Closed String Tachyons, AdS/CFT, and Large N QCD

TL;DR

This paper investigates the instabilities caused by tachyons in AdS/CFT models with orbifolds, analyzing how these lead to gauge theory decays and the resulting IR theories, using effective potentials and large-N behavior.

Contribution

It generalizes previous calculations to show how tachyonic instabilities in orbifold AdS/CFT correspond to gauge theory decays, involving double-trace operators and large-N analysis.

Findings

Small radius instabilities correspond to gauge theory decays.

Decays lead to IR theories with fewer degrees of freedom.

No such instability in freely acting orbifolds.

Abstract

We find that tachyonic orbifold examples of AdS/CFT have corresponding instabilities at small radius, and can decay to more generic gauge theories. We do this by computing a destabilizing Coleman-Weinberg effective potential for twisted operators of the corresponding quiver gauge theories, generalizing calculations of Tseytlin and Zarembo and interpreting them in terms of the large-N behavior of twisted-sector modes. The dynamically generated potential involves double-trace operators, which affect large-N correlators involving twisted fields but not those involving only untwisted fields, in line with large-N inheritance arguments. We point out a simple reason that no such small radius instability exists in gauge theories arising from freely acting orbifolds, which are tachyon-free at large radius. When an instability is present, twisted gauge theory operators with the quantum numbers of the large-radius tachyons aquire VEVs, leaving a gauge theory with fewer degrees of freedom in the infrared, analogous to but less extreme than ``decays to nothing'' studied in other systems with broken supersymmetry. In some cases one is left with pure glue QCD plus decoupled matter and U(1) factors in the IR, which we thus conjecture is described by the corresponding (possibly strongly coupled) endpoint of tachyon condensation in the M/String-theory dual.