[Sorry this is so long, but the way the question was expressed suggested that some background might be helpful. If it's not, if this starts too basic, skip down five paragraphs (to "Now, back to your driveway . . . ")].

First of all, oil doesn't mix with water, it lies on the top (the old adage about oil and water not mixing is correct), and that's the crucial fact, but let's back up a bit, just in case.

You probably know that light travels in waves, and that different wavelengths are what we're perceiving when we see different colors, just as different wavelengths traveling through the air are what we're perceiving when we hear different pitches of sound.

But what if there's more than one wave at a time being generated? Then things get interesting. If the distance from wave peak to wave peak of the two (or more) is not exactly alike, then they are (to a greater or lesser degree) "out of phase" -- the peaks of the two waves will sometimes match up and sometimes not, and ditto the troughs. As the peaks get closer they reinforce each other and the collective outcome, the "resultant," gets stronger ("constructive interference"). When they get farther apart --- the peaks of one wave occuring at the same time as the troughs of the other -- the resultant gets weaker ("destructive interference"). And of course everything in between happens as well. (That the waves may also be different sizes, trough to peak, matters also, but let's ignore that.) And of course there are always, outside a laboratory, many more than two waves involved.

If these are sound waves, the result will be variations in the pitch of what we hear. If they are radio waves, we get the possibility of transmitting sound by managing the difference between the carrier wave and its sidebands, which are at minutely different frequencies. (And I assume, but don't actually know, that TV works the same.) And if they're light waves? Then we may see changing colors or a moire pattern or a spectrum of some sort, or a spectrum with dark lines here and there which will tell us things about the molecular makeup of whatever's giving off the light to begin with -- such as a distant star.

Now, back to your driveway: an oil slick lying on water doesn't appear to have any thickness at all, but of course it does. It's a thin film, with the top essentially flat (because of surface tension), and the bottom also flat, because it's lying on the water surface. When light impinges on the oil from the top, some of it is reflected from the top surface, but some penetrates and instead bounces off the bottom of the film -- or more accurately, bounces off the top of the water. Now the waves reaching your eye are no longer in synch, partly because of having traveled different distances, and partly because the wave that bounced off the top of the oil underwent a phase shift that the bottom reflected wave did not (never mind why). Light waves out of synch are perceived by us in different ways, and one of the nicest is as a spectrum -- a rainbow -- in this case a floating rainbow that will change shape if you fan the top of the oil gently, minutely changing the thickness of the film.

The effect can be obtained other ways also: you may have seen lapel pins for sale, about the size and shape of a quarter and very smooth to the touch, that reflect a shifting rainbow similar to an oil slick. It's the same effect, but here the reflecting light waves have been put out of synch by a "diffraction grating" -- circular microscopic grooves etched in the surface, which feels perfectly smooth to your finger, but is not smooth to a light wave.

REFERENCES:

There should be a wealth of web sites dealing with this stuff. Try "diffraction" or "diffraction grating," "interferometry," maybe "thin film diffraction," and bop around until you find something that works for you.

And anyone with a glint of interest in the physical world should immediately rush out and obtain "Thinking Physics," by Lewis Carroll Epstein (Insight Press, SF; 3d ed. 2002). It's a compendium of one- and two-page questions and answers about how real world stuff behaves -- pool balls, icebergs, flowing water, rolling cars, falling bricks, ascending balloons -- the kind of deceptively simple little problems that can reduce a dinner party of intelligent people to warring factions. [Disclaimer: I'm not Epstein, don't even know him, and have no interest. Except that I'm interested -- you know what I mean.]

First of all, oil and water don't mix - they agree to peacefully coexist! Tiny drops of oil can become suspended in water, but they will eventually find their way to the surface and float.

What you're seeing is light being refracted by the oil diffently from water.

What is refraction you say?

An interesting pheomenon occurs when light travels from one transparent material to another; light beams actually change directions. This bending at the surface of transparent objects is called refraction. Refraction can be looked at much the same way as reflection.

Rainbows are caused by another aspect of refraction called dispersion. Light waves of different frequencies (colors) bend different amounts. In most cases, this is not noticable, but prisms make use of this to spread out the spectrum so we can see all of the colors. Violet light bends the most, with each color bending a little less up to red, which bends the least.

- http://library.thinkquest.org/13405/index.html


Anyway - I think this is right because how we see things is all about how light reflects/refracts off/through stuff?

To all the above answers, I'd like to add the most interesting fact: It only gives a rainbow when the oil has spread out so that it's only one molecule thick. For proof, you have to understand the answers above.

This phenomenon, called iridescence, occurs because the thickness of the film is comparable to the wavelengths of visible colors of light. Light reflecting from the rear surface of the film combines with the light that reflected from the front surface, but since the light traveled at a different velocity in the film (it has a different refractive index), the light that has traversed the oil has been retarded in phase. When it combines with the light reflecting from the front surface, the two parts interfere with each other. Interference can be constructive (they are in phase, and simply add together) or destructive (they are out of phase, and cancel each other). The color you see corresponds to the color (wavelength) that has constructive interference. If you look at the same spot from a different angle, the color changes, because the back-reflected light took a different-length path through the oil. Different thicknesses produce different colors, whether it is because the oil film is a different thickness, or because you are look through it at an angle.