We’re working on a self-guided nature trail brochure for kids, and one of the stops currently says:
A friend of mine looked it over and said that (what is presumably) Rayleigh scattering doesn’t refract light the same way a prism would, but he isn’t sure. I looked into it myself but still don’t understand how the phenomenon works at the microscopic level.
Can any armchair physicists help clarify this? 
I’m not sure on exactly what phenomenon the poem is supposed to refer to. There are some effects caused by refraction due to the atmosphere, so the answer to the title of the thread is “no”.
But rayleigh scattering (which makes the sky blue and sunsets red) is a separate effect. So if your specifically saying that refraction is why the sky is blue, then that would be inaccurate.
Sorry, I should’ve clarified: we’re talking specifically about why the sky appears blue.
If that doesn’t actually happen, what is the difference between “scattering” and “refracting”?
“Like” is a very ambiguous word. How alike do you mean?
But no, scattering is not the same as refracting, and if the atmosphere was splitting light like a prism, you would see a whole spectrum (as you do with a rainbow, where the droplets are splitting light much as a prism does).
However, that stuff about greens, silvers and greys worries me more than the claim that it is like a prism. Who sees greens, silvers and greys in the sky. (OK the are clouds, and the green flash, but I doubt you mean them, and if you do, orange is a much more common sky color.)
Your own link describes scattering pretty well. Refracting is when light gets bent when passing from one transparent medium to another, and will split colours because some wavelengths get bent more than others.
Thanks for catching that about the weird colors – they were left over from a previous iteration of the writing, and we’ll fix that!
As for the scattering vs refracting, if I understood the Wikipedia article I wouldn’t bothered to have asked
Some other diagrams I’ve seen make it seem like “scattering” is where translucent molecules (namely water vapor) refract blue wavelengths in random directions while ignoring longer wavelengths. Like in this picture, I am led to believe (I think…?) that orange light is pretty much going directly through it while blue light is bouncing around, like a pinball machine, inside the rock and its internal air pockets. The end effect is that it gets crazy unpredictable and produces a diffuse blue glow.
But isn’t that just a whole bunch of refractions, one after another, of a certain wavelength? The picture I have in my head is a spectrum of sunlight hitting billions of tiny molecules in the air, much of which is water… the orange light’s wavelength lets it bypass most of the molecules, but the blue light hits one, then another, then another, then another, etc. until the entire sky is a diffuse blue from all the scattered light.
So does that mean scattering = a bunch of microscopic refractions? Your posts here seem to suggest that, no, they are entirely different phenomena. So then what does “scattering” mean? Are airborne atoms absorbing photos and then re-emitting them as visible light in the blue bands? That makes no sense 
As for the prism’s rainbow effect, isn’t that due to sunlight’s wide set of wavelengths and their different refractive indices through glass? If you shoot a laser through a prism, it still gets refracted, but not separated into different wavelengths, no?
What is happening to the light microscopically when it encounters airborne molecules?
A prism works because the material it is made of is dispersive - that is it has a different refractive index at different wavelengths. Thus when light passes through it, different wavelengths are bent different amounts and the colours separate into a fan. In the sky, a rainbow is due to dispersive refraction in water droplets. This why you get a colour spectrum in the bow. The actual geometry of the lightpath is on the other hand pretty interesting, and not obvious.
Refracting in a prism is all about QED (quantum electrodynamics) and thus all about how electrons and photons interact, and how the paths the photons take are modified due to this interaction, making them appear to travel slower than c to varying extents.
Rayleigh scattering is a slightly more macroscopic effect. It involves photons taking a range of possible directions as they pass though a medium due to the presence of very tiny particles. The particles are smaller than the wavelength of the light. The wikipedea article does summarise it, but perhaps a bit teresly. Light famously can be wave and particle. The explanation is all about considering it as an electromagnetic wave. Very classical. As it passes by a particle - say a nitrogen molecule in the air, the electrical field of the wave can affect the electrical field around the molecule. The field around the molecule distorts and rebounds in concert with the field of the light. As it does so it radiates energy at the same frequency as the exiting energy, but does so in all directions. (Imagine a large ship in a big sea, with big long waves driving it up and down. The inertia of the ship means that it will keep moving at the tops and bottoms of the waves for a little bit, and thus it will create waves that radiate out in a circular pattern from the ship. The ship is taking energy from the transverse waves it is riding in, and passing it to the circular waves. These circular waves have the same frequency as the transverse ones, and so it appears that the wave has been scattered.) A quantum explanation might draw a comparison with light as particles, and the occasional bouncing off scattering particles, more like a shooting billiard ball though a random scattering of ball bearings.
The critical point is that the frequency/wavelength of the incident energy has not been changed. However the relative size of the scattering particle (or ship) and the wave determines the split of energy between what remains in the incident wave and the scattered component. The shorter the wavelength the more energy is scattered. But it isn’t an absoulte thing. Not like dispersive refraction where all the energy at a given wavelength goes at exactly one angle. Scattering is just a slight bias. But enough that the sky goes a bit blue, and the sunsets go a bit red. It isn’t as if the sky is a solid pure saturated colour, nor sunsets.
The atmosphere http://en.wikipedia.org/wiki/Atmospheric_refraction refracts light as well, but (I’m guessing) the refractive index is so close to that of vacuum there’s no dispersion to speak of, so the result is pretty much just that the Sun actually sets before it appears to do so.
refraction happens when light travels through a boundary between two things with different, well, “refractive indexes.” Think water/air (why a spoon or drinking straw looks bent), or air/glass (a la a prism). The rays of light are all bent in the same direction. Sunlight going through the atmosphere isn’t refracted because there is no interface for refraction. Instead, the light rays are scattered randomly; the degree of scattering depends on the wavelength.
I don’t see anything especially wrong with this. Both refraction and scattering occur because light encounters a region of different refractive index. The difference is that in scattering, these little regions, which correspond to density fluctuations in the air, are strewn about randomly, while in the prism, for example, they are organized into a crystal structure. Hence refraction has a coherence that scattering lacks.
But from a very crude point of view, I don’t see anything really wrong with saying that scattering is kind of like having a bunch of tiny glass beads strewn randomly throughout the atmosphere. A key fact you’d have to emphasize, however, is that the size distribution of the beads is not uniform – there are far more smaller beads than larger beads, which is why blue light is preferentially scattered. (The strength of the interaction between a light wave and a scattering region will drop significantly with mismatch between the wavelength of the light and the size of the region, for the same reason you need the right antenna size to interact strongly with a radio wave).
As for why there are more small “beads” (density fluctuations) than large ones, you can look at this two ways. Mechanically, the deviations of a large collection of random variables from their average behaviour is more common for smaller deviations. For example, if I flip 10,000 coins, the probability that the average number of heads deviates by 1% from 5000 is much higher than that it deviates by 5%. If I throw 10,000 rocks into a box, I will observer more deviations from the average density in a 1cm x 1cm area than in a 10cm x 10cm area. (Parenthetically, this is why you observe greater deviations from averages like, e.g. cancer statistics, in smaller sample sizes, which is why you never hear about ‘cancer clusters’ in anything other than small towns. They never happen in New York City.)
From a thermodynamic point of view, you can expand the probability of deviations from equilibrium around equibrium, and the probability of deviations in the entropy will go something like exp(-(dS/kT)^2), so a small deviation in the density (dS small) is much more likely than a large one (dS big).
That’s true, but you don’t need QED to come up with a reasonable model of refraction (just like you don’t need QM to model Rayleigh scattering). A model where electrons behave as if they’re attached to atoms with tiny springs is sufficient for most purposes. The electrons behave as small oscillators, and there’s no need to treat light as photons. It’s just a wave that undergoes a continuous phase shift as it passes through a material such that it appears to slow down.
I’m not the definitive expert by any means, but this doesn’t seem right.
With refraction, photons are moving from one medium to another, such as from air to glass in a prism and at the boundary they refract, or bend. My optics professor who wrote the book on fiber optics (true story, he had us use his text book for the class and then gave us back his $5.00 royalty to avoid a conflict of interest. My, ah, er, friend who wasn’t as ethical as now had shoplifted the textbook, so he actually made money on the deal.) gave an analogy of a farmer driving his tractor from a dry field to a muddy field at an angle. When crossing the boundary, the wheels on the dry side will be moving faster, so the angle changes. Once completely in the new field, the tractor goes straight.
With scattering, the photon encounters a molecule and does a random movement. This is not at the border of a change in medium and as such should not be described as the same mechanisms.
This. I know reality is much more complicated than the pictures in my head, but refraction involves a wave changing directions as it passes through media of differing velocities. Ocean waves, for instance, bend or refract as they move into shallow water. Scattering happens when the wave is reflected (bounces off) a medium it can’t easily penetrate (i.e, an water wave reflecting from a sea wall).
When light encounters a transparent medium with a significantly different index of refraction, some of the light enters the medium and is bent, with the degree of bending inversely proportional to the wavelength.
But, depending on the incidence angle, some of the light will reflect from the surface of this new medium, changing direction with no preference for wavelength. This is why snow, salt, and sugar all look white even though they are colorless. The fine granules reflect white light in multiple directions at once – this is called multiple scattering.
As stated above, the earth’s atmosphere contains particulate matter that is mostly very fine. Particles heavier than air will eventually settle to the ground but smaller particles settle slowly so that particle concentration is inversely related to particle size. When light encounters these particles it will reflect or scatter from them. Waves are mostly unaffected by particles much smaller than their wavelength, so longer waves (reds, oranges) are scattered only a bit, but blues and violets are scattered a lot. The scattering of light by wavelength is known as Rayleigh Scattering.
Important to note though that the “particles” here include the gas molecules that make up the atmosphere. Anyone know how blue the sky would be if it was all gas and zero aerosols?
