Dans cette thèse, on étudie les C2-feuilletages de codimension 1, dans les 3-variétés compactes connexes et orientables. Il est bien connu que l’on peut construire explicitement sur de telles variétés un feuilletage qui possède des composantes de Reeb. Vient alors le problème crucial d’existence des feuilletages tendus (toujours ouvert).Rappelons qu’un feuilletage tendu n’admet pas de composante de Reeb, mais que la réciproque est fausse.La première partie de ce travail, consiste à comprendre la différence entre un feuilletage non-tendu sans composante de Reeb et un feuilletage tendu. On verra que l’orientation transverse des feuilles toriques joue un rôle crucial, en donnant une condition nécessaire et suffisante sur cette orientation transverse pour qu’un feuilletage soit tendu. Pour cela on étudiera de près les procédés géométriques de tourbillonement et de spiralement, et on montrera qu’ils apparaissent toujours au voisinage d’une feuille torique.La seconde partie de ce travail se concentre sur le problème d’existence de feuilletages tendu. Rappelons que depuis les travaux de D. Gabai [1983], on sait que si une 3-variété admet une homologie non-triviale, alors elle admet un feuilletage tendu. Mais le problème d’existence est toujours ouvert parmi les sphères d’homologies, et on s’intéresse ici à celles qui sont fibrées de Seifert. On montre que toutes les sphères d’homologie entière fibrées de Seifert sauf S3 et la sphère d’homologie de Poincaré admettent un feuilletage tendu. Par contre, parmi les sphères d’homologie rationnelle non-entière, fibrées de Seifert, il existe une infinité de telles variétés qui admettent un feuilletage tendu, et une infinité qui n’en admettent pas. / In this thesis, we study codimension 1, C2-foliations, in compact, connected and orientable 3-manifolds. It is well known that we can explicitly construct on such manifolds a foliation admitting Reeb components. Then comes the crucial problem of existence of taut foliation (still opened).Recall that a taut foliation does not admit a Reeb component, but the converse is false. The first step of this work focuses on the difference between a non-taut and Reebless foliation, and a taut foliation. We will understand that the transverse orientation of the torus leaves plays a key-role, by giving a necessary and sufficient condition on the transverse orientation, for a foliation to be taut. For this, we will study the geometric processes of turbulization and spiraling with generalizations, and we see that they always appear in a neighborhood of a torus leaf.The second step of this work is concentrated on the problem of existence of taut foliations. Recall that since the work of D. Gabai [1983], we know that if a 3-manifold has non-trivial homology, then it admits a taut foliation. This problem is still opened among homology spheres and we focus here on Seifert fibered ones. We show that all Seifert fibered integral homology spheres (but S3 and Poincar ́e homology sphere) admit a taut foliation. Nevertheless, among Seifert fibered rational (and non-integral) homology spheres, there exist infinitely many which admit a taut foliation and infinitely many which do not admit one.

Made available in DSpace on 2014-07-29T16:02:15Z (GMT). No. of bitstreams: 1 Dissertacao Wender J de Souza.pdf: 1008440 bytes, checksum: b2d3405f265353a21b9eaaad1c91d71f (MD5) Previous issue date: 2010-03-29 / In this work we are interested in the solution of the following problem: Let Y = ( f ,g) be a vector field of class C1 in R2. Suppose that (x, y) = (0,0) is a singular point of Y and assume that for any q &#8712; R2, the eigenvalues of DY have negative real part, this is, det(DY) > 0 and tr(DY) < 0. Then, the solution (x, y) = (0,0) of Y is globally asymptotically stable. To this end, we show that this problema is equivalent to the following: Let Y : R2 &#8594;R2 be a C1 vector field. If det(DY) > 0 and tr(DY) < 0, then Y is globally injective. This equivalence was proved by C. Olech [1]. So we show the injectivity of the vector field Y under the conditions det(DY) > 0 and tr(DY)<0. In fact, we present a more stronger result, which was obtained by C. Gutierrez and can be found in [4]. This result is given by: Any planar vector field X of class C2 satisfying the r-eigenvalue condition for some r &#8712; [0,¥) is injective. / Neste trabalho, estamos interessados em estudar a solução do seguinte problema: Seja Y = ( f ,g) um campo de vetores, de classe C1, em R2. Suponha que (x, y) = (0,0) é um ponto singular de Y e suponha que, para todo q &#8712; R2, os autovalores de DY tem parte real negativa, isto é, det(DY) > 0 e tr(DY) < 0. Então, a solução (x, y) = (0,0) de Y é globalmente assintoticamente estável. Para este fim, mostramos que este problema é equivalente ao seguinte: Seja Y : R2 &#8594;R2 uma campo de vetores de classe C1. Se det(DY) > 0 e tr(DY) < 0, então Y é globalmente injetora. Esta equivalência foi demonstrada por C. Olech em [1]. Desta forma, a estratégia é estudar a injetividade do campo Y sob as condições det(DY)> 0 e tr(DY) < 0. Na verdade, apresentamos um resultado um pouco mais forte, o qual foi obtido por C. Gutierrez e pode ser encontrado em [4]. Este resultado é dado por: Qualquer campo de vetores X : R2 &#8594;R2 de classe C2 satisfazendo a condição de r-autovalor, para algum r &#8712; [0,¥), é injetora.