Diffuse Reflection vs. Scattering?

[iridescence]

Is there a fundamental difference between specular and diffuse reflection, or between diffuse reflection and scattering? I have heard various explanations that I am trying to sort out. (Please note that these "explanations" are not necessarily an accurate record of what was presented to me-- rather a reconstruction of what I understood them to mean. I am not a proper physics student, just an artist wanting to do my best to make sense of the interactions of light and matter.)


Explanation #1 (what I initially picked up from art training): Diffuse reflection is the color of the object that we see when light bounces off a rough surface in random directions. Specular reflection (highlights) will be the color of the light source, and its angle of reflection is equal to the angle of incidence. The amount of specular reflection depends on surface smoothness. Sub-surface scattering occurs only in translucent material.

Explanation #2 (after looking into it a bit more): The amount of specular reflection depends on index of refraction and not on surface polish. Light reflecting from the microfacets of a rough surface is still specular (though fuzzy) reflection of the light source. Diffuse reflection is actually a form of subsurface scattering and does not depend on the roughness of a surface but rather on the microtranslucency of the material. Even if they appear to be macroscopically opaque, all colored materials aside from metals have some amount of translucency. Light penetrates to some degree and is partially absorbed and reflected around within the material before re-emerging as colored. Scattering is simply a form of reflection in which the direction of the light seems to be random, but actually obeys the law of angle of incidence = reflection... it's just that the microfacet interfaces (either surface or within the structure of the material) tilt in so many unpredictable directions.

Explanation #3 (after looking into--and probably grossly misunderstanding--quantum electrodynamics): Reflection is one possible average/aggregate of scattering probablilities. Scattering, compared to reflection, is a more fundamental interaction of light with matter-- the energy of light perturbs the electrons in the matter... if it is not a useful amount to jump energy states, it is quickly re-emitted in all directions, or at least having the probability of any direction. In certain circumstances, for reasons that I have failed so far to understand, the probabilities of certain paths cancel out and others add up, and it looks like reflection on a macro scale.

Ok, so although quantum things do seem to be quite helpful to me in understanding why materials absorb certain wavelengths, with regard to the actions of light in general, I'm not sure that attempting QED really helps my already shaky understanding of classical/wave optics... anyway, for now I would primarily like to know if I am on the right track with anything in Explanations 2 & 3, and what are the misconceptions that jump out? One thing that confuses me is that in Explanation #2, scattering seems to be a type of reflection; and in #3, reflection seems to be a type of scattering.

I appreciate any thoughts-- thanks!

 

[Simon Bridge]

All these explanations are correct within their contexts.
The QED description is the most detailed - but still involves interaction at a surface.
Light propogates through a material by directly traveling between atoms, and being absorbed and re-emmitted by them. Some atoms will release energy from the absorbed photon in stages rather than all at ones - perhaps re-emmitting in IR so you don't see it, perhaps partly IR and partly visible so you see a change in color.

I think that's your missing bit.

The other descriptions are understood in terms of averaging over many effects for a particular material.

Thus you get colored materials (like paint) absorb particular wavelengths ... blue paint absorbs mostly red for eg. ... which gives you the standard rules for mixing paint being different from that for mixing light (blue+yellow=green for paint and white for light). Some of the effects are due to the way the brain interprets input from the eyes - so magenta does not have it's own wavelength.

Translucent materials would let most light through but shift the wavelength or selectively absorb some of the light incident.

On the right track - looks like it to me.
Have fun.

 

[sophiecentaur]

Translucent materials would let most light through but shift the wavelength or selectively absorb some of the light incident.

.

I'm not sure what you mean by this. Are you suggesting that the light coming through 'frosted' glass would have a different frequency on exit? Or can you suggest a material for which this would happen?

 

[Simon Bridge]

Material properties is a big subject. Even with just "how stuff affect light" it's hard to cover all the bases properly in a short post.

Since some materials can absorb a photon of one wavelength and relax by emmitting a photons of other wavelengths, it stands to reason that an otherwise transparent material with some of this material diffusely embedded would have a tint to it.

Not all materials that let light through are glass.

Frosted glass, as you know, has strong random scatter off surfaces giving it the frosted look. Whether a portion of transmitted light is also color shifted depends on the glass.

One can also eliminate or enhance a particular color by controlling the thickness of the material.

I notice I also didn't cover iridescence - but OP probably has that down.

 

[sophiecentaur]

I used ordinary glass as an example where this wavelength shift clearly doesn't take place but are there many translucent materials that actually behave as you suggest? I should have thought that if there were, they would be used all over the place - for aesthetic effect.

I know that UV - Optical wavelength shifting is common enough but I have never come across optical - optical shifting.

You say, earlier that
"Light propogates through a material by directly traveling between atoms, and being absorbed and re-emmitted by them. "
but, way back in my solid state physics course, I seem to remember that light travels through 'transparent' materials without interacting with individual atoms - rather with the structure as a whole. Any interaction with individual sites would, surely, involve re-radiation away from the original direction (scattering).

 

[Simon Bridge]

I used ordinary glass as an example where this wavelength shift clearly doesn't take place but are there many translucent materials that actually behave as you suggest? I should have thought that if there were, they would be used all over the place - for aesthetic effect.

I'm being very general ... you've seen green bottle glass? A transparent material need not be all one kind of substance.

You've seen liquid dyes of different colors, yet you can see through them?

Can't think of any transparent materials that do the phosphor thing though.

I know that UV - Optical wavelength shifting is common enough but I have never come across optical - optical shifting.

Don't believe I mentioned it myself either.

I seem to remember that light travels through 'transparent' materials without interacting with individual atoms - rather with the structure as a whole. Any interaction with individual sites would, surely, involve re-radiation away from the original direction (scattering).

So how do the individual photons know that the atoms they approach are part of a transparent medium and so should be ignored? How do the photons know the extent of the medium so as to know how much to interact with "as a whole"?

I vaguely remember doing random-walk paths through different materials and deriving refractive index - but I could be thinking of something else. I'm not going to really push the idea of photons bouncing around a solid like balls in a pinball machine. See...
https://www.physicsforums.com/showthread.php?t=243463
... now I'm wondering if I have not allowed my language to get too relaxed.
For instance, metals are very shiny - the QM picture is more like a big tub of electrons rather than a bunch of individual atoms.

In the electric field description, where incoming light is modeled as EM waves, the materials electric field changes the local permittivity, changing the local speed of light.

It's like your sig says: it's all models.
And I've messed up before.

 

[sophiecentaur]

Does "green bottle glass" do any wavelength shifting? I seem to remember the concept of 'F (farbe) Centres' to account for selective absorption in a transparent medium.

You write "Translucent materials would let most light through but shift the wavelength or selectively absorb some of the light incident.". The word 'translucent' means letting through light, does it not? That implies visible light and a shift in wavelength to something else visible. If you really didn't mean that then you might have included a caveat. I was getting all excited about some new material you'd heard about.

You write "So how do the individual photons know that the atoms they approach are part of a transparent medium and so should be ignored? How do the photons know the extent of the medium so as to know how much to interact with "as a whole"?"
All a photon can do is transfer energy from the arriving wave into the system it hits. Why insist that the interaction should be with an individual atom? The outer electrons of an atom in a dense medium are very strongly associated with the other surrounding atoms (the structure) and it is naive to think of them as behaving like isolated (monomolecular gas) atoms. A dense medium has a band energy structure and not line structure. That has to be because of the fact that individual atoms are not the entities that are interacting with em waves; it is the whole, structure or significant region in it. If you shine a light through a gas, the energy that does actually interact with the individual atoms re-radiates in all directions (see 'absorption spectroscopy'). If that were how it worked in solids, you would also get this effect but to a much greater degree because of the greater density. But there is also a very small effect of changing the light speed, due to the density of the gas, which must loosely involve all the atoms. That effect does not affect the (shape of the) wavefront passing through.

Also
"For instance, metals are very shiny - the QM picture is more like a big tub of electrons rather than a bunch of individual atoms."
Yes, and when you bear in mind that the interaction of EM is not just with an electron but a charge system, with specific energy levels, the interaction with a metal surface must be with the system and not just one electron. Any model that you use must satisfy the classical / macroscopic observations and that requires phase coherence across a wave front (that goes for transmission in a transparent medium too) so that any interaction with a particular 'notional' isolated electron would need to be exactly in phase with interactions with any other isolated electron. Only this could produce specular reflection.

 

[Simon Bridge]

I want to break this up into two parts.
This part is about the manner I chose to describe things - the language I chose.
I believe you have a point and I could have been clearer. eg.

You write "Translucent materials would let most light through but shift the wavelength or selectively absorb some of the light incident.". The word 'translucent' means letting through light, does it not? That implies visible light and a shift in wavelength to something else visible. If you really didn't mean that then you might have included a caveat. I was getting all excited about some new material you'd heard about.

I see the problem.
I was thinking of what sort of things would produce a visible effect on interaction with a material - re. the original post.

"having it's wavelength shifted" is not a good description anyway - kinda implies something like the doppler shift doesn't it? (You do get a wavelength shift in materials but it won't give a tint since it shifts back when it exits.)

So if you shone light outside the visible spectrum and it fluoresced in the visible that would count in the above statement as a color change. Color observed out is not the color observed in.

Having said that - the different mechanisms to produce a color change (fluorescence, selective absorbtion etc) could be described in different models. Since the lesson I wanted OP to take away was that each of the descriptions supplied were "good" (in their places) - I should probably have listed common effects and described them in each model.

I had hoped to illustrate that what is being called "light" in each case is a bit different - since the description are based on different models for light. I'll have to have a rethink and come back to it - maybe be more specific to the exact points in the first post?

Or, maybe you could demonstrate a better answer for me :)

 

[sophiecentaur]

Wavelength change but not frequency change when light passes between media, actually: boundary conditions apply.

We have a whole raft of ideas in these posts and it may be better to discuss some of them in their own thread - to avoid confusion.

As for explanations involving photons, I always find they tend to be really fraught because they remain as 'little bullets' in the minds of so many people.

 

[Simon Bridge]

This second part I want to talk about what I mean about photons interacting at a point in QED and what this model seems to mean for interaction of photons with materials.
(I talked about interaction with atoms, which is a bit general - so let me tidy up my language a bit - I'll try not to write a textbook though.) It may help to realize that I am using a kind of Feynmanesque public-lecture language. This is the sort of thing he'd say:

“Photons do nothing but go from one electron to another. [...] Reflection and transmission [of light] are really the result of an electron picking up a photon, ‘scratching its head,’ so to speak, and emitting a new photon.”
--- Richard Feynman[1]

The specular effects and so on across the surface come from lots of these interactions all averaged out over the paths available to the photons that get detected - by your eye in this case.

In the metal, says Feynman, and QED, the photon interacts with but one electron.

What that means depends on the electron's state, which depends on the potentials for the entire metal. The photon only needs to know what that one electron is doing and that electron only needs to know about it's immediate surroundings and so on. Everything is connected so local interactions described wholly locally are still affected by non-local ... stuff.

The light we see from the metal is made up of many photons - the effect of which can be predicted by adding all the amplitudes for the effects from all the possible ways a photon may get from the source to your eye via the metal.

The result provides all the observed phenomena - angle of incedence equals angle of reflection being the easy one to extract. The yellowy color of copper or gold is because not all wavelengths have the same chance to be scattered. Irregularities in the surface mess with the sum creating a diffuse shine rather than a clear reflection. Since the sum of amplitudes is across the entire material - and in detail must extend within the material[2] too - then the entire material contributes to the final spectrum seen (not to mention mixing in the eye and processing in the brain). This sum tells us about the distribution of the photons likely to be received, but each individual photon still only interacts with one electron and is detected as only one photon in one go.

In the same way we can say that the photons in Young's interference pass through only one slit or the other one even though the resulting spectrum includes contributions from both of them, we can say that the photons interacting with a material do so only with individual electrons even though the resulting spectrum contains contributions from the whole material.

The caveat here is that I don't intend the reader to understand that any of this is "really" happening. It's a way to think about things that helps the math come out right. It is a model.

-----------------------
[1] He said that a lot in his public lectures - see him quoted http://news.discovery.com/space/fun-with-feynman-diagrams.html for eg because I don't want to hunt through the vids to locate the exact times. The lectures are on YouTube: if you've not seen them, they are worthwhile.
[2] within the classical boundary - language gets tricky.

 

[sophiecentaur]

That last sentence is music to my ears (eyes?). One has to be so careful here and your caveat should be printed on the side of every cigarette packet and textbook.

 

[Simon Bridge]

We have a whole raft of ideas in these posts and it may be better to discuss some of them in their own thread - to avoid confusion.

I'll second that. Though, if we are not careful, we'll end up just writing chapters to a textbook.

I should probably tidy up the answer here - with your assistance.
The original question actually managed to encompass a lot of ideas and I think I was trying to talk about too many of them at the same time :/

The "second part" above really needs it's own thread.

That last sentence is music to my ears (eyes?). One has to be so careful here and your caveat should be printed on the side of every cigarette packet and textbook.

- probably should put it in my sig :) Feynman puts it at the start of each lecture series - all that talk about Mayans, and people still miss it.

 

[Simon Bridge]

OP must be bemused by now :)

#1 is the kind of model you need if you are going to paint something - especially if you do texturing for 3D computer models. eg.

"light" is loosly applied to refer to illumination effects ... what the movie/photography industry would call "lighting" iirc. Here there is one, whitish, light-source to the front and somewhat above the model pictured.

The "diffuse" part is what gives the base colors
The "specular intensity" makes it shiny or dull
and there is actually a specular color added too - not so obviously - which is a bit orangy.
The model also includes glows - where light originates from.
(There are errors in the rendering which the alert observer should pick up.)

#2 the main difference here is that random scatter is replaced with regular reflection - previously though of a specular reflection. It's all specular - it's just that reflection from deep within the material also color the reflected light. I'd think of this as a wave-based model.

The slight translucence is why the undercoat color matters - painting miniatures or theatre/movie sets I'd often use a dark matt color as an undercoat since it makes surfaces look more like what they are painted up to be (rocks, bits of alien technology etc).

#3 QED which does away with regular reflection, to replace with only random scattering. This is a particle-based model.

Which is why you get the observation

in Explanation #2, scattering seems to be a type of reflection; and in #3, reflection seems to be a type of scattering.

... as for art, I just use whatever model the renderer uses, and that usually means OpenGL. IRL: #2 gets me the furthest.