The principle of a laser is based on two separate features: a) a light emitting/amplifying medium and b) an optical resonator (usually defined by two parallel mirrors).
The physical process responsible for the light amplification is supposed to be the stimulated (induced) emission process which is assumed to occur in case of a population inversion between two atomic states in a radiation field of the corresponding frequency. This would lead to the atomic emission amplifying the incident radiation field exactly in phase (coherently) which would vastly increase the amplitude of the resultant wavetrain compared to an incoherent (random) superposition (because for N emitters it would be proportional to N rather than √N ).
Now there is in my opinion a fundamental conceptual problem with the assumption of the existence of a stimulated (induced) emission: atomic physics distinguishes two different mechanisms for radiative transitions between two levels i,k of an atom: a) spontaneous emission that occurs with a probability given by the decay constant Ai,k
, and b) induced emission or absorption due to an external radiation field. Resonant scattering is for instance usually considered as an absorption of a photon which lifts an electron to a higher energy level followed by the re- emission of a photon when the electron falls spontaneously back again. However, both a theoretical consideration and experimental evidence shows that this picture of a two-step process is not correct and that resonant scattering has to be described as a coherent process (i.e. a forced oscillator with damping constant Ai,k
). Unlike photoionization or excitation by electron/ion impact, scattering involves therefore no atomic energy changes as no work as being done.
The existence of an induced absorption process is therefore implausible, as the same physical cause (i.e. the external radiation field) can not result in two different effects. By means of symmetry arguments, this questions also the reality of the induced emission process.
With regard to the optical resonator (e.g. mirrors), classical laser theory assumes now that it (apart from focusing the light) serves as a means to enable all atoms in the light emitting medium to radiate in phase, namely if its length is a multiple of half the wavelength. In this case of optical resonance, a standing wave will be set up for a purely sinusoidal signal, and it is assumed that this circumstance enables the proposed process of stimulated emission to amplify all emissions in phase. It is usually argued here that the intrinsically spontaneous and random emission develops into a wave with a uniquely defined phase because one of the initial emissions 'overwhelms' all the others and hence defines the phase of the radiation field. However, this is only a hand waving argument which in fact defies common sense (as it would violate the superposition principle for instance). Even if a stimulated emission process exists, it is only possible that the initial 'photons' present before the situation of a population inversion are being amplified in phase separately when encountering an atom in an excited state. Although a standing wave will be set up for each of these coherently amplified 'photons' separately in the optical resonator, this has actually no effect on getting further emissions in phase as this would happen anyway (laser or maser effects do apparently also occur in natural media without any optical resonator as long as there is a significant population inversion). The only effect the optical resonator has here is to amplify each of the wavetrains by folding it back into itself if it is long enough. This, not surprisingly, is easily the case for all laser transitions as these arise typically from (quasi-) metastable states with very long life times. For a wave train with length L the resultant amplitude would therefore be L/l if l is the distance between the mirrors (provided the mirror reflectivity allows that many reflections; otherwise the amplitude would be determined by the reflectivity). The resultant wave train is now accordingly shorter (in effect) but this will in many cases not be so important for the subsequent interaction with matter as the increased wave amplitude. Nevertheless, as all N standing waves are randomly out of phase, the amplification is obviously much less than for a coherent superposition of all emitters (L/l is usually much smaller than N).
In view of these conceptual problems with the laser process, it may be of interest that may own work on airglow excitation (see http://www.plasmaphysics.org.uk/papers/airglow2.htm#34
(Chpt.3.4)) has revealed a hitherto unknown radiative enhancement mechanism associated with the line broadening by plasma field fluctuations with the enhancement factor given by the ratio of this broadening to the natural line broadening. As the latter is usually very small for typical laser transitions (the lifetime of the upper state has to be very long in order to enable the population inversion), this mechanism could, at least for some types of lasers, be very effective and important. This in turn would then obviously lend further weight to the above criticism in particular with regard to the stimulated emission process.