Spontaneous emission of energy is a radiative process involving the release of a photon and typified by phenomena such as fluorescence and phosphorescence. Einstein showed that the probability of spontaneous emission is strongly frequency-dependent (proportional to f³). In the visible region of the spectrum, f ≈ 10¹² Hz, and spontaneous emission is a dominant process. In the radiofrequency range where NMR energies are found, however, f ≈ 100 MHz, and so spontaneous emission becomes extremely improbable.
Here is another way to think about this concept: The energy gap between the spin-up and spin-down states in NMR is really quite small by atomic emission standards — at 1.5T it is only about 2 x 10−7 eV (electron-volts). By comparison, visible light photons have energies of about 2 eV, or 10 million times higher. There is thus a considerable "advantage" for a high-energy light photon to be emitted by phosphorescence, but relatively little "motivation" for an already low energy nuclear spin to switch states spontaneously.
Advanced Discussion (show/hide)»
From Einstein's theory, the probability (P) of spontaneous energy emission of an ensemble of N> spin-1/2 magnetic moments in free space is given by
where μo is the permeability of free space, ω is the angular frequency, γ is the gyromagnetic ratio, h is Planck's constant, and c is the speed of light. For 1 ml of water at 1 T this translates into spontaneous emission of only 0.024 photons per second, an extremely low number.
There is still considerable controversy as to exactly how the energy absorbed by nuclear spins during NMR is released and induces a current in the receiver coil. The traditional answer "photon emission with radio waves" is almost certainly incorrect. See the article by David Hoult below for speculations about "near field effects", virtual photons, and explanations within the field of quantum electrodynamics (QED).
Hoult DI. The origins and present status of the radio wave controversy in NMR. Concepts Magn Reson Part A 2009; 34A:193-216
If a system seeks to minimize its total energy level, why don't all the protons simply fall into the lower energy state?