The particle-number fluctuations originated from collective excitations are investigated for a three-dimensional, repulsively interacting Bose-Einstein condensate (BEC) confined in a harmonic trap. The contribution due to the quantum depletion of the condensate is calculated and the explicit expression of the coefficient in the formulas denoting the particle-number fluctuations is given. The results show that the particle-number fluctuations of the condensate follow the law $〈\ensuremath{\delta}{N}_{0}^{2}〉\ensuremath{\sim}{N}^{22/15}$ and the fluctuations vanish when temperature approaches the BEC critical temperature.
In this paper we show that a normal total number-of-particle fluctuation can be obtained consistently from the static thermodynamic relation and dynamic compressibility sum rule. In models using the broken U(1) gauge symmetry, in order to keep the consistency between statics and dynamics, it is important to identify the equilibrium state of the system with which the density response function is calculated, so that the condensate particle number
N0, the number of thermal depletion particles $\stackrel{\ifmmode \tilde{}\else \~{}\fi{}}{N}$, and the number of noncondensate particles
NNC can be unambiguously defined. We also show that the chemical potential determined from the Hugenholtz-Pines theorem should be consistent with that determined from the equilibrium equation of state. The
N4∕3 anomalous fluctuation of the number of noncondensate particles is an intrinsic feature of the broken U(1) gauge symmetry. However, this anomalous fluctuation does not imply the instability of the system. Using the random phase approximation, which preserves the U(1) gauge symmetry, such an anomalous fluctuation of the number of noncondensate particles is completely absent.
We calculate numerically and analytically the grandcanonical fluctuation of the number of particles both, in the fermionic condensate and above it, for ideal Fermi systems of constant density of states. We compare the canonical fluctuations, obtained from the equivalent Bose condensate fluctuation, with the grandcanonical fermionic calculation. The fluctuations of the condensate are almost the same in the two ensembles with a small correction coming from the total particle number fluctuation in the grandcanonical ensemble. On the other hand, well below the condensation temperature, the number of particles above the condensate and its fluctuation are insensitive to the choice of ensemble.