Models of inflation and the spectral index of the adiabatic density perturbation

Abstract

If an adiabatic density perturbation is responsible for large scale structure and the cmb anisotropy, its spectral index $n$ will be measured in the forseeable future with an accuracy $\Delta n\sim .01$. This is precisely the kind of accuracy required to distinguish between many models of inflation. Most of them have an inflationary potential $V\simeq V_0(1\pm\mu\phi^p)$ with the constant term dominating. Except for $0\lsim p\lsim 2$, the prediction is $n=1\pm (2/N)(p-1)/(p-2)$, where $N$ is the number of $e$-folds of inflation after cosmological scales leave the horizon. It typically lies in the range $0.9\lsim n\lsim 1.1$. For $p=2$ one has $n=1\pm 2\Mpl^2 \mu$ where $\Mpl=(8\pi G)^{-1/2}$. A generic supergravity theory gives contributions of order $\pm H^2$ to the inflaton mass-squared $m^2$ whereas inflation requires $|m^2| << H^2$. Proposals for keeping $m^2$ small are surveyed with emphasis on their implications for $n$. Finally, the case of a multi-component inflaton (as in `double' inflation) is discussed. In all cases, the best method of calculating the spectrum of the adiabatic density perturbation starts with the assumption that, after smoothing on a super-horizon scale, the evolution of the universe along each comoving worldline will be practically the same as in an unperturbed universe with the same initial inflaton field. The possible isocurvature density perturbation is briefly discussed, with a simple derivation of the fact that the low multipoles of the cmb anisotropy are six times as big as for an adiabatic density perturbation.