Quoting Birrell and Davies on this point. (They are actually talking about a very closely related situation - radiation from a moving mirror.)
"One trajectory of interest is uniform acceleration... it follows that <Tμν>ren = 0, which implies that no energy at all is radiated... Clearly the mirror emits particles... and it is also clear that a particle detector would respond to the presence of quanta. Nevertheless no energy is radiated. This example beautifully illustrates the looseness of the relation between particles and energy-momentum. The presence of quanta need not imply the presence of energy.
There appears to be a paradox concerning how the particle detector can, in the absence of field energy, absorb quanta and make a transition to an excited state. The resolution is that in so doing, the detector emits negative energy into the field to compensate. The emission of negative energy by mirrors and detectors is not without precedent in quantum field theory..."
Elsewhere they say:
"The emission of negative energy is a purely quantum phenomenon. It opens up the possibility of unusual new physical processes not encountered in classical theory. Negative energy fluxes play a role in the quantum evaporation of black holes. It might be supposed that a moving mirror could be used to violate the second law of thermodynamics by radiating negative energy into a hot body, thereby cooling it. However, if [the flux] is integrated over time between periods of stasis, it is always positive definite."