Quantization of EM waves I am trying to work through some problems on how the spins of Spin-Correlated Radical Pairs are influenced by external electromagnetic radiation. I am aware that EM radiation can be written in a quantum mechanical form using electric and magnetic field operators instead of the definite fields that we work with in classical mechanics. However, if the radiation is continuous and has high luminous flux, can I get a satisfying approximation by using the classical description of EM waves?
To put it succinctly - if continuous and bright radiation is being applied, and we want to measure its effects on a microscopic system, is it necessary to use the quantized description of light to get accurate results?
 A: If you are dealing with bright light, meaning that the number of photons within the shortest time interval of interest is large, then a classical model of light is usually sufficient for treating the internal dynamics of the spins. Quantum mechanically, light energy is discretized with energy $h c/\lambda$, but when the number of photons is large, you can approximate the discrete allowed values of energy as a continuum and still get correct answers.
Note that there are situations when intense light might need a quantum mechanical description (such as when you want to understand the properties of scattered radiation, or when dealing with non-classical states of light like squeezed states). But you typically don't need to deal with quantized light fields to model the internal dynamics of atoms, even if the radiation is weak. You just need that the incident radiation is classical i.e., in a thermal or coherent state (the output of a light bulb or laser, respectively).
Technically deriving spontaneous emission requires a quantum mechanical treatment of light, but the results of such derivations can be added to a model by hand.
There are all sorts of caveats and specific situations in which you might not be able to get away with classical light after all; it depends a lot on what you are trying to calculate and how accurate you want to be. But a classical approximation certainly sounds like a good starting point in your case.
