Two other heating mechanisms

In addition to the earlier addressed
dipole relaxation and ionic conductivity there are also:

Electronic conductivity – of the kind possessed by metals and other substances such as carbon (graphite, and in “black particulates”). An example is an object which has caught fire or begun to char: it can then be heated very efficiently by the microwaves. If the conductivity is high such as in homogeneous metals the electric field will, however, be short-circuited so that no heating occurs. But some self-heating of e.g. stainless steel plates carrying large currents may occur, in e.g. waveguides. Aluminium metal has several times higher conductivity and is heated much less, if at all.
Several solid substances – and increasingly many at temperatures > 200 °C – possess a combination of ionic and electronic conductivity. Semiconductors have partially different conduction mechanisms. All these kinds of conductivities increase with temperature, often quite much over a short temperature range.

Direct heating of magnetic materials can occur and is then by the same mechanism as the so-called hysteresis in e.g. iron at mains frequency. Magnetic materials which can absorb microwaves significantly are called ferrites and have a low electric conductivity. They convert the magnetic field directly into heat and are used as fillers in e.g.seals and wavetraps against microwave leakage. Direct magnetic heating occurs also in microwave processing of certain ores.

Non-thermal effects – or “just” heating?
So-called athermal microwave effects are often discussed, by media and groups of people who are against “electricity”. In human beings, the nerve signals are electricllay conducted in the whole body. Serious research has shown that there are situations when disturbances many occur in certain tissues, at exposure levels comparable to those which may cause inappropriate overheating. Almost all such disturbances, however, disappear entrirely when the the field is shut off. Microwaves are non-ionising, as opposed to e.g X-rays: any possible injuries are not successively added in time. However, there may still be a very small risk buildup during many years of direct exposure of the brain by intense mobile phoning. But the safety standards for microwave heating have always been more stringent than for mobile phones.  
As said earlier, the molecule groups move incessantly and then collide. The collisions are fully elastic, i.e. no energy is consumed – if that were the case, the movement would come to and end in a thermally insulated system which would then heat up “by itself”! In liquid water, the groups collide about once every picosecond (a millionth of a millionth second). Water molecules are exchanged between the groups at about every 1000th collision. Microwaves at 2450 MHz change direction every 200 picoseconds. Hence, collisions occur about 200 times between every field reversal!
The “geometric” effect of the turning energy, i.e. the actual alignment of the molecule groups, can be calculated and one then finds that 1000 watts in 100 grams of water (corresponding to 2,3 °C temperature rise per second) at room temperatuare causes a relative alignment of only 0,003 %! Hence, the energy that “a microwave” can add is almost insignificant in comparison with the inherent, own, and permanent energy of movement of the molecule groups.
It is concluded that the field strengths used in microwave heating result in only temperature increases.