Title:Is the excitation spectrum of an interacting Bose gas really gapless ?

Abstract: In Bogoliubov's theory of an interacting Bose gas, one does not diagonalize the part $\hat{H}_2$ of the Hamiltonian that is quadratic in the boson creation and annihilation operators, but instead one diagonalizes the "grand canonical" Hamiltonian $\tilde{H}_2 = \hat{H}_2 - \mu \hat{N}_1$, where $\mu$ is the chemical potential and the operator $\hat{N}_1$ counts the number of bosons outside the condensate. In this paper, we argue that the ground state energy and other thermodynamic quantities should be derived by diagonalizing $\hat{H}_2$ rather than $\tilde{H}_2$, which results in a gapped excitation spectrum. Gapped excitation spectra were derived previously, by a number of authors, and were deemed unphysical on the grounds that such spectra violate Goldstone's theorem. Here, we argue that such an interpretation stems from an improper identification of the {\em collective} Goldstone modes of the system, which have to do with the phase of the order parameter, whereas the spectrum of $\hat{H}_2$ represents single-particle excitations.