Can energy be taken out of the QFT vacuum? There have been recent questions about the vacuum. In my simplified knowledge the vacuum is like a ground state energy level, and also that there might even exist other lower energy levels than the vacuum we find ourselves in. The sea is a soup of created and annihilated pairs of virtual particles with virtual energies and momenta.
In a normal sea on earth, which also represents a ground state of water, energy can be taken out of random waves by the clever construction of valves that allow only one way motion of water. Is it conceivable that a  gadget of similar function could be found for the vacuum sea, or is it forbidden by conservation laws?
My intuition tells me that it might be possible if GR is taken into account, but my physics knowledge does not stretch to support this.
Edit : An explanation of why I am asking this question:
Let me expand on the example of the sea. The energy from the waves comes from either tides, i.e. gravitational forces, or wind (temperature differentials). If these were missing the oceans would be like glass representing a unique ground state of the gravitational well of the earth.
In an analogy, a gravitational wave going through the vacuum  would be supplying energy to the vacuum sea.
I have been thinking of this analogy ever since cold fusion surfaced and refuses to die out, the most recent one being discussed here too. Approached from nuclear physics orders of magnitude the claims seem preposterous. There are people though who believe they have  results of extra energy over input energy, much more than chemical reactions could supply.
This set me thinking on vacuum energy and the analogy with getting energy from the sea. A crystal is a prime candidate for any exploration of such concepts and in all cold fusion "successful" results crystals have been used. Now if the effect depended on the vacuum and how much distorted it was by a passage of a gravitational wave at the time of the experiment, or the exact orientation of the crystal, or the type of impurities in the crystal ( F centers etc) one would expect to get haphazard results, and non repeatable by other experimenters.
Of course this would be the first experimental evidence of gravitational waves :).
Edit 20/7/12
Maybe I should clarify that an acceptable answer in the negative would be one based on conservation laws. I believe that data trumps theory, and next in line are conservation laws,because they are the distillation of an enormous amount of data. Some people seem to think that theoretical definitions can substitute for proof in physics, but physical theories change, solid data do not, and this is physics, not axiomatic mathematics.
 A: No you cannot take energy out of the vacuum, BY DEFINITION.
Using this analogy with a "sea" is nonsense, because the statement "vacuum is a sea of virtual particles" is also an ambiguous statement not to be taken literally.
The definition of vacuum is ground-state energy, and so if you could take energy away from it, then the energy of the vacuum would be lowered, contradicting the fact that it's already in the lowest possible state.
This is also the reason why spontaneous absorption does not occur with electrons/atoms (whereas spontaneous emission can occur).

I want to point out why this question is not closed yet:
Anna's train of thought is "The answer does not satisfy me because it is dependent on definitions which change as theories change."
BUT, the question itself is dependent on definitions... you can't ask about apples and then say 'well maybe they can be oranges'. If you can pose a question about 'energy' and 'vacuum' then they have to be defined.
I now vote to close!
A: Peter Milonni has devoted a lot of work to this area, i will recommend this book: The Quantum Vacuum, it basically gives a overview on lot of physical systems where vacuum field effects are dominant, and how certain boundary conditions can effect them to produce unexpected effects
the expansion of fields around the vacuum in harmonic modes with quantized particle population of the modes is only valid when the physics can be fairly approximated in terms of the vacuum eigenmodes (which are the regular field wave functions). Of course, when there are special boundary conditions, this doesn't change too much since we still have eigenmodes (which correspond to the boundary geometry) but still can talk about particle populations
when the system (meaning, the boundary conditions) are dynamic (moving mirrors, or superconductor walls moving over the superconductivity phase) i don't think there is an authoritative answer about the validity of these approximations. So in short, there could be interesting things to say about the vacuum dynamics when a more friendly framework exists to make computations in highly dynamical regimes
I made a somewhat relevant question a while ago about casimir walls that melted and froze back with an oscillatory transversal magnetic field.
i'm sorry, i'm not addressing your larger question, as to what relevant things change when taking GR in consideration. Hopefully someone more knowledgeable on the matter will. 
A: Well, let us be honest here. This question was not supposed to be answered from the very beggining. 
First of all we don't know how to mix quantum mechanics and gravity. There is no good consistent theory for that.
Another thing is that today "the Dirac sea" analogy is not considered to be a very good thing. It is a pre-QFT naiive picture.
Finally we are supposed to talk about "waves" in this "sea"... While the "sea" itself is an obsolete analogy... And all that is in a context of non-existing theory... Come on...
Now. There is actually a formal way to answer the question, because the question is about the "QFT vaccuum". 
And the QFT vacuum have a precise definition. Which basically says that it is "something you cannot take energy from".
Actually, we start from that definition to build QFT. So the answer is: "you cannot take energy out of the QFT vacuum by definition".  
Maybe we are wrong to start from this definition. Maybe for quantum gravity we need another starting point. 
But then it wouldn't be a QFT -- it would be a new theory which will have QFT as a limiting case.
And in the range of validity of QFT that "formal" answer will hold.
A: In theory, QFT vacuum state is isotropic and invariant for all observers, as the base of all ladder operator algebras. It does not have any features other than the fluctuations. Hence, you cannot extract any more energy from it.
In practice, extracting energy out of the vacuum is not only possible, but it has been already achieved in laboratory using optical parametric amplifiers (NOPA); they are used for squeezing input light, but when there is nothing in the input, the output is a field that has lower average standard deviation of the energy than the normal vacuum in the range of frequencies where the amplifier is active. If we assign to the normal vacuum zero energy, then this "squeezed vacuum" field must have negative energy.
Of course, let's make perfectly clear that this mechanism cannot be used to extract useful energy, since the energy spent pumping the amplifiers greatly exceed anything that you could extract.
“In theory, theory and practice are the same. In practice, they are not.”
― Albert Einstein
A: There has been a proposal to use Casimir force and meta materials to build a vacuum energy extractor. Casimir force applied on parallel plates made of "normal" material pushes the plates outwardly and inwardly for metamaterials. I do not know if the idea has yet been tested.
http://physicsworld.com/cws/article/news/2007/may/02/casimir-force-could-drive-tiny-ratchets
A: The ground state (=vacuum) is an eigenstate of the system Hamiltonian. The wave function is unique, no energy uncertainty exist. In this sense there is no fluctuation of energy to harvest. "Fluctuations" exist for non commuting with the Hamiltonian variables but it cannot be used to get some energy for the reason given above.
In GR the energy is not conserved and this is not connected to the vacuum.
Edit: in heat machines one tries to increase the temperature difference $T_{hot}-T_{cold}$ to increase the machine efficiency. Vacuum is the coldest body ever known. So the vacuum "cleaner" efficiency cannot exceed zero.
A: The vacuum state has a property named passivity, which means that any local operation onto the vacuum state does not extract but inject energy to the system. This implies that  energy cannot be taken out of the vacuum only by local operations. However, if we adopt local operations and classical communication (LOCC), a part of zero-point energy of the vacuum state can be extracted. The scheme is called quantum energy teleportation. More information is available in wikipedia and a review article by Hotta, who first proposed the concept:
http://www.tuhep.phys.tohoku.ac.jp/~hotta/extended-version-qet-review.pdf .  
