calabi–Yau threefolds X 3 . In this formalism, the toric web-diagram of X 3 isthought of as resulting from gluing copies of planar 3-vertices C λμν along their edges.Recall that the topological vertex C λμν has three legs <strong>en</strong>ding on stacks of Lagrangian D-branes(C × S 1 ) repres<strong>en</strong>ted by 2d partitions λ, μ and ν.The partition function Z X3 (q) dep<strong>en</strong>ds on the following quantities:(i) The parameter q which reads in terms of the string coupling as e −g s; it plays the role of theBoltzmann weight.(ii) The Kähler parameters {t i } of the local Calabi–Yau threefold X 3 (i = 1,...,h 1,1 (X 3 )).Below, we will consider simple examples whereh 1,1 (X 3 ) = 1.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 311Fig. 1. The toric web-diagram of C 3 . It appears as local patches in toric Calabi–Yau threefolds. The three edges <strong>en</strong>d onstacks of Lagrangian D branes. λ, μ and ν are 2d partitions which, in QFT set up, may be thought of as the externalmom<strong>en</strong>ta.(iii) The boundary conditions (op<strong>en</strong> strings) described by 2d partitions μ (g<strong>en</strong>eric repres<strong>en</strong>tationsof U(∞)). In the QFT language where Feynman graphs play a quite similar roleas the toric web-diagrams, the 2d partition μ corresponds to the “external mom<strong>en</strong>tum” ofFeynman graph. Recall that a 2d partition μ is a Young diagram with columns(μ 1 ,μ 2 ,...), μ i μ i+1 , μ i ∈ Z + .Columns of the 2d partition are associated with Lagrangian D-branes and rows with Lagrangiananti-D-branes.(iv) Lagrangian D-brane/anti-D-brane pairs are needed for the gluing of the vertices. The gluingoperation is achieved by inserting 2d partitions ν and their transpose ν T at the cuts andsumming over all possible ν’s. In QFT language, ν corresponds to “internal mom<strong>en</strong>ta”.The topological 3-vertex 3 method for computing the partition function Z X3 (q) is illustrated onthe three examples giv<strong>en</strong> below.2.2. ExamplesExample 1 (The 3-vertex of C 3 ). The toric graph of C 3 is giv<strong>en</strong> by Fig. 1. Following [28], thepartition function of the 3-vertex, with a stack of Lagrangian D-branes <strong>en</strong>ding on its legs capturedby the boundary conditions (λ,μ,ν), is giv<strong>en</strong> byZ X3 (q) = ∑C λμν (q)(Tr λ V Tr μ V Tr ν V).(2.2)λ,μ,νIn this relation, the trace Tr λ of the holonomy matrix V, with eig<strong>en</strong>values x = (x 1 ,x 2 ,...),isgiv<strong>en</strong> by the Schur function S λ (x). The latter dep<strong>en</strong>ds on the 2d partition λ = (λ 1 ,λ 2 ,...) andthe x i = q i−1/2−λ i. The rank three t<strong>en</strong>sorC (3) = C λμν ,is the topological 3-vertex whose explicit expression reads asC λμν (q) = q[S κ(λ) (ν T q−ρ ) ∑with ρ = (ρ 1 ,ρ 2 ,...)and ρ i = 1/2 − i.2d partitions η(S λ T /η q−ν−ρ ) (S μ/η q−ν T −ρ )](2.1)(2.3)(2.4)3 For simplicity, we use 3-vertex to refer to the planar 3-vertex.


312 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341Fig. 2. Resolved conifold O(−1) ⊕ O(−1) → P 1 made of two planar 3-vertices.Fig. 3. Toric web-diagram of O(−3) → P 2 made of three planar vertices.Eq. (2.4) involves the pro<strong>du</strong>ct of skew-Schur functions S μ/η . It re<strong>du</strong>ces, for the closed topologicalstring case, to∞∏ (Z C 3(q) = C ∅∅∅ (q) = 1 − qn ) −nn=1(2.5)which is nothing but the 3d MacMahon function.Example 2 (Resolved conifold X 3 = O(−1) ⊕ O(−1) → P 1 ). The resolved conifold has oneKähler parameter t parameterizing the size of the projective line P 1 . Fig. 2 describes its toricweb-diagram.This local threefold X 3 is obtained by gluing two 3-vertices along one edge leaving th<strong>en</strong> fourop<strong>en</strong>ed external legs.In the simplest case where there is no boundary terms on the external legs, the partition functionof the closed topological string on the resolved conifold is giv<strong>en</strong> by,∑Z X3 (q, t) = C ∅∅ν (q)(−1) |ν| e −|ν|t C ∅∅ν T (q).(2.6)2d partitions νIn this relation, ν T is the transpose of the 2d partition ν with |ν| boxes and C ∅∅ν (q) is as inEq. (2.4) by setting the boundary conditions as λ =∅and μ =∅.Example 3 (Local P 2 : X 3 = O(−3) → P 2 ). The local P 2 has one Kähler mo<strong>du</strong>lus t parameterizingthe size of the projective plane P 2 . Fig. 3 describes its toric web-diagram.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 313This local threefold is obtained by gluing three 3-vertices. For simplicity, we consider herealso the case where there is no boundaries. The corresponding partition function reads as follows:Z X3 (q, t) = ∑C ∅μν T (q)C λμ T ∅ (q)C λ T ∅ν (q)( −e −t) |ν|+|λ|+|ν| qκ(λ)+κ(μ)+κ(ν)(2.7)withλ,μ,νκ(λ) = 2 [( ‖λ‖ 2 −|λ| ) − (∥ ∥λ T ∥ ∥ 2 − ∣ ∣λ T ∣ ∣ )] ,being the Casimir of the 2d partition.(2.8)3. Beyond the planar vertex methodFirst, we describe briefly the field theory setting of the local elliptic curve geometry leavingtechnical details for next sections. Th<strong>en</strong> we give our first results regarding the topological nonplanar3-vertex formalism and the explicit expression of the partition function associated withEq. (1.2).3.1. Field theory set upThe local CY3 examples we have described above have toric web-diagrams involving planar3-vertices; see Figs. 1–3. These toric threefolds X 3 have a very remarkable field theory setup; they describe supersymmetric vacua of 2D N = 2 linear sigma model with U r (1) gaugesymmetry and (r + 3) matter multiplets Φ i ,U r (1): Φ j ≡ e iqa j Φ j , a = 1,...,r.The defining equation of X 3 is giv<strong>en</strong> by the field equation of motion of the D a auxiliary fields,X 3 :r+3δLδD a = ∑qi a |z i| 2 = 0,i=1where the field coordinates z i are the leading compon<strong>en</strong>ts of the chiral superfields Φ i and wherer+3∑qi a = 0,i=1a = 1,...,r,stands for the Calabi–Yau condition.Toric Calabi–Yau threefolds can be also realized as hypersurfaces H 3 in higher d dim<strong>en</strong>sioncomplex Kähler toric manifolds Y d ,H 3 ⊂ Y d , d 4.Locally, the Kähler toric d-fold Y d may be imagined as giv<strong>en</strong> by the toric fibration R d × T d or asR d × F d with fiber F d = R × T d−1 . The toric web-diagram of Y d involve d-dim<strong>en</strong>sional verticeswhere shrink all the 1-cycles of the toroidal fibration (see Fig. 4).The toric CY3 hypersurfaces H 3 have also a supersymmetric field theory setting. It will bedeveloped in details in forthcoming sections, see also the analysis of [47]. TheH 3 ’s describe aswell supersymmetric vacua of 2D N = 2 linear sigma model with gauge {V a } and matter {Φ i }multiplets, i = 1,...,r + d.(3.1)(3.2)(3.3)(3.4)


314 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341Fig. 4. A g<strong>en</strong>eric toric web-diagram of a 4-vertex. In toric geometry, this web-diagram corresponds to the real base of thelocal patch C 4 .The defining equation of the toric CY hypersurface H 3 is giv<strong>en</strong> by the field equations ofmotion of both the D- and the F -auxiliary fields,{ δLδDH 3 :a = 0, a = 1,...,r(3.5)δLδF α = 0, α= 1,...,m∣ ,with m = d − 3 and d 4. The first r equations, which are similar to Eq. (3.2), re<strong>du</strong>ce thedim<strong>en</strong>sion down to (d −r). The second equations, which are gauge invariant constraint relations,⎧f ⎪⎨ α (z 1 ,...,z d ) = 0f α (λ 1 z 1 ,...,λ d z d ) = 0, α= 1,...,m,(3.6)⎪⎩λ j = e iqa j α a∣re<strong>du</strong>ce the number of free field variables down to 3; say (w 1 ,w 2 ,w 3 ). Up to solving Eqs. (3.5),one can express all the z i field variables in terms of the w’s as shown belowz i = z i (w 1 ,w 2 ,w 3 ), i = 1,...,r + d.In the next subsection, we study in details the case d = 4.3.2. Results on non-planar vertex formalismThe results we will give below concern the following:(1) the toric realization of the local deg<strong>en</strong>erate elliptic curve (1.2),(2) the set up of the non-planar 3-vertex formalism and(3) the computation of the partition function Z H3 .3.2.1. Local deg<strong>en</strong>erate elliptic curveConsider the local Calabi–Yau threefold (1.2) and focus on the particular local deg<strong>en</strong>erateelliptic curve,H 3 = O(+3) ⊕ O(−3) → E (t,∞) , m= 3.(3.8)The deg<strong>en</strong>erate elliptic curve E (t,∞) is giv<strong>en</strong> by the toric boundary (divisor) of the complexprojective plane P 2 ,E (t,∞) = ∂ ( P 2) .This is just a compact divisor (hyperline) of P 2 . The toric web-diagram associated to (3.8) isgiv<strong>en</strong> by Fig. 5.The non-compact toric 4-fold Y 4 of Eq. (3.4) is giv<strong>en</strong> byY 4 = O(−3) → WP 3 1,1,1,3 ,(3.7)(3.9)(3.10)


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 315Fig. 5. Non-planar toric web-diagram of O(+3) ⊕ O(−3) → E (t,∞) . This is a toric CY3 divisor of the four dim<strong>en</strong>sioncomplex Kähler manifold O(+3) ⊕ O(−3) → P 2 . The hollow triangle ABC refers to the deg<strong>en</strong>erate elliptic curveE (t,∞) . The full triangles ABD, ZCD, BCD refer to the three other projective planes.Fig. 6. Toric web-diagram of O(+3) → E (t,∞) . This figure looks like a “toric cap” obtained by gluing three triangles asshown on the figure.where WP1,1,1,3 3 stands for the complex 3 dim<strong>en</strong>sion weighted projective space. To keep in touchwith the Calabi–Yau condition, we promote Y 4 to the toric Calabi–Yau 4-foldX 4 = O(−6) → WP1,1,1,3 3 ,and in g<strong>en</strong>eral toX 4 = O(−3 − m) → WP1,1,1,m3with m 1.(3.11)(3.12)3.2.2. Toric cap and toric cylinderFrom Eq. (3.8), one distinguishes two special divisors of the local deg<strong>en</strong>erate ellipticcurve H 3 :(1) “Toric cap”: see Fig. 6This divisor corresponds to H 3 tak<strong>en</strong> as the fibration O(−3) → Y 2 . The base Y 2 is a compactcomplex surface giv<strong>en</strong> byY 2 = O(+3) → E (t,∞) .(3.13)The toric web-diagram of the complex surface Y 2 is exhibited in Fig. 5.ThefirstChernclassc 1 (T ∗ Y 2 ) is equal to +3. The toric web-diagram of E (t,∞) is giv<strong>en</strong> bythe boundary of the triangle and O(+3) is a compact line.


316 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341Fig. 7. On left, the real skeleton of the toric web-diagram of the fibration of O(−3) → E (t,∞) . It could be interpreted asa triangulation of the cylinder of base E (t,μ) , giv<strong>en</strong> by the boundary of a triangle, and a non-compact line as a fiber. Onright, the usual cylinder R × S 1 .The toric web-diagram of Y 2 is non-planar and can be thought of as the triangulation of thetopological cap of [28]. Below, we will refer to Y 2 as the “toric cap”. Notice also the followingfeatures:(a) The complex surface Y 2 is a compact divisor of H 3 ; it is made of the union of three complexprojective planes, which we d<strong>en</strong>ote as P 2 1 , P2 2 and P2 3 .The projective planes P 2 i belong to three differ<strong>en</strong>t C3 spaces of the ambi<strong>en</strong>t C 4 .(b) The toric web-diagram of Y 2 is made of three triangles as shown 4 in Fig. 6. Recall that thetoric web-diagram of a g<strong>en</strong>eric projective plane is giv<strong>en</strong> by a triangle (Fig. 3).The projective planes (triangles) have mutual intersections I ij along complex projective lines(edges) and a non-planar tri-intersection vertex I 123 .(2) Toric cylinderThis divisor corresponds to think about H 3 as the fibration O(+3) → Ỹ 2 .Here the base Ỹ 2 is a non-compact complex surface giv<strong>en</strong> byỸ 2 = O(−3) → E (t,∞) .(3.14)Its first Chern class c 1 (T ∗ Ỹ 2 ) is equal to −3. Here also the toric web-diagram of E (t,∞) is theboundary of the triangle and O(−3) is a non-compact line. The toric web-diagram of Ỹ 2 is nonplanar;it can be thought of as the triangulation of the cylinder R × S 1 . We will refer to Ỹ 2 as the“toric cylinder” whose toric web-diagram is shown in Fig. 7.The Ỹ 2 divisor of H 3 is made of the union of three sheets O(−3) → P 1 ibelonging to threediffer<strong>en</strong>t C 3 spaces of the ambi<strong>en</strong>t C 4 .From Eqs. (3.13) and (3.14), it is clear that the elliptic curve E (t,∞) is an intersecting curveof the complex surfaces Y 2 and Ỹ 2 .3.3. Topological partition functionTo build the non-planar 3-vertex formalism for the local elliptic curve H 3 , we will follow theconstruction used in the derivation of the usual topological 3-vertex method [28].4 Y 2 can be imagined as the triangulation of the cap.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 317However since the local elliptic curve is a CY3 hypersurface in X 4 ,withH 3 ⊂ X 4 ,(3.15)X 4 = O(−3 − m) → WP 3 1,1,1,m ,a conv<strong>en</strong>i<strong>en</strong>t way to achieve the goal is to proceed as follows:(3.16)(1) develop the 4-vertex formalism for the ambi<strong>en</strong>t toric CY4-fold X 4 ,(2) compute the partition function for X 4 . Actually this step may be also viewed as alternativeway to get the 4d g<strong>en</strong>eralization of the MacMahon function [50],(3) impose the appropriate constraint relations to get the non-planar 3-vertex and the topologicalpartition function for the local elliptic curve H 3 .It is interesting to note here the emerg<strong>en</strong>ce of a 4-vertex formalism in the construction. This isnot surprising since the toric web-diagrams of H 3 and X 4 have quite similar skeletons. In thefirst case the tetrahedron is hollow and in the second it is full.The first step to realize these objectives is specify the special Lagrangian fibration of the toricCY4-fold like X 4 ∼ R 4 × F 4 with fiber tak<strong>en</strong> asF 4 ∼ R × T 3 .(3.17)On the hypersurface H 3 in X 4 , real 1-cycles of F 4 shrink and one is left with the usual specialLagrangian fibration of the toric CY3-folds H 3 ∼ R 3 × F 3 withF 3 ∼ R × T 2 .(3.18)In [49], we give the explicit expressions of the various Hamiltonians and the values of the verticesof the web-diagrams solving the Calabi–Yau conditions.The next step is to study the 4-vertex formalism of X 4 and its re<strong>du</strong>ction down to the torichypersurface H 3 .3.3.1. Toric web-diagrams and g<strong>en</strong>eralized partitionsThe toric web-diagrams of X 4 and H 3 have be<strong>en</strong> described above (Fig. 5). For the case X 4 ,the toric web-diagram can be decomposed into four local patchesU 1 , U 2 , U 3 , U 4 .(3.19)To each patch U i ∼ C 4 with fibration R 4 × F 4 it is associated a topological 4-vertex C (4) .Thisvertex dep<strong>en</strong>ds on the boundary conditions on its external legs. We will see that, using 3d g<strong>en</strong>eralizedYoung diagrams, it can be either defined asC (4) = C ΛΣΥ Γ ,or equival<strong>en</strong>tly by using 2d partition like(3.20)C (4) = C (αβγ )(δɛε)(ζ ηθ)(λμν) .(3.21)The toric web-diagram of H 3 is in<strong>du</strong>ced from the one of X 4 . It can be also decomposed into fourlocal patches as follows,U ∗ 1 , U ∗ 2 , U ∗ 3 , U ∗ 4 ,(3.22)


318 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341where the asterix refers to the projection∗ : X 4 → H 3 , U i → U ∗ i .(3.23)To each patch U ∗ i∼ C 3 with fibration R 3 × F 3 it is associated a non-planar topological 3-vertexC ∗ (3) , (3.24)∗ : C (4) → C ∗ (3) .To get the 4-vertex C (4) , the partition function Z X4 and the topological partition function Z H3 ofthe local elliptic curve, we need first intro<strong>du</strong>cing some key tools.In the standard 3-vertex formalism of [28,30], one uses a set of basic objects; in particular 2dand3d-partitions. In the 4-vertex formalism, we have to build the analogue of these mathematicalingredi<strong>en</strong>ts.α) 3d partitions Roughly, a 3d partition Π can be thought of as an integral N 1 × N 2 rank twot<strong>en</strong>sor (Π ia ) with the property,Π ={Π i,a ∈ Z + ,Π i,a Π i+j,a+b 0},(3.25)where i, j = 1, 2,...,N 1 and a,b = 1, 2,...,N 2 .The 3d partition, which has be<strong>en</strong> used for various purposes, has a set of remarkable combinatorialfeatures. Below, we give useful ones.(i) 3d partitions are g<strong>en</strong>eralizations of the usual 2d partitions λ = (λ 1 ,λ 2 ,...) with λ 1 λ 2 ··· 0 and the integers λ i ∈ Z + .By setting N 3 = Π 1,1 , the 3d partitions can be imagined as a cubic sublattice of Z 3 +[1,N 1 ]×[1,N 2 ]×[1,N 3 ].(3.26)The cubic diagram of Π can be considered as a set of unit cubes (i,j,k)with integer coordinatessuch that (i, j) ∈ λ and 1 k Π(i,j). The integers Π(i,j) define the height of the stack ofcubes on the (x 1 ,x 2 ) plane. The projection of Π on the (x 1 ,x 2 ) plane is just the 2d partition λ.(ii) A subclass of 3d partitions solving the conditions (3.25) is giv<strong>en</strong> by the particular repres<strong>en</strong>tationΠ = λ ⊗ μ, Π ia = λ i μ a , λ i μ a λ i+j μ a+b ,where λ and μ are 2d partitions as in Eq. (2.1).(3.27)We also have the following associated ones:Π T = λ T ⊗ μ, ˜Π = λ ⊗ μ T , ˜Π T = λ T ⊗ μ T ,where λ T stands for the usual transpose of the Young diagram λ.(3.28)(iii) Like in the case of 2d partitions, one may associate to each 3d partition Π a Fock spacestate |Π ia 〉 with norm 〈Π|Π〉≡‖Π‖ 2 ,‖Π‖ 2 = ∑ ( ∑)ia ) (3.29)i1 a1(Π 2 = ∑ ( ∑)ia )a1 i1(Π 2 .


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 319〈Π ia | stands for the <strong>du</strong>al state associated with the <strong>du</strong>al partition Π + = ˜Π T .Wealsohavethefollowing relationI id =∑|Π ia 〉〈Π ia |, 〈Π ia |Π jb 〉=δ ij δ ab ,(3.30)3d partitionsdefining the resolution of the id<strong>en</strong>tity operator I id .(iv) The number |Π| of unit boxes (cubes) of the 3d partition is defined as|Π|= ∑ i,aΠ i,a .(3.31)(v) The boundary (∂Π) of the 3d partition Π is giv<strong>en</strong> by the 2d profile of the correspondingg<strong>en</strong>eralized Young diagram. As this property is important for the pres<strong>en</strong>t study, let give somedetails.Giv<strong>en</strong> a 3d partition Π,theboundary term on the plane x i = N i is a Young diagram (2d partition).On the planes x 1 = N 1 , x 2 = N 2 and x 3 = N 3 , the boundary of Π is composed of by three 2dpartitions λ, μ and ν. So we th<strong>en</strong> have:∂Π = (λ,μ,ν).(3.32)Particular boundaries are giv<strong>en</strong> by the case where a 2d partition is located at infinity; that is thereis no boundary. We distinguish the following situations:∂Π = (∅,μ,ν),∂Π = (∅, ∅,ν),∂Π = (∅, ∅, ∅),where ∅ stands for the vacuum.(3.33)(vi) A conv<strong>en</strong>i<strong>en</strong>t way to deal with 3d partitions is to slice them as a sequ<strong>en</strong>ce of 2d partitionswith interlacing relations. We mainly distinguish two kinds of sequ<strong>en</strong>ces of 2d partitions:perp<strong>en</strong>dicular and diagonal. We will not need this property here; but for details on this mattersee for instance [30] and refer<strong>en</strong>ces therein.β) 4d partitions The 4d partitions P are ext<strong>en</strong>sions of the 3d partitions Π considered above.They can be imagined as 4d g<strong>en</strong>eralized Young diagrams described by the typical integral rank3-t<strong>en</strong>sor,P iaα ∈ Z + , with P iaα P (i+j)(a+b)(α+β) ,(3.34)with 1 i N 1 ,1 a N 2 and 1 α N 3 .Several properties of 2d and 3d partitions ext<strong>en</strong>d to the 4d case; there are also specific propertiesin particular those concerning their slicing into lower-dim<strong>en</strong>sional ones. Below we describesome particular properties of 4d partitions by considering special repres<strong>en</strong>tations.Sub-classes of 4d partitions are giv<strong>en</strong> by:(i) the pro<strong>du</strong>ct of a 2d and a 3d partitions μ and Π likeP = μ ⊗ Π, (P iaα ) = (μ i Π aα ),(3.35)


320 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341with i = 1,...,N 1 ; a = 1,...,N 2 and α = 1,...,N 3 ,(ii) the pro<strong>du</strong>ct of three kinds of 2d partitionsP = λ ⊗ μ ⊗ ν, (P iaα ) = (λ i μ a ν α ).(3.36)The boundary ∂P of a g<strong>en</strong>eric 4d partition P can be defined in two ways. First in terms of 3dpartitions as follows∂P = (Λ,Π,Σ,Υ).We also have the following particular boundary conditionscase I:case II:case III:case IV:∂P = (∅,Π,Σ,Υ),∂P = (∅, ∅,Σ,Υ),∂P = (∅, ∅, ∅,Υ),∂P = (∅, ∅, ∅, ∅),where ∅ stands for the “3d vacuum” (no boundary condition).Second by using 2d partitions to define boundary of P like(3.37)(3.38)∂P = ( [a, b, c]; [d, e, f]; [g, h, i]; [j, k, l] ) ,where [a, b, c],...and [j, k, l] stand for the boundaries of the 3d partitions Λ,...and Υ .Notice that the second repres<strong>en</strong>tation is more richer since along with the configuration(3.39)∅ =[∅, ∅, ∅],we have moreover the two following extra configurations(3.40)[∅, b, c],[∅, ∅, c].For the case I of Eq. (3.38) correspond th<strong>en</strong> the three following boundary configurations⎧⎪⎨case i: ([∅, b, c]; [d, e, f]; [g, h, i]; [j, k, l])∂P = case ii: ([∅, ∅, c]; [d, e, f]; [g, h, i]; [j, k, l]),⎪⎩case iii: ([∅, ∅, ∅]; [d, e, f]; [g, h, i]; [j, k, l]) ∣(3.41)(3.42)where the last one (case iii) is the case I described by the first relation of Eq. (3.38).This property indicates that one disposes of differ<strong>en</strong>t ways to deal with 4d partitions eitherthe simplest one using 3d-partitions or the more refined on involving 2d partitions. Below weconsider both repres<strong>en</strong>tations.Notice moreover that giv<strong>en</strong> a 4d partition P, we can associate to it various kinds of transposepartitions. Using the particular realization Eq. (3.36), the corresponding transposes read asλ T ⊗ μ ⊗ ν, λ ⊗ μ T ⊗ ν, λ ⊗ μ ⊗ ν T ,λ T ⊗ μ T ⊗ ν, λ ⊗ μ T ⊗ ν T , λ T ⊗ μ ⊗ ν T , λ T ⊗ μ T ⊗ ν T ,(3.43)where λ T stands for the usual transpose of the Young diagram λ.The exact mathematical definitions and the full properties of 4d partitions are not our immediateobjective here; they need by themselves a separate study. Here above we have giv<strong>en</strong> just th<strong>en</strong>eeded properties to set up the structure of the 4-vertex formalism and its restricted non-planar3-vertex method.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 3213.3.2. Tetra-vertex C (4) and Z X4α) The 4-vertex C (4) The 4-vertex C (4) of the toric Calabi–Yau X 4 can be built by ext<strong>en</strong>ding the3-vertex construction Eq. (2.3).In the 3d partition set up, the vertex C (4) has four external legs L Λ ,L Σ ,L Υ ,L Γ with boundaryconditions as in Eq. (3.37). The 4-vertex C (4) with boundary conditions (ΛΣΥ Γ ) can be definedas a function of the Boltzmann weight q = e −β as follows:C ΛΣΥ Γ ≡ C ΛΣΥ Γ (q).(3.44)In the case where Λ = Σ = Υ = Γ = ∅, the corresponding 4-vertex C ∅∅∅∅ should be equal tothe g<strong>en</strong>erating function Z C 4 of the 4d g<strong>en</strong>eralized Young diagramsC ∅∅∅∅ = Z C 4.Recall that the g<strong>en</strong>erating functional Z C 4 can be defined as a power series like,∑Z C 4 =q |P| ,4d partitions P(3.45)(3.46)where |P| is the number of hypercubes in P.In the g<strong>en</strong>eric case, C ΛΣΥ Γ should be giv<strong>en</strong> by the g<strong>en</strong>eralization 5 of the 3-vertex (2.4) andcould a priori be expressed in terms of the pro<strong>du</strong>ct of some hypothetic g<strong>en</strong>eralized Schur functionsS Λ (q).In the 2d partition set up, one can C (4) in quite similar manner. Using Eqs. (3.37)–(3.39), wecan g<strong>en</strong>erally define it as in Eq. (3.21). This is a kind of rank 12 objectC (abc)(def)(ghi)(jkl) ,(3.47)dep<strong>en</strong>ding on the Boltzmann weight and the boundary conditions (external mom<strong>en</strong>ta) a,...,l.To obtain its explicit expression, we use the following:(i) the relation betwe<strong>en</strong> the 4-vertex and composites of planar 3-vertices,(ii) the results on the usual 3-vertex formalism.The first property follows by noting that 4-vertices of the toric web-diagram of X 4 with thespecial Lagrangian fibrationR 4 × R × T 3 ,(3.48)corresponds to the intersection of the planar 3-vertices of three triangles. To fix the ideas, considerFig. 5 and focus on the point A repres<strong>en</strong>ting a 4-vertex of the toric web-diagram of X 4 .Thepoint A is the intersectionA = Δ 1 ∩ Δ 2 ∩ Δ 3 ,(3.49)of the triangles,5 The partition function Z C 3 can be also defined as the g<strong>en</strong>erating function of 3d partitions [30].


322 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341Fig. 8. A typical 4-vertex in O(−3 − m) → WP1,1,1,m 3 using 2d partitions. This is a spatial vertex made of three planar3-vertices: (abc), (def) and (ghi).Δ 1 = triangle ABC,Δ 2 = triangle ABD,Δ 3 = triangle ACD.These triangles are boundary faces of the tetrahedronABCD.Inside of the tetrahedron, the toric fiber isT 3 = S 1 × S 1 × S 1 .(3.50)(3.51)(3.52)On each triangle face, a circle shrinks leaving T 2 .On each edge of a triangle, one more circle shrinks leaving S 1 .At the vertex A, all 1-cycles of T 3 shrinks down to zero.The property captured by Eqs. (3.49) means that we may relate the 4-vertex C ΛΣΥ Γ tothree planar vertices of the triangles (3.50). This can be done by expressing the 3d partitions(Λ,Σ,Υ,Γ)in terms of 2d partitions (a, b, c), (d, e, f), (g, h, i) and (j, k, l) as followsΛ = (a, d, g),Υ = (e, ∅, f),Σ = (b, c, ∅),Γ = (∅, h, i).(3.53)The decomposition (3.53) is illustrated on the formal Fig. 8 where the three triangles are repres<strong>en</strong>tedin differ<strong>en</strong>t colors.Substituting (ΛΣΥ Γ ) as in Eq. (3.53), we can first rewrite C ΛΣΥ Γ like C (adh)(bc∅)(e∅f)(∅hi) .The latter reads immediately from Fig. 8 and is giv<strong>en</strong> byC (adg)(bc∅)(e∅f)(∅hi) = C abc C def C ghi ,where C abc , C def and C ghi are topological 3-vertices with the explicit expression Eq. (2.4).(3.54)β) The function Z X4 The toric web-diagram of the X 4 4-fold is giv<strong>en</strong> by Fig. 9.The corresponding partition function Z X4 can be computed by specifying the 4-vertices, propagators,framings and using Feynman like rules. Notice that in the 2d partition set up, the toricwebs of X 4 and H 3 are as in Fig. 10.Using the mom<strong>en</strong>ta prescriptions described by the Young diagrams of the Fig. 10 as wellas trivial boundary conditions for the external legs, the partition function reads in terms of theKähler mo<strong>du</strong>lus t of X 4 as follows:Z H3 = ∑ [(Aτν T ωϕ T ρσ T)(B ετ T χρ T ψς T)(F υɛ T ιψ T κω T)(G ϕκ T χσ T ις T)H τυωϕρσεχικψς],{ϰ} (3.55)


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 323Fig. 9. A typical the toric web-diagram of X 4 with boundary conditions (ΛΠΣΥ ).Fig. 10. Toric web-diagram of the local elliptic curve using 2d partitions. External and internal mom<strong>en</strong>ta have be<strong>en</strong>expressed in terms of 2d partitions.where the sum over {ϰ} stands for the collective sum over the 2d partitions ϰ = τ , υ, ω, ϕ, ρ, σ ,ε, χ, ι, κ, ψ, ς and where we have setA τυ T ωϕ T ρσ T = C ∅τυ TC ∅ωϕ TC ∅ρσ T,B ετ T χρ T ψς T = C ∅ετ TC ∅χρ TC ∅ψς T,andF υε T ιψ T κω T = C ∅υε TC ∅ιψ TC ∅κω T, (3.56)G ϕκ T χσ T ις T = C ∅ϕκ TC ∅χσ TC ∅ις T,H τυωϕρσεχικψς =∏ϰ=τ,υ,ω,ϕ,ρ,σ,ε,χ,ι,κ,ψ,ς(−e−t ) |ϰ| q κ(ϰ) ,(3.57)where κ(μ) is the second Casimir of the 2d partition μ (2.8). The factors C αβγ are giv<strong>en</strong> byEq. (2.4).


324 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–3413.3.3. Partition function for the local 2-torusThe partition function Z H3 of the local elliptic curve may be obtained by implem<strong>en</strong>ting inZ X4 the constraint relations (3.24) capturing the projection X 4 → H 3 .Choosing trivial boundary conditions for the external legs and using:(i) the expression of the 4-vertex (3.54),(ii) the rules of the planar vertex formalism of [28],we can write down directly the expression of the partition function Z H3 . We find∑ (Z H3 =A∗ωϕ T ρσ T B ∗ χρ T ψς T F ∗ ιψ T κω T G ∗ ϕκ T χσ T ις T Hωϕρσχικψς) ∗ ,withξ,ρ,σ,η,υ,τ,ς,θA ∗ ωϕ T ρσ T = C ∅ωϕ TC ∅ρσ T,B ∗ χρ T ψς T = C ∅χρ TC ∅ψς T,F ∗ ιψ T κω T = C ∅ιψ TC ∅κω T,G ∗ ϕκ T χσ T ις T = C ∅ϕκ TC ∅χσ TC ∅ις T,(3.58)(3.59)together withH ∗ ωϕρσχικψς =∏μ={ω,ϕ,ρ,σ,χ,ι,κ,ψ,ς}(−e−t ) |μ| q κ(μ) ,(3.60)where the factors C αβγ areasinEq.(2.4).In what follows, we turn to study the field theory set up of the local 2-torus by starting bylocal P 2 model.4. Sigma model for local P 2In this section, we first review briefly the supersymmetric sigma linear model realization ofthe local P 2 model. This model is useful for the purpose of this paper.We also use this field realization to fix conv<strong>en</strong>tion notations and to intro<strong>du</strong>ce some mathematicalobjects and their physical interpretations.The local P 2 model is nicely formulated in the language of 4D, N = 1 supersymmetry whichis, roughly, equival<strong>en</strong>t to the usual 2D, N = 2 supersymmetry. The complex two dim<strong>en</strong>sionprojective plane P 2 has one Kähler parameter t, interpreted as the Fayet–Iliopoulos (FI) couplingconstant in the supersymmetric gauge theory.The U(1) gauged linear sigma theory describing the local P 2 target space geometry involvesthe following 4D, N = 1 superfields (supersymmetric repres<strong>en</strong>tations):(1) A U(1) gauge superfield V = V(x,θ,¯θ) which reads, in the Wess–Zumino gauge, as follows:V =−θσ μ ¯θA μ − i ¯θ 2 θλ+ iθ 2 ¯θ ¯λ + 1 2 θ 2 ¯θ 2 D,(4.1)where (x μ ,θ a , ¯θȧ) stands for the 4D, N = 1 superspace coordinates.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 325In this relation, A μ (x) and (λ a (x), ¯λȧ(x)) are respectively the U(1) gauge vector and gauginofields.The scalar field D is the usual auxiliary field capturing the local Calabi–Yau geometry. It capturesas well as part of the scalar field pot<strong>en</strong>tial V of the gauge theoryV = D 2 + ∑ i|F i | 2 ,(4.2)where the F i terms will be intro<strong>du</strong>ced below.(2) Four chiral superfields {Φ 0 ,Φ 1 ,Φ 2 ,Φ 3 }, with θ-expansionΦ i = z i + θψ i + θ 2 F i ,(4.3)with z i being the field coordinates of local P 2 , ψ i the Weyl spinors and F i the so-calledF -auxiliary fields.The Φ i complex superfields carry the following q i -charges under the U(1) gauge symmetry,(q 0 ,q 1 ,q 2 ,q 3 ) = (−3, 1, 1, 1).(4.4)The q i ’s add exactly to zero as required by the Calabi–Yau condition3∑q i = 0,i=0(4.5)of local P 2 .The superfield Lagrangian d<strong>en</strong>sity L local P 2 = L(Φ, V ) of the local P 2 model reads, in theN = 1 D = 4 formalism, as follows,∫L(Φ, V ) =d 4 θ3∑i=0∫¯Φ i e 2qiV Φ i − 2td 4 θV + L gauge (V ).Here L gauge (V ) is the superspace Lagrangian d<strong>en</strong>sity for the U(1) vector multiplet that can befound in [47].The equation of motion of the auxiliary field D leads to,(4.6)|z 1 | 2 +|z 2 | 2 +|z 3 | 2 − 3|z 0 | 2 = t.(4.7)It is nothing but the defining equation of the local projective plane O(−3) → P 2 .The compact part P 2 of this threefold is a complex plane giv<strong>en</strong> by the divisor z 0 = 0; it isparameterized by the complex coordinates (z 1 ,z 2 ,z 3 ) describing a complex surface embeddedin C 3 and has a U(1) gauge symmetry rotating the phases of the coordinates variables.By setting x i =|z i | 2 , the complex surface z 0 = 0 can be repres<strong>en</strong>ted by the planar triangle[48],x 1 + x 2 + x 3 = t(4.8)with Kähler mo<strong>du</strong>lus t; see also Fig. 1.Because of the symmetry under permutation of the x i ’s, the triangle is equilateral. The l<strong>en</strong>gthof its edges are equal to t and th<strong>en</strong> it has an area giv<strong>en</strong> by A = t2√ 32.


326 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–3415. Field model for O(−3) → E (t,∞)We first study the gauge invariant supersymmetric field model with target space giv<strong>en</strong> by thecurve E (t,∞) = ∂P 2 . For details on the deg<strong>en</strong>erate elliptic curve E (t,∞) ,seeApp<strong>en</strong>dix A. Th<strong>en</strong>,we consider the ext<strong>en</strong>sion to O(−3) → E (t,∞) .5.1. Divisors of local P 2To begin, notice that the local P 2 Eq. (4.7) has several divisors; i.e., codim<strong>en</strong>sion one subspacesdescribing boundary patches of the normal bundle of the projective plane.The standard ones are obtained by setting one of the z i ’s to zero; z i = 0 with i = 0, 1, 2, 3.5.1.1. Toric boundary of P 2In this paragraph, we consider the three following complex surfaces [D 1 ], [D 2 ] and [D 3 ],[D 1 ]: |z 2 | 2 +|z 3 | 2 − 3|z 0 | 2 = t ⇔ z 1 = 0,[D 2 ]: |z 3 | 2 +|z 1 | 2 − 3|z 0 | 2 = t ⇔ z 2 = 0,[D 3 ]: |z 1 | 2 +|z 2 | 2 − 3|z 0 | 2 = t ⇔ z 3 = 0,and their union [D]=[D 1 ]∪[D 2 ]∪[D 3 ].To see what this local geometry describes precisely; let us set |z 0 | 2 = 0 in above equationsfrom where one sees that each relation describes a complex one dim<strong>en</strong>sion projective line P 1 .Todistinguish betwe<strong>en</strong> these complex projective lines, we use the conv<strong>en</strong>tion notation P 1 iwhere thesubindex i refers to z i = 0. Thus we have(5.1)P 1 1 : |z 2| 2 +|z 3 | 2 = t,P 1 2 : |z 3| 2 +|z 1 | 2 = t,P 1 3 : |z 1| 2 +|z 2 | 2 = t.(5.2)As we see, these projective lines have the following intersection matrix 6( )−2 1 1P 1 i ∩ P1 j = 1 −2 1 ,(5.3)1 1 −2from which one sees that the complex curveE (t,∞) = P 1 1 ∪ P1 2 ∪ P1 3 ,is elliptic (E (t,∞) ∼ T 2 ). Indeed, computing3∑E (t,∞) ∩ E (t,∞) = P 1 i ∩ P1 i + 2( P 1 1 ∩ P1 2 + P1 2 ∩ P1 3 + P1 3 ∩ 1) P1 ,i=1we get, up on using (5.3),E (t,∞) ∩ E (t,∞) =−3 × 2 + 2 × 3 = 0.(5.4)(5.5)(5.6)6 D<strong>en</strong>oting by {α 1 ,α 2 ,α 3 },abasisofH 2 (P 2 ,R), and by {A 1 ,A 2 ,A 3 }, the <strong>du</strong>al basis of H 2 (P 2 ,R)with ∫ A i α j = δ i j ,the intersection matrix Eq. (5.3) is giv<strong>en</strong> by I ij = ∫ P 2 α i ∧ α j .


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 327Fig. 11. (a) Left: Toric graph of P 2 . (b) Right: Toric graph of E (t,∞) where we have added a hole to avoid confusion.From the toric diagram of P 2 , one also see that E (t,∞) describes indeed a toric complex onedim<strong>en</strong>sionalcurve defining the toric boundary of P 2 ; i.e.:E (t,∞) ≡ ∂P 2 .It is this curve that will be used to deal with the local 2-torus in the large complex structure limit.5.1.2. Elliptic curve E (t,∞)An interesting question concerns the derivation of the defining equation describing the ellipticcurve E (t,∞) . From the above analysis, it is not difficult to see that E (t,∞) is giv<strong>en</strong> by thefollowing system of equations (see Fig. 11),⎧⎨ |z 1 | 2 +|z 2 | 2 +|z 3 | 2 = t,z (5.8)⎩ i ≡ z i e iqiα ,z 1 z 2 z 3 = 0.In these relations, we have three complex variables (z 1 ,z 2 ,z 3 ) subject to three constraint equations.The two first equations, which are real, are just the defining linear sigma model equationof P 2 . They will be interpreted as the field equation of motion of the auxiliary D-field in supersymmetricsigma model.The third relation, which is covariant under U(1) gauge symmetry, is an extra complex conditionimplem<strong>en</strong>ted in order to restrict P 2 geometry to its toric boundary ∂P 2 . It will be interpretedlater as the equation of motion of the auxiliary F-fields.Notice that, the implem<strong>en</strong>tation of the boundary condition is a new feature. It can be th<strong>en</strong>viewed as:(i) a g<strong>en</strong>eralization of the usual approach for dealing with sigma model realization of toricmanifolds,(ii) a way to approach g<strong>en</strong>us g Riemann surfaces,(iii) a method that can be used to describe the toric boundary of more g<strong>en</strong>eral complex n-dim<strong>en</strong>sionaltoric Calabi–Yau manifolds. We will make a comm<strong>en</strong>t regarding this point in theconclusion section.Notice finally that for t ≠ 0 the three complex variables cannot vanish simultaneously, i.e.,(5.7)(z 1 ,z 2 ,z 3 ) ≠ (0, 0, 0).(5.9)


328 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341In the particular case t = 0, the geometry collapses to the origin (0, 0, 0) whereliveaP 2 singularityand an elliptic one.5.1.3. Divisor O(−3) → E (t,∞)Using the above result on the toric realization of the elliptic curve, one can immediately writedown the defining equation of the divisor O(−3) → E (t,∞) of the local P 2 .Wehave⎧⎨ |z 1 | 2 +|z 2 | 2 +|z 3 | 2 − 3|z 0 | 2 = t,z (5.10)⎩ i ≡ z i e iqiα , i = 0, 1, 2, 3,z 1 z 2 z 3 = 0,where (q 0 ,q 1 ,q 2 ,q 3 ) areasinEq.(4.4) and where the complex variable z 0 parameterizes th<strong>en</strong>on-compact direction O(−3).If we do not worry about the Calabi–Yau condition, the first relation can be ext<strong>en</strong>ded as|z 1 | 2 +|z 2 | 2 +|z 3 | 2 − m|z 0 | 2 = t,where m is an arbitrary positive integer.(5.11)5.2. Superfield actionHere we give the supersymmetric field description of (5.11). We start by studying the fieldrealization of the toric curve E (t,∞) . Th<strong>en</strong> we consider its ext<strong>en</strong>sion to the local geometry.5.2.1. Gauge invariant model for the elliptic curveTo build the supersymmetric model describing the toric curve E (t,∞) , we start from the superfieldcont<strong>en</strong>t Eq. (4.7) of local P 2 theory and implem<strong>en</strong>t the constraint equation (5.10) by usingLagrange multiplier method together U(1) gauge invariance.The appropriate Lagrange superfield multiplier is giv<strong>en</strong> by a chiral superfield Υ with chargeq Υ =−3 under U(1) gauge symmetry so that the chiral superfield monomialW(Φ,Υ)= Φ 1 Φ 2 Φ 3 Υ,(5.12)is gauge invariant. Thus the supersymmetric Lagrangian super-d<strong>en</strong>sity with target space E (t,∞)is giv<strong>en</strong> by the d<strong>en</strong>sity,( ∫)L E = L P 2 + g d 2 θW(Φ,Υ)+ hc ,(5.13)where g is a complex coupling constant. Since Υ has no kinetic term, its elimination through theequation of motiongivesδL EδΥ = 0(5.14)gΦ 1 Φ 2 Φ 3 = 0,(5.15)whose lowest term is precisely z 1 z 2 z 3 = 0.Notice that contrary to P 2 , the superfield realization of the curve E (t,∞) has a non-trivial chiralsuperpot<strong>en</strong>tial. As we see, this result is a particular situation that can be ext<strong>en</strong>ded to build toricrealization of other toric manifolds.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 329Notice also that the Lagrange superfield multiplier Υ can be giv<strong>en</strong> a geometric interpretation.This superfield has no kinetic term ῩΥ nor couplings to the gauge superfield V ; i.e., no termtype∫d 4 θ Ῡe 2qV Υ,(5.16)in the Lagrangian super-d<strong>en</strong>sity. The lack of (5.16) can be interpreted as corresponding tofreezing the supersymmetric gauge invariant dynamics of Υ . This property explains why theCalabi–Yau condition for the complex toric curve E (t,∞) should read as,3∑q i + q γ =i=13∑q i − 3 = 0.i=1(5.17)We will turn to this property wh<strong>en</strong> we consider the local threefold O(m) ⊕ O(−m) → E (t,∞) .Notice also that the chiral superpot<strong>en</strong>tial (5.12) is not the unique gauge invariant term onemay have. The g<strong>en</strong>eral form of W(Φ,Υ) is giv<strong>en</strong> byW(Φ,Υ)=∑n 1 +n 2 +n 3 =3g n1 ,n 2 ,n 3Φ n 11 Φn 22 Φn 33 .(5.18)We will discuss this point in Section 5 wh<strong>en</strong> we study the g<strong>en</strong>eralization to higher dim<strong>en</strong>sion CYmanifolds.For the mom<strong>en</strong>t, let us complete this discussion by giving the gauged superfield realization ofthe complex surface O(−m) → E (t,∞) .5.2.2. Field model for the divisor O(−m) → E (t,∞)In addition to the U(1) gauge superfield V , this model involves five chiral superfields(Φ 0 ,Φ 1 ,Φ 2 ,Φ 3 ,Υ)with charges(q 0 ,q 1 ,q 2 ,q 3 ,q γ ) = (−m, 1, 1, 1, −3).The Lagrangian super-d<strong>en</strong>sity L divisor is giv<strong>en</strong> by,∫L divisor =d 4 θ( ∫+ g3∑i=0∫¯Φ i e 2qiV Φ i + L gauge (V ) − 2t)d 2 θW(Φ,Υ)+ hc ,d 4 θV(5.19)(5.20)where the chiral superpot<strong>en</strong>tial W(Φ,Υ)is as in Eq. (5.12). Here also the first Chern class of thecomplex surface has a contribution coming from Υ and reads as3∑q i + q γ =−mi=0showing, as expected, that O(−m) → E (t,∞) is not a Calabi–Yau surface.(5.21)5.3. Mo<strong>du</strong>li space of supersymmetric vacuumHere we study the supersymmetric vacuum of the field model (5.20). We show that the surface(5.10) corresponds to a particular vacuum giv<strong>en</strong> by the vev z γ = 0.


330 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–3415.3.1. Mo<strong>du</strong>li space of vacuaIn the supersymmetric vacuum, the vanishing condition of the scalar pot<strong>en</strong>tial V = V(z) ofthe model (5.13) reads as|D| 2 +|F 0 | 2 +|F 1 | 2 +|F 2 | 2 +|F 3 | 2 +|F γ | 2 = 0.(5.22)The dep<strong>en</strong>d<strong>en</strong>ce of the scalar pot<strong>en</strong>tial V in the scalar fields z is obtained by replacing theauxiliary fields D and F i by their explicit expressions in terms of the matter fieldsD = D(z 0 ,z 1 ,z 2 ,z 3 ,z γ ),F i = F i (z 0 ,z 1 ,z 2 ,z 3 ,z γ ).(5.23)These expressions are obtained by using the equations of motionδLδD = 0,δLδ ¯F i= 0.Eq. (5.22) is solved as follows,D = 0, F i = 0.(5.24)(5.25)As noted before, D = 0 leads to|z 1 | 2 +|z 2 | 2 +|z 3 | 2 − m|z 0 | 2 = t(5.26)and describes local P 2 for the particular case m = 3.F i = 0 involves five terms: F 0 which is trivial, and the remaining F γ , F 3 , F 2 , and F 1 lead to:z 1 z 2 z 3 = 0,z 1 z 2 z γ = 0,z 3 z 1 z γ = 0,z 2 z 3 z γ = 0,(5.27)where z γ stands for the lowest compon<strong>en</strong>t field of the chiral superfield Υ . There are severalsolutions of these relations. These solutions may be classified into two sets:(1) The first set is giv<strong>en</strong> byz γ = 0.(5.28)Consequ<strong>en</strong>tly Eqs. (5.27) re<strong>du</strong>ce to the first equation z 1 z 2 z 3 = 0.The mo<strong>du</strong>li background associated with this solution describes exactly the complex surfaceO(−3) → E (t,∞) .(2) The second set corresponds toz γ ≠ 0(5.29)and two variables amongst the three z 1 , z 2 and z 3 vanish.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 331So Eqs. (5.26) becomez 2 = z 3 = 0: |z 1 | 2 − m|z 0 | 2 = t,z 1 = z 3 = 0: |z 2 | 2 − m|z 0 | 2 = t,z 1 = z 2 = 0: |z 3 | 2 − m|z 0 | 2 = t.(5.30)5.3.2. Case z γ = 0Let us now consider the interesting case z γ = 0 and study the solution of constraint equationz 1 z 2 z 3 = 0. Here also there are several solutions which we list below:(1) Case z 3 = 0but(z 1 ,z 2 ) ≠ (0, 0)In this case the geometry re<strong>du</strong>ces to|z 1 | 2 +|z 2 | 2 − 3|z 0 | 2 = t.(5.31)It describes the local complex projective lineO(−3) → P 1 ,(5.32)which we d<strong>en</strong>ote as O(−3) → P 1 3 where the sub-index 3 on P1 3 refers to z 3 = 0.The same thing is valid for z 1 = 0 and z 2 = 0.They describe respectively the local surfaces O(−3) → P 1 1 and O(−3) → P1 2 .(2) Case z 1 = 0, z 2 = 0, z 3 = √ tThis solution describes one of the three possible vertices of O(−3) → E (t,∞) ; the two othervertices are associated with the points:(i) z 1 = 0, z 2 = √ t, z 3 = 0 and,(ii) z 1 = √ t, z 2 = 0, z 3 = 0.(3) Case z 1 = z 2 = z 3 = 0isaP 2 singularityThis solution corresponds to the limit t = 0 where both E (t,∞) and so P 2 collapse down to apoint.The above analysis can be viewed as an interesting step towards the study of topologicalvertex of local g<strong>en</strong>us g-Riemann surfaces (in particular g = 1) by using toric diagrams based onthe curve E (t,∞) . To that purpose, one first has to build the toric realization of basic objects ofthe topological vertex method. For instance, the complex coordinates associated with the verticesof the elliptic curve E (t,∞) are giv<strong>en</strong> by the local patchesU 1 : z (1)0 , z(1) 1 , z(1) 2 , z(1) 3= 0,U 2 : z (2)0 , z(2) 1 , z(2) 3 , z(2) 2= 0,U 3 : z (3)0 , z(3) 2 , z(3) 3 , z(3) 1= 0.(5.33)


332 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341The upper index of z (i)jrefers to the corresponding chart U i . Note that on each chart, we have therelationz (i)1 z(i) 2 z(i) 3= 0, i = 1, 2, 3.(5.34)The patches U i could be interpreted as the three “toric pants” needed for building E (t,∞) andmay be related to the topological pant considered in [30]. By gluing these three patches, onerepro<strong>du</strong>ces E (t,∞) .6. Local 2-torusThe local complex surface X 2 = O(−m) → E (t,∞) we have considered so far is not a Calabi–Yau 2-fold. The first Chern class c 1 (T ∗ X 2 ) of this variety is equal to −m. For our concern, thissurface is thought of as a divisor of the local Calabi–Yau threefold,O(m) ⊕ O(−m) → E (t,∞) .(6.1)6.1. Field modelThe supersymmetric field model relations describing the toric Calabi–Yau threefold (6.1) canbe easily derived from previous study. They are giv<strong>en</strong> by the following system of compon<strong>en</strong>tfield equations,{|z1 | 2 +|z 2 | 2 +|z 3 | 2 + 3|z 4 | 2 − 3|z 0 | 2 = t,(6.2)z 1 z 2 z 3 = 0.Here z 0 and z 4 parameterize the non-compact directions and (z 1 ,z 2 ,z 3 ) are as before. The toricgraph of this local threefold is shown on Fig. 12; it has three tetra-val<strong>en</strong>t vertices.To get the superfield Lagrangian d<strong>en</strong>sity L locT 2, we think about Eqs. (6.2) as the field equationsof motion of the D and F i auxiliary fields.The first relation is associated withδL locEδD = 0,while the second follows from,δL locE= 0.δF iThe result is∫ 3∑∫L locE = d 4 θ ¯Φ i e 2V Φ i +i=1∫+ L gauge (V ) − 2td 4 θ ( ¯Φ 0 e −2mV Φ 0 + ¯Φ 4 e 2mV Φ 4)( ∫d 4 θV + g)d 2 θΦ 1 Φ 2 Φ 3 Υ + hc ,(6.3)(6.4)(6.5)where Υ is a Lagrange superfield multiplier capturing the constraint restricting the field variablesto the boundary of P 2 .In addition to the U(1) gauge multiplet V , the chiral superfields of this model are(Φ 0 ,Φ 1 ,Φ 2 ,Φ 4 ,Φ 5 ,Υ)(6.6)


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 333Fig. 12. Toric graph of O(−3) → E (t,∞) . The compact part is E (t,∞) = (∂P 2 ) with the usual three vertices. Its toricfrontier consists of three intersecting P 1 ’s in the homology class of 2-torus. (a) Figure (on left) repres<strong>en</strong>ts real skeleton.(b) Figure (on right) gives its fatt<strong>en</strong>ing.and carry the following q i -charges under the U(1) gauge symmetry,(q 0 ,q 1 ,q 2 ,q 3 ,q 4 ,q γ ) = (−m, 1, 1, 1,m,−3),where m is a priori equal to 3; but in g<strong>en</strong>eral can take any integral value.(6.7)6.2. G<strong>en</strong>eralizationThe construction we have developed here above can be g<strong>en</strong>eralized to other Calabi–Yau manifolds.Below we make a comm<strong>en</strong>t on two kinds of g<strong>en</strong>eralizations. The first ext<strong>en</strong>sion deals withthe gauged sigma model realization of local g<strong>en</strong>us g-Riemann surfaces for g 2. The secondg<strong>en</strong>eralization concerns sigma model approach for higher complex dim<strong>en</strong>sional compact toricCalabi–Yau manifolds.6.2.1. Local g<strong>en</strong>us g-Riemann surfacesSo far we have se<strong>en</strong> that for each local elliptic curve, it is associated a U(1) gauge symmetry.This gauge symmetry is inherited from the P 2 model. Since local g<strong>en</strong>us g-Riemann surfaces canbe <strong>en</strong>gineered by gluing several local elliptic curves, we conclude that a class of local g<strong>en</strong>usg-Riemann surfaces could be described by higher rank Abelian U n (1) gauged supersymmetricfield model type. The rank n of the gauge symmetry dep<strong>en</strong>ds on the way the gluing is done.To illustrate the idea, let us give the example of g = 2-Riemann surface described by a 2DU 2 (1) gauged N = 2 supersymmetric sigma model.The local g = 2-Riemann surface in the large complex structures limits can be <strong>en</strong>gineered bygluing two local elliptic curves with compact base E (t,∞)1= ∂P 2 1and E(t,∞)2= ∂P 2 2. In the sigmamodel approach, we distinguish differ<strong>en</strong>t repres<strong>en</strong>tations according to whether P 2 1 and P2 2 havean intersection point or edge.In the first case, the sigma model involves five complex field variables,(z 1 ,z 2 ,z 3 ,z 4 ,z 5 )and a U 2 (1) gauge invariance under which these complex variables have the following gaugecharges(q1i)= (1, 1, 1, 0, 0),(6.9)(q2i)= (0, 0, 1, 1, 1).(6.8)


334 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341The field theoretic equations describing the compact part of the mo<strong>du</strong>li space of supersymmetricvacua are giv<strong>en</strong> by,{ {|z1 | 2 +|z 2 | 2 +|z 3 | 2 = t 1 , |z3 | 2 +|z 4 | 2 +|z 5 | 2 = t 2 ,(6.10)z 1 z 2 z 3 = 0,z 3 z 4 z 5 = 0,where t 1 and t 2 are respectively the Kähler mo<strong>du</strong>li of the projective planes P 2 1 and P2 2 .Theholomorphic constraint equations z 1 z 2 z 3 = 0 and z 3 z 4 z 5 = 0 are implem<strong>en</strong>ted in the gauged supersymmetricsuperfield model by two chiral superfields Υ 1 and Υ 2 with gauge charges (q 1 γ ,q2 γ )equal to (−3, 0) and (0, −3) respectively.Notice that Eq. (6.10) describes indeed a complex curve with g<strong>en</strong>us g = 2. The toric threefoldbased on this g<strong>en</strong>us g = 2 curve is parameterized by sev<strong>en</strong> complex variables,(z 0 ,z 1 ,z 2 ,z 3 ,z 4 ,z 5 ,z 6 ),(6.11)with gauge charges as(q1i)= (−3, 1, 1, 1, 0, 0, −3),(6.12)(q2i)= (−3, 0, 0, 1, 1, 1, −3).The gauged supersymmetric field theoretical equation−m|z 0 | 2 +|z 1 | 2 +|z 2 | 2 +|z 3 | 2 + m|z 6 | 2 = t 1 , z 1 z 2 z 3 = 0,−n|z 0 | 2 +|z 3 | 2 +|z 4 | 2 +|z 5 | 2 + n|z 6 | 2 = t 2 , z 3 z 4 z 5 = 0,(6.13)where m and n are in g<strong>en</strong>eral arbitrary integers; but can be set equal to 3 to keep in touch withthe first Chern class of the complex two-dim<strong>en</strong>sional projective plane.These relations involve sev<strong>en</strong> complex variables constrained by four complex constraint equationsleaving th<strong>en</strong> three complex variables free. Note also that the first relation of the above equationdescribes O(m) ⊕ O(−m) → E (t,∞)1while the second describes O(n) ⊕ O(−n) → E (t,∞)2.6.2.2. Higher-dim<strong>en</strong>sional toric CY manifoldsThe gauged supersymmetric sigma model for the boundary surface of local P 2 that we haveconsidered in this paper can be ext<strong>en</strong>ded for compact divisors of local P n−1 . The latter is giv<strong>en</strong>by the following U(1) gauge invariant complex dim<strong>en</strong>sion n hypersurfac<strong>en</strong>∑n|z 0 | 2 + |z i | 2 = t,(6.14)i=1embedded in C n+1 parameterized by the local coordinates {z 0 ,z 1 ,z 2 ,...,z n } with gauge charge(q 0 ,q 1 ,q 2 ,...,q n ) = (−n, 1, 1,...,1).(6.15)In Eq. (6.14), t is the usual Kähler parameter of P n−1 . To describe the compact (divisor) boundary∂(P n−1 ) of the toric n-fold, we supplem<strong>en</strong>t the hypersurface equation by the following extragauge covariant constraint relation,n∏z i = 0.(6.16)i=1


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 335Ext<strong>en</strong>ding the analysis of Section 4, theU(1) gauged supersymmetric sigma model describing∂(P n−1 ) reads as follows( ∫)L ∂P n−1 = L P n−1 + g d 2 θW(Φ,Υ)+ hc(6.17)with chiral superpot<strong>en</strong>tialW(Φ,Υ)= Υn∏Φ i .i=1(6.18)The gauge charge q γ of the Lagrange multiplier superfield Υ is equal to (−n). Here also the firstChern class of ∂(P n−1 ) is id<strong>en</strong>tically zero. As noted before, this property is not a new featuresince the most g<strong>en</strong>eral gauge invariant chiral superpot<strong>en</strong>tial ext<strong>en</strong>ding Eq. (6.18) is giv<strong>en</strong> by( )W(Φ,Υ)=∑g {mi }m 1 +···+m n =nΥn∏i=1Φ m ii,(6.19)where g {mi } are complex coupling constants.The equation of motion of Υ gives a degree n homog<strong>en</strong>eous polynom describing a complex(n − 2) dim<strong>en</strong>sion holomorphic CY hypersurface with complex structures g {mi }.7. ConclusionIn this paper, we have set up the basis of the non-planar topological 3-vertex method to computethe topological string amplitudes for the family of local elliptic curvesO(m) ⊕ O(−m) → E (t,μ) , m∈ Z,in the limit of large complex structure μ; i.e.,(7.1)|μ|→∞.G<strong>en</strong>erally speaking, the base E (t,μ) stands for an elliptic curve with Kähler parameter t andcomplex structure μ embedded in the projective plane P 2 . In the large limit μ; the correspondingelliptic curve E (t,∞) is realized as the toric boundary of P 2 ;seeApp<strong>en</strong>dix A for more details; inparticular Eqs. (A.5), (A.12), (A.17).First, we have reviewed the main idea of the usual (planar) topological 3-vertex method fornon-compact toric threefolds.Th<strong>en</strong>, we have drawn the first lines of the non-planar topological 3-vertex method for the localdeg<strong>en</strong>erate 2-torus. The latter is a non-compact toric Calabi–Yau threefold giv<strong>en</strong> by a hypersurfacein a complex Kähler 4-fold.The key idea in getting the particular toric repres<strong>en</strong>tation of the local 2-torus with large complexstructure is based on thinking about E (t,∞) as giv<strong>en</strong> by the toric boundary of the complexprojective plane P 2 .Inthisview,O(m) ⊕ O(−m) → E (t,∞) becomes a toric threefold and soone may ext<strong>en</strong>d the results of the topological 3-vertex method of [28] to the case of the local(deg<strong>en</strong>erate) 2-torus. Obviously, to compute the topological amplitudes, we have to use the nonplanar3-vertex method rather than the usual planar 3-vertex one. Regarding this matter, wehave giv<strong>en</strong> first results concerning the local deg<strong>en</strong>erate elliptic curve O(m) ⊕ O(−m) → E (t,∞) .More analysis is however still needed before getting the complete explicit results.(7.2)


336 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341We have also developed the gauged supersymmetric sigma model realization that underliesthe geometry the local 2-torus with |μ|→∞and exhibited explicitly the role of D- and F-terms.We have discussed as well how this construction could be ext<strong>en</strong>ded to local g<strong>en</strong>us g-Riemannsurfaces O(m) ⊕ O(2 − 2g − m) → Σ g in the limit of large complex structures.The results obtained in the field theory part of the paper may also be viewed as an explicitanalysis regarding implem<strong>en</strong>tation of F-terms in the Witt<strong>en</strong>’s original work on phases of N = 2supersymmetric theories in two dim<strong>en</strong>sions [47].Acknowledgem<strong>en</strong>tsThe authors thank the International C<strong>en</strong>tre for Theoretical Physics, and S. Randjabar Daemifor kind hospitality at ICTP. This research work is supported by Protars III CNRST-D12/25.App<strong>en</strong>dix AIn this app<strong>en</strong>dix, we give useful properties on the complex projective plane and on particularaspects on the complex curves in P 2 .More precisely d<strong>en</strong>oting by P 2 t , the projective plane with Kähler parameter t and by E(t,μ)the following elliptic curve in P 2 t ,E (t,μ) : z 3 1 + z3 2 + z3 3 + μz 1z 2 z 3 = 0we want to show that the boundary ∂(P 2 t ) is nothing but the deg<strong>en</strong>erate limit μ →∞of E(t,μ) ;that is∂ ( P 2 )t ≃ E (t,∞) .(A.1)This question can be also rephrased in other words by using the fibration,P 2 = B 2 × T 2 ,(A.2)where B 2 is real 2-dim<strong>en</strong>sional base (an equilateral triangle). The boundary ∂(P 2 ) is a toricsubmanifold with fibration∂ ( P 2) = Δ 1 × S 1 ,(A.3)where Δ 1 = (∂B 2 ) is the boundary of a triangle.Clearly, thought not exactly the standard 2-torus S 1 × S 1 , the boundary ∂(P 2 ) has somethingto do with it. It is the large complex structure μ of the elliptic curve E (t,μ) ;say|μ|→+∞.(A.4)As we need both Kähler and complex structures to answer the question (A.1), let us first givesome useful details and th<strong>en</strong> turn to derive the id<strong>en</strong>tity ∂(P 2 t ) ≃ E(t,∞) .Projective plane P 2There are differ<strong>en</strong>t ways to deal the complex projective plane P 2 . Below, we give two <strong>du</strong>aldescriptions by using the so-called type IIA and type IIB geometries [51].Type IIA geometryIn this set up, known also as toric geometry, the projective plane P 2 is defined by the followingreal 4-dim<strong>en</strong>sional compact hypersurface in C 3 ,|z 1 | 2 +|z 2 | 2 +|z 3 | 2 = t,(A.5)


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 337where (z 1 ,z 2 ,z 3 ) stand for local complex coordinates. In the above relation, the complex variablesobey the gauge id<strong>en</strong>tificationsz ′ k ≡ eiϕ z k ,(A.6)where the real phase ϕ is the parameter of the U(1) gauge symmetry. The phase ϕ can be used tofix one of the three phases of the z k =|z k |e iϕ kcomplex coordinates leaving th<strong>en</strong> two free phases;say ϕ 1 and ϕ 2 . These free phases are precisely the ones used to parameterize the 2-torus in thefibration (A.2).The positive parameter t is the Kähler mo<strong>du</strong>lus of the projective plane; it controls the sizeof P 2 . Indeed, in the singular limit t → 0, we have the two following:(i) the size of the complex surface P 2 vanishes[ (lim vol P2 )] = 0,t→0(A.7)in agreem<strong>en</strong>t with both the relation vol(P 2 ) ∼ t 2 (footnote 7) and Eq. (A.5) which becomesth<strong>en</strong> singular.(ii) the size of the complex boundary ∂(P 2 ) vanishes as well[ [ (lim vol ∂ P2 )]] = 0.t→0(A.8)The two above relations show that the Kähler parameter t of P 2 and the Kähler parameter r of itboundary ∂(P 2 ) are intimately related. We will show later that they are the same. 7Notice in passing that in the field theory language, the relation (A.5) has an interpretation asthe field equation of motion of the D-auxiliary field in the U(1) gauged sigma model realizationof P 2 . There, the Kähler parameter t is interpreted as the Fayet–Iliopoulos coupling constantterm. This description is well known; some of its basic aspects have be<strong>en</strong> studied in Section 4 ofthis paper; we will th<strong>en</strong> omit re<strong>du</strong>ndant details.Type IIB geometryIn the type IIB geometry, one thinks about the complex projective surface P 2 as a complexholomorphic algebraic surface obtained by taking the coset of the complex space C 3 \{(0, 0, 0)}by the complex Abelian group C ∗ ;P 2 = [ C 3 \ { (0, 0, 0) }] /C ∗ .(A.9)The C ∗ group action allows to make the following id<strong>en</strong>tifications,(z 1 ,z 2 ,z 3 ) ≡ (λz 1 ,λz 2 ,λz 3 )(A.10)with λ being an arbitrary non-zero complex number. This id<strong>en</strong>tification re<strong>du</strong>ces the complex 3dim<strong>en</strong>sion down to complex 2 dim<strong>en</strong>sions. Here also, one can make gauge choices by workingin a particular local coordinate patch. A standard gauge choice is the one giv<strong>en</strong> by the conditionλz 3 = 1.7 From Eq. (A.5), it is not difficult to see that the volume of P 2 is proportional to t 2 while the volume of its boundary∂(P 2 ) is proportional t.


338 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341Complex curves in P 2Complex curves (real Riemann surfaces) in P 2 are complex codim<strong>en</strong>sion one submanifoldsobtained by imposing one more complex constraint relation f(z i ) = 0 on the projective complexvariables z 1 ,z 2 and z 3 . The most common curves in P 2 are obviously the projective lines P 1 ,conics and elliptic curves.G<strong>en</strong>erally speaking, the constraint equation f(z i ) = 0 can be stated as follows,f(λz 1 ,λz 2 ,λz 3 ) = λ n f(z 1 ,z 2 ,z 3 ) = 0,(A.11)where n stands for the degree of homog<strong>en</strong>eity of the curve. The case n = 3 is giv<strong>en</strong> by thefollowing typical cubicz 3 1 + z3 2 + z3 3 + μz 1z 2 z 3 = 0.(A.12)This relation describes an elliptic curve E of degree 3 with a complex structure μ. This curve Ehas be<strong>en</strong> ext<strong>en</strong>sively used in physical literature; in particular in the geometric <strong>en</strong>gineering of4D superconformal field theories embedded in 10D type IIB superstring on elliptically fiberedCalabi–Yau threefolds [44–46] and [51,52].Before proceeding further, it is interesting to notice that the elliptic curve E is a g<strong>en</strong>us oneRiemann surface having a real 3d mo<strong>du</strong>li space; parameterized by(μ 1 ,μ 2 ; r)(A.13)with μ = μ 1 + iμ 2 is the complex structure and r is its Kähler mo<strong>du</strong>lus. So, elliptic curves maybe g<strong>en</strong>erally d<strong>en</strong>oted as followsE (r,μ) .(A.14)Regarding the complex parameter μ of the elliptic curve E (r,μ) , it is explicitly exhibited in typeIIB geometry set up as shown on Eq. (A.12).However, it is interesting to notice that the Kähler parameter r cannot be exhibited explicitlysince E (r,μ) has no standard type IIA geometry realization 8 of the type giv<strong>en</strong> by Eq. (A.5).The construction developed in this paper gives a way to circumv<strong>en</strong>t this difficulty by usingthe deg<strong>en</strong>erate repres<strong>en</strong>tation (A.3).With the above features in mind, we turn now to the derivation of Eq. (A.1).∂(P 2 ) as the deg<strong>en</strong>erate elliptic curve E (r,∞)Here we would like to show that ∂(P 2 ) is E (r,μ) but with a large complex structure μ; that is|μ|→∞.To get the key point behind the id<strong>en</strong>tity (A.1) as well as the deg<strong>en</strong>eracy of the elliptic curveE (r,μ) , we give the two following properties:(a) the projective plane P 2 has three particular intersecting divisors D i . These are associatedwith the hyperlinesz i = 0in P 2 which, up on using Eq. (A.5), lead to the following relations(A.15)8 In toric geometry the real base of the fibration B n × T n of complex n-dim<strong>en</strong>sional toric manifolds involves projectivelines. Torii appear rather in the fiber.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 339D 1 : |z 2 | 2 +|z 3 | 2 = t,D 2 : |z 1 | 2 +|z 3 | 2 = t,D 3 : |z 1 | 2 +|z 2 | 2 = t.(A.16)From these equations, we see that each divisor D i is a projective line with Kähler parameter t.The equality of the Kähler parameters t 1 = t 2 = t 3 = t of these projective lines may be also interpretedas <strong>du</strong>e to the permutation symmetry of the {z i } projective coordinate variables of P 2 .Thisfeature translates, in the language of toric geometry, as corresponding to having an equilateraltriangle for the real base B 2 .(b) The divisors {D i } are precisely the ones we get by taking the large complex structure limit(A.4) of the complex curve Eq. (A.12). Under this condition, Eq. (A.12) re<strong>du</strong>ces th<strong>en</strong> to thedominant monomialμz 1 z 2 z 3 = 0.(A.17)Notice that the above relation is obviously invariant under the C ∗ transformations (A.10) sinceμ(λz 1 )(λz 2 )(λz 3 ) = λ 3 (μz 1 z 2 z 3 ) = 0.(A.18)To have more insight about the elliptic curve E (t,∞) with large complex structure; |μ|→∞,itis interesting to solve Eq. (A.17). There are three solutions classified as follows:(i) z 1 = 0, what ever the two other complex variables z 2 and z 3 are; provided that(z 2 ,z 3 ) ≠ (0, 0),(z 2 ,z 3 ) ≡ (λz 2 ,λz 3 ).(A.19)But these relations are nothing but the definition of the divisor D 1 in type IIB geometry.(ii) z 2 = 0, what ever the other complex variables z 1 and z 3 are; provided that(z 1 ,z 3 ) ≠ (0, 0),(z 1 ,z 3 ) ≡ (λz 1 ,λz 3 ),(A.20)describing th<strong>en</strong> the divisor D 2 .(iii) z 3 = 0, what ever the other complex variables z 2 and z 1 are; provided that(z 1 ,z 2 ) ≠ (0, 0),(z 1 ,z 2 ) ≡ (λz 1 ,λz 2 ),(A.21)associated with the divisor D 3 .To conclude the boundary (∂P 2 t ) of the projective plane P2 t is indeed described by an ellipticcurve with a Kähler parameter t inherited from the P 2 t one; but with a large complex structureμ; see also footnote 7. The limit μ →∞explains the deg<strong>en</strong>eracy property in the baseΔ 1 (A.3).


340 L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341Refer<strong>en</strong>ces[1] M. Marino, Chern–Simons theory and topological strings, Rev. Mod. Phys. 77 (2005) 675–720, hep-th/0406005.[2] R. Dijkgraaf, E. Verlinde, H. Verlinde, Notes on topological string theory and two-dim<strong>en</strong>sional topological gravity,in: String Theory and Quantum Gravity, World Sci<strong>en</strong>tific, 1991, p. 91.[3] M. Bershadsky, S. Cecotti, H. Ooguri, C. Vafa, Kodaira–Sp<strong>en</strong>cer theory of gravity and exact results for quantumstring amplitudes, Commun. Math. Phys. 165 (1994) 311, hep-th/9309140.[4] G.L. Cardoso, B. de Wit, J. Kappeli, T. Mohaupt, Stationary BPS solutions in N = 2 supergravity with R 2 -interactions, JHEP 0012 (2000) 019, hep-th/0009234.[5] A. Ceresole, R. D’Auria, S. Ferrara, The symplectic structure of N = 2 supergravity and its c<strong>en</strong>tral ext<strong>en</strong>sion, Nucl.Phys. B (Proc. Suppl.) 46 (1996) 67–74, hep-th/9509160.[6] T. Graber, E. Zaslow, Op<strong>en</strong> string Gromov–Witt<strong>en</strong> invariants: Calculations and a mirror ‘theorem’, hep-th/0109075.[7] M. Aganagic, A. Klemm, C. Vafa, Disk instantons, mirror symmetry and the <strong>du</strong>ality web, hep-th/0105045.[8] J. Bryan, R. Pandharipande, Curves in Calabi–Yau and topological quantum field theory, Duke Math. 126 (2005)369–396;J. Bryan, R. Pandharipande, On the rigidity of stable maps to Calabi–Yau threefolds, Geom. Topol. Monogr. 8(2006) 97–104.[9] J. Bryan, R. Pandharipande, The local Gromov–Witt<strong>en</strong> theory of curves, math.AG/0411037.[10] D. Karp, C. Liu, M. Marino, The local Gromov–Witt<strong>en</strong> invariants of configuration of rational curves, Geom. Topol.Monogr. 10 (2006) 115–168.[11] H. Ooguri, A. Strominger, C. Vafa, Black hole attractors and the topological string, Phys. Rev. D 70 (2004) 106007,hep-th/0405146.[12] C. Vafa, Two dim<strong>en</strong>sional Yang–Mills, black holes and topological strings, hep-th/0406058.[13] A. Dabholkar, Exact counting of black hole microstates, Phys. Rev. Lett. 94 (2005) 241–301, hep-th/0409148.[14] H. Ooguri, C. Vafa, E. Verlinde, Hartle–Hawking wave-function for flux compactifications, Lett. Math. Phys. 74(2005) 311–342, hep-th/0502211.[15] R. Dijkgraaf, R. Gopakumar, H. Ooguri, C. Vafa, Baby universes in string theory, Phys. Rev. D 73 (2006) 066002,hep-th/0504221.[16] A. Belhaj, L.B. Drissi, E.H. Saidi, A. Segui, N = 2 supersymmetric black attractors in six and sev<strong>en</strong> dim<strong>en</strong>sions,Nucl. Phys. B 796 (2008) 521–580, arXiv: 0709.0398.[17] E.H. Saidi, M.B. Sedra, Topological string in harmonic space and correlation functions in S 3 stringy cosmology,Nucl. Phys. B 748 (2006) 380–457, hep-th/0604204.[18] M. Aganagic, A. Neitzke, C. Vafa, BPS microstates and the op<strong>en</strong> topological string wave function, hep-th/0504054.[19] E. Witt<strong>en</strong>, Topological sigma models, Commun. Math. Phys. 118 (1988) 411.[20] M. Bershadsky, S. Cecotti, H. Ooguri, C. Vafa, Holomorphic anomalies in topological field theories, Nucl. Phys.B 405 (1993) 279–304.[21] H. Ooguri, C. Vafa, Knot invariants and topological strings, Nucl. Phys. B 577 (2000) 419, hep-th/9912123.[22] J. Walcher, Ext<strong>en</strong>ded holomorphic anomaly and loop amplitudes in op<strong>en</strong> topological, string, arXiv: 0705.4098[hep-th].[23] S. Yamaguchi, S.T. Yau, Topological string partition functions as polynomials, JHEP 0407 (2004) 047, hep-th/0406078.[24] M. Alim, J.D. Lange, Polynomial structure of the (op<strong>en</strong>) topological string partition function, JHEP 0710 (2007)045, arXiv: 0708.2886 [hep-th].[25] M. Aganagic, R. Dijkgraaf, A. Klemm, M. Marino, C. Vafa, Topological strings and integrable hierarchies, Commun.Math. Phys. 261 (2006) 451, hep-th/0312085.[26] V. Bouchard, B. Florea, M. Marino, Topological op<strong>en</strong> string amplitudes on ori<strong>en</strong>tifolds, JHEP 0502 (2005) 002,hep-th/0411227.[27] M. Aganagic, C. Vafa, Mirror symmetry, D-branes and counting holomorphic discs, hep-th/0012041.[28] M. Aganagic, A. Klemm, M. Marino, C. Vafa, The topological vertex, Commun. Math. Phys. 254 (2005) 425–478,hep-th/0305132.[29] A. Iqbal, A.-K. Kashani-Poor, The vertex on a strip, hep-th/0410174;A. Iqbal, A.-K. Kashani-Poor, Instanton counting and Chern–Simons theory, Adv. Theor. Math. Phys. 7 (2004)457–497, hep-th/0212279.[30] A. Iqbal, C. Kozcaz, C. Vafa, The refined topological vertex, hep-th/0701156.[31] P. Sulkowski, Crystal model for the closed topological vertex geometry, JHEP 0612 (2006) 030, hep-th/0606055.[32] A. Iqbal, C. Kozcaz, C. Vafa, The refined topological vertex, hep-th/0701156.


L.B. Drissi et al. / Nuclear Physics B 804 [PM] (2008) 307–341 341[33] M. Taki, Refined topological vertex and instanton counting, hep-th/0710.1776.[34] N.A. Nekrasov, Seiberg–Witt<strong>en</strong> prepot<strong>en</strong>tial from instanton counting, Adv. Theor. Math. Phys. 7 (2004) 831, hep-th/0206161.[35] N. Nekrasov, A. Okounkov, Seiberg–Witt<strong>en</strong> theory and random partitions, hep-th/0306238.[36] H. Nakajima, K. Yoshioka, Instanton counting on blowup. I. 4-dim<strong>en</strong>sional pure gauge theory, Inv<strong>en</strong>t. Math. 162 (2)(2005) 313, math.AG/0306198.[37] A. Braverman, P. Etingof, Instanton counting via affine Lie algebras II: From Whittaker vectors to the Seiberg–Witt<strong>en</strong> prepot<strong>en</strong>tial, math.AG/0409441.[38] A. Iqbal, A.K. Kashani-Poor, Instanton counting and Chern–Simons theory, Adv. Theor. Math. Phys. 7 (2004) 457,hep-th/0212279;A. Iqbal, A.K. Kashani-Poor, SU(N) geometries and topological string amplitudes, Adv. Theor. Math. Phys. 10(2006) 1, hep-th/0306032.[39] T. Eguchi, H. Kanno, Topological strings and Nekrasov’s formulas, JHEP 0312 (2003) 006, hep-th/0310235.[40] T.J. Hollowood, A. Iqbal, C. Vafa, Matrix models, geometric <strong>en</strong>gineering and elliptic g<strong>en</strong>era, hep-th/0310272.[41] C. Vafa, Two-dim<strong>en</strong>sional Yang–Mills, black holes and topological strings, hep-th/0406058.[42] N. Caporaso, M. Cirafici, L. Griguolo, S. Pasquetti, D. Seminara, R.J. Szabo, Topological strings, two-dim<strong>en</strong>sionalYang–Mills theory and Chern–Simons theory on torus bundles, hep-th/0609129.[43] R. Ahl Laamara, A. Belhaj, L.B. Drissi, E.H. Saidi, Black holes in type IIA string on Calabi–Yau threefolds withaffine ADE geometries and q-deformed 2d quiver gauge theories, Nucl. Phys. B 776 (2007) 287–326, hep-th/0611289.[44] S. Katz, P. Mayr, C. Vafa, Mirror symmetry and exact solution of 4D N = 2 gauge theories I, Adv. Theor. Math.Phys. 1 (1998) 53–114, hep-th/9706110.[45] M. Ait B<strong>en</strong> Haddou, A. Belhaj, E.H. Saidi, Geometric <strong>en</strong>gineering of N = 2CFT 4 s based on indefinite singularities:Hyperbolic case, Nucl. Phys. B 674 (2003) 593–614, hep-th/0307244.[46] R. Ahl Laamara, M. Ait B<strong>en</strong> Haddou, A. Belhaj, L.B. Drissi, E.H. Saidi, RG cascades in hyperbolic quiver gaugetheories, Nucl. Phys. B 702 (2004) 163–188, hep-th/0405222.[47] E. Witt<strong>en</strong>, Phases of N = 2 theories in two dim<strong>en</strong>sions, Nucl. Phys. B 403 (1993) 159–222, hep-th/9301042.[48] N.C. Leung, C. Vafa, Branes and toric geometry, Adv. Theor. Math. Phys. 2 (1998) 91–118, hep-th/9711013.[49] L.B. Drissi, H. Jehjouh, E.H. Saidi, Topological string on local elliptic curve with large complex structure, Afr. J.Math. Phys. 6 (2008) 83–91.[50] L.B. Drissi, H. Jehjouh, E.H. Saidi, G<strong>en</strong>eralized MacMahon G(q) as q-deformed CFT correlation function, Nucl.Phys. B 801 (2008) 316, arXiv: 0801.2661.[51] S. Katz, P. Mayr, C. Vafa, Mirror symmetry and exact solution of 4D N = 2 gauge theories I, Adv. Theor. Math.Phys. 1 (1998) 53–114, hep-th/9706110.[52] A. Belhaj, A. Elfallah, E.H. Saidi, On the non-simply laced mirror geometries in type II strings, Class. QuantumGrav. 17 (2000) 515–532;A. Belhaj, A. Elfallah, E.H. Saidi, On the affine D(4) mirror geometry, Class. Quantum Grav. 16 (1999) 3297–3306.


Chapitre 6<strong>Vertex</strong> <strong>Topologique</strong> RaffinéGrâce au vertex topologique C µνλ (q), nous avons pu calculé explicitem<strong>en</strong>t les amplitudesde la corde topologique de type A sur les variétés de CY3 torique. Dans cette partie,nous continuons l’étude <strong>en</strong>treprise dans le chapitre précéd<strong>en</strong>t <strong>en</strong> insérant les résultats baséssur le vertex topologique usuel dans un cadre plus général. Il s’agit de généraliser le formalismeprécéd<strong>en</strong>t dans le but d’obt<strong>en</strong>ir des invariants topologiques au delà des invariantsGromov-Witt<strong>en</strong> et Donaldson-Thomas [174].Dans ce chapitre, nous prés<strong>en</strong>tons le formalisme <strong>du</strong> vertex topologique raffiné dont le 3-vertex topologique raffiné C λµν (t, q) dép<strong>en</strong>d de deux paramètres q = e −iɛ 1, et t = e iɛ 2. Ceformalisme qui a été intro<strong>du</strong>it dans [177, 178, 179, 180] et qui a pour origine les travaux deIqbal et al [151, 159], a une intérprétation <strong>en</strong> terme de l’instanton de Nekrasov [55, 56, 177].Signalons égalem<strong>en</strong>t qu’<strong>en</strong> utilisant le formalisme <strong>du</strong> vertex topologique raffiné, on obti<strong>en</strong>tla fonction de partition raffinée de la corde topologique sur les variétés de CY3 toriques.Dans la section 1, Nous intro<strong>du</strong>isons le vertex topologique raffiné. La section 2 est consacréeaux exemples des fonctions de partitions des variétés de Calabi-Yau toriques et nousexhibons dans la section 3 l’invariance de la fonction de partition raffinée des cordes topologiquessous la transition de flop. Dans la section 4, on développe le li<strong>en</strong> <strong>en</strong>tre le vertextopologique raffiné et les invariants topologiques.6.1 Formalisme <strong>du</strong> vertex raffinéNotre objectif consiste à trouver l’expression <strong>du</strong> vertex topologique raffiné qui provi<strong>en</strong>t<strong>du</strong> calcul microscopique de la solution de Seiberg-Witt<strong>en</strong> [57, 79]. De ce fait, le vertexraffiné C λµν (t, q) a la même interprétation combinatoire <strong>en</strong> termes de partitions 3D210


<strong>Vertex</strong> topologique raffinéνq tλ µFig. 3-1 – <strong>Vertex</strong> topologique raffiné C λµν (t, q)Rappelons que les tranches diagonales d’une partition 3D sont des diagrammes de Young2D qui s’<strong>en</strong>chaîn<strong>en</strong>t les unes aux autres. Ces partitions 2D apparaiss<strong>en</strong>t dans les plansx − y = a avec a ∈ Z. Dans le cas <strong>du</strong> vertex usuel, la ième tranche est pondérée par q |π| où|π| est le nombre de boîtes.Fig. 3-2 – les tranches diagonales 2D d’une partition plane.∏q |πa| = q P a∈Z |πa| = q≠a∈Zde boites dans πAlors que dans le cas <strong>du</strong> vertex raffiné, la partition 3D est représ<strong>en</strong>tée de différ<strong>en</strong>tesmanières. Pour le cas où a < 0, les tranches sont représ<strong>en</strong>tées par le paramètre t et poura ≥ 0, celles-ci sont paramétrisées par le paramètre q∏t ∏ |πa| q |πa| = t P a≺0 |πa| q P a≥0 |πa|a≺0 a≥0211


5.1 Raffinem<strong>en</strong>t <strong>du</strong> vertex topologiqueLa fonction génératrice des partitions 3D est une généralisation de la fonction de MacMahondonnée par :M (q, t) = ∑ πq P ∞i=1 |π(−i)| t P ∞j=1 |π(j−1)| =∞∏ (1 − q i−1 t j) −1.i,j=1L’intro<strong>du</strong>ction des paramètres q et t dans les tranches est représ<strong>en</strong>tée de la manière suivante.Si nous comm<strong>en</strong>çons par a > 0, le mouvem<strong>en</strong>t de la tranche est davantage ori<strong>en</strong>tévers la gauche <strong>en</strong> passant par l’origine. En fait, chaque fois que nous déplaçons la tranchevers la gauche, il faut compter avec le paramètre t. Or si la tranche se déplace vers le haut(qui se pro<strong>du</strong>it quand nous allons de a = i à a = i − 1, i = 0, 1, 2), nous comptons avec leparamètre q. Après avoir pris <strong>en</strong> considération le framing, le vertex raffiné est donné par[91] :C λµν (t, q) = G λµν(t,q)avecM(t,q)= ( ) ‖µ‖ 2 +‖ν‖ 2q 2t k(µ)2 Ptν t (t −ρ ; q, t) ∑ ηG λµν (t, q) =( qt∑η) ‖µ‖ 2 +‖ν‖ 22t k(µ)( qt) |η|+|λ|−|µ|2s λ t /η (t −ρ q −ν ) s µ/η(t−ν t q −ρ) .(3.1)2 M (t, q) P ν t (t −ρ ; q, t) ×( q) |η|+|λ|−|µ|(2st λ t /η (t −ρ q −ν ) s µ/η t−ν t q−ρ) (3.2)et P ν t (t −ρ ; q, t) est la fonction de Macdonald donnée par l’expression suivante [175] :P ν t (t −ρ ; q, t) = (t) ‖ν‖22 ˜Zν (t, q)∏ (= 1 − t a(i,j)+1 q l(i,j)) −1, avec a (i, j) = νtj − i, l (i, j) = ν i − j.(i,j)∈ν(3.3)6.1.1 <strong>Vertex</strong> raffiné et fonction de partition de la corde ouverteLe vertex raffiné a égalem<strong>en</strong>t une interprétation <strong>en</strong> termes des amplitudes topologiquesdes cordes ouvertes généralisées <strong>en</strong> prés<strong>en</strong>ce des piles des A-branes. En fait, les résultatsde [176] suggèr<strong>en</strong>t que raffinem<strong>en</strong>t des amplitudes de la théorie des cordes ouvertes sontliées aux invariants de noeud de Khovanov.En utilisant le formalisme <strong>du</strong> vertex raffiné, la fonction de partition de la corde ouvertedép<strong>en</strong>d de l’arête sur laquelle la pile des branes est mise. Cela est <strong>du</strong> à l’abs<strong>en</strong>ce des212


<strong>Vertex</strong> topologique raffinépropriétés de symétrie cyclique dans le vertex raffiné. En effet, nous avons trois choix(3.4)correspondants aux trois arêtes de C 3 . Les résultats sont les suivants :I : C λ∅∅ (t, q) = ( ) |λ|q 2st λ t (t −ρ ) ,II : C ∅µ∅ (t, q) = ( )‖ µt ‖ 2 −|µ|q 2stµ t (q −ρ )III : C ∅∅ν (t, q) =‖ν‖ 2q 2Qs∈ν (1−t1+a(s) q l(s) ) .Dans le cas I : La fonction de partition de la corde ouverte est comme suit :Z (t, q, x) = ∑ C λ∅∅ (t −1 , q −1 ) s λ (x) = ∑ ( ) |λ|t 2sq λ t (t ρ ) s λ (x)λλ∏= ∞ ( √ ) ∏t1 + xq t−i+ 1 2 = ∞ ( )√ (3.5)1 − Qt −i+ 1 t2 , Q = −x . qi=1Dans le cas I : La fonction de partition de la corde ouverte est donnée par :i=1Z (t, q, x) = ∑ C ∅µ∅ (t −1 , q −1 ) s µ (x) = ∑µµ∏= ∞ ( )∏1 + xq −i+ 1 2 = ∞i=1i=1(tq)‖µ t ‖−|µ|2(1 − Qq −i+ 1 2s µ t (q ρ ) s µ (x)), Q = −x.(3.6)Dans les deux résultats ci-dessus, la fonction de partition est celle obt<strong>en</strong>ue à partir <strong>du</strong>vertex ordinaire sauf qu’ici elle dép<strong>en</strong>d de t ou q selon l’arête sur laquelle la brane est finie.Un cas intéressant est celui <strong>du</strong> troisième exemple dans lequel la brane finit sur l’axe préféré.Dans ce cas, l’amplitude de la corde ouverte <strong>en</strong> utilisant le vertex raffiné est donnée par :Z (t, q, V ) = ∑ λ= ∑ λC ∅∅ν (t −1 , q −1 ) T r ν VC ∅∅ν (t −1 , q −1 ) s ν (x)(3.7)où x = {x 1 , x 2 , · · · } et le vertex raffiné s’écrit sous la forme suivante :C ∅∅ν (t, q) =Q‖ν‖ 2q 2s∈ν ( 1−t 1+a(s) q l(s) )‖ νt ‖ 2C ∅∅ν (t −1 , q −1 ) = (−1)|ν| t 2 ( q) t |ν|/2Q( 1−t 1+a(s) q l(s) )s∈νPar conséqu<strong>en</strong>t, pour le cas x = {−Q, 0, 0, · · · } , le résultat est la suivante :Z (t, q, Q) = ∑ ∑C ∅∅ν (t −1 , q −1 ) s ν (x) = ∞ C ∅∅k (t −1 , q −1 ) (−Q) kνk=0∑= ∞ ( ) k ∏ kQ √ tk(1 − tq n−1 ) −1 .k=0n=1(3.8)(3.9)213


5.2 Fonctions de partitions <strong>du</strong> vertex raffinésC’est exactem<strong>en</strong>t la fonction de partition <strong>en</strong> utilisant la série de Hilbert raffinée de pro<strong>du</strong>itsymétrique C [151].6.2 Fonctions de partitions raffinéesDans cette section, nous explicitons l’expression de la fonction de partition de la variétéde Calabi-Yau locale <strong>en</strong> utilisant le vertex topologique raffiné :6.2.1 O(−1) ⊕ O(−1) ↦→ P 1La théorie des cordes de type-IIA compactifiée sur une variété de Calabi-Yau donne lieuà une théorie de jauge N = 2 sur la direction transversale de C 2 . En utilisant le formalismede vertex topologique, la fonction de partition de la corde topologique est donnée par :Z(q, Q) = ∑ ν= ∑ νQ |ν| (−1) |ν| C ∅ ∅ ν (q) C ∅ ∅ ν t(q)Q |ν| (−1) |ν| s ν t(q −ρ )s ν (q −ρ ) (3.10)=∞∏ ( ) 1 − Q qi+j−1=i,j=1k=1∞∏ (1 − q k Q ) koù T = ln(Q) est à la fois le paramètre de Kahler et la taille de P 1 .Par la suite, on considère le vertex topologique raffiné pour déterminer la fonction departition. Dans ce cadre, le diagramme torique de X et le collage des vertex raffinés sontprés<strong>en</strong>tés dans la Fig. 3-3. On obti<strong>en</strong>t alors l’expression explicite de la fonction de partitionFig. 3-3 – Diagramme torique O(−1) ⊕ O(−1) ↦→ P 1 . Les vertex sont collés autour de leurdirection préférée.raffinée de la corde topologique de type AZ(t, q, Q) := ∑ νQ |ν| (−1) |ν| C ∅ ∅ ν (t, q) C ∅ ∅ ν t(q, t). (3.11)214


<strong>Vertex</strong> topologique raffinéAvecC ∅ ∅ ν (t, q) = q ||ν||22 ˜Zν (t, q) = q ||ν||22∏(1 − t a(s)+1 q l(s) ) −1s∈νIl existe une autre formule <strong>du</strong>e à la combinaison de deux vertex selon leur arête x (directionnon préférée), où le paramètre sur chaque arête de collage doit être différ<strong>en</strong>t l’un de l’autre.Cela a permis, <strong>en</strong>tre autre d’écrire la fonction de partition raffinée dont le diagrammetorique correspondant est schématisé dans la figure (3-4)Fig. 3-4 – Diagramme torique O(−1) ⊕ O(−1) → P 1 . Les vertex sont collés autour leursdirections impréférées.Dans ce cas la fonction de partition est sous la forme :Z(t, q, Q) = ∑ λ Q|λ| (−1) |λ| C λ ∅ ∅ (t, q) C λ t ∅ ∅(q, t)= ∑ ( ) |λ| ( ) |λ|λ (−Q)|λ| q 2st λ t(t −ρ t 2) sq λ (q −ρ )= ∑ λ (−Q)|λ| s λ t(t −ρ ) s λ t(q −ρ )= ∏ ∞i,j=1 (1 − Q qi− 1 2 t j− 1 2 ).(3.12)6.2.2 O(0) ⊕ O(−2) → P 1Nous avons déjà vu deux manières de combiner deux vertex mais il <strong>en</strong> existe beaucoupd’autres. Par exemple, dans le cas de P 1 ×P 1 locale, on peut avoir deux géométries obt<strong>en</strong>ues<strong>en</strong> pr<strong>en</strong>ant la taille d’un des P 1 très grande. Cette limite donne deux copies de O(0) ⊕O(−2) → P 1On s’intéresse premièrem<strong>en</strong>t au cas de vertex topologique usuel. Le formalisme de la215


5.2 Fonctions de partitions <strong>du</strong> vertex raffinésFig. 3-5 – Deux choix possibles de deux directions préférées (a) la ligne interne et b) leslignes externes parallèlesfonction de partition est donné par :Z(q, Q) = ∑ νQ |ν| (−1) |ν| C ∅ ∅ ν (q) (−1) |ν| q κ(ν)2 C ∅ ∅ ν t(q)= ∑ νQ |ν| s ν t(q −ρ ) q κ(ν)2 s ν (q −ρ ) = ∑ νQ |ν| s ν t(q −ρ ) s ν t(q −ρ ) (3.13)=∞∏ ( ) ∏ ∞1 − Q qi+j−1 −1 ( )= 1 − Q qk −k.i,j=1k=1Dans le cas de vertex raffiné, nous avons deux choix concernant la direction préférée commeil a été figuré dans (3-5) : pour le cas (a) la fonction est donnée par :Z (t, q, Q) = ∑ ν= ∑ ν= ∑ νQ |ν| (−1) |ν| C ∅ ∅ ν (t, q)f ν (t, q) C ∅ ∅ ν t(t, q)(−Q) |ν| ˜Zν (t, q) ˜Z ν t(q, t) q ‖ν‖22 t ‖νt ‖ 22 f ν (t, q)(Q √ qt ) |ν| t ‖ ν t ‖ 2Qs∈ν (1−ta(s)+1 q l(s) )(1−t a(s) q l(s)+1 )216(3.14)


<strong>Vertex</strong> topologique raffinéavec T r λ V 1 = s ν (x) où x = (x 1 , x 2 , · · · ) sont les valeurs propres de la matrice d’holonomieV . La fonction Zλ,µ I (q, Q) est donnée par la formule suivante :Zλµ I (q1, Q) = ∑ (−Q) |ν| C λµν (q)C ∅∅ν t(q)ν∏= s λ t(q −ρ )s µ t(q −ρ−λ , Qq ρ ) ∞ (3.20)(1 − Qq i+j−1−λt j )On obti<strong>en</strong>t ainsi la fonction de partition de la corde ouverte normalisée <strong>en</strong> divisant par lafonction de partition de la corde ferméei,j=1˜Z λµ(q, I Q) := ZI λµ (q, Q)Z∅∅ I (q, Q) = s λ t(q−ρ )s µ t(q −ρ−λ , Qq ρ ) ∏(i,j)∈λ(1 − Qq j−i ) (3.21)Par la suite, on va suivre le même calcul de la fonction de partition après la transition deflop que celle décrite ci-dessus. A partir <strong>du</strong> diagramme torique <strong>du</strong> conifold résolu (voir lafigure ci-dessus),Fig. 3-8 – Deux différ<strong>en</strong>tes résolutions de conifold sont liées les unes aux autres par latransition flop.la fonction de partition après la transition de flop est donnée par :Z IIλµ(q, ˆQ) = ∑ ν(− ˆQ) |ν| C ∅µν (q)C λ∅ν t(q) (3.22)<strong>en</strong> termes de fonction de schurZλµ(q, II ˆQ)∑= q κ(µ)2 (− ˆQ) |ν| s ν (q −ρ )s ν t(q −ρ )s λ t(q −ρ−νt )s µ (q −ρ−νt ) (3.23)ν219


5.4 <strong>Vertex</strong> raffiné et homologie des <strong>en</strong>trelacsAinsi, nous remarquons que les deux fonctions de partition ne sont égales que si les deuxparamètres de Kahler avant et après la transition flop sont liés par l’équation :Q = ˆQ −1 (3.24)En effet, la fonction de partition de la corde ouverte <strong>en</strong> utilisant le vertex topologique usuelest invariant sous la transition de flop. La même procé<strong>du</strong>re sera appliquée cette fois aucas de vertex topologique raffiné. La fonction de partition correspondante au diagrammetorique <strong>du</strong> conifold résolu <strong>en</strong> appliquant les règles de collage est exprimée sous cette forme :Z (a) (Q, t, q) = ∑ α (−Q)|α| C α1 αβ 1(t, q)C β2 α t α 2(q, t)= ∑ α,τ,σ (−Q)|α|( ) 1q 2 (‖α‖2 +‖β 1 ‖ 2 ) 1tt(× s α t1/τ(t −ρ q −β 1 )sα/τ (q −ρ t −βt 1 ) ×(× P α t2(q −ρ ; t, q)tq2 κα P β t1(t −ρ ; q, t) ( q) 12 (|τ|+|α 1|−|α|)t) 1t 2 (||αt || 2 +||α 2 || 2 ) (3.25)qq− 1 2 κα) 12 (|σ|+|β 2 |−|α|) s β t2 /σ(q −ρ t −α 2)s α t /σ(t −ρ q −αt 2 ).Un calcul analogue à celui qui a donné la fonction de partition avant la transition de flop,donne la fonction de partition après la transition correspondante (voir la figure 3-8b)Z (b) (Q, t, q) = ∑ α (−Q)|α| C α,β2 ,β 1(t, q)C α t ,α 1 ,α 2(q, t)= ∑ α,τ,σ (−Q)|α|( ) 1q 2 (||β 2 ||2 +||β 1 || 2 ) 1tt(× s α t /τ(t −ρ q −β 1 )sβ2 /τ(q −ρ t −βt 1 ) ×(× P α t2(q −ρ ; t, q)tq) 12 κ β 2 P β t1(t −ρ ; q, t) ( qtq) 12 (||α 1|| 2 +||α 2 || 2 )q12 κα 1t) 12 (|τ|+|α|−|β 2 |)2 (|σ|+|α|−|α 1|)sα/σ (q −ρ t −α 2)s α1 /σ(t −ρ q −αt 2 ).(3.26)Le formalisme <strong>du</strong> vertex topologique que nous avons étudi donne le même résultat pour lesdeux géométries (avec branes) liées par la transition de flop. Cette ”symétrie”, cep<strong>en</strong>dant,n’est pas conservée dans le formalisme <strong>du</strong> vertex topologique raffiné. La section suivantemontre que les fonctions de partition que nous avons étudié, ont un li<strong>en</strong> très étroit avec lesinvariants des cordes topologiques.6.4 <strong>Vertex</strong> raffiné et homologie des <strong>en</strong>trelacsL’une des propriétés les plus intéressantes <strong>du</strong> vertex topologique raffiné est la relationdes invariants d’homologie d’<strong>en</strong>trelacs de Hopf [127] [176] avec les fonctions des partitionsraffinées [159]. Dans cette section, nous déterminons les fonctions de partition des théoriesdes cordes topologiques ouvertes raffinées d’une configuration de deux branes sur le conifold220


<strong>Vertex</strong> topologique raffinérésolu et leurs relations avec les invariants polynomials d’<strong>en</strong>trelacs de Hopf. Le point dedépart sera le vertex topologique raffin ayant l’expression suivante :C λµν (t, q) = ( ) ‖µ‖ 2 −|µ|q 2t k(µ)2 (t) ‖ν‖22 ˜Zν (q, t) ∑ tη( qt) |η|+|λ|−|µ|2s λ t /η (t −ρ q −ν ) s µ/η(t−ν t q −ρ)(3.27)L’expression de la fonction de partition raffinée <strong>en</strong> utilisant la figure (3-8), est donnée sousla forme suivante :Z λµ (q, t, Q) = ∑ ν(−Q) |ν| C ∅µν (q, t) C λ∅ν t (q, t) . (3.28)avecC ∅µν (q, t) = ( ) ‖µ‖ 2q 2t k(µ)2 (t) ‖ν‖2 ( )2 ˜Zν (q, t)stµ t −ρ q −νtC λ∅ν t (q, t) = ( ) |λ|q 2(q) ‖ ν t ‖ 2( )2 ˜Zν t(q, t)stλ t t −ρ q −νt(3.29)En utilisant les deux expressions de l’éq(3.29) , la fonction de partition <strong>en</strong> termes desfonctions de Schur s’écrit de la manière suivante :Z λµ (q, t, Q) = h λµ (q, t) ∑ ν(−Q) |ν| (q) ‖ ν t ‖ 22(t) ‖ν‖22 ˜Zν (q, t) ˜Z ν t(q, t)s λ t(t −ρ q −νt )sµ(t −ρ q −νt )h λµ (q, t) = ( ) ‖µ‖ 2 −|λ|q 2t k(µ)2tLa fonction de partition normalisée est donnée par la formule suivante :(3.30)Ẑ λµ (q, t, Q) = Z λµ(q,t,Q)Z ∅∅ (q,t,Q)(3.31)avecZ ∅∅ (q, t, Q) = ∑ (−Q) |ν| (q) ‖ ν t ‖ 22(t) ‖ν‖22 ˜Zν (q, t) ˜Z ν t(q, t)ν∏= ∞ () (3.32)1 − Qq i− 1 2 t j− 1 2i,j=1Nous évoquons par la suite les invariants d’<strong>en</strong>trelacs de Hopf et nous prés<strong>en</strong>tons la relationavec la fonction de partition de corde ouverte <strong>en</strong> utilisant le vertex topologique raffiné.En donnant un <strong>en</strong>trelacs coloré par une collection de représ<strong>en</strong>tations R 1 · · · , R l de sl(N),l’invariant polynomial correspondant est donné par :¯P sl(N)R 1···R k(q) (3.33)221


5.4 <strong>Vertex</strong> raffiné et homologie des <strong>en</strong>trelacsPour chaque R a , a = 1, · · · , l qui est la représ<strong>en</strong>tation fondam<strong>en</strong>tale de sl(N), nous avonsl’expression suivante 1¯PN (q) =sl(N) ¯P□···□(q) (3.34)Les invariants polynomials (3.34) sont liés aux valeurs moy<strong>en</strong>nes des opérateurs de bouclede Wilson W (L) = W R1 ,···R l(L) dans la théorie de Chern-Simons. Par exemple, l’invariantsl(N) polynômial P N (q) est associé avec les valeurs moy<strong>en</strong>nes de boucle de Wilson W (L) =W □···□ (L),¯P N (L) = q −2Nlk(L) 〈W (L)〉 (3.35)où lk(L) = ∑ a


<strong>Vertex</strong> topologique raffinéraffinée [159] :¯P ( λµ (Q 1 , Q 2 , a) = (−1) |λ|+|µ| qt[ ∑=ν) |λ|+|λ||µ|fλ (q, t)(Q −1 ( qt) )1|ν|+|µ|22Ẑ λµ (q, t, Q) (3.39)(−Q) |ν| (t) ‖ν‖22(q) ‖ ν t ‖ 22 ˜Zν (q, t) ˜Z ν t(q, t)s λ(t −ρ q −νt) s µ(t −ρ q −νt)]( √ ) q |λ|+|µ|[Z ∅∅ (q, t, Q)] −1 Q −1 2 ( q) |λ||µ|× (−1) |λ|+|µ| .ttC’est l’un des principaux résultats dans l’étude <strong>du</strong> vertex topologique raffiné. La relation<strong>en</strong>tre la théorie des noeuds paramétrisée par (Q 1 , Q 2 , a) et le vertex avec les paramètres(q, t, Q) est donnée par :√ q = Q1 (3.40)√t = −Q1 Q 2Q = −Q 2 a −2où a = Q N 1 . Dans le cas Q 2 = −1 et q = t, on obti<strong>en</strong>t le vertex topologique usuel.6.4.1 UnknotOn calcule le polynôme de Poincaré (3.34) d’homologie des représ<strong>en</strong>tations (λ, µ) =(□, ∅), qui est exactem<strong>en</strong>t le super-polynôme unknot suivant :P □,∅ (Q 1 , Q 2 , a) = -a= -a(= aPν (−Q)|ν| t ‖ν‖22∞Q “1−Qq i− 1 2 t j− 1 2(i,j=1 √t √− Q √t t1−t q 1−t1Q 1 −Q −11−”q ‖ ν t ‖ 22 ˜Zν (q, t) ˜Z ν t(q, t)s □(t −ρ q −νt ))a−2Q 1 −Q −11)= a−a−1Q 1 −Q −11avec a = Q N 1 .,,(3.41)La fonction de partition dénouée (unknot) dép<strong>en</strong>d de deux paramètres Q 1 et Q 2 .223


5.4 <strong>Vertex</strong> raffiné et homologie des <strong>en</strong>trelacs6.4.2 Entrelac de HopfConsidérons maint<strong>en</strong>ant l’<strong>en</strong>trelacs de Hopf coloré par la représ<strong>en</strong>tation suivante :(R 1 , R 2 ) = (□, □). Dans ce cas, <strong>en</strong> utilisant les équations (3.40) et (3.39), nous obt<strong>en</strong>onsP □,□ (Q 1 , Q 2 , a) = Q −1 √ qtqt∑ν (−Q)|ν| t ||ν||22 t ||νt || 22 ˜Zν (q, t) ˜Z ν t(t, q) s □ (t −ρ q −νt ) 2∏ ∞i,j=1 (1 − Q qi− 1 2 t j− 1 2 )= a 2 q(1 − t) − Q t 12 1 + q − t + qt+ Q 2 t 2 q (1 − t) 2 q1 − t + tq(1 − t) 2= a 2 q 1 + q − t + qt −4 1 − t + qt− a−2 + a (3.42)(1 − t)2(1 − t) 2 (1 − t) 2= a −2 1 − Q2 1 + Q 4 1 Q 2 2(1 − Q 2 1) 2 − a 2 1 + Q2 1 Q 2 2 − Q 2 1 + Q 4 1 Q 2 2(1 − Q 2 1) 2 + a 4 Q 2 1 Q 2 2(1 − Q 2 1) 2Ce résultat obt<strong>en</strong>u est <strong>en</strong> accord avec le super-polynôme d’<strong>en</strong>trelacs de Hopf calculé dans[159]. Le succès <strong>du</strong> vertex topologique raffiné a permis d’ouvrir une nouvelle directiondans l’étude des invariants d’<strong>en</strong>trelacs de Hopf. La fonction de partition de la corde a étéreliée aux invariants sl(N) d’<strong>en</strong>trelac de Hopf coloré par d’autres représ<strong>en</strong>tations sl(N).Cep<strong>en</strong>dant, même si ces résultats sont <strong>en</strong>core bi<strong>en</strong> loin d’avoir des interprétations dans lathéorie des noeuds. Il s’agit donc d’un nouveau terrain <strong>en</strong>core à explorer.224


H. Jehjouh6.5 Contribution : Refining the Shifted 3-<strong>Vertex</strong>J.Math.Phys.50 :013509,2009, arXiv : 0812.0513RésuméDans ce travail, notre objectif principal consistait à l’étude des propriétés raffinées et «shiftées » <strong>du</strong> vertex topologique usuel C λµν . Pour ce fait, nous avons étalé différ<strong>en</strong>tespropriétés de base sur le vertex topologique standard C λµν et sa version raffinée R λµν dansle cadre de la corde topologique. Nous avons aussi consacré une partie <strong>du</strong> papier à lafonction de MacMahon « shiftée » S 3 utilisée dans l’hiérarchie BKP. Nous avons développéles aspects des propriétés de la fonction de MacMahon à 3 dim<strong>en</strong>sions raffinée. Ceci fait,nous avons dérivé les expressions explicites <strong>du</strong> vertex topologique « shifté » Sˆλˆµˆν(q) <strong>en</strong>prés<strong>en</strong>ce des conditions de bords données par les partitions 2d ainsi que de sa versionraffinée Tˆλˆµˆν(q, t).225


226H. Jehjouh


JOURNAL OF MATHEMATICAL PHYSICS 50, 013509 2009Refining the shifted topological vertexL. B. Drissi, 1,2,a H. Jehjouh, 1,2,b and E. H. Saidi 1,2,c1 Faculté des Sci<strong>en</strong>ces, Laboratory/UFR-Physique des Hautes Energies,Rabat, 1014 Morocco2 Groupem<strong>en</strong>t National de Physique des Hautes Energies (GNPHE), Siège focal:FS,Rabat, 1014 MoroccoReceived 29 April 2008; accepted 6 November 2008; published online 7 January 2009We study aspects of the refining and shifting properties of the 3d MacMahonfunction C 3 q used in topological string theory and BKP hierarchy. We derive theexplicit expressions of the shifted topological vertex S q and its refined versionT q,t. These vertices complete results in literature. © 2009 American Instituteof Physics. DOI: 10.1063/1.3040186I. INTRODUCTIONIn the past few years, there has be<strong>en</strong> some interest in the study of the topological vertexformalism of toric Calabi–Yau threefold CY3. 1 This interest has followed the basic result, accordingto which the topological vertex C is a powerful tool to compute toric CY3 topologicalstring amplitudes. 1–4,23 It also came from the remarkable relation betwe<strong>en</strong> C and the Gromov–Witt<strong>en</strong> invariants of g<strong>en</strong>us g-curves in toric CY3s. 5–7Rec<strong>en</strong>tly it has be<strong>en</strong> shown that the standard topological three-vertex C q may have twokinds of g<strong>en</strong>eralizations: one known as the refining of C q and the other as its shifting.In the first case, the refined topological vertex R q,t is a two parameter function computingthe refined topological string amplitudes of toric CY3s. 8–10 It has be<strong>en</strong> found also thatR q,t computes as well the Nekrasov’s 11 instantons of the topological string free <strong>en</strong>ergyFX I , 1 , 2 of four dim<strong>en</strong>sional SUN gauge theories. 12 In Nekrasov’s ext<strong>en</strong>sion, the usualtopological string coupling constant g top =ln1/q gets replaced by the pair of parameters 1 and 2 . 13 In the second case, the standard MacMahon function C 3 q Refs. 3 and 14 has be<strong>en</strong> ext<strong>en</strong>dedto the so-called shifted partition function S 3 q. This is the g<strong>en</strong>erating function of the shifted planepartitions and it is used in the study of BKP hierarchy. 15–18The aim of this paper is to contribute to this matter by combining both the refining and theshifting operations to get the refined-shifted topological vertex T q,t ext<strong>en</strong>ding R q,t andS 3 q obtained rec<strong>en</strong>tly in literature. More precisely, we want to complete the missing relationspres<strong>en</strong>ted in the two following tables:iFirst, we determine the refined version of the shifted MacMahon function S 3 q obtainedby Foda and Wheeler. The refined version of S 3 q, d<strong>en</strong>oted below as T 3 q,t, is missing.It is a two parameter function g<strong>en</strong>erating shifted 3d-partitions needed to complete the table,C 3 q = known: Eq. 2.4refining→shifting ↓ ↓ shiftingS 3 q = known: Eq. 2.9refining→R 3 q,t = known: Eq. 2.8T 3 q,t = ... ?. 1.1a Electronic mail: drissilb@gmail.com.b Electronic mail: jehjouh@gmail.com.c Electronic mail: h-saidi@fsr.ac.ma.0022-2488/2009/501/013509/12/$25.0050, 013509-1© 2009 American Institute of PhysicsDownloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-2 Drissi, Jehjouh, and Saidi J. Math. Phys. 50, 013509 2009iiSecond, we ext<strong>en</strong>d the g<strong>en</strong>eralized MacMahon functions S 3 q and T 3 q,t by implem<strong>en</strong>tingboundary conditions captured by the strict notice that refers to a 2d-partition and ˆ toa strict 2d-partition. The hat is sometimes dropped out for simplicity of notations 2dpartitions ˆ , ˆ , and ˆ. The resulting S ˆ ˆ ˆq and T ˆ ˆ ˆq,t are also needed to complete thefollowing table:C q = known Eq. 2.1refining→shifting ↓ ↓ shiftingS ˆ ˆ ˆq = ... ?refining→R q,t = known: Eq. 2.6T ˆ ˆ ˆq,t = ... ?. 1.2The organization of this paper is as follows: In Sec. II, we give g<strong>en</strong>eralities on topologicalvertices. In particular, we review briefly the expression of the constructions of the standardtopological vertex C q: the refined one R q,t and the shifted MacMahon functionS 3 q. In Sec. III, we derive the explicit expression of the shifted topological vertex S ˆ ˆ ˆqof Eq. 1.2. In Sec. IV, we compute the its refined version T ˆ ˆ ˆq,t. In Sec. V we give aconclusion, and in Sec. VI we collect some useful tools as an App<strong>en</strong>dix.II. TOPOLOGICAL VERTEX: A REVIEWIn this section, we review briefly some basic tools; in particular, the explicit expressions of thethree following topological vertices:1 The standard topological vertex d<strong>en</strong>oted as C q.23The refined topological vertex R q,t. This is a two parameter g<strong>en</strong>eralization of C q.The standard 3d-MacMahon function C 3 q, its refined version R 3 q,t as well as the shifted3d-MacMahon function S 3 q obtained in Ref. 15.These objects have interpretations in a the topological string A-model in which q=e −g s withg s being the topological string coupling constant. b The statistical mechanical models in whichthe parameter q describes the Boltzmann weight q=e −1/KT with T being absolute temperature. 14,19A. <strong>Vertex</strong> C „q…Following Ref. 3, the standard topological vertex C q, with boundary conditions in thex i ,x j orthogonal planes of Z 3 lattice giv<strong>en</strong> by the 2d partitions ,,, can be defined in thetransfer matrix method as follows:C q = f t q L 0 + q − q 0 Lt=0−1t=− − q −t q L 0,where is a g<strong>en</strong>eric 2d partition state a Young diagram, t its transpose and L 0 = with being the number of boxes of the Young diagram. The operators x are vertex operators ofthe c=1 bosonic CFT 2 , whose explicit expressions can be found in Ref. 14, and f is a realnumber giv<strong>en</strong> byf = q1/22 + 2 + 2 q n+nn=01−q n n.2.12.2For the particular case where there is no boundary condition, i.e., ===, the vertex C qcoincides exactly with the 3d-MacMahon function C 3 ,Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-3 Refining the shifted topological vertex J. Math. Phys. 50, 013509 2009C 3 q = q L 0 + 1q 0 Lt=0−1t=−The function C 3 , which reads explicitly as, see also Eq. 1.1,C 3 q = n=11 nn1−q − 1q L 0 .2.32.4has several interpretations. It is the amplitude of the A-model topological closed string on C 3 ; i.e.,C 3 =C . It is also the g<strong>en</strong>erating function of 3d partitions,C 3 q =3d-partitions Likewise, the vertex C q inherits the interpretations of C 3 q with a slight g<strong>en</strong>eralization andmore power since it allows gluing 1 to topological amplitudes of all non compact toric CY3s.It is the partition function of A-model topological string of C 3 with op<strong>en</strong> strings on boundariesand, up on using the gluing method, 1 it allows to compute the partition function of toric CY3s.C q has also a combinatorial interpretation in terms of the g<strong>en</strong>erating function of the planepartitions with the boundary conditions ,, where , , and are 2d-partitions.The explicit expression of C can be exhibited in differ<strong>en</strong>t, but equival<strong>en</strong>t forms. Its expressionin terms of the pro<strong>du</strong>ct of three Schur functions can be found in Refs. 3 and 8.q .2.51. Ext<strong>en</strong>sionsTwo kinds of g<strong>en</strong>eralizations of the topological vertex C have be<strong>en</strong> considered in literature.These are as follows:i The refined vertex R q,t having a connection with Nekrasov’s 11 partition function ofSUN gauge theories and with the link invariants. 13ii The shifted MacMahon function S 3 q used in BKP hierarchy. 15 The S q ext<strong>en</strong>sion ofthe shifted vertex by implem<strong>en</strong>ting boundary conditions was not computed before; it is aresult of the pres<strong>en</strong>t paper.Let us give some details on R and S 3 ; th<strong>en</strong> we turn back to the computation of S q.B. Refined vertex: R „q,t…The refined topological vertex R q,t is a two parameter ext<strong>en</strong>sion of C q. As notedbefore, it has a topological string interpretation in connection with Nekrasov’s instantons. Itsexplicit expression has be<strong>en</strong> first derived by Iqbal et al. and can be expressed in differ<strong>en</strong>t butequival<strong>en</strong>t ways. It is giv<strong>en</strong>, in the transfer matrix method, byR q,t =r t t L 0 + q − it 0 Lt=0where , q L 0, and x are as before, and−1t=− − q − j t q L 0,2.6r = q1/22 + 2 + 2 1−q k−1 t l .2.7q nt t n k,l=1For the particular case ===, the vertex R q coincides exactly with the refined3d-MacMahon function m<strong>en</strong>tioned in Sec. I 1.1,Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-4 Drissi, Jehjouh, and Saidi J. Math. Phys. 50, 013509 2009R 3 q,t = 1−q k−1 t l .k,l=12.8By setting t=q back in 2.8, we get the standard C 3 q relation. The explicit expression of therefined R q,t in terms of Schur functions can be found in Ref. 8.C. Shifted S 3 „q…The shifted 3d MacMahon S 3 q is the g<strong>en</strong>erating functional of strict plane partitions. Theexplicit expression of S 3 q has be<strong>en</strong> derived by Foda and Wheeler by using transfer matrixmethod. It reads asS 3 q = n=1 n1+qn n, 2.91−qand has an interpretation in the BKP hierarchy of the so-called neutral free fermions. It wasclaimed in Ref. 15 that S 3 q could be relevant to the topological string <strong>du</strong>al to ON Chern–Simon theory in the limit N→.III. SHIFTED TOPOLOGICAL VERTEXThe expression Eq. 2.9 of the shifted topological vertex has be<strong>en</strong> derived in the abs<strong>en</strong>ce ofany kind of boundary conditions. Here, we want to complete this result by considering the derivationof the shifted topological vertex S with g<strong>en</strong>eric boundary conditions with the property,S = S q, S = S 3 q. 3.1Notice that S q g<strong>en</strong>erates the shifted 3d partitions with boundary conditions giv<strong>en</strong> by the strict2d partitions ,, along the axis x 1 ,x 2 ,x 3 . For the definitions of the shifted 3d and strict 2dpartitions, see the App<strong>en</strong>dix.The main result of this section is collected in the following proposition where some terminologyhas be<strong>en</strong> borrowed from: 8Proposition 1: The perp<strong>en</strong>dicular shifted topological vertex S q with g<strong>en</strong>eric boundaryconditions, giv<strong>en</strong> by three strict 2d-partitions ,,, reads as follows:L q = f S 1−qn1+q nn, 3.2which can be also put in the normalized formn=1In relation 3.2, the numerical factor f isS = S L ,L q =1.3.3f =2 l+l+l q 1/22 + 2 + 2 ,f q =1,where 2 = i i 2 and l is the l<strong>en</strong>gth of the strict partition that is the number parts of the strict2d partition = 1 ,..., l ,0,....The function S q is the perp<strong>en</strong>dicular partition function of shifted 3d-partitions. It reads interms of Schur functions P t / and Q / as follows:3.4Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-5 Refining the shifted topological vertex J. Math. Phys. 50, 013509 2009withS = P t /q −− Q / q −t− Z h ,strict 2d = 1 , ..., i , ... ,,3.5 k = 1 2 − k,3.6as well asandn = 1 2 t 2 − ,h =2 −l−l q −nt −n q −/2−/2 ,q −− = q − 1 − 1, ...,q − i − i,...,3.7Z = 2−l P tq − q +/2+nt n=11+q n1−q nn .The function P x is the Schur function associated with the strict 2d partition ; it is defined as3.8 − x =strict 2d partition P / x,where / is the complem<strong>en</strong>t of in . We also have the orthogonality relation3.9Q x =2 −l P x,Q x,P x = ,3.10where = i i i.To establish this result, we consider shifted 3d-partitions 3 inside of a cube with sizeN 1 N 2 N 3 and boundary conditions giv<strong>en</strong> by the strict 2d-partitions ,,. More precisely, thestrict 2d-partition belongs to the plane x 2 ,x 3 of the ambi<strong>en</strong>t real three-dim<strong>en</strong>sional space, belongs to the plane x 3 ,x 1 and to the plane x 1 ,x 2 ,Th<strong>en</strong> proceed by steps as follows:Step 1: Compute the perp<strong>en</strong>dicular partition function S q by using the transfer matrixapproach. 3 This method has be<strong>en</strong> used for calculating the topological vertex C of the A-modeltopological string on C 3 which lead to Eqs. 2.1–2.4. S reads in terms of pro<strong>du</strong>cts x asfollows:S = q−n−nt q N 2 +N 1 t N 1 N 2q L 0 + q − jq 0L − q − i t q L 0.j=1i=13.11By using the relation q −kL 0 zq kL 0= zq k and q L 0=2 l q , the function S can bebrought towithS = t j=1N 1 + q − j − jN 2 − q − i t − i ,i=13.12Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-6 Drissi, Jehjouh, and Saidi J. Math. Phys. 50, 013509 2009The vertex operators z are giv<strong>en</strong> by = q−/2+n−/2+nt 2 l+l . 3.13 + z = exp mNodd2m zm m, − z = exp mNodd2m zm −m,3.14with m being operators satisfying the following commutation relations: m , n =− m 2 n+m,0, m, n Z odd . 3.15Notice that in the particular limit N 1 , N 2 , and N 3 →, the vertex S gets id<strong>en</strong>tified withS 3 q =shifted 3d 2 p q , 3.16whose explicit expression is precisely 2.9.Step 2: Commuting the vertex operators + z i to the right of − z j by using the commutationrelations + x − y = 1+xy1−xy − y + x, x y = y x,3.17we getS = strict 2d t N 1 N 2 − q − t j− j + q − i − i,j=1i=13.18with =Z q−/2+n−/2+nt 2 l+l . 3.19This relation can be simplified further by using the Schur functions P t /x j = t − x j andQ t /x i = t + x j following from the id<strong>en</strong>tities t N 2 j=1 − x j = P / x 1 ,...,x N , t N 1 j=1 = x j =Q / x 1 ,...,x N , as well asDownloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-7 Refining the shifted topological vertex J. Math. Phys. 50, 013509 2009 x j =strict 2d L / x j ,N 1 x j = L / x 1 ,x 2 ,...,j=1strict 2d with x=x 1 ,x 2 ,... and where we have set L −/ = P / and L + / =Q / .Th<strong>en</strong>, the partition function S becomesS q = Z q−/2+n−/2+nt 2 l+l P t /q −t− Q / q −− .strict 2d 3.203.21To determine the factor Z, we first use the id<strong>en</strong>tity S =Z, th<strong>en</strong> the cyclic propertyS =S =S which implies in turns that S =S ; from which we learn the followingresult:Z = q−/2−nt 1+qn2 l 1−q nnP tq − . 3.22n=1Step 3: The shifted MacMahon function S 3 q can be recovered from the above analysis byusing the id<strong>en</strong>tity S 3 q=S =Z.This <strong>en</strong>ds the proof of Eq. 3.2. Notice that L q can be also put in the formL q = q k/2 P tq − strict 2d P t /q −− Q / q −t − ,where k=2 2 − t 2 is the Casimir associated to strict 2d partition.3.23IV. REFINING THE SHIFTED VERTEXIn this section, we derive the explicit expression of the refining version T q,t of theshifted topological vertex S q. This is a two parameters q and t with boundary conditionsgiv<strong>en</strong> by the strict 2d partitions , and .Notice that like for R of Eq. 2.6, the function the refined-shifted topological vertex T is noncyclic with respect to the permutations of the strict 2d partitions ,,; T T T . It obeys, however, the propertiesT q,q = S q, T q,t = T 3 q,t. 4.1Proposition 2: The explicit expression of the refined-shifted topological vertex T reads asfollows:K t,q =f T 1+qj−1 t i1−q j−1 t i,4.2where the factor f =f q is the same as in Eq. 3.4.T =T q,t is the refining version of S q of Eq. 3.5; it is the perp<strong>en</strong>dicular partitionfunction g<strong>en</strong>erating the strict plane partitions. Its explicit expression reads in terms of the skewSchur functions P t / and Q / as follows:i,j=1T =h˜q,t Z t,qstrict 2d partitions q,tP t /q − t − Q / t −t q − ,4.3where h˜ is the refinem<strong>en</strong>t of h 3.7 and it is giv<strong>en</strong> byDownloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-8 Drissi, Jehjouh, and Saidi J. Math. Phys. 50, 013509 2009We also haveh˜ q,t =2 −l−l q −nt −/2 t −n−/2 .Z t,q = q t 2 /22 −l P tt − i,j=1 1+qj−1 t i1−q j−1 t i.as well as q,t=q/t /2 . Notice that for q=t=1, q,t=1.To establish this result, we use the following steps.Step 1: Compute the refined expression T 3 q,t of the shifted 3d MacMahon’s function S 3 qin terms of the two parameters q and t. To that purpose, we start from the defining relation ofT 3 q,t by using strict 2d-partitions,T 3 q,t =strict 2d partition 4.44.52 p q a=1 −at a=1a−1, 4.6where we have used the diagonal slicing of shifted 3d-partitions in terms of the strict 2d-onesa as shown below = a,aZa = i,i+a .iNotice that the slices with a0 are weighted by the factor q a while the slices with a0 areweighted by t a .Th<strong>en</strong>, we use the transfer matrix method which allows to express T 3 q,t as the amplitudeT q,t; that is,T q,t = 0 t L 0 + 1t 0 La=0−a=−1 − 1q L 00.By using q −kL 0 zq kL 0= zq k , we can also put T in the formT q,t = 0 + t i − q j−1 0.i0j0Next commuting the − ’s to the left of the + ’s by help of relations 3.17, we obtain 1+qj−1 t i1−q j−1 t i, T 3 q,t = T q,t = j=1i=14.74.84.94.10which re<strong>du</strong>ces to S 3 q Eq. 3.16 by setting t=q.Step 2: To get the expression of the perp<strong>en</strong>dicular partition function T for arbitrary boundaryconditions, we mimic the approach in Ref. 1 and factorize T as follows:T q,t = g t,q T diag ,where T diag stands for the diagonal partition function and g t,q giv<strong>en</strong> by4.11g t,q = q −nt t −n4.12describing the change from diagonal slicing to perp<strong>en</strong>dicular one. To compute T diag , we use thetransfer matrix method. We first haveT diag = t t L 0 + q i 0− tL − t − t jq 0 L .i0j0By using q −kL 0 zq kL 0= zq k , we can bring it to the form4.13Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-9 Refining the shifted topological vertex J. Math. Phys. 50, 013509 2009T diag = t + t i q i− − q j−1 t − j ti0j0Th<strong>en</strong> using q L 0=2 l q and Eq. 3.17, we <strong>en</strong>d with.4.14withdiag = Z T t − q − t − + t −t q − ,j0i04.15 q,t =2 −l−l q −/2 t −/2 , q,t = q t /2 .4.16Using Eq. 3.20 and the skew Schur functions P t / and Q / , the partition function 4.11 readsaswhereT =h˜ Z stricttq /2P t /q − t − Q / t −t q − ,h˜ q,t = q,t g t,q,=2 −l−l q −nt −/2 t −n−/2 .4.174.18To determine the factor Z, we need two data: first we use the id<strong>en</strong>tity T =Z q,t andsecond, we require that Z = =T 3 q,t, asinEq.4.1. We findZ q,t = q t 2 /22 −l P tt − 1+qj−1 t i1−q j−1 t i.4.19j,i=1This <strong>en</strong>ds the proof of Eq. 4.2.Notice that K q,t can be also put in the closed formK = q t 2 + 2 /2t k/2 P tt − tq +−/2P t /x , Q / y , ,4.20with x , =q − t − , y , =t −t q − and the property T =T K as well as the normalizationK =1.V. CONCLUSIONIn this paper we have studied the refining and the shifting properties of the standard topologicalvertex C . After having reviewed some basic properties on the following:1 The standard vertex C q and its refined version R used in the framework of topologicalstrings.2 The shifted MacMahon function S 3 q used in BKP hierarchy,we have completed the missing relations in Eqs. 1.1 and 1.2. In particular, we havederived the explicit expressions of the following:aThe shifted topological vertex S ˆ ˆ ˆq with boundary conditions giv<strong>en</strong> by g<strong>en</strong>eric strict 2dpartitions. The shifted MacMahon function S 3 q, giv<strong>en</strong> by Eq. 2.9 and first obtained inRef. 15, follows by putting ˆ =ˆ =ˆ =.Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-10 Drissi, Jehjouh, and Saidi J. Math. Phys. 50, 013509 2009FIG. 1. Shifted Young diagram of =4,2,1.bThe topological vertex T ˆ ˆ ˆq,t describing the refined version shifted topological vertexS ˆ ˆ ˆq. Putting ˆ =ˆ =ˆ =, wegetT 3 q,t = j=1 □i=1 i1+qj−1 t i1−q j−1 t5.1describing the refined version of Foda–Wheeler relation recovered by setting t=q.In the <strong>en</strong>d, notice that it would be interesting to seek whether T ˆ ˆ ˆq,t could be associatedwith some gauge theory instantons as does R q,t with the Nekrasov’s ones.ACKNOWLEDGMENTSThis research work was supported by Protars III D12/25.APPENDIX: STRICT PARTITION AND SCHUR FUNCTIONIn this App<strong>en</strong>dix, we give some useful tools on the strict 2d-partitiions, the shifted planepartitions and on Schur functions.1. Strict 2d- and shifted 3d-partitionA2d-partition, or a Young diagram, d<strong>en</strong>oted as = 1 , 2 ,..., r ,... is a sequ<strong>en</strong>ce of decreasingnon-negative integers 1 2¯ r ¯.A strict 2d-partition is a sequ<strong>en</strong>ce of strictly decreasing integers 1 2 ¯. The sum of theparts i of the 2d-partition is the weight of d<strong>en</strong>oted by = 1 + 2 + ¯ + r + ¯ .A2d strict partition is said a partition of n if =n and is repres<strong>en</strong>ted by its shifted Youngdiagram obtained from the usual diagram by shifting to the right the ith row by i−1 squares asshown on Fig. 1.The shifted Young diagram is giv<strong>en</strong> by a collection of boxes with coordinatesA1i, ji =1, ...,l, i j i + i −1. A2A shifted plane partition of shape is determined by the sequ<strong>en</strong>ce ...,. −1 , 0 , 1 ,..., where 0 the 2d-partition on the main diagonal and k is the 2d-partition on the diagonal shifted by aninteger k. All diagonal partitions are strict 2d-partitions forming altogether a shifted plane partitionwith the propertyDownloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-11 Refining the shifted topological vertex J. Math. Phys. 50, 013509 2009FIG. 2. A strict plane partition.For illustration, see the example Fig. 2... −1 0 1 ... A3 −2 = 3, −1 = 4,3, 0 = 5,3, 1 = 3,2, 2 = 2, 3 = 1.2. Property of Schur function for strict partitionThe shifted topological vertex is defined by using skew Schur P and Q functions. 20–22 Theseare symmetric functions that appear in topological amplitudes and are defined by a sequ<strong>en</strong>ce ofpolynomials P x 1 ,x 2 ,...,x n , nN, with the property= P / x 1 , ...,x n Tx T , 0, otherwise, A4where the sum is over all shifted Young tableaux of shape /. The skew Schur function Q / isrelated to P / as in Eqs. 3.9 and 3.10. We also havestrict Q xP y = i,j1+x iy j1−x i y j.The relation betwe<strong>en</strong> the Schur function P for strict partition that we have used here above andthe usual Schur functions S ˜ for the double partition ˜ is giv<strong>en</strong> byA5S ˜t =2 −l P 2t 2,A6where t/2 is t 1 /2,t 3 /2,t 5 /2,... and P t/2= P t 1 /2,t 3 /2,t 5 /2,.... Notice that the doublepartition ˜ in Frob<strong>en</strong>uis notation reads in terms of the strict partition =n 1 ,n 2 ,...,n k as˜ = n1 ,n 2 , ...,n k n 1 −1,n 2 −1, ...,n k−1 −1.A71 M. Aganagic, A. Klemm, M. Marino, and C. Vafa, Commun. Math. Phys. 254, 425 20052 A. Iqbal and A.-K. Kashani-Poor, Adv. Theor. Math. Phys. 7, 457 2004.3 A. Okounkov, N. Reshetikhin, and C. Vafa, Prog. Math., 244, 597 2006.4 L. B. Drissi, H. Jehjouh, and E. H. Saidi, Nucl. Phys. B to be published.5 T. Graber and E. Zaslow, e-print arXiv:hep-th/0109075.6 M. Aganagic, A. Klemm, and C. Vafa, Z. Naturforsch., A: Phys. Sci. 57, 12002.7 D. Karp, C. Liu, and M. Marino, Geom. Topol. 10, 1152006.8 A. Iqbal, C. Kozcaz, and C. Vafa, e-print arXiv:hep-th/0701156.Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


013509-12 Drissi, Jehjouh, and Saidi J. Math. Phys. 50, 013509 20099 M. Taki, e-print arXiv:hep-th/0710.1776.10 P. Ginsparg, e-print arXiv:hep-th/9108028.11 N. A. Nekrasov, Adv. Theor. Math. Phys. 7, 8312004.12 N. Nekrasov and A. Okounkov, e-print arXiv:hep-th/0306238.13 S. Gukov, A. Iqbal, and C. Kozçaz, e-print arXiv:hep-th/07051508.14 L. B. Drissi, H. Jehjouh, and E. H. Saidi, Nucl. Phys. B 801, 316 2008.15 O. Foda and M. Wheeler, e-print arXiv:math-ph/0612018.16 E. Ramos and S. Stanciu, Nucl. Phys. B 427, 3381994.17 E. H. Saidi and M. B. Sedra, J. Math. Phys. 35, 3190 1994.18 E. H. Saidi, M. B. Sedra, and J. Zerouaoui, Class. Quantum Grav. 12, 1567 1995.19 H. N. Temperley, Proc. R. Soc. London, Ser. A 199, 361 1949.20 A. Y. Orlov, e-print arXiv:math-ph/0302011.21 M. Vuleti, e-print arXiv:math.Co/0707.0532.22 M. Vuleti, e-print arXiv:math-ph/0702068.23 J. J. Heckman and C. Vafa, e-print arXiv:hep-th/0610005.Downloaded 24 Mar 2009 to 140.105.16.64. Redistribution subject to AIP lic<strong>en</strong>se or copyright; see http://jmp.aip.org/jmp/copyright.jsp


Chapitre 7Conclusion et PerspectivesLe travail prés<strong>en</strong>té dans ce mémoire de thèse est ess<strong>en</strong>tiellem<strong>en</strong>t articulé autour <strong>du</strong>calcul des fonctions des partitions Z top et des amplitudes A top des modèles des cordestopologiques sur une classe de variétés de CY3 à savoir les variétés CY toriques X 3 . Généralem<strong>en</strong>tparlant, ces amplitudes topologiques apparaiss<strong>en</strong>t sous forme de prépot<strong>en</strong>tiels F<strong>en</strong> théorie des supercordes type II compactifiées sur les variétés X 3 ; et d’ou l’importance<strong>du</strong> calcul des quantités. Rappelons aussi que les amplitudes A top des cordes topologiques,<strong>en</strong> particulier la fonction de partition Z top , ont été approchées de deux façons :(1) <strong>en</strong> utilisant le formalisme <strong>du</strong> 3- vertex topologique qui est considéré comme un outileffici<strong>en</strong>t pour étudier les invariants topologiques. Cette méthode, inspirée des diagrammesde Feynman de la théorie quantique des champs, permet de calculer de manière simple etrigoureuse les fonctions de partitions Z top et les amplitudes A top des cordes topologiques.(2) <strong>en</strong> exploitant la relation de <strong>du</strong>alité <strong>en</strong>tre la théorie de cordes topologiques type A surC 3 et le modèle statistique de fusion des cristaux à 3D basé sur la matrice de transfert.Cette représ<strong>en</strong>tation cristalline puise sa force dans les théories de champs conformes c = 1avec ses réalisations bosonique et fermionique ainsi que son interpretation <strong>en</strong> terme desdiagrammes et tableaux de Young à 2D et leurs généralisations à des espaces à 3D.La <strong>du</strong>alité cristal/3-vertex nous a permis de réaliser <strong>en</strong>tre autres les amplitudes de lathéorie de cordes topologiques de modèle A vivant dans une classe de variétés CY toriquesqui sont interprétés <strong>en</strong> terme de cristaux dont O (−3) → CP 2 et le conifold résoluO (−1) ⊕ O (−1) → CP 1 constitu<strong>en</strong>t deux exemples explicites parmi plusieurs d’autres.Par ailleurs et dans l’objectif de mieux expliciter ces techniques, nous avons jugé utiled’organiser ce mémoire de thèse <strong>en</strong> cinq chapitres <strong>en</strong> plus d’une intro<strong>du</strong>ction générale,cette conclusion et une annexe sur des outils mathématiques et qui est reportée à la fin238


Conclusion et perspectivesde ce docum<strong>en</strong>t. Dans ce qui suit, nous survolons rapidem<strong>en</strong>t quelques idées de base quiont été exposées au long de ce manuscrit tout <strong>en</strong> dressant un bilan général de nos travauxoriginaux et mettant <strong>en</strong> exergue quelques perspectives.Dans le but de faciliter la lecture de ce mémoire de thèse, nous avons initié ce manuscritpar une étude synthèse de quelques aspects fondam<strong>en</strong>taux des variétés de CY3 <strong>en</strong> mettantl’acc<strong>en</strong>t sur les variétés de CY3 toriques qui nous ont intéressé <strong>en</strong> premier lieu. Après avoirdécrit certains aspects des supercordes sur des CY à trois dim<strong>en</strong>sions, nous avons montré<strong>en</strong>tre autres comm<strong>en</strong>t ces diagrammes toriques permett<strong>en</strong>t le calcul des amplitudes de lathéorie des cordes topologiques.Avec ces outils <strong>en</strong> main, nous nous sommes p<strong>en</strong>chés sur l’étude de la théorie des cordestopologiques avec ses deux versions type A et type B et sur la <strong>du</strong>alité <strong>en</strong>tre cordes ouverteset cordes fermées. Par la suite, nous avons étudié les invariants de la théorie des cordestopologiques. Dans le contexte <strong>du</strong> modèle B, nous avons prés<strong>en</strong>té nos contributions danscette matière à savoir : ”De la théorie des cordes twistorielles à la Supergravité ConformeN =4, D=4 ” et ”Pure fermionic twistor like model & target space supersymmetry”. Dansle premier travail, nous avons étudié la théorie des cordes twistorielles. Nous sommes intéressésau modèle B dans la super-variété de Calabi-Yau CP 3|4 et sa relation avec la théoriede supergravité conforme N =4, D=4. Dans la deuxième contribution, nous avons dévelopéun modèle twistoriel purem<strong>en</strong>t fermionique. Nous avons égalem<strong>en</strong>t construit l’action deschamps twistoriels décrivant les champs fermioniques. Finalem<strong>en</strong>t, nous avons discuté leli<strong>en</strong> <strong>en</strong>tre le modèle twistoriel purem<strong>en</strong>t fermionique et l’espace cible R (d,4−d) à 4 dim<strong>en</strong>sions.Ensuite, nous nous sommes intéressés à l’étude <strong>du</strong> modèle statistique <strong>du</strong> cristal fon<strong>du</strong>. Lesconfigurations pour ce modèle sont des collections de cubes qui ont une interprétation <strong>en</strong>terme <strong>du</strong> 3- vertex topologique C ∅∅∅ représ<strong>en</strong>tant le diagramme torique de la variété deCalabi-Yau C 3 . Ce vertex a une remarquable interprétation <strong>en</strong> terme de diagramme deYoung à trois dim<strong>en</strong>sions. En utilisant la matrice de transfert T de la physique statistique,nous avons calculé les fonctions de partition <strong>du</strong> modèle cristallin avec des conditions debords. Nous avons montré <strong>en</strong>tre autres que ces amplitudes sont compatibles avec la fonctiongénératrice des invariants topologiques à l’instar des invariants de noeud de la théorie deChern-Simons sur la 3-sphère. Ces études ont été couronnées par la contribution ”G<strong>en</strong>eralizedMacMahon G(q) as q-deformed CFT Correlation Function” utilisant des techniquesde la matrice de transfert et des méthodes de la théorie conforme c = 1.Puis, nous avons porté une att<strong>en</strong>tion particulière au vertex topologique. Nous avons d’abord239


construit les amplitudes A top de la théorie des cordes topologiques type A <strong>en</strong> utilisant leformalisme de 3-vertex topologique. Après avoir rappelé certaines propriétés sur les 3- vertex,notamm<strong>en</strong>t <strong>en</strong> ce qui concerne le framing, leurs symétries et leurs règles de collage,nous avons étudié des exemples de calcul des amplitudes des cordes topologiques sur desvariétés de Calabi-Yau toriques. Parmi nos contributions dans cette matière, nous citonsles travaux ”Non Planar Topological 3-<strong>Vertex</strong> Formalism”, et ”Topological String on ToricCY3s in Large Complex Structure Limit”, où nous avons développé le formalisme de vertextopologique non planaire. Nous avons remarqué que le vertex topologique non planaireC (np) va au delà <strong>du</strong> 3- vertex usuel C µνλ de ”Aganagic, Klemm, Marino, et Vafa” faisantinterv<strong>en</strong>ir des diagrammes de Young à 2D. La différ<strong>en</strong>ce <strong>en</strong>tre les deux vertex réside ess<strong>en</strong>tiellem<strong>en</strong>tdans les deux faits suivants : (a) C (np) est associé avec les modèles sigma linéaireN = 2 à 2D avec superpot<strong>en</strong>tiel W ≠ 0 et (b) semble décrire une des configurationsparticulières d’une représ<strong>en</strong>tation plus générale faisant interv<strong>en</strong>ir des vertex topologiquestetra-val<strong>en</strong>t C ΣΥΓϜ où Σ, Υ, Γ, et Ϝ sont des diagrammes de Young généralisés.Motivés par la fonction de partition de Nekrasov de la théorie de jauge N = 2 SU(N), nousnous sommes égalem<strong>en</strong>t intéréssés au 3-vertex topologique raffiné Cµνλ R . Dans ce contexte,nous avons calculé la fonction de partition de la théorie des cordes topologiques ouvertes<strong>en</strong> utilisant le formalisme de vertex topologique raffiné. Nous avons donné des exemplesillustratifs et discuté le formalisme de vertex topologique utilisant de la ”flop transition”.Nous avons aussi étudié la relation <strong>en</strong>tre le vertex raffiné et les invariants topologiques dela théorie de Chern-Simons sur la sphère S 3 . Cette étude a été couronnée par la contribution”Refining the Shifted Topological <strong>Vertex</strong>” où nous avons dérivé l’expression explicite<strong>du</strong> vertex topologique shifté et sa version raffinée.A la fin de cette prés<strong>en</strong>tation, de nombreuses questions sans réponses émerg<strong>en</strong>t et dontles études pourrai<strong>en</strong>t faire l’objet de développem<strong>en</strong>ts ultérieurs. Grâce au formalisme <strong>du</strong>vertex topologique non planaire que nous avons développé dans [69, 70], il serait aussiutile de calculer la fonction de partition pour d’autre classes des variétés de Calabi-Yautoriques et dériver les invariants des noeuds. La notion de vertex topologique tetraval<strong>en</strong>tque nous avons intro<strong>du</strong>it et qui généralise le 3- vertex topologique mérite aussi davantaged’exploration. D’autres aspects que nous avons r<strong>en</strong>contré le long de ce mémoire de thès<strong>en</strong>écessit<strong>en</strong>t égalem<strong>en</strong>t des études approfondies ; c’est le cas par exemple de la fonctionde MacMahon M d (q) à d dim<strong>en</strong>sions. Il serait intéressant de dériver le raffinem<strong>en</strong>t de lafonction de MacMahon M d (q) <strong>en</strong> termes des argum<strong>en</strong>ts t et q ; c.à.d trouver l’expressionde l’ext<strong>en</strong>sion M d (q, t), et d’élaborer leurs interpretations à la fois <strong>en</strong> théorie de cordes240


Fonctions de Schur et MacMahontopologiques et <strong>en</strong> théorie des champs conformes utilisant la matrice de transfert. Tousces problèmes nécessit<strong>en</strong>t des études approfondies, nous espérons les examiner dans desoccasions futures.241


242


Chapitre 8Annexe : Fonctions de Schur etMacMahonLes diagrammes et les tableaux de Young à deux dim<strong>en</strong>sions constitu<strong>en</strong>t un des ingrédi<strong>en</strong>tsprincipaux dans ce mémoire de thèse. Ils jou<strong>en</strong>t un rôle crucial dans la dérivationde plusieurs résultats de la théorie des cordes topologiques et les théories conformes sousjac<strong>en</strong>tes. Ce sont des objets combinatoires qui ont été intro<strong>du</strong>its par Young au début <strong>du</strong>20 ème siècle pour l’étude des représ<strong>en</strong>tations irré<strong>du</strong>ctibles des groupes de symétries. Lesdiagrammes et les tableaux de Young à 2D sont représ<strong>en</strong>tés sous forme d’empilem<strong>en</strong>t deboites de longueur décroissante et ont été généralisés récemm<strong>en</strong>t pour des dim<strong>en</strong>sions supérieures; <strong>en</strong> particulier à trois dim<strong>en</strong>sions.Dans cette annexe, nous intro<strong>du</strong>isons dans un premier temps les diagrammes et tableauxde Young à 2D. Puis nous examinons les fonctions symétriques et les propriétés des polynômesde Schur s(X). Grâce à leur interprétation combinatoire <strong>en</strong> terme de tableaux deYoung, ces fonctions de Schur permett<strong>en</strong>t de déterminer <strong>en</strong>tre autres les fonctions génératricesde nombreuses familles de partitions planaires, <strong>en</strong> particulier fonction de MacMahonM 3d (q) qui sera considérée dans la dernière section de cette annexe.8.1 Diagrammes et tableaux de YoungCette section prés<strong>en</strong>te les notions de base pour les partitions d’<strong>en</strong>tiers ainsi que plusieursdéfinitions s’y rattachant [182, 183, 184]. En donnant les propriétés de l’<strong>en</strong>semble destableaux de Young, un <strong>en</strong>tier naturel n peut être décomposé de différ<strong>en</strong>tes façons <strong>en</strong> somme243


Diagrammes et tableaux de Youngd’<strong>en</strong>tiers naturels non nuls.8.1.1 Définitions et propriétésPr<strong>en</strong>ons un nombre <strong>en</strong>tier positif n fixé. Une partition de n est une suite d’<strong>en</strong>tiers nonnégatifs décroissante, tels que λ r ≥ λ r+1 pour tout r ≥ 1. Elle admet la notation classiqueλ = (λ 1 , λ 2 , ..., λ r )et représ<strong>en</strong>té par un <strong>en</strong>semble de points de coordonnées (i, j) avec 1 ≤ i ≤ n et 1 ≤ j ≤ λ i .On associe à chaque partition de poids n une représ<strong>en</strong>tation graphique donnée par undiagramme de Young, ou diagramme de Ferrer. La taille d’une partition est la somme des<strong>en</strong>tiers qu’on la note :|λ| = ∑ iλ iLes <strong>en</strong>tiers λ i , pour i ∈ {1, ..., r} sont les parts de λ. On dira que λ est paire lorsque toutesses parts le sont. La longueur d’une partition est sous la forme l (λ) = r = max{i ∈ N|λ i > 0}. Il est souv<strong>en</strong>t pratique de décrire une partition par ses ordres de multiplicité, onnotem i = m i (λ) = Card {j : λ i = i} .Ainsi la partition s’écrit aussi sous forme expon<strong>en</strong>tielle comme⎛⎞λ =⎝1, · · · 1, 2, · · · 2, · · · n, · · · n⎠} {{ } } {{ } } {{ }m 1 fois m 2 fois m n foisλ = (1 m 1, 2 m 2..., r mr )Dans ce cas, le poids de la partition associée est représ<strong>en</strong>té par|λ| = ∑ k≥1km kLa partition conjuguée de λ, notée λ ′ est celle dont le diagramme de Young est symétriqueau diagramme de λ par rapport à la diagonale principale ; avec :λ ′ i = Card {j : λ j ≥ i}On ne fait pas la distinction <strong>en</strong>tre une partition et son diagramme.Exemple1 : La suite d’<strong>en</strong>tiers (5, 4, 2, 2, 1, 1, 1) est une partition de 17 dont l’écriture expon<strong>en</strong>tielleest (1 3 2 2 3 0 4 1 5 1 ). Son diagramme de Young est repres<strong>en</strong>té dans la figure (3-1)244


Fonctions de Schur et MacMahonFig. 3-1 – Diagramme de Young à 2 dim<strong>en</strong>sionsIl est constitué d’un <strong>en</strong>semble de cases justifiées à gauche et <strong>en</strong> bas et le nombre de cases dechaque ligne correspond aux élém<strong>en</strong>ts de la partition associée. Etant donné un diagrammede Young λ et une case s = (i, j) de celui-ci, on définit alors les quantités suivantes :n(λ) = ∑ i≥1(i − 1) λ iAutrem<strong>en</strong>t dit, c’est la somme obt<strong>en</strong>ue :n(λ) = ∑ i≥1(λ′i2si on intro<strong>du</strong>it des 0 dans la première ligne <strong>du</strong> diagramme de Young associé à λ et des 1dans la seconde ligne, et ainsi de suite. De même pour le partage conjugué, on a :)n(λ ′ ) = ∑ s∈λ ′ l ′ (s) = ∑ s∈λ,a ′ (s)Par la suite, on donne quelques définitions, associées au diagramme de Young.On définit le nombre de cases de λ droite d’une case s qui est appelé le bras de s, notéa(s) tandis que le nombre de cases à sa gauche est appelé co-bras de s et il est noté a′(s).Le nombre de cases de λ au dessus d’une case s est appelé la jambe de s et il est notél(s) pour le cas <strong>du</strong> nombre de cases au dessous de s est appelé co-jambe de s et il est notél′(s). La longueur de bras a(s), la co-longueur de bras a′(s), la longueur de jambe l(s) etla co-longueur de jambe l′(s) sont données par les formules suivantes :a(s) = λ i − j, l(s) = λ j − ia′(s) = j − 1, l′(s) = i − 1Elles sont représ<strong>en</strong>tées dans la figure (3-2)245


Diagrammes et tableaux de YoungFig. 3-2 – Une case dans le diagramme de YoungFig. 3-3 – a) la longueur de bras b) la co-longueur de bras c) la longueur de jambe d) la colongueurde jambeCes valeurs a(s) = 1, a ′ (s) = 1, l(s) = 1, l ′ (s) = 2 peuv<strong>en</strong>t être visualisées de la façonsuivante :Le cont<strong>en</strong>u de s, noté c(s), correspond à la diagonale de la case s et il est donné parc(s) = j − i. La longueur d’équerre de s est notée commeh(i, j) = λ i + λ t j − j − i + 1où l’équerre d’un élém<strong>en</strong>t d’un diagramme de Young est l’<strong>en</strong>semble des cases à droiteFig. 3-4 – la longueur d’équerresur la ligne (bras), au-dessus dans la colonne (jambe). Etant donné une partition µ, la246


Fonctions de Schur et MacMahondim<strong>en</strong>sion quantique correspondante est donnée par :dim q R =∏(i,j)∈Rq κ(R)/4[h(i, j)]Une partition gauche est un couple de deux partitions λ, µ telle que le diagramme de µ soitcont<strong>en</strong>u dans celui de λ. On la note λ/µ, et la taille d’une partition gauche est donnée par :|λ/µ| = |λ| − |µ|. Par exemple, une partition gauche (8, 7, 6, 4, 1) / (4, 2, 1) est représ<strong>en</strong>téepar la figure(3-5).Fig. 3-5 – a) Diagramme de Young de la partition (8, 7, 6, 4, 1) . b) Diagramme gauche(8, 7, 6, 4, 1) / (4, 2, 1) .La frontière d’une partition λ est une suite de cases noires et blanches : la case noirecorrespond à un pas horizontal, tandis que la case blanche correspond à un pas vertical.Par exemple, la frontière de (6, 4, 4, 3, 1) est montrée dans la figure(3-6)==Fig. 3-6 –247


Diagrammes et tableaux de YoungTableaux de YoungUn tableau de Young T est un diagramme de Young dont les cases sont remplies par des<strong>en</strong>tiers compris <strong>en</strong>tre 1 et m. Une forme standard d’un tableau répond aux règles suivantes :dans une même ligne les nombres sont croissants et ils sont strictem<strong>en</strong>t croissants dans unecolonne. La partition est appelée la forme <strong>du</strong> tableau. Il existe deux types de tableaux :• Un tableau standard T est un tableau cont<strong>en</strong>ant les <strong>en</strong>tiers (1, 2, · · · , n) de sorte que lesnombres soi<strong>en</strong>t placés de façon strictem<strong>en</strong>t croissante de gauche à droite dans les lignes,et de bas <strong>en</strong> haut dans les colonnes où le nombre de tableaux standard de la forme λ estaussi la dim<strong>en</strong>sion d’une représ<strong>en</strong>tation irré<strong>du</strong>ctible et il est donné par :n!∏h(i, j)i,j(3.1)Exemple de tableau standard de forme λ = (5, 4, 3, 2)13 1410 11 126 7 8 91 2 3 4 5• Un tableau semi-standard est un tableau dans lequel les nombres sont placés de façoncroissante de gauche à droite dans les lignes et strictem<strong>en</strong>t croissante de bas <strong>en</strong> haut dansles colonnes. Les nombres de tableaux semi-standard d’une partition λ = (2, 1) s’écriv<strong>en</strong>tsous la forme :d (2,1) = ∏(i,j)∈λm + j − ih(i, j)= 3 + 1 − 1h(1, 1) = 3 × 3 + 2 − 1h(1, 2) = 1 × 3 + 1 − 2h(2, 1) = 1 = 8Exemple de tableau semi-standard de forme λ = (5, 4, 3, 2)7 84 5 62 2 3 51 1 2 3 4Il existe aussi un tableau semi standard de forme λ/µ qui est une application T de λ/µdans l’<strong>en</strong>semble des <strong>en</strong>tiers strictem<strong>en</strong>t positifs tels que les indices vont <strong>en</strong> décroissant248


Fonctions de Schur et MacMahonstrictem<strong>en</strong>t de bas <strong>en</strong> haut dans les colonnes et <strong>en</strong> décroissant au s<strong>en</strong>s large de gauche àdroite dans les lignes. C’est l’exemple <strong>du</strong> tableau semi standard de forme (5, 4, 3, 2) /(2, 1)7 84 5 62 3 52 3 48.1.2 Diagrammes de Young et groupes de symétriea) Relations <strong>en</strong>tre une partition d’ordre n et les classes des élém<strong>en</strong>ts degroupe de permutationLes états physiques d’un système selon la nature fermionique ou bosonique à N particulessont des états de base des représ<strong>en</strong>tations irré<strong>du</strong>ctibles complètem<strong>en</strong>t symétriquesou complètem<strong>en</strong>t antisymétriques de groupe des permutations S n . Les représ<strong>en</strong>tations irré<strong>du</strong>ctiblesde S n sont liées aux différ<strong>en</strong>tes écritures possibles de l’ordre n comme sommesd’<strong>en</strong>tiers décroissants et non-nuls, appelées ”partitions” et peuv<strong>en</strong>t être caractérisées parun diagramme de Young [191, 192, 193, 194]. Par exemple le nombre n = 3 admet 3 partitions: {3} {2, 1} et {1, 1, 1} et il y a ainsi exactem<strong>en</strong>t 3 représ<strong>en</strong>tations irré<strong>du</strong>ctibles <strong>du</strong>groupe symétrique d’ordre 3.Fig. 3-7 – Diagramme de Young shiftéLes repres<strong>en</strong>tations irré<strong>du</strong>ctibles de S prouv<strong>en</strong>t une utilité pour analyser les représ<strong>en</strong>tationsirré<strong>du</strong>ctibles des groupes classiques GL(n) et SU(n).b) Relations <strong>en</strong>tre les diagrammes de Young et les poids de SU(n)Chaque diagramme de Young avec n lignes représ<strong>en</strong>te une représ<strong>en</strong>tation irré<strong>du</strong>ctiblede groupe SU(n). Dans ce cas, il suffit de limiter le nombre des lignes <strong>du</strong> diagramme deYoung à k ≤ n − 1 pour obt<strong>en</strong>ir toutes les représ<strong>en</strong>tations irré<strong>du</strong>ctibles. Le i me poidsfondam<strong>en</strong>tal est représ<strong>en</strong>té par un diagramme de Young fait d’une colonne de hauteur i.249


Diagrammes et tableaux de YoungLa correspondance <strong>en</strong>tre le plus haut poids avec ses indices de Dynkin f i et le tableau Yavec ses lignes de longueur λ i est donnée par :∑n−1f i = (λ 1 , · · · , λ n−1 ) ↔ Y = (λ = λ j ).Autrem<strong>en</strong>t dit, λ k est le nombre de colonnes de Y de hauteur k.Par exemple : pour le groupe SU(3), il y a 5 types de repres<strong>en</strong>tations qui sont donnéescomme suit :j=28.1.3 Notation de Frob<strong>en</strong>iusOn retrouve aussi une autre notation d’une partition, appelée notation de Frob<strong>en</strong>ius[181]. On suppose premièrem<strong>en</strong>t une partition λ qui est un diagramme de Young de poidsn. On note d la longueur de la « diagonale » <strong>du</strong> diagramme, <strong>en</strong>suite on suppose que ladiagonale principale <strong>du</strong> diagramme de Young λ est constituée de r noeuds, avec α i est l<strong>en</strong>ombre des noeuds dans le iéme ligne de λ à droite de (i, i) (1 ≤ i ≤ r) par contre β i est l<strong>en</strong>ombre des noeuds dans le iéme colonne de λ au dessous de (i, i) pour (1 ≤ i ≤ r) . Si λ estun diagramme de Young, alors on peut, de même, lui associer une configuration de pointsdans R ∗ via les coordonnées modifiées de Frob<strong>en</strong>ius :α i = λ i − i,β i = λ ′ i − iIl existe un <strong>en</strong>tier unique r, à savoir la longueur de la diagonale, tel que :on note alors la partition λ comme suit :α 1 > · · · α r ≥ 0 et β 1 > · · · β r ≥ 0(α 1 , · · · , α r | β 1 , · · · , β r )oùα 1 > α 2 > · · · > α r ≥ 0 et β 1 > β 2 > · · · > β r ≥ 0Notons que la longueur de λ est le nombre d’élém<strong>en</strong>ts non nuls de la suite, noté par l(λ)<strong>en</strong> termes de coordonnées de Frob<strong>en</strong>ius et il est donné par deux formules :l(λ) = α 1 et |λ| = ∑ α i + ∑ β i + r250


Fonctions de Schur et MacMahonFig. 3-8 – Coordonnée de Frob<strong>en</strong>ius modifiée d’un diagramme de YoungFinalem<strong>en</strong>t, il est facile de voir que la conjuguée de (α | β) est (β | α) .8.1.4 Diagramme de MayaLe diagramme de Maya est une suite de cases noires et blanches, étiquetées par desnombres <strong>en</strong>tiers et arrangées sur une ligne horizontale. Les carrés sont tous noirs (blancs)et ils s’interprèt<strong>en</strong>t comme une bijection m : Z → Z, où m(j) j> 0Correspondance <strong>en</strong>tre les Diagrammes de Maya et YoungSoit M e l’<strong>en</strong>semble des diagrammes de Maya de charge e et Y l’<strong>en</strong>semble des diagrammesde Young telle qu’il existe une bijection <strong>en</strong>tre les deux diagrammesY → M eAlors, à chaque diagramme de Young, on peut associer un diagramme de Maya M = (M × ,M 0 ) où les cases blanches M × (noires M 0 ) sont les positions j ∈ Z des pas horizontales( pas verticals). Autrem<strong>en</strong>t dit, si on considère un diagramme de Young d’une partitionλ = (λ 1 , · · · , λ s ) où λ r ≥ · · · ≥ λ s > 0, alorsM 0 = ∪ s j=0 {k ∈ Z | λ j+1 − (j − 1) < k < λ j − j} (λ 0 = ∞, λ s+1 = 0).Inversem<strong>en</strong>t, si on a M 0 = {m(j) j≥1 } avec m 1 < m 2 < · · · , alors le diagramme de Youngassocié estλ = (1 r 1, 2 r 2, 3 r 3...) r j = m j+1 − m j − 1.251


Diagrammes et tableaux de YoungFig. 3-9 – a) diagramme de Young b) diagramme de Maya8.1.5 Partitions planesUne partition plane π est un réseau fini de points de R 3 , considéré comme une matriceπ = (π ij ) dont les <strong>en</strong>trées non nulles sont des nombres non négatifs (i, j) et vérifie π ij ≤ πi+1j et π ij ≤ π ij+1 pour i, j ≥ 1⎛⎞π 11 π 12 π 13 · · ·π 21 π 22 π 23 · · ·⎜ π⎝ 31 π 32 π 33 · · · ⎟⎠.. . . ..La représ<strong>en</strong>tation graphique des partitions planes est donnée par un empilem<strong>en</strong>t de cubesπ(x) sur chaque case dans le diagramme de Young x ∈ π. Elle s’interprète aussi comme undiagramme à trois dim<strong>en</strong>sions composé des points (i, j, k) ∈ π tel que :(i, j) ∈ λ 1 ≤ k ≤ π (i, j)L’exemple suivant représ<strong>en</strong>te une partition plane de forme R = (7, 5, 3, 1)11 11 12 2 13 3 1 1 16 3 3 2 1 1 1252


Fonctions de Schur et MacMahonFig. 3-10 – Partition plane 3DLa mise des cubes sur les cases (i, j) dans le diagramme de gauche λ/µ, donne unepartition gauche à 3 dim<strong>en</strong>sions. Les composantes de π sont des parts π (x) et le nombrede cubes est donné par :|π| = ∑ π (x)x∈λIl est appelé le poids d’une partition plane π. En effet, dans le plan, on peut représ<strong>en</strong>terune partition plane par un diagramme de Young où chaque case est munie d’un <strong>en</strong>tierreprés<strong>en</strong>tant le nombre de cubes correspondants dans l’axe vertical, c’est-à-dire qu’on aun tableau constitué d’un diagramme de Young rempli d’<strong>en</strong>tiers de manière décroissante(mais pas nécessairem<strong>en</strong>t strictem<strong>en</strong>t) sur les lignes et les colonnes. Toute partition planese décompose <strong>en</strong> tranches diagonales, notées π (m) où m ∈ Z, qui satisfont aux conditionssuivantes :· · · π(−2) < π(−1) < π(0) > π(1) > π(2) > · · · ,et plus précisém<strong>en</strong>t la relation d’<strong>en</strong>trelace <strong>en</strong>tre deux partitions λ et µ à 2dim<strong>en</strong>sions esttelle queµ > λ ⇔ µ 1 ≥ λ 1 ≥ µ 2 ≥ λ 2 ≥ µ 3 ≥ · · ·Chaque tranche diagonale d’une partition plane π est une partition linéaire π (m) à (2dim<strong>en</strong>sions). Par exemple la diagonale principale de π est une partition π (0) = (π 11 , π 22 · · · ).8.1.6 Partitions strictes et diagramme de Young shiftéUne partition stricte λ est une suite d’<strong>en</strong>tiers strictem<strong>en</strong>t positifs non nuls dont lescomposantes de λ sont des parts strictem<strong>en</strong>t distincts [190]. Plus précisém<strong>en</strong>t, une partition253


Diagrammes et tableaux de Youngstricte λ de degré ou de poids n se définit comme suit :λ = (λ 1 , λ 2 , · · · , λ r ) λ 1 > λ 2 > · · · > λ r ≥ 0La somme des parts λ i est une partition de poids n, notée|λ| = λ 1 + λ 2 + · · · + λ r + · · ·Pour une partition stricte λ = (n 1 , n 2 , · · · , n r ) avec n 1 > n 2 > · · · > n r ≥ 0 où r est paire.Pour n k = 0, la partition correspondante est λ = (n 1 , n 2 , · · · , n k−1 ) et si n k >0, nous avonsλ = (n 1 , n 2 , · · · , n k ).On définit une partition double λ” de poids 2n dont la notation de Frob<strong>en</strong>uis est la suivanteλ” = (n 1 , n 2 , · · · , n k | n 1 − 1, n 2 − 1, · · · , n k−1 − 1)Une partition stricte λ = (λ 1 , λ 2 , · · · , λ r ) à 2 dim<strong>en</strong>sions peut se représ<strong>en</strong>ter dans le réseaucartési<strong>en</strong> Z × Z par un <strong>en</strong>semble de points de coordonnées (i, j) avec :{(i, j) | i = 1 · · · l(λ), i ≤ j ≤ λ i + i − 1}En pratique, à partir de cet <strong>en</strong>semble de points, on construit un diagramme de cases obt<strong>en</strong>ude manière à déplacer dans le diagramme usuel, la (i − 1) case à droite dans la i iéme ligne.Un tel diagramme s’appelle un diagramme de Young modifié.Exemple : La suite d’<strong>en</strong>tiers (7, 4, 3, 1) est une partition stricte de 7 dont le diagramme deYoung shifté associé est donné par la figure (3-11).Fig. 3-11 – Diagramme de Young shiftéL’exemple suivant, montre le diagramme d’une partition (ν) de 2n. L’aire A 1 dans cediagramme est le diagramme de la partition modifiée et l’aire A 2 est la partition conjuguéelorsque λ = (6, 3, 1) pour une partition double ν = (7, 5, 4, 2, 1, 1).254


Fonctions de Schur et MacMahonFig. 3-12 – Diagramme de Young shifté à partir d’une partition 2dL’analogue de la formule (3.1) est <strong>du</strong>e à I.G.Macdonald [198]. Pour les partitions strictesλ est définit comme suit|λ|!∏h(i, j)i,j(3.2)avec h(i, j) la formule de hook pour la partition double ν. Une partition plane stricte π deforme λ est une suite (..., .λ −1 , λ 0 , λ 1 , ...) qui vérifie la propriété suivante :· · · ⊂ λ −1 ⊂ λ 0 ⊃ λ 1 ⊃ · · ·où λ 0 correspond à une partition ordinaire sur la diagonale principale.Exemple : Soit la partition plane ˆπ = {(3) , (4, 3) , (5, 3) , (3, 2) , (2) (1)}. Chaque partitiondiagonale π (x) est une partition stricte (2d). Le poids d’une partition plane stricte estnoté comme suit :|π| = ∑ π (x)x∈λavec la condition que π 1 (x) > π 2 (x) .8.2 Fonctions de SchurLes fonctions de Schur jou<strong>en</strong>t un rôle trés important dans ce qui va suivre. Nous <strong>en</strong>donnons ici diverses définitions ainsi que certaines propriétés qui les concern<strong>en</strong>t.A toute représ<strong>en</strong>tation <strong>du</strong> groupe symétrique est associée une fonction symétrique. C’estainsi que les fonctions de Schur S λ sont associées aux représ<strong>en</strong>tations irré<strong>du</strong>ctibles. Lestableaux de Young fourniss<strong>en</strong>t un moy<strong>en</strong> élégant pour exprimer ces polynômes. Par ailleurs,il existe une règle purem<strong>en</strong>t combinatoire qui fait appel aux tableaux de Young et quipermet de décomposer le pro<strong>du</strong>it de deux polynômes de Schur. Ceci implique <strong>en</strong> particulierque les tableaux permett<strong>en</strong>t de décomposer le pro<strong>du</strong>it t<strong>en</strong>soriel de deux représ<strong>en</strong>tationsirré<strong>du</strong>ctibles de GL(E) <strong>en</strong> somme directe de représ<strong>en</strong>tations irré<strong>du</strong>ctibles.255


Fonctions de SchurBi<strong>en</strong> qu’elles ont été définies par Jacobi, les fonctions de Schur jou<strong>en</strong>t un rôle fondam<strong>en</strong>taldans la théorie des repres<strong>en</strong>tations. Nous comm<strong>en</strong>çons par rappeller la définition originale.La fonction de Schur est définie par :s λ = a λ+δa δavec a δ est le déterminant de Vandermondea δ =∏(x i − x j ) = det(x n−ji ).1≤i


Fonctions de Schur et MacMahonProposition (Cauchy) : Etant donné deux <strong>en</strong>sembles de variables x = (x 1 , x 2 , · · · ) et y =(y 1 , y 2 , · · · ) , les séries génératrices des fonctions de Schur vérifi<strong>en</strong>t les conditions suivantes :∑s λ (x)s λ (y) = ∏ (1 − x i y j ) −1 ∑, s λ (x)s λ t(y) = ∏ (1 + x i y j )i,ji,jλet la somme s’effectue sur tous les tableaux de Young. Pour le cas de la fonction de Schurgauche (ou <strong>en</strong> anglais skew Schur, elle vérifie les conditions suivantes :∑s λ/η (x)s µ/η (y) = ∏ (1 − x i y j ) ∑ s η/λ (x)s η/µ (y)ηi,jη∑s λ/η (x)s µ/η (y) = ∏ (1 + x i y j ) −1 ∑ s η t /λ(y)s η/µ (x)ηi,jη∑s η/λ (x)s η (y) = s λ (y) ∑ s µ (y)s µ (x)ηµ∑s η t /λ(x)s η (y) = s λ (y) ∑ s µ (y)s µ t(x)ηµ∑s λ/η (x)s η/µ (y) = s λ/µ (x, y)η∑s λ/η (x)s η (y) = s λ (x, y)ηPour une représ<strong>en</strong>tation fondam<strong>en</strong>tale λ = •, la fonction de Schur est définie pars λ/• = s λet si λ ⊂ µ alors s λ/µ = 0. Les fonctions de Schur form<strong>en</strong>t une base orthonormée pour lepro<strong>du</strong>it scalaire 〈., .〉{0, si λ ≠ µ〈s λ (x), s ν (y)〉 = δ (λ,ν) =1, si λ = µoù δ (.,.) est le symbole de Kronecker. C’est une conséqu<strong>en</strong>ce de la formule de Cauchy <strong>en</strong>termes des variables x = (x 1 , x 2 · · · ) et y = (y 1 , y 2 · · · )∑s λ (x)s λ (y) = ∏ (1 − x i y j ) −1i,jλNotons que l’une des motivations principales à la sommation des fonctions de Schur sur des<strong>en</strong>sembles de partitions est la détermination de fonctions génératrices de partitions planes.Considérons une autre forme de fonction de Schur gauche ”fonction Schur gauche”, avecλ et µ sont deux partitions telles que le diagramme gauche λ/µ soit une représ<strong>en</strong>tationprincipale de sa classe d’équival<strong>en</strong>ce :λs λ/µ = ∑ νc λ µνs ν257


Fonctions de Schuroù c λ µν est le nombre de tableaux gauche semi-standard de type λ − µ qui exprime ladécomposition <strong>du</strong> pro<strong>du</strong>it de deux fonctions de Schur :s µ s ν = ∑ λc λ µνs λPar exemple, pr<strong>en</strong>ons λ = (5, 4, 3), µ = (3, 1), ν = (4, 3, 1), alors c λ µν = 2. En effet, <strong>en</strong>choisissant un tableau de Young T , il y a deux tableaux de forme λ/µ indiqués dans cettefigure.T =52 6 7λ/µ :5 6 72 3 45 6 72 3 81 3 4 81 81 4En fonction des variables (x 1 , . . . , x n ), la fonction de Schur est exprimée comme :s λ/µ (x 1 , . . . , x n ) =plus particulièrem<strong>en</strong>t,∑λ≻λ(1)≻···≻λ(n−1)≻µx |λ|−|λ(1)|1 x |λ(1)|−|λ(2)|2 . . . x |λ(n−1)|−|µ|n⎧⎨x |λ|−|µ|1 , µ < λ ,s λ/µ (x 1 , 0, . . . ) =⎩0 , µ ⊀ λ .La proposition précéd<strong>en</strong>te peut donc être réécrite comme suit :s λ/µ (x 1 , . . . , x n ) = ∑ λ≻λ(1)≻···≻λ(n−1)≻µ s λ/λ(1)(x 1 )s λ(1)/λ(2) (x 2 ) . . . s λ(n−1)/µ (x n )Alors, la fonction de Schur est le caractère de la représ<strong>en</strong>tation irré<strong>du</strong>ctible de GL(n).8.2.1 Propriétés de la fonction de SchurConcluons cette section <strong>en</strong> donnant quelques propriétés de la fonction de Schur, quiseront très utiles pour la suite :s λ (cx) = c |λ| s λ (x)s λ/µ (cx) = c |λ|−|µ| s λ/µ (x)s λ (q −ρ ) = q − k λ2 s λ t(q −ρ )Dans le cas de fonctions de Schur-gauche, elles vérifi<strong>en</strong>t les conditions suivantes :s λ/µ (cx) = c |λ|−|µ| s λ/µ (x)s λ/µ (q ρ ) = (−1) |λ|−|µ| s λ t /µ(q −ρ )258


Fonctions de Schur et MacMahonOn peut aussi représ<strong>en</strong>ter la fonction de schur-gauche <strong>en</strong> termes d’élém<strong>en</strong>ts de matricesdes opérateurs vertexS R/Q (x) = 〈R| V − (x i ) |Q〉 = 〈Q| V + (x i ) |R〉Finalem<strong>en</strong>t, nous terminons par les formules qui seront très utiles dans le calcul de la fonctionde partition de la corde topologique <strong>en</strong> utilisant le formalisme de vertex topologique.On considère un diagramme de Young λ et une case d’une partition λ (i, j). On rappelleque le cont<strong>en</strong>u c(i, j) de la case est c(i, j) = j − i, la longueur de crochet h(i, j) estλ ′ i + λ j − i − j + 1 et le poids minimum d’une partition est donné par m (λ) = ∑ iles fonctions de Schur sont données par les formules suivantes :∏n+c(i,j)s λ (1, 1; · · · , 1) =} {{ }h(i,j)(i,j)∈λn fois∏s λ (q, q 2 ; · · · , q n ) = q m(λ)s λ (q −ρ ) = q |λ|(i,j)∈λ∏2 +n(λ)(i,j)∈λ1−q c(i,j)+n1−q h(i,j)11−q h(i,j)iλ i . AlorsExemple : Pr<strong>en</strong>ons la partition λ = (2, 2). La fonction de Schur est sous la forme suivante8.2.2 Opétareurs vertexs (2,2) (q, q 2 , q 3 ) = q 6 (1−q 3 )(1−q 4 )(1−q 2 )(1−q 3 )(1−q 3 )(1−q 2 )(1−q 2 )(1−q 1 )= q 6 + q 7 + 2q 8 + q 9 + q 10 .Cette partie vise à définir la relation <strong>en</strong>tre les états de la théorie des champs conformes(champs fermioniques libres à 2d) et le diagramme de Young. La représ<strong>en</strong>tation irré<strong>du</strong>ctiblede SU(N) peut être caractérisée par les états dans l’espace de Hilbert qui est généré parles états d’espace de Fock. En effet, soit une partition à 2dλ = (λ 1 , λ 2 , λ 3 , · · · ) , λ i ≥ λ i+1 , (3.3)nous pouvons associer à un état quantique |λ〉 , l’énergie E = ∑ r i qui est le nombre total deboîtes dans le diagramme de Young (n 1 , n 2 , · · · , n r ). Nous comm<strong>en</strong>çons par définir l’espacedes états de Fock fermionique :V = V 0 ⊕ V 1 ⊕ V 2 · · ·Le seul état de V 0 est le vide |0〉, qui par définition ne conti<strong>en</strong>t aucune particule. Pourcompléter cette définition sur V n , nous allons maint<strong>en</strong>ant définir les générateurs ψ i et ψ ∗ i259


Fonctions de Schurd’algèbre de Clifford qui sont des fermions libres qui vérifi<strong>en</strong>t les relations suivantes :{ψ i , ψ ∗ j} = δ ij{ψ i , ψ j } = 0{ψ ∗ i , ψ ∗ j} = 0Ces opérateurs de création et d’annihilation sont définis par ses actions sur les états debase de l’espace de Fock qui pass<strong>en</strong>t de V n à V n+1 .ψ j |x 1 , x 2 , · · · x n > = |x 1 , x 2 , · · · x n >ψ ∗ j|x 1 , x 2 , · · · x n > = |x 1 , x 2 , · · · x n >On construit V <strong>en</strong> choisissant un vecteur |0〉 et <strong>en</strong> agissant par des opérateurs de créationet d’annihilation. Ceci in<strong>du</strong>it à une représ<strong>en</strong>tation irré<strong>du</strong>ctible de l’algèbre de Cliffordψ ∗ i s · · · ψ ∗ i 1ψ jr · · · ψ j1|0〉 avec j r > · · · > j 1 > 0 ≥ i 1 > · · · > i sψ j |0 >= 0 j ≤ 0 et ψ ∗ j|0 >= 0 j > 0Les fermions libres vérifi<strong>en</strong>t les séries génératrices :ψ(z) = ∑ ψ n+1 z −n−1 ,2n∈Zψ ∗ (z) = ∑ n∈Zψ ∗ n+z −n−112Le courant de Noether U(1) est donné par :J(z) =: ψ(z)ψ ∗ (z) := ∑ m∈Zz −m−1 J met il vérifie les relations de commutation :[J n , J m ] = −nδ n+m,0 [J n , ψ k ] = ψ k+n [J n , ψ ∗ k] = −ψ ∗ k−nOn pr<strong>en</strong>d l’ordre normal de champ fermionique qui est défini par :ψ(z)ψ ∗ (z) =: ψ(z)ψ ∗ (z) : + 1z − w ,|z| > |w|Dans le calcul des amplitudes de la corde topologique, le boson libre est donné <strong>en</strong> termesdes oscillateurs par :∑L 0 = ∞ α −n α nn=1[L 0 , α n ] = nα nQ L 0α n Q −L 0= Q −n α n260


Fonctions de Schur et MacMahona) Bosonisation dans l’espace de HilbertPar la suite, nous allons voir comm<strong>en</strong>t l’espcae de Fock fermionique est isomorphe àl’espace de Fock bosonique des champs scalaires. Mais tout d’abord, considérons la sériegénératrice de champs chiral à deux dim<strong>en</strong>sions :φ(z) = i ∑ n≠0α nnz n .On peut donc exprimer les bosons à l’aide des fermions. Cette correspondance est <strong>du</strong>eà la bosonisation qui est une construction propre à une dim<strong>en</strong>sion d’espace permettantd’exprimer les champs fermioniques d’une théorie <strong>en</strong> fonction de champs de Bose. Il s’agitd’une relation d’équival<strong>en</strong>ce quantique exacte <strong>en</strong>tre deux théories (une de fermions et unede bosons), connue sous le nom de la correspondance bosons-fermions telles que les bosonslibres α n sont issus de la bosonisation des fermions complexes qui vérifi<strong>en</strong>t les relationssuivantes :∑α n = ψ k+n ψ ∗ k, n = ±1, ± 2, · · ·k∈Z+ 1 2i∂φ = : ψψ :ψ = : e iφ :ψ ∗ = : e −iφ : .L’algèbre de l’oscillateur ou algèbre de Heis<strong>en</strong>berg est l’algèbre de Lie de base n vérifiantles relations de commutation suivantes :[α n , α m ] = −nδ n+m,0[α n , ψ k ] = ψ k+n[α n , ψ ∗ k] = −ψ ∗ k−nIl se trouve que l’action de l ’opérateur de création α −n sur le diagramme de Young estdonnée <strong>en</strong> ajoutant une seule bande de longueur n dans tous les chemins possibles. De lamême manière, <strong>en</strong> faisant agir l’opérateur d’annihilation α n sur le diagramme, cela con<strong>du</strong>ità éliminer une bande de frontière <strong>du</strong> diagramme voir la figure (3-13).Cette équival<strong>en</strong>ce s’établit <strong>en</strong> utilisant les opérateurs de vertex qui ont les mêmes dim<strong>en</strong>sionsconformes que le fermion complexe :Γ ± (z) = exp ∑ n>0261z ±nn α ±n .


Fonctions de SchurFig. 3-13 – Appliquation de α −3 à une partition.Ils vérifi<strong>en</strong>t les conditions suivantes :avecΓ + (x)Γ − (y) =11− x yΓ − (y)Γ + (x) ,Γ + (x) |0〉 = |0〉 ,〈0| Γ − (y) = 〈0| .Γ ± (1)q nL 0= q nL 0Γ ± (q n ) q nL 0Γ ± (1) = Γ ± (q −n )q nL 0Γ ± (x)q nL 0= q nL 0Γ ± (xq n ) q nL 0Γ ± (x) = Γ ± (xq −n )q nL 0b) Connection <strong>en</strong>tre le diagramme de Young et les fermions libresIl est bi<strong>en</strong> connu que les états de Fock fermioniques sont étiquetés par les diagrammesde Young (2D) indiqués par µ = (µ 1 ≥ µ 2 .. ≥ µ n ) dont la notation de Frob<strong>en</strong>ius est donnéepar les coordonnéesa i = µ i − i + 1 2b i = µ t i − i + 1 2alors l’état fermionique associé à une partition µ est donné par :|µ〉 =d∏ψ ∗ −a iψ −bi |0〉i=1où d est le nombre des cases sur la diagonale de µ.Γ + (1) |λ〉 = ∑ λ≻µ |µ〉Γ − (1) |λ〉 = ∑ µ≻λ |µ〉262


Fonctions de Schur et MacMahonLe mode zéro de Virasoro est L 0 qui compte le nombre |µ| de boîtes dans le diagramme deYoung µ :∞∑L 0 |µ〉 = |µ| |µ〉 . avec L 0 = α −n α nPlus particulièrem<strong>en</strong>t, le pro<strong>du</strong>it des opérateurs vertex agit sur l’état fermionique qui setra<strong>du</strong>it par la fonction de Schur gauche :n=1Γ − (x n )|ν >= ∑ s µ/ν (x n )|µ >µ⊃νΓ − (x n )|ν >= ∑ s ν/µ (x n )|µ >µ⊃ν(3.4)8.3 Fonction de MacMahonLes fonctions génératrices, dites fonctions de Macmahon, intervi<strong>en</strong>n<strong>en</strong>t dans plusieursdomaines que ce soit <strong>en</strong> physique, dans l’étude de la théorie de corde topologique ou <strong>en</strong>mathématiques [194, 195, 196, 197, 198].Dans cette section, on prés<strong>en</strong>te diverses fonctions génératrices des partitions standard, raffiné,coloré et shifté via le formalisme de matrice de transfert de Okounkov–Reshetikhin–Vafa [199]. Ces fonctions posséd<strong>en</strong>t aussi une interprétation physique dans la théorie descordes, les théories des noeuds et théories de Doladson-Thomas.Le premier paragraphe concerne l’expression de la fonction génératrice associé aux diagrammesde Young standard ; le second paragraphe évoque les différ<strong>en</strong>tes expressions desz02yxFig. 3-14 – Une partition 2D (3, 2, 1, 1) dans la tranche diagonale à t = −2. L’état fermionique|µ〉 = ψ ∗ −7/2ψ ∗ −5/2ψ ∗ −1/2ψ −1/2 |0〉 .263


Fonction de MacMahonfonctions génératrices associées aux diagrammes de Young standard, raffiné, shifté, et coloré.8.3.1 Conjecture de MacMahonLes partitions planes ont été énumérées par Percy Macmahon <strong>en</strong> 1912, qui a donné labelle fonction génératrice [92, 194]. Alors si on considère s l’<strong>en</strong>semble des partitions planes,la fonction génératrice de s est le polynôme ou la série suivante :∑q |π| (3.5)π∈sLa fonction de Macmahon est la fonction génératrice pour l’<strong>en</strong>semble des fonctions indexéespar les partitions∞∏M(q) = (1 − q n ) −n = 1 + q + 3q 2 + 6q 3 + 13q 4 + ...n=1Fig. 3-15 –dont le coeffici<strong>en</strong>t de q n est le nombre de partitions planes de volume n. Cep<strong>en</strong>dant, onutilise une fonction génératrice pour calculer le nombre de partitions de n qui est noté p(n),qui veut dire le nombre de manières distinctes (et d’ordre indép<strong>en</strong>dant) de représ<strong>en</strong>ter ncomme une somme d’<strong>en</strong>tiers. Par conv<strong>en</strong>tion p(0) = 1, p(n) = 0 pour n négatif. On vacomm<strong>en</strong>cer par une partition à 2-dim<strong>en</strong>sions qui est une partition linéaireλ = (λ 1 , λ 2 , ...λ m )264


Fonctions de Schur et MacMahonavec λ 1 ≥ λ 2 ≥ · · · . La fonction génératrice <strong>du</strong> nombre de partitions s’écrit∑p 1 (n)q n = 1 + x + 2x 2 + 3x 3 + 5x 4 + 7x 5 + 11x 6 + 15x 7 + 22x 8 + · · ·n=0= ∏ (1 − x i ) −1i≥1où q est une variable indéterminée et p 1 (n) indique le nombre de partitions de taille n. Lapreuve de cette fonction de partition est donnée <strong>en</strong> terme d’une partition λ et satisfait àla condition suivante :|λ| = λ 1 + λ 2 + ...λ m = ∑ iid i d i = λ i − λ i+1et la fonction génératrice <strong>en</strong> terme de l’indéterminé est donné par∑q |λ| =λ∑d 1 ,d 2 ,···q d 1+2d 2 +··· = ∏ i≥1∑q id i= ∏ ( ) 1 − xi −1i≥1Une partition à trois dim<strong>en</strong>sions est souv<strong>en</strong>t appelée une partition plane avec la fonctiongénératrice s’écrivant sous la forme suivante :∑p 2 (n)q n = 1 + x + 3x 2 + 6x 3 + 13x 4 + 24x 5 + 48x 6 + 86x 7 + 160x 8 + · · ·n=0= ∏ (1 − x i ) −ii≥1d ioù p 2 (n) est le nombre des partitions planes. MacMahon a conjecturé l’id<strong>en</strong>tité suivantedans la terminologie des partitions planes : la fonction génératrice d’une partition à ddim<strong>en</strong>sions (d-partition) <strong>du</strong> cube à d-dim<strong>en</strong>sion est donnée parG d (q) = ∏ (1 − q i ) − (i+d−3)!(i−1)!(d−2)!i≥1(3.6)Il est important de souligner que cette expression qui a été conjecturé par Macmahon, futdémontrée via le formalisme de matrice de transfert [93]. Et elle possède une interprétationphysique <strong>en</strong> termes des théories des champs conformes.8.3.2 Fonctions génératricesLa fonction de Macmahon développée <strong>en</strong> mathématique semble avoir trouvé aussi d’intéressanteapplication dans le domaine de la physique, notamm<strong>en</strong>t dans la physique statistique,théorie des cordes. Grâce au formalisme de la matrice de transfert, il existe un li<strong>en</strong>étroit <strong>en</strong>tre la fonction de Macmahon et le modèle de cristal fon<strong>du</strong> qui est exactem<strong>en</strong>t la265


Fonction de MacMahonfonction de partition de la théorie des cordes topologiques sur la variété de CY.Dans cette dernière section, nous énonçons et démontrons le résultat analogue à ceux quiprécèd<strong>en</strong>t, mais dans le contexte de formalisme de matrice de Transfert permettant d’atteindreles expressions des fonctions de Macmahon standard, raffiné, shifté et coloré.Le formalisme de matrice de transfert 1 peut désormais être exprimé selon deux opérateursde création et d’annihilation qui consiste à décomposer une partition 3d <strong>en</strong> tranches diagonales,ce qui permet d’effectuer une suite des partitions 2d. Alors la fonction génératrice<strong>du</strong> nombre de partitions s’écrit :∑Z (q) :=q |π| (3.7)π−3d partitionsFig. 3-16 – une suite des partitions standard 2D et partition 3D.La décomposition de partition 3D sera une suite des partitions 2D rangées par ordre croissantde a = −∞ jusqu’à 0 et décroissant de a = 1 jusqu’à ∞. Okounkov–Reshetikhin–Vafa,montr<strong>en</strong>t que la fonction (3.7) peut s’exprimer <strong>en</strong> fonction des opérateurs de création etd’annihilation. Ce qui donne une expression simple de la série génératrice correspondante,c’est à dire la fonction de Macmahon généralisée à 3D〈 ( ∞) (∏∞) 〉Z (q) : = 0 | q L 0∏Γ + (1) q L 0Γ − (1) q L 0| 0〈 n=1m=1Z (q) : = 0 | ∏ ) ∏ (Γ +(q n− 1 2 Γ − q − (m− 1 2) ) 〉| 0(3.8)n>0m>0= ∏ ∞( )k=1 1 − qk −k= M (q) .Ce résultat était principalem<strong>en</strong>t obt<strong>en</strong>u <strong>en</strong> utilisant le formalisme de matrice de transfert(voir chapitre cristal Calabi-Yau), on peut de même trouver la fonction de Macmahon de1 Ce formalisme est détaillé dans le chapitre I.266


Fonctions de Schur et MacMahonla manière suivante : Etant donnés deux partitions λ et ν, une partition gauche est unepartition 3D de la forme λ/ν avec {π i,j | (i, j) ∈ λ/ν}. La fonction génératrice correpondantaux partitions planes de la forme ν c est donnée parZ ν (q) := ∑ π(ν c )q |π| =∏ ( )1 − qĥ(i,j) . (3.9)(i,j)∈ν coù ĥ (i, j) = j − ν i + i − ν t j − 1 est la longueur d’equerre d’une boîte (i, j) ∈ ν c . Si unepartition ν est égale à ∅, alors la fonction génératrice devi<strong>en</strong>t la fonction de Macmahon à3DM (q) := Z ∅ (q) =∞∏ ( ) 1 − qk −k. (3.10)Cette fonction est exactem<strong>en</strong>t la fonction de partition de la corde topologique de type Asur C 3 . Nous avons eu l’occasion de détailler ce résultat au chapitre 1.k=1Fonctions de MacMahon raffinéDans la version raffiné, la fonction de Macmahon à 3 dim<strong>en</strong>sions est donnée <strong>en</strong> termesdes opérateurs de création et d’annihilation par l’expression suivante :M (q, t) = ∑ q P ∞i=1 |π(−i)| t P ∞i=1 |π(j−1)|〈 π= 0 | ∏ ) ∏ (Γ +(q n− 1 2 Γ − t −(m− 1 2) ) 〉| 0(3.11)n>0m>0∏M (q, t) . = ∞ (1 − q i−1 t j ) −1i,j=1Ce résultat constitue un premier pas vers la composition de vertex topologiques raffiné etpermet d’avoir une relation avec les invariants des noeuds de Khovanov.Fonction de MacMahon shiftéNous proposons maint<strong>en</strong>ant comm<strong>en</strong>t associer un diagramme de Young shifté à unefonction de Macmahon via le formalisme de matrice de transfert [185, 186, 188, 189]. Avantde démontrer la fonction de partition, il est utile de rappeler quelques outils, à savoir lesopérateurs vertex, opérateurs de création et annihilation <strong>en</strong> fonction des diagrammes deYoung shifté.( ∑ )2Γ + (x) = expm∈N impaire m zm λ m( ∑ )2Γ − (y) = expm∈N impaire m zm λ −m267,.


Fonction de MacMahonavec[λ m , λ n ] = − m 2 δ n+m,0 m, n ∈ Z impaireles relations de commutations <strong>en</strong>tre les opérateurs vertex(Γ + (x)Γ − (y) =1+xy1−xyΓ ± (x)Γ ± (y) = Γ ± (y)Γ ± (x))Γ − (y)Γ + (x)Pour établir la forme explicite de la fonction de Macmahon shifté, nous estimons de décomposerles partitions 3d stricte comme suite des partitions shifté 2d. En utilisant leformalisme de matrice de transfert, la fonction génératrice des partitions est donnée parM shif (q) = ∑ 3d shifté 2p(π) q〈|π|〉∏= 0 | ∞ ) ∏ ∞ )Γ +(q −j+ 1 2 Γ −(q k− 1 2 | 0(3.12)j=1k=1∏M shif (q) = ∞ ( ) n1+q n1−q nn=1Il s’agit <strong>en</strong> effet d’une fonction génératrice d’une partition stricte [189].Fonction génératrice pour les diagrammes de Young coloré à 3 dim<strong>en</strong>sionPassons maint<strong>en</strong>ant aux fonctions de Macmahon qui apparaiss<strong>en</strong>t comme des fonctionsgénératrices des diagrammes de Young colorés. Cette fonction est liée à la fonction departition Donaldson-Thomas de l’orbifold C 3 /Z 3 .Fig. 3-17 – Diagramme de Young coloréPr<strong>en</strong>ons une partition colorée 3d, les tranches diagonales sont des partitions colorées à 2d268


Fonctions de Schur et MacMahonayant les mêmes couleurs voir la figure (3-18) une séqu<strong>en</strong>ce d’<strong>en</strong>trelacem<strong>en</strong>t des diagrammesde Young. En d’autres termes, si, pour k ≥ 1, nous définissons les diagrammes de Young µµ −k = { π k,0 , π (k+1),1 , π (k+2),2 , · · · }µ 0 = {π 0,0 , π 1,1 , π 2,2 , · · · }µ k = { π 0,k , π 1,(k+1) , π 2,(k+2) , · · · }Nous aurons besoin d’attribuer une couleur à chaque diagramme de Young avec la conditionsuivante :π = { ∅ = µ −m < · · · < µ −1 < µ 0 > µ 1 > · · · > µ n = ∅ }pour certains m, n ≥ 1. Fixons un <strong>en</strong>semble des couleurs et remplaçons la variable q avecun <strong>en</strong>semble des variables :Q = {q g | g ∈ C}où C est l’<strong>en</strong>semble des couleurs. En particulier, C est le sous groupe abéli<strong>en</strong> fini G = Z n degroupe SO(3). Il se trouve que la variable Q compte le nombre des boites d’un diagrammede YoungQ |λ〉 = q |λ| |λ〉Fig. 3-18 – Séqu<strong>en</strong>ce de diagrammes de Young à 2d coloréPour chaque g ∈ G, on peut définir les opérateurs diagonaux par :Q g |λ〉 = q g|λ| |λ〉269


Fonction de MacMahonavec les opérateurs de création et d’annihilation qui satisfont les conditions suivantes :Γ + (x)Q g = Q g Γ + (q g x) , Q g Γ − (x) = Γ − (q g x)Q g . (3.13)Le problème se ré<strong>du</strong>it alors à trouver l’expression de l’opérateur vertex qui calcule lafonction de partition Z ZnavecQ = {q 0 , · · · , q n−1 } , q = q 0 q 1 · · · q n−1et l’opérateur de création et d’annihilation sont comme suit :Ā ± (x) = Γ ± (x)Q 1 Γ ± (x)Q 2 · · · Q n−1 Γ ± (x)Q 0 . (3.14)Il se trouve que le pro<strong>du</strong>it des opérateurs compte la fonction de partition <strong>du</strong> diagrammede Young Z n coloré à 3 dim<strong>en</strong>sion. Nous utilisons les relations suivantesA + (x) = Q · Γ + (xq 1 q 2 q 3 · · · q n−1 q 0 ) Γ + (xq 2 q 3 · · · q n−1 q 0 ) · · · Γ + (xq 0 )A − (x) = Γ + (x)Γ + (xq 1 ) · · · Γ + (xq 1 q 2 · · · q n−1 ) · Q= Γ +(xqq−11 q −12 · · · q −1n−1q0−1la relation de commutation est définie comme suit :)Γ+(xqq−12 q −13 · · · q −1n−1q0−1) ( )· · · Γ+ xqq−10 · Q(3.15)A + (x)A − (y) = C(x, y) · A − (y)A + (x) A + (x) = Q −1 A + (x); A − (x) = A − (x)Q −1avecC(x, y) =(11−qxy) n ∏(· ∏0≤a≤b


Bibliographie[1] J. Polchinski, String theory, Cambridge University Press, 1998.[2] C. Vafa, Lectures on Strings and Dualities, hep-th/9702201.[3] S. Randjbar-Daemi, Intro<strong>du</strong>ction to Chiral Anomalies, lectures pres<strong>en</strong>ted at Intro<strong>du</strong>ctorySchool on String Theory, ICTP, Trieste, Italy, (1998).[4] K. S. Narain, Toroidal compactification and heterotic string, lectures pres<strong>en</strong>ted atIntro<strong>du</strong>ctory School on String Theory, ICTP, Trieste, Italy (1998).[5] C. Gomez and R. Hernandez, Fields, strings and branes, hep-th/9711102.C. Gomez, lectures pres<strong>en</strong>ted at the Workshop on Noncommutative Geometry, Superstringsand Particle Physics. Rabat -Morocco, (May 11-12 2001).[6] J. Polchinski, What is String Theory ?, hep-th/9411028.[7] E. Kiritsis, Intro<strong>du</strong>tion to Superstring Theory, hep-th/9709062.E. Kiritsis, Intro<strong>du</strong>ction to Non-perturbative string theory, hep-th/9708130.[8] R. Dijkgraaf, R. Gopakumar, H. Ooguri, C. Vafa, Baby Universes in String Theory,Phys. Rev. D73 (2006) 066002, hep-th/0504221.[9] M. Bershadsky, S. Cecotti, H. Ooguri and C. Vafa, “Kodaira-Sp<strong>en</strong>cer theory of gravityand exact results for quantum string amplitudes,” Commun. Math. Phys.165, 311(1994), hep-th/9309140.[10] M. Gre<strong>en</strong>, J. Schwarz and E. Witt<strong>en</strong>, Superstring Theory, vol 1 and 2, CombridgeUniversity Press, 1987.[11] B.Diu, C. Guthmann, D. Lederer, B. Roulet, Physique statistique (Hermann, Paris,1989).[12] H N Temperley, Statistical mechanics and the partition of numbers I : the transitionto liquid helium, Proc. Royal Soc. London A 199 361,V (1949). Temperley H N V,271


BIBLIOGRAPHIEStatistical mechanics and the partition of numbers II : the form of crystal surfaces,Proc. Cambr. Phil. Soc. 48 683 1952[13] Weiss C and Holthaus M From number theory to statistical mechanics : Bose-Einsteincond<strong>en</strong>sation in isolated traps Chaos, Solitons and Fractals 10 795 1999.[14] P. Di Francesco, P. Mathieu, D. S<strong>en</strong>echal, Conformal Field Theory, Springer-Verlag,New York, 1997.[15] Juan M. Maldac<strong>en</strong>a, ”The Large N limit of superconformal field theoriesand supergravity”, published in Adv.Theor.Math.Phys.2 :231-252,1998,Int.J.Theor.Phys.38 :1113-1133,1999. Also in *Bu<strong>en</strong>os Aires 1998, Tr<strong>en</strong>ds in theoreticalphysics 2* 51-63, hep-th/9711200.[16] A. Dabholkar, Exact counting of black hole microstates, Phys. Rev. Lett. 94 (2005)241-301, hep-th/0409148.[17] B. Pioline, Lectures on Black Holes, Topological strings and Quantum Attractors,Clas. Quant. Grav. 23 (2006) S981, hep-th/0607227.[18] L. Kauffman, Knots and physics, Teaneck, NJ, World Sci<strong>en</strong>tific Press, 1991.[19] W. Lickorish, An intro<strong>du</strong>ction to knot theory, Springer-Verlag, 1998.[20] A. M. Polyakov, Phys. Lett. B103(1981)207-211.[21] A. M. Polyakov, Phys. Scr. 15 (1987)191.[22] Adil Belhaj, ”Ingénierie Géométrique des Théories Quantiques des Champs dans lesModèles de Supercordes”, Thèse de Doctorat, Lab/UFR-PHE, Faculté des Sci<strong>en</strong>ces,Rabat.[23] P. Ramond, Dual Theory for fermions, Phys. Rev. D3(1971)2415.A. Neveu, J. H. Schwarz, Nucl. Phys. B31 86.[24] G. V<strong>en</strong>ieziano, An intro<strong>du</strong>ction to <strong>du</strong>al models of strong interactions and their physicalmotivations, Phys. Rep. C9 199.J. H. Schwarz, Dual resonance theory, Phys. Rep. C8(1973)269.A. Neveu, J. Schwarz, Nucl. Phys. B31(1971)86.P. Ramond, Phys. Rev. D3 (1971)2415.[25] F. Gliozzi, J. Scherk and D. Olive, Phys. Lett B65(1976)282.[26] F. Gliozzi, J. Scherk and D. Olive, Nucl. Phys. B122 (1977)282.272


BIBLIOGRAPHIE[27] A.A. Belavin, A. M. Polyakov and A. B. Zamolodchikov, J. Stat. Phys. 34, (1984),763 ; Nucl. Phys. B241(1984)333.[28] D. Friedan, Z. Qiu and S. Sh<strong>en</strong>ker, Phys. Rev. Lett.52(1984)1575.[29] S. Fubini and G. V<strong>en</strong>eziano, Ann. Phys. 63(1970)12.[30] A. Neveu, J. H. Schwarz and C. B. Thorn, Nucl. Phys. B35(1971)529.[31] R. Dijkgraaf, E. Verlinde, H. Verlinde, “Notes on topological string theory and twodim<strong>en</strong>sional topological gravity in String theory and quantum gravity, World Sci<strong>en</strong>tificPublishing, p. 91, (1991).[32] R. Dijkgraaf, Intro<strong>du</strong>ction to Topological Field Theories, Rec<strong>en</strong>t Directions in ParticleTheory (1993) 689-743.[33] R. Dijkgraaf, Intersection Theory, Integrable Hierarchies and Topological FieldTheory, hep-th/9201003.[34] E. Witt<strong>en</strong>, Phases of N = 2 theories in two dim<strong>en</strong>sions, Nucl. Phys. B403 (1993)159222, hep-th/9301042.[35] E. Witt<strong>en</strong>, Topological Sigma Models,” Commun. Math. Phys. 118, 411 (1988).[36] A. Schwarz, Sigma models having supermanifolds as target spaces, Lett. Math. Phys.38 (1996) 9196, hep-th/9506070.[37] A. Neitzke and C. Vafa, Topological Strings and their Physical Applications, hepth/0410178.[38] C. Vafa, Superstrings and topological strings at large N, J. Math. Phys. 42 (2001)2798-2817, hep-th/0008142.[39] Marcel Vonk, A mini-course on topological strings, arXiv :hep-th/0504147.[40] E. Witt<strong>en</strong>, Anti-de Sitter Space and Holography, Adv. Theor. Math. Phys. 2 (1998)253-291, hep-th/9802150.[41] J.Maldac<strong>en</strong>a, The large N limit of superconformal field theories and supergravity,Adv.Theor.Math. Phys 2 (198) 231 hep-th/ 9711200.[42] R. Ahl Laamara, ”Correspondance Jauge/Gravité <strong>en</strong> Théorie des Cordes et Modèle<strong>Topologique</strong> de Fluide des Cordes”, Thèse de Doctorat, Lab/UFR-PHE, Faculté desSci<strong>en</strong>ces, Rabat (2007-2008).[43] R. Gopakumar and C. Vafa, Topological gravity as large N topological gauge theory,Adv.Theor. Math. Phys. 2, 413 (1998), arXiv :hep-th/9802016.273


BIBLIOGRAPHIE[44] C. Vafa, ”Superstrings and topological strings at large N”, J. Math.Phys. 42 (2001)2798, hep-th/0008142.[45] M. Aganagic, A. Neitzke, C. Vafa, BPS Microstates and the Op<strong>en</strong> Topological StringWave Function, hep-th/0504054.[46] H. Ooguri, A. Strominger, and C. Vafa, Black hole attractors and the topologicalstring, Phys. Rev. D70 (2004) 106007, hep-th/0405146.[47] M. Aganagic, D. Jafferis, N. Saulina, Branes, Black Holes and Topological Stringson Toric Calabi-Yau Manifolds, Journal-ref : JHEP0612 :018,2006, arXiv :hepth/0512245.[48] L. B. Drissi, ”<strong>Contributions</strong> à l’étude des Branes, Trous Noirs Attracteurs et testesde la Conjecture OSV ”, Thèse de Doctorat, Lab/UFR-PHE, Faculté des Sci<strong>en</strong>ces,Rabat (2006-2007).[49] R. Ahl Laamara, A. Belhaj, L.B. Drissi, E.H. Saidi, ”Black Holes in type IIA Stringon Calabi-Yau Threefolds with Affine ADE Geometries and q-Deformed 2-D QuiverGauge Theories” hep-th/0611289.[50] El Hassan Saidi, ” Computing the Scalar Field Couplings in 6D Supergravity” Journalref: Nucl.Phys.B803 :323-362,2008, arXiv :0806.3207. ” On Black Hole EffectivePot<strong>en</strong>tial in 6D/7D N=2 Supergravity ”, Journal-ref : Nucl.Phys.B803 :235-276,2008,arXiv :0803.0827.[51] I. Antoniadis, S. Hoh<strong>en</strong>egger, Topological Amplitudes and Physical Couplings inString Theory, Journal-ref : Nucl.Phys.Proc.Suppl.171 :176-195,2007, arXiv :hepth/0701290.[52] I. Antoniadis, E. Gava, K. Narain, T. Taylor Topological amplitudes in string theory,Nucl. Phys. B413 (1994) 162 hep-th/9307158.[53] A. Iqbal, All g<strong>en</strong>us topological string amplitudes and 5-brane webs as Feynman diagrams,hep-th/0207114.[54] A. Iqbal, A. Kashani-Poor, SU(N) geometries and topological string amplitudes, Adv.Theor. Math. Phys. 10 (2006) 1-32, hep-th/0306032.[55] N. Nekrasov and A. Okounkov, Seiberg-Witt<strong>en</strong> theory and random partitions,arXiv :hepth/0306238.[56] H. Nakajima and K. Yoshioka, Instanton counting on blowup. I. 4-dim<strong>en</strong>sional puregauge theory, Inv<strong>en</strong>t. Math 162, no. 2, 313 (2005), arXiv :math.A.G/0306198.274


BIBLIOGRAPHIE[57] A. Braverman and P. Etingof, Instanton counting via affine Lie algebras II : fromWhittaker vectors to the Seiberg-Witt<strong>en</strong> prepot<strong>en</strong>tial, arXiv :math.AG/0409441.[58] M. Kreuzer, Toric Geometry and Calabi-Yau Compactifications, arXiv :hepth/0612307.[59] V. Bouchard, Lectures on complex geometry, Calabi-Yau manifolds and toric geometry,arXiv :hep-th/0702063.[60] Marcos Marino, Chern-Simons Theory and Topological Strings, Rev.Mod.Phys. 77(2005) 675-720, arXiv :hep-th/0406005.[61] Gerald V. Dunne, Christian Schubert, Bernoulli Number Id<strong>en</strong>tities from QuantumField Theory, arXiv :math/0406610.[62] M. Rossi, Geometric Transitions, math.AG/0412514.[63] Piotr Sulkowski, Calabi-Yau Crystals in Topological String Theory, arXiv :hepth/0712.2173.[64] Jun Li, Chiu-Chu Melissa Liu, Kef<strong>en</strong>g Liu, Jian Zhou, A Mathematical Theory ofthe Topological <strong>Vertex</strong>, arXiv :math/0408426.[65] J. Bryan, D. Karp, The closed topological vertex via the Cremona transform,math.AG/0311208.[66] Piotr Sulkowski, Crystal Model for the Closed Topological <strong>Vertex</strong> Geometry, JHEP0612 (2006) 030, arXiv :hep-th/0606055.[67] A. Iqbal, A-K Kashani-Poor, The <strong>Vertex</strong> on a Strip, hep-th/0410174, InstantonCounting and Chern-Simons Theory, Adv. Theor. Math. Phys. 7 (2004) 457-497,hep-th/0212279.[68] R. Dijkgraaf, A. Sinkovics, M. Temurhan Universal correlators from geometry, JHEP0411 (2004) 012, hep-th/0406247.[69] Lalla Btissam Drissi, Houda Jehjouh, El Hassan Saidi, Topological String onToric CY3s in Large Complex Structure Limit, Nucl.Phys.B813 :315-348, 2009,arXiv :0812.0526.[70] Lalla Btissam Drissi, Houda Jehjouh, El Hassan Saidi, Non Planar Topological 3-<strong>Vertex</strong> Formalism, Nucl.Phys.B804 :307-341,2008, arXiv :0712.4249.[71] Mina Aganagic, Albrecht Klemm, Marcos Marino, Cumrun Vafa , The Topological<strong>Vertex</strong>,Commun. Math. Phys. 254 (2005) 425-478, hep-th/0305132.275


BIBLIOGRAPHIE[72] Yukiko Konishi, Satoshi Minabe, Flop invariance of the topological vertex,arXiv :math/0601352.[73] N. Saulina, C. Vafa, D-branes as defects in the Calabi-Yau crystal, hep-th/0404246.[74] T. Okuda, Derivation of Calabi-Yau Crystals from Chern-Simons Gauge theory,JHEP 0503 (2005) 047 hep-th/0409270.[75] N. Halmagyi, A. Sinkovics, P. Sułkowski, Knot invariants and Calabi-Yau crystals,JHEP 0601 (2006) 040, hep-th/0506230.[76] Takuya Okuda, BIons in topological string theory, Journal-ref : JHEP0801 :062,2008,arXiv :0705.0722.[77] Dominic Joyce, Lectures on Calabi-Yau and special Lagrangian geometry,arXiv :math/0108088.[78] Riccardo Ricci, Super Calabi-Yau’s and Special Lagrangians, JHEP 0703 (2007) 048,arXiv :hep-th/0511284.[79] N. A. Nekrasov, Seiberg-Witt<strong>en</strong> prepot<strong>en</strong>tial from instanton counting, Adv. Theor.Math. Phys. 7, 831 (2004), arXiv :hep-th/0206161.[80] A. Iqbal and A. K. Kashani-Poor, Instanton counting and Chern-Simons theory,Adv. Theor.Math. Phys. 7, 457 (2004), arXiv :hep-th/0212279, SU(N) geometriesand topological string amplitudes, Adv. Theor. Math. Phys. 10, 1 (2006), arXiv :hepth/0306032.[81] T. Eguchi and H. Kanno, Topological strings and Nekrasov’s formulas, JHEP 0312,006 (2003), arXiv :hep-th/0310235.[82] Pan P<strong>en</strong>g, A simple proof of Gopakumar-Vafa conjecture for local toric Calabi-Yaumanifolds arXiv :math.AG/0410540.[83] J. Bryan and R. Pandharipande, The local Gromov-Witt<strong>en</strong> theory of curves,arXiv :math/0411037.[84] D. Maulik, N. Nekrasov, A. Okounkov, R. Pandharipande, Gromov-Witt<strong>en</strong> theoryand Donaldson-Thomas theory, I-II, math.AG/0312059, math.AG/0406092.[85] C. Faber, R. Pandharipande, Hodge integrals and Gromov-Witt<strong>en</strong> theory,math.AG/9810173.[86] T. Graber and E. Zaslow, Op<strong>en</strong> string Gromov-Witt<strong>en</strong> invariants : Calculations anda mirror ’theorem’, arXiv :hep-th/0109075.276


BIBLIOGRAPHIE[87] Yukiko Konishi, Integrality of Gopakumar-Vafa invariants of toric Calabi–Yau threefolds,Publ. RIMS, Kyoto Univ. 42 (2006), 605–648.[88] A. Iqbal, A-K Kashani-Poor, Instanton Counting and Chern-Simons Theory, Adv.Theor. Math. Phys. 7 (2004) 457-497, hep-th/0212279.[89] R. Gopakumar, C. Vafa, M-Theory and Topological Strings–I, hep-th/9809187, R.Gopakumar, C. Vafa, M-Theory and Topological Strings–II, hep-th/9812127.[90] T. Hollowood, A. Iqbal, C. Vafa, Matrix models, geometric <strong>en</strong>gineering and ellipticg<strong>en</strong>era, hep-th/0310272.[91] Andrei Okounkov, Nikolai Reshetikhin, Cumrun Vafa, Quantum Calabi-Yau andClassical Crystals, arXiv :hep-th/0309208.[92] P. McMahon, Combinatory analysis I, II, Cambridge University Press, 1915-1916,reprinted by Chelsea, New York, 1960.[93] Lalla Btissam Drissi, Houda Jehjouh, El Hassan Saidi, G<strong>en</strong>eralized MacMahon G(q)as q-deformed CFT Correlation Function, arXiv :0801.2661, Nuclear Physics B 801(2008) 316–345.[94] J. Heckman, C. Vafa, Crystal Melting and Black Holes, hep-th/0610005.[95] P. Candelas, “Lectures on Complex Geometry” in Trieste 1987, Proceedings, Superstrings’87. 1987.[96] Vinc<strong>en</strong>t Bouchard, Lectures on complex geometry, Calabi–Yau manifolds and toricgeometry, arXiv :hep-th/0702063v1.[97] B. R. Gre<strong>en</strong>e, “String theory on Calabi-Yau manifolds,” hep-th/9702155.[98] S.-T. Yau, “On Calabi’s conjecture and some new results in algebraic geometry,”Proceedings of the National Academy of Sci<strong>en</strong>ces of the U.S.A. 74 (1977) 1798–1799.S.-T. Yau, “On the Ricci curvature of a compact Kahler manifold and the complexMonge-Ampère equations. I.,” Communications on pure and applied mathematics 31(1978) 339–411.[99] K. Hori and C. Vafa, Mirror symmetry, hep-th/0002222.[100] M. Aganagic, C. Vafa, Mirror symmetry, D-branes and counting holomorphic discs,hep-th/0012041.[101] M. Aganagic and C. Vafa, Mirror Symmetry, D-Branes and Counting HolomorphicDiscs, arXiv :hep-th/0012041.277


BIBLIOGRAPHIE[102] E.Witt<strong>en</strong>, Mirror manifolds and topological field theory, arxiv : hep-th 9112056.[103] K. Hori, S. Katz, A. Klemm, R. Pandharipande, R. P. Thomas, C. Vafa, R. Vakil,and E. Zaslow, Mirror symmetry, vol. 1 of Clay Mathematics Monographs. AmericanMathematical Society, Provid<strong>en</strong>ce, RI, 2003.[104] M. Aganagic, A. Klemm, and C. Vafa, Disk Instantons, Mirror Symmetry and theDuality Web, hep-th/0105045.[105] M. Aganagic and C. Vafa, Mirror symmetry and supermanifolds, hep-th/0403192.[106] A. Grassi, M. Rossi, Large N <strong>du</strong>alities and transitions in geometry,math.AG/0209044.[107] H. Ooguri, C. Vafa, Worldsheet Derivation of a Large N Duality, Nucl. Phys. B641(2002) 3-34 hep-th/0205297.[108] J. Gomis, T. Okuda, Wilson Loops, Geometric Transitions and Bubbling Calabi-Yau’s, JHEP 0702 (2007) 083 hep-th/0612190.[109] V. V. Batyrev, Dual Polyhedra and Mirror Symmetry for Calabi-Yau Hypersurfaces,J. Alg. Geom. 3 (1994) 493, alg-geom/931000.[110] R. Gopakumar and C. Vafa, On the gauge theory/geometry correspond<strong>en</strong>ce, Adv.Theor. Math. Phys. 3 (1999) 14151443, hep-th/9811131.[111] P. Candelas and X. C. de la Ossa, “Comm<strong>en</strong>ts on Conifolds,” Nucl. Phys. B 342(1990) 246.[112] P. Griffiths, J. Harris, Principles of algebraic geometry, Wiley-Intersci<strong>en</strong>ce, New York,1978.[113] N. Leung, C. Vafa, Branes and toric geometry, Adv. Theor. Math. Phys. 2 (1998) 9,hep-th/9711013.[114] S. Franco, A. Hanany, D. Martelli, J. Sparks, D. Vegh, B. Wecht, Gauge Theoriesfrom Toric Geometry and Brane Tilings, JHEP 0601 (2006) 128, hep-th/0505211.[115] D. Cox, Rec<strong>en</strong>t developm<strong>en</strong>ts in toric geometry, arXiv :alg-geom/9606016.[116] H. Skarke, String <strong>du</strong>alities and toric geometry : An intro<strong>du</strong>ction, hep-th/9806059.[117] W. Fulton, Intro<strong>du</strong>ction to Toric Varieties. Annals of Mathematics Studies. PrincetonUniversity Press, 1993. 157p.[118] V. Bouchard, Toric Geometry and String Theory. PhD thesis, University of Oxford,2005, hep-th/0609123.278


BIBLIOGRAPHIE[119] D. A. Cox, The Homog<strong>en</strong>eous Coordinate Ring of a Toric Variety, alg-geom/9210008.[120] Kristian D. K<strong>en</strong>naway, A Geometrical Construction of Rational Boundary States inLinear Sigma Models, arXiv :hep-th/0203266v3.[121] K. Hori, Linear Models of Supersymmetric D-branes, hep-th/0012179.[122] S. Hellerman, S. Kachru, A. Lawr<strong>en</strong>ce, J. McGreevy, Linear Sigma Models for Op<strong>en</strong>Strings, hep-th/0109069.[123] R. Harvey, H. B. Lawson, Calibrated geometries, Acta Mathematica 148 (1982) 47.[124] A.Strominger, S.TYau et E.Zaslow Mirror symmetrie is T-<strong>du</strong>ality, Nuclear Phys, B479, 1996, p. 243-259.[125] R. K. Kaul, T. R. Govindarajan, Three Dim<strong>en</strong>sional Chern-Simons Theory as aTheory of Knots and Links, Journal-ref : Nucl.Phys. B380 (1992) 293-336, arXiv :hepth/9111063.[126] Hirosi Ooguri and Cumrun Vafa, Knot Invariants and Topological Strings, J.r :Nucl.Phys. B577 (2000) 419-438, arXiv :hep-th/9912123.[127] M. Khovanov, sl(3) link homology, Algebr. Geom. Topol. 4 (2004) 1045,math.QA/0304375.[128] D. Karp, C. M. Liu, M. Marino, The local Gromov-Witt<strong>en</strong> invariants of configurationsof rational curves, Geometry Topology Monographs 10 (2006) 115-168,math.AG/0506488.[129] T. Graber and E. Zaslow, Op<strong>en</strong> string Gromov-Witt<strong>en</strong> invariants : Calculations anda mirror ’theorem’ arXiv :hep-th/0109075.[130] Johanna Knapp, D-Branes in Topological String Theory, arXiv :0709.2045.[131] K. Hori, A. Iqbal, C. Vafa, D-Branes And Mirror Symmetry, hep-th/0005247.[132] H. Ooguri, Y. Oz, Z. Yin, D-Branes on Calabi-Yau Spaces and Their Mirrors, Nucl.Phys. B477 (1996) 407-430, hep-th/9606112.[133] L.B Drissi, H. Jehjouh, E.H Saidi, Topological String on Local Elliptic Curve withLarge Complex Structure, Afr Journal Of Mathematical Physics, Volume 6 (2008)95-103.[134] E. Witt<strong>en</strong>, Quantum field theory and the Jones polynomial, Commun. Math. Phys.121 (1989) 351-399.279


BIBLIOGRAPHIE[135] C.Faber and R. Pandharipande, Hodge Intergrabls and Gromov-Witt<strong>en</strong> Theory,math-AG/ 9810173.[136] Marcos Marino, Les Houches lectures on matrix models and topological strings,arXiv :hep-th/0410165.[137] G. ’t Hooft, A planar diagram theory for strong interactions, Nucl. Phys. B72 461(1974).[138] R. Dijkgraaf, E. Verlinde, H. Verlinde, “Notes on topological string theory and twodim<strong>en</strong>sionaltopological gravity,” in String theory and quantum gravity, World Sci<strong>en</strong>tificPublishing, p. 91, (1991)[139] S. Yamaguchi and S. T. Yau, “Topological string partition functions as polynomials,”JHEP 0407, 047 (2004) arXiv :hep-th/0406078.[140] M. Bershadsky, S. Cecotti, H. Ooguri, and C. Vafa, Holomorphic Anomalies in TopologicalField Theories,” Nucl. Phys. 405B (1993) 279-304 ;[141] M. Aganagic, R. Dijkgraaf, A. Klemm, M. Marino and C. Vafa, Topological stringsand integrable hierarchies,” Commun. Math. Phys. 261, 451 (2006), arXiv :hepth/0312085.[142] V. Bouchard, B. Florea and M. Marino, “Topological op<strong>en</strong> string amplitudes on ori<strong>en</strong>tifolds,”JHEP 0502, 002 (2005), arXiv :hep-th/0411227.[143] M. Alim and J. D. Lange, Polynomial Structure of the (Op<strong>en</strong>) Topological StringPartition Function, JHEP 0710, 045 (2007), arXiv : 0708.2886.[144] A. Moujib, ”Modèles Cosmologiques Supersymétriques”, Thèse de Doctorat,Lab/UFR-PHE, Faculté des Sci<strong>en</strong>ces, Rabat (2008-2009).[145] A. Iqbal, N. Nekrasov, A. Okounkov, C. Vafa, Quantum foam and topological strings,hep-th/0312022.[146] N. Dunfield, S. Gukov, J. Rasmuss<strong>en</strong>, The Superpolynomial for Knot Homologies,Experim<strong>en</strong>t. Math. 15 (2006) 129, math.GT/0505662.[147] J. M. F. Labastida and M. Marino, A new point of view in the theory of knot andlink invariants, math.qa/0104180.[148] J. M. F. Labastida, Knot Invariants and Chern-Simons Theory, arXiv :hepth/0007152.[149] E. J van R<strong>en</strong>sburg, The Statistical Mechanics of Interacting Walks, Polygons, Animalsand Vesicles (Oxford : Oxford University Press) (2000).280


BIBLIOGRAPHIE[150] C. Weiss and M. Holthaus, From number theory to statistical mechanics : Bose-Einstein cond<strong>en</strong>sation in isolated traps Chaos, Solitons and Fractals 10 795, (1999).[151] Amer Iqbal, Can Kozcaz, Cumrun Vafa , The Refined Topological <strong>Vertex</strong>, hepth/0701156.[152] Andrei Okounkov, Nicolai Reshetikhin, Random skew plane partitions and the Pearceyprocess, arXiv :math/0503508.[153] D. Ghoshal and C. Vafa, c =1 string as the topological theory of the conifold, Nucl.Phys. B 453, 121 (1995), hep-th/9506122.[154] A. Iqbal, All g<strong>en</strong>us topological string amplitudes and 5-brane webs as Feynman diagrams,hep-th/0207114.[155] R. Gopakumar, C. Vafa, On the gauge theory/geometry correspond<strong>en</strong>ce, Adv. Theor.Math. Phys. 3 (1999) 1415-1443, hep-th/9811131.[156] E. Witt<strong>en</strong>, Chern-Simons gauge theory as a string theory, Prog. Math. 133 (1995)637-678, hep-th/9207094.[157] J. Labastida, Chern-Simons theory : t<strong>en</strong> years after, Tr<strong>en</strong>ds in Theoretical PhysicsII, AIP Confer<strong>en</strong>ce Proceedings, eds. H. Falomir, R. Gamboa, F. Schaposnik, vol.484, hep-th/9905057.[158] Mina Aganagic, Albrecht Klemm, Marcos Marino, and Cumrun Vafa, Matrix Modelas a Mirror of Chern-Simons Theory, arXiv : hep-th/0211098.[159] Sergei Gukov, Amer Iqbal, Can Kozcaz, Cumrun Vafa , Link Homologies and theRefined Topological <strong>Vertex</strong>, arXiv :0705.1368.[160] J. M. F. Labastida, M. Marino, Polynomial invariants for torus knots and topologicalstrings, Commun. Math. Phys. 217 (2001) 423-449, arXiv :hep-th/0004196.[161] J. M. F. Labastida, M. Marino, C. Vafa, Knots, links and branes at large N, JHEP0011 (2000) 007, arXiv :hep-th/0010102.[162] L.B Drissi, H. Jehjouh, E H Saidi, De la théorie des cordes twistorielles à la SupergravitéConforme N=4, D=4 , African Journal of Mathematical physics Volume 4Number (2007) pages 65-96[163] R. Ahl Laamara, L.B Drissi, H. Jehjouh, E H Saidi, Pure fermionic twistor like model& target space supersymmetry, arXiv :hep-th/0605167.[164] Christian Saemann, Aspects of Twistor Geometry and Supersymmetric Field Theorieswithin Superstring Theory, arXiv :hep-th/0603098.281


BIBLIOGRAPHIE[165] Nathan Berkovits, Edward Witt<strong>en</strong>, Conformal Supergravity in Twistor-StringTheory, JHEP 0408 (2004) 009 arXiv :hep-th/0406051.[166] Edward Witt<strong>en</strong>, Perturbative Gauge Theory As A String Theory In Twistor Space,Commun.Math.Phys. 252 (2004) 189-258, arXiv :hep-th/0312171.[167] Mina Aganagic, Hirosi Ooguri, Natalia Saulina, Cumrun Vafa Black Holes, q-Deformed 2d Yang-Mills, and Non-perturbative Topological Strings, Nucl.Phys. B715(2005) 304-348, hep-th/0411280.[168] Cumrun Vafa, Two Dim<strong>en</strong>sional Yang-Mills, Black Holes and Topological Strings,hep-th/0406058.[169] H. Morton, S. Lukac, The HOMFLY polynomial of the decorated Hopf link, Journalof Knot Theory and Its Ramifications 12 (2003) 395, math.GT/0108011.[170] M. Aganagic, M. Marino and C. Vafa, All loop topological string amplitudes fromChern-Simons theory, hep-th/0206164.[171] M. Marino and C. Vafa, Framed knots at large N, hep-th/0108064.[172] Seungjoon Hyun and Sang-Heon Yi, Non-compact Topological Branes on Conifold,arXiv :hep-th/0609037v2.[173] C. Vafa, Superstrings and topological strings at large N, J. Math. Phys. 42 (2001)2798-2817, hep-th/0008142.[174] Dagan Karp, Chiu-Chu Melissa Lui and Marcos Marino, The Local Gromov-Witt<strong>en</strong>Invaraints of Configurations Of Rational Curves, arXiv :math/0506488v2.[175] I. G. Macdonald, Symmetric functions and Hall polynomials, Oxford MathematicalMonographs, 1979.[176] S. Gukov, A. S. Schwarz and C. Vafa, Khovanov-Rozansky homology and topologicalstrings, Lett. Math. Phys. 74, 53 (2005), arXiv :hep-th/0412243.[177] Masato Taki, Refined Topological <strong>Vertex</strong> and Instanton Counting,JHEP0803 :048,2008 arXiv :hep-th/0710.1776.[178] Masato Taki, Flop Invariance of Refined Topological <strong>Vertex</strong> and Link Homologies,arXiv :0805.0336v1.[179] A. Iqbal, C. Kozcaz and K. Shabbir, Refined Topological <strong>Vertex</strong>, Cylindric Partitionsand the U(1) Adjoint Theory, arXiv :0803.22600.282


BIBLIOGRAPHIE[180] H. Awata and H. Kanno, Refined BPS state counting from Nekrasov’s formula andMacdonald functions,” arXiv :0805.0191.[181] G. E. Andrews, G<strong>en</strong>eralized Frob<strong>en</strong>ius partitions, Memoirs of the AMS (301) 1984.[182] Andrews G E, The Theory of Partitions Encyclopedia of Mathematics and its Applications,Vol 2 (Reading : Addison-Wesley), 1976.[183] A. Okounkov, N. Reshetikhin, Correlation functions of Schur process with applicationto local geometry of a random 3-dim<strong>en</strong>sional Young diagram, math.CO/0107056.[184] Ville Muston<strong>en</strong> and R. Rajesh, Numerical Estimation of the Asymptotic Behaviourof Solid Partitions of an Integer, cond-mat/0303607v1.[185] O. Foda, M Wheeler, BKP plane partitions, Journal-ref : JHEP 0701 (2007) 075,math-ph/0612018.[186] A. Yu. Orlov, Hypergeometric functions related to Schur Q-polynomials and BKPequation, math-ph/0302011.[187] Mirjana Vuleti, A G<strong>en</strong>eralization of macmahon formula, math.Co/0707.0532.[188] Mirjana Vuleti, Shifted schur process and asymptotics of large random strict planepartitions math-ph/0702068.[189] Lalla Btissam Drissi, Houda Jehjouh, El Hassan Saidi, Refining the Shifted Topological<strong>Vertex</strong>, J.Math.Phys.50 :013509,2009, arXiv : 0812.0513.[190] M. Vuletić, Shifted schur process and asymptotics of large random strict plane partitionsmath-ph/0702068.[191] B<strong>en</strong>jamin Young, with an app<strong>en</strong>dix by Jim Bryan, G<strong>en</strong>erating functions forcolored 3D Young diagrams and the Donaldson-Thomas invariants of orbifolds,arXiv :0802.3948.[192] Atkin A O L, Bratley P, Macdonald I G andMcKay J K S, Some computations form-dim<strong>en</strong>sional partitions, Proc. Cambridge Phil. Soc. 63 1097, 1967.[193] Sagan, Bruce E, The Symmetric Group Repres<strong>en</strong>tations Combinatorial Algorithmsand Symmetric Functions Brooks(Cole Publishing Company. 1991).[194] Stanley R P, Enumerative Combinatorics, Vol. 2 (Cambridge : Cambridge UniversityPress), 199.[195] MacMahon, P. A, Partitions of numbers whose graphs possess symmetry, Trans. CambridgePhil. Soc. 17, 149-170, (1898).283


BIBLIOGRAPHIE[196] George E.Andrews, MacMahon’s conjecture on symmetric plane partitions, Proc.Nati. Acad. Sci. USA Vol. 74, No. 2, pp. 426-429, February 1977.[197] Mirjana Vuletić, A G<strong>en</strong>eralization of MacMahon’s Formula,arxiv.org/abs/0707.0532v1.[198] Macdonald, IG Symmetric Functions and Hall Polynomials, Oxford Sci<strong>en</strong>ce Publication,1995.[199] A. Okounkov, Random partitions and instanton counting, math-ph/0601062.284


UNIVERSITÉ MOHAMMED V – AGDALFACULTÉ DES SCIENCESRabatDOCTORATRésumé de la ThèseDiscipline :PhysiqueSpécialité :Physique des Hautes EnergiesUFR :Physique des Hautes EnergiesResponsable de l’UFR : El Hassan SaidiPériode d’accréditation : 2005/2009Titre de la thèse :<strong>Contributions</strong> à l’Etude <strong>du</strong> <strong>Vertex</strong> <strong>Topologique</strong><strong>en</strong> Théorie des CordesPrénom, Nom : Houda JehjouhRésumé :Le travail de recherche prés<strong>en</strong>té dans ce mémoire de thèse porte sur l'étude<strong>du</strong> vertex topologique <strong>en</strong> théorie des cordes et la relation de <strong>du</strong>alité « théoriede cordes topologiques type A sur C 3 / modèle statistique de fusion descristaux utilisant la matrice de transfert. Nous comm<strong>en</strong>çons par intro<strong>du</strong>ire lesgéométries toriques des variétés de Calabi-Yau, la théorie des cordestopologiques et les invariants topologiques. Par suite, nous étudions lemodèle de fusion des cristaux et son li<strong>en</strong> avec la théorie des cordestopologiques tout <strong>en</strong> prés<strong>en</strong>tant une de nos contributions originales <strong>en</strong>relation avec la fonction de MacMahon généralisée. Nous exposons <strong>en</strong>suiteles vertex topologique standard et raffiné qui permet de trouver exactem<strong>en</strong>tl'amplitude de la théorie des cordes topologiques que nous illustrons àtravers des exemples avec un acc<strong>en</strong>t sur des invariants topologiques bi<strong>en</strong>connus à l’instar des invariants de Gromov-Witt<strong>en</strong> et Gopakumar-Vafa.Nous terminons par exposer nos résultats concernant le vertex topologiqu<strong>en</strong>on planaire ainsi que le vertex topologique raffiné-shifté.Mots-clefs (5) : Théorie des cordes, Variétés de Calabi-Yau et Géométrietorique, Modèle de fusion des cristaux, <strong>Vertex</strong> topologiques standard etraffiné, Fonction de MacMahon et Diagrammes de Young.Faculté des Sci<strong>en</strong>ces, 4 Av<strong>en</strong>ue Ibn Battouta B.P. 1014 RP, Rabat – MarocTel +212 (0) 37 77 18 34/35/38, Fax : +212 (0) 37 77 42 61, http://www.fsr.ac.ma

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!