Hi Jeroen,
I don't want to argue, but rather try to understand.
You have obviously put some thought on the
light-matter interaction puzzles. However, as much
as I dislike the whizzing-ball picture of photons, I'm
wondering whether it is possible to discard the photon
concept (in the sense used by eg. Louisell, or Glauber)
in as radical a way as you seem to suggest.
For example, in superconducting transition edge
sensors the smallest excitation is *much* smaller
than the energy of absorbed photons. Then it must
be some other mechanism than the physics of the
*detector* which forces the EM energy being
absorbed in the lumps of the size omega hbar.
An example is the work of Aaron Miller et al
http://people.bu.edu/alexserg/APL_2003.pdf
From the detector physics point-of-view, they
*would be* capable to see two 0.4eV events, or
three 0.27eV events when shining a 1.55um laser
on the detector - but they never see those, only
single 0.8eV events. Or, more accurately: they
see the Poisson distribution, characteristic
of a coherent state being received.
Now, assume that a detector with a different
composition (say, an APD - although I don't
know offhand if they have sufficient energy resolution)
is detecting the same phenomenon. If it, too, sees
only single 0.8eV events and not any double
0.4eV events, wouldn't it be more natural
to ascribe the 'lumpedness' as a property of
the EM field rather than the property of the detector?
In particular, if you shorten the wavelength of the
illuminating laser (i.e. change something in the
transmitter side), you see that the size of absorbed
energy lumps goes up. Now, detector physics (including
the detectors excitation spectrum) supposedly does
not change, still the lump size it sees changes.
Wouldn't it in this case be more natural to associate
the lump size to either the transmitter or the EM field,
rather than to the detector? (OK, I suppose you can
answer 'it is the transmitter', but then a similar argument
about a transmitter of any physical composition
can be constructed).
I do agree that the photon absorption phenomena *could*
be explained by a postulated physical phenomenon which
always occurs at the light-matter interaction, and which
always gives the lump size of hbar-omega regardless
of the detector physics (solid state, gas ...) and regardless
of the range of energy excitations available in the detector
(semiconductor gap, superconducting gap, continuous
spectrum of thermal excitations...). But isn't such a
postulate much more awkward than accepting the
discreteness to be a feature of the EM field itself (in
the Louisell sense)?
I think single-photon sources *do* exist, but I better not
make that claim too strongly before refreshing my memory
on the subject. There has been a review, I must dig it out...
At least: non-poissonian photon sources *do* exist (e.g
bunched or anti-bunched); whether there are externally
triggerable ones nowadays I'm not sure.
I remember being intriguiged in late-90's by claims that
an ordinary LED, when driven by sub-poissonian current,
would act effectively as single photon emitter. The question
is of course how to create that sub-poissonian current,
in particular when there is the junction capacitance present.
I should dig out those old papers as well...
Regards,
Mikko