Torsion in gauge theory

The potential conflict between torsion and gauge symmetry in the Riemann-Cartan curved spacetime was noted by Kibble in his 1961 pioneering paper and has since been discussed by many authors. Kibble suggested that, to preserve gauge symmetry, one should forgo the covariant derivative in favor of the ordinary derivative in the definition of the field strength $F_{\mu \nu }$ for massless gauge theories, while for massive vector fields, covariant derivatives should be adopted. This view was further emphasized by Hehl et al. in their influential 1976 review paper. We address the question of whether this deviation from normal procedure by forgoing covariant derivatives in curved spacetime with torsion could give rise to inconsistencies in the theory, such as the quantum renormalizability of a realistic interacting theory. We demonstrate in this paper the one-loop renormalizability of a realistic gauge theory of gauge bosons interacting with Dirac spinors, such as the SU(3) chromodynamics, for the case of a curved Riemann-Cartan spacetime with totally antisymmetric torsion. This affirmative confirmation is one step toward providing justification for the assertion that the flat-space definition of the gauge-field strength should be adopted as the proper definition.