I’ve read, that essentially there exist three types of reflections in Physics:
- Total Internal Reflection (See Below)
Metallic reflections (almost) tend to preserve the polarization of the light (except for what’s written below), while non-metallic reflections tend to polarize the light. The latter are also the basis for “polarizing mirrors”.
Beam-splitters are essentially polarizing mirrors:
- When randomly-polarized light hits them, the deflected beam will be plane-polarized in one direction, while the transmitted beam will contain, what the deflected beam does not contain.
- When circularly-polarized light hits them, nothing really prevents them from splitting the beam.
- When plane-polarized light hits them, depending on the angle of polarization, the amplitude of one emerging beam can become much lower, than that of the other. This is probably also why, linear polarizers can interfere with the physical auto-focus of a DSLR-camera.
(Edit 02/25/2018 :
Even though the articles I gave above ‘seem complete’, only today I’ve learned that they need to be modified. Specifically, the deflected beam is only polarized perfectly, when the incident beam strikes a non-metallic mirror at Brewster’s Angle. And I have no reason to think, that this account is wrong. )
From what I read, reflection, according to the particle depiction, takes place, because photons couple with plasmons, to form surface-polaritons.
From what I read, refraction takes place, according to the particle depiction, because photons couple with excitons, to form photon-excition polaritons.
(Updated 02/27/2018 : )
(As of 02/20/2018 : )
Today, according to the WiKi-article I linked to above, about “plasmons”, during metallic reflection, those quasi-particles couple with “electrostatic fields”. Personally, I have no qualms visualizing, that those electrostatic fields still form a photon.
But, unless the WiKi-article is also willing to suggest – which they don’t – that one photon could be coupled with both a plasmon and an exciton at the same time, they might find it hard to explain, based on what they’ve written so far, why metallic reflections in particular, tend to preserve the polarization of the light.
Further, I think that the concept is understood, that if, according to some theory, one fundamental particle needs to be coupled with more than one quasi-particle at the same time, a dilemma generally ensues, as to what would happen if only one of the quasi-particles was absorbed by the medium. And so there would be a strong bias towards formulating a hypothesis, according to which 1 photon only needs to be coupled with 1 quasi-particle, at any one point in time.
BTW, It’s interesting to observe that according to modern Physics, metallic reflections take place, because the medium has a negative instead of a positive dielectric constant. When I was studying Physics, this was still thought to be impossible.
According to the WiKi on plasmons, this is actually needed in order for plasmons to form, the “linear” polarization of which tends to be into versus out of the plane of the surface – i.e., ‘normal’ to the surface.
(Edit 02/21/2018 : )
In fact, if the WiKiPdeia article was true, according to which the plasmon possesses only fields normal to the surface, and if the assumption is also applied, that the wave-function determines the polarization of the reflected beam by itself, it would also be hard to explain why a metallic reflection is equally strong, when the incident beam itself is perfectly normal to the surface, since then, all its dipole-moments are parallel to the same surface, regardless of how the incident beam was polarized.
Yet, this is something metallic reflections famously do, and which non-metallic reflections won’t do.
(Edit 02/22/2018 : )
Total Internal Reflection:
When studying Refraction, from the classical wave-theory, it’s taught that a beam of light can pass from a side of a boundary, from which there is a lower index of refraction, to a side with a higher index of refraction, and that the angle of the incident beam (with respect to the normal vector) will be greater than the angle of the refracted beam (with respect to the normal vector). And, as the angle of the incident beam in such an example approaches 90⁰, Non-Metallic Reflection usually replaces the phenomenon of Refraction progressively, all the way until Grazing Incidence Reflection takes over completely. By that time, the reflected beam is Not Inverted. ( :1 )
But, the angle of the refracted beam approaches a definite maximum angle, the reciprocal of the sine-function of which, is also the relative index of refraction on that side. As logic would have it, there is nothing which prevents a beam from originating inside the region of higher refractive index – for example, from inside a glass prism – at an angle to the normal, which exceeds this maximum angle, so that there is no longer a matching angle on the side with the lower index of refraction.
When that happens, we obtain ‘Total Internal Reflection’. This form of reflection is often viewed as spectacular, because it has a supposed efficiency of 100%. By contrast, even high-quality metallic mirrors, usually only achieve an efficiency of 80-90% tops. For this reason, a Pentaprism is often used in the design of SLR cameras, ‘to make the viewfinder brighter’. Because light in a Penta-Mirror is reflected numerous times, the viewfinder of SLR cameras that use those, is usually only about 70% bright. Yet, because a Penta-Mirror is also hollow, those tend to make a camera lighter, than cameras with Pentaprisms.
If the reader questions whether this claim of 100% efficiency is accurate, I can attest, that if there is contamination on the outside of such a reflective surface – i.e., on the outer surface of the prism, where the internal beam
totally reflects – then some reduction of energy in the beam can take place. I think this is due to the electrostatic field, of a certain kind of quasi-particle, emanating on the opposite side?
If the only subject of interest is how this affects polarization, then Total Internal Reflection behaves as Non-Metallic Reflection would,
based on the fact that light is attempting to cross, from a more-positive, to a more-negative dielectric index / refractive index.
If the subject of partial reflection comes up, because due to the angle, an inversely-refracted beam also emerges on the lower-index side, then the deflected beam on the higher-index side will be as if Non-Metallic,
and with inverted phase. Such internal beams are often of lower amplitudes, than most, partially-reflected beams. If the goal of the exercise became, that the partially-deflected beam on the high-index side was supposed to be of similar amplitude, to that of the refracted beam, on the low-index side, then I’d say that the angle of the incident beam would need to be moved close, to its analog of Brewster’s angle. Either way, close to grazing the boundary, there should be no inversion.
BTW, By same token, I’d suggest that Metallic Reflection
is inverted in phase as well.
(Edit 02/25/2018 : )
This last comment on metallic reflection, needs to be modified. Specifically, I learned that when circularly-polarized light is reflected off a metallic mirror, the direction of polarization reverses. This suggests that, probably, the field-moment perpendicular to the surface is inverted, while the field-moment parallel to the surface is not inverted.
(Edit 02/27/2018 : )
The following diagram is meant to illustrate, how Non-Metallic Reflection takes place on the low-index-of-refraction side, with the part of the polarization highlighted, which would be ‘normal’ to the surface:
While this situation could cause some initial confusion, it can be explained. Since at Brewster’s angle, the normal component of the deflected beam’s polarization is zero, the most natural inference to make would be, that it would be positive on one side of Brewster’s angle, and negative to the other side… But then the question must also be answered, on which side of Brewster’s angle this component is positive, and on which side negative.
As I’ve diagrammed it, the inversion takes place, with angles less than Brewster’s angle. And the reason I would propose, is that this leads to consistency, if the angle of incidence approaches 0⁰. At that point, all the arrows in the diagram above will point parallel to the surface, and in the same direction. I would suggest that the same type of consistency should exist, on the higher-index-of-refraction side, if the beam originated from there.
At the same time, the real-world observation needs to be confirmed, that light can travel down an optical fiber without causing problems, where both refraction and total internal reflection take place. By same token it would seem to follow, that if an optical fiber was replaced with a very narrow metallic tube, the inside surface of which has been made mirror-like, ‘issues would ensue’, in the fact that upon each reflection, the circular direction of polarization would need to reverse…
At the same time it should be noted, that the “Chandra X-Ray Telescope” uses grazing incidence reflection. The composition of its ‘mirror’ has high-atomic-number elements that can interact with X-Rays. But although to mundane inspection, these elements are often metallic, the fact that they work with X-Rays in the intended way, is unrelated to how they would interact with visible light.