A gluon propagator is a singular quantity (function, distribution) in general (say, as a function of momentum, for instance). In some cases one knows how to handle those singularities -- in covariant gauges (like Feynman or Landau) the prescription for making the singularities integrable (=treatable) is a causal (Feynman) interpretation of the poles. The choice of a particular gauge depends on the problem at hand and can simplify the solution a lot. For instance, covariant gauges are very convenient for perturbation theory computations but they are not explicitly unitary and, in this sense, are not "physical". A more "physical" one is a Coulomb gauge -- but it makes perturbation theory computation less convenient though can decrease the number of diagrams sometimes. There is a class of "fixed" gauges determined by the requirement nA=0, with n being a fixed four-vector. For n^2=0 it is light-cone gauge, n^2>0 is Hamiltonian (temporal), at n^2
I have not seen any gauge in which any vertex depends upon arbitrary parameter e.g., gauge fixing paramter. But what if some of the vertex say four gluon vertex depends on it? Pls comment on the perturbative applicability of the gauge. Can we 'tune' that parameter such that the expression for the vertex goes to zero identically?
For example: In covariant gauges four gluon vertex has a standard form as in common literature. Now suppose there is a gauge in which this four gluon vertex modifies to (standard form + arbitrary parameter).
Can we 'tune' that parameter such that (standard form + arbitrary parameter) = 0 ?