To analyse the diffractive properties of photons, one is forced to use a wave model. Photons, of course, exhibit both wave-like and particle-like properties, depending on the kind of observation that is being made and depending on the particular kind of interaction they are involved in. The same can be said for literally any particle, of course.Yes it's interesting to think about what it can actually mean to speak of the "size" of a photon. It's obviously true that EM waves diffract round objects and the degree to which they can is a function of wavelength. And, in the extreme case of a single photon (e.g. in the double slit experiment), even one individual photon can diffract in this way.
That is true for a monochromatic wave, but a monochromatic wave cannot behave like a particle. If we want to model a photon as a particle that has a localised position in space then we must model it not as a monochromatic wave but as a wave packet consisting of a bunch of different frequencies. The spread of frequencies used to construct the wave packet is then related to how "wide" the wavepacket is in space - which is how the Heisenberg uncertainty principle comes about.But surely, according to the Uncertainty Principle, a single photon that is truly monochromatic can have no particular position in space: it can be detected with equal, infinitesimally low probability, anywhere along its direction of travel.
Yes. This is the same as what I just wrote, above.So its "size", in that sense, is sort of infinite. Whereas a photon whose probability of being detected has a limited spatial extent must be non-monochromatic, being composed of a Fourier sum of different wavelengths that interfere constructively only in a limited region of space (i.e. wave packet) and thereby must have an uncertainty in its momentum (and energy).