Following World War II, physicists involved in the development of radar for the war efforts turned their attentions to peacetime applications of pulsed radio frequency measurements. Indeed, MRI and NMR spectroscopy still share many central concepts with radar including pulsed RF energy, transmit/receive through the same antenna, and phased array reception. (NMR is Nuclear Magnetic Resonance, the same physical phenomenon used in MRI. The preceding "N" for nuclear was dropped by the medical profession to produce an acronym "MRI" without the supposed negative reference to nuclear radiation, nuclear energy, nuclear power plants, and so forth). Much progress was made toward basic science applications of NMR, but in the late 1960s, the convergence of three newly-developed technologies, superconducting magnets, digital computers, and the fast Fourier transform (FFT) set the stage for the emergence of an application having profound social and humanitarian impact - clinical MRI.
The physical basis of NMR was well understood by the late 1960s - in the presence of a strong magnetic field, specific nuclei can absorb and emit energy in quantities and at frequencies characteristic of the nucleus and proportional to the strength of the magnetic field. Several generations of NMR spectrometers had been used by chemists to resolve the differences in nuclear resonant frequency between like atoms in differing chemical combinations. The central paradigm of NMR was "chemical shift" - due to differing molecular configurations, nuclei experience differing magnetic fields and precess at resonant frequencies characteristic of their molecular composition. The NMR experiment consisted of placing the sample in a strong magnetic field, giving a burst of RF energy at the resonant frequency, sampling the absorbed RF as it was reradiated by the sample, and transforming the samples into a spectrum where each peak corresponds to a characteristic frequency.
From here on, we will consider only water protons as the resonant nucleus. From a chemist's point of view, water protons are not very interesting in themselves. Their relaxation rates (i.e. how quickly they return to equilibrium after RF irradiation), described by the phenomenological spin-lattice and spin-spin relaxation times T1 and T2, held less interest to chemists than chemical shift and quantum couplings. However, in the human body, water exists at a higher concentration than any other substance, and hydrogen (the proton) has the highest nuclear resonant frequency of commonly occuring atoms. This means that, when we choose to examine water protons, there are plenty of them around and, because of their high resonant frequency, they absorb and emit lots of energy when they resonate (due to the applied RF). Water signal can be created in abundance, although it's not so interesting if viewed in bulk.
With the insightful addition of some simple hardware - magnetic field gradients - to the NMR spectrometer, a machine designed to resolve chemical information was converted to a machine that resolves spatial information. For a more complete discussion of how the MRI scanner works, go to How the heck do they do that?
Even though the concentration of protons in the human body is high, protons will not resonate unless they experience a magnetic field. The stronger the magnetic field, the more energy that can be absorbed and then released as the MR signal. This brings us to the first of the three technologies, the superconducting magnet (pick the link for a more quantitative discussion). Prior to the discovery of superconduction, all conductors were thought to have a resistance; that is, conductors release part of the energy flowing through them as heat. Resistance is a characteristic of all materials, with high-resistance materials called insulators and low-resistance materials called conductors.
You can easily make a magnet from a spool of wire and a battery... connect both ends of the wire to the battery terminals and an electric current runs through the wire, creating a magnetic field aligned with the axis of the spool. Eventually the battery will run down and the magnetic field will collapse; the chemical energy in the battery has been converted to heat by the resistance of the wire. The magnets of early NMR spectrometers worked this way - they consumed large quantities of electricity, created lots of heat, and were only magnetic while electric current flowed. However, some materials at very low temperatures become superconducting... make the wire from a superconducting alloy, such as niobium-tin, place the wire spool in a tub of liquid helium, connect the battery, and a similar but stronger magnetic field will occur. Connect the ends of the wires to each other instead of to the battery (easier said than done), and the magnetic field will persist indefinitely, or at least until the helium boils away. The wire is now superconducting; it has no resistance and the current continues to flow in the wire indefinitely.
Clearly, the main benefit of superconduction is efficiency. In a resistive magnet, resistance limits the field strength by the heat that can be dissipated by the coil while it is at field (and, for those of you familiar with Ohm's law, by the voltages that can practically be generated). Such efficiency means that superconducting coils can be driven to much much higher current densities and proportional magnetic field strengths without self-destructing.
More to come ...
Sections on digital computers, and the fast Fourier transform (FFT) coming soon.