36. Eye on ArXiv: 6 May 2008 – Where does the cosmological perturbation theory actually break down?

I am going to briefly discuss today the paper by Cristian Armendariz-Picon, Michele Fontanini, Riccardo Penco, and Mark Trodden “Where does cosmological perturbation theory break down?”

The issue of possibility to describe inflation in terms of effective QFT keeps bothering people’s minds Let me remind you that two weeks ago we had a paper by Steven Weinberg discussing the similar subject. What is the essence of the problem that Christian et al. found? Naively, it seems that as long as , dynamics of inflaton + metric perturbations can be described in terms of QFT (canonically normalized scalar field with some potential). Well, not quite so, as the authors say.

If you start to take effective QFT tree level corrections into account, the correlator for Mukhanov variable will have the form

(1)

etc., so the actual smallness parameter in the effective QFT expansion should not be , but something like instead.

I believe this fact by itself is not new. Take a correlation function for curvature perturbation . Apart from the standard term there are corrections to this correlator small with respect to the parameter

(2)

where is the number of efoldings (see our paper for example as well as recent works by David Seery, David Lyth etc.) These corrections grow while the number of efoldings grows and the expansion breaks down sooner or later. Now, if we take a look at the formula (1), we see that the smallness parameter there looks remarkably similar to (2). Indeed, the number of efoldings is defined as

where is the slow roll parameter.

What is the physical meaning of the effective QFT breakdown the authors found? The point is that the number of efoldings cannot be very large, and if it is, then inflation enters eternal self-reproducing regime. There the bare notion of homogeneous isotropic FRW backround breaks down (eternally inflating Universe consists of different Hubble patches, and the distribution of among them is highly inhomogeneous). Still, the theory is under control even in this regime (instead of effective QFT, we can describe dynamics of the inflaton field by stochastic formalism).

By the way, authors say on the p. 7 that by in-vacuum they mean Bunch-Davies. This is probably a misprint: in-vacuum corresponds to the absence of particles at , while Bunch-Davies is Euclidean vacuum corresponding to the linear combination of in- and out-modes.

P.S. There is also very nice description of the Schwinger-Keldysh formalism in the paper.