We consider a family of non-supersymmetric MQCD five-brane configurations introduced by Witten, and discuss the dependence of the curves on the microscopic theta angle and its relation with CP. We find evidence for a non-trivial spectral flow of the curves (vacua) and for the level-crossing of adjacent curves at a particular value of the theta angle, with spontaneous breaking of CP symmetry, providing an MQCD analogue of the phase transitions in theta proposed by 't Hooft.
November 1997The modelling of gauge dynamics via brane configurations of weakly coupled string theory [1,2,3,4], or M-theory [5,6,7,8,9], provides a geometrical interpretation of various field theoretical strong coupling phenomena. In some cases, the geometrical viewpoint can be used to discover new exact solutions [5] of N = 2 theories. In the context of N = 1 theories, it provides a semiclassical approach to such thorny problems as confinement and chiral symmetry breaking [6,7]. These constructions are similar in spirit to previous work on geometric engineering of gauge theories (see for example [10]). Here the role of a nontrivial string compactification is played by some complicated brane configuration sitting in flat space at small string coupling, and carefully adjusted to provide a weakly coupled supersymmetric Yang-Mills theory at intermediate scales. Since the gauge theory lives on the branes world-volume, one can relate many classical and semiclassical features of gauge theories to some properties of brane dynamics. The real power of the method arises when taking the strong coupling limit of the brane configuration. If a strong coupling dual is available, one can describe many non-perturbative, infrared properties of the gauge theory, just by reading the tree-level data of the dual brane configuration.In four-dimensional models one uses the duality between type IIA strings and Mtheory [5]. Here one maps the type IIA brane configuration to a single smooth five-brane, whose world-volume is appropriately embedded in the flat eleven-dimensional background, as a product M 4 × Σ, with M 4 the four-dimensional Minkowski space, and Σ a holomorphic curve with respect to a given complex structure of the background. In general, the detailed physics of the resulting M-theory model differs from the Yang-Mills theory we are interested in; however, to the extent that some observables are protected by supersymmetry, we can calculate them in the deformed strong coupling model. This is the case, for example, of all holomorphic quantities determined by the Seiberg-Witten curve in N = 2 models [5], including BPS spectra [11]. For N = 1 models the agreement is only qualitative in principle, but some observables can be accurately matched, like superpotentials and gaugino condensates [6,8].In the absence of some unbroken supersymmetry on the world-volume, we cannot use holomorphy to accurately match observables, and a working assumption must be made that no phase transitions occur in the way to strong coupling. Still, qualitative features based o...