A Prism Splits White Light: The Dispersion Experiment

Send a beam of white light at an angle through a triangular prism and it spreads out on a sheet of paper as a band running from red to violet. This is dispersion. What it proves is not that prisms conjure colours, but that white light is a mixture of coloured light to begin with — the conclusion Newton reached in 1666 with a darkened room and two prisms.

It is very easy to run this as a spectacle. What turns it into an experiment is what comes next: putting the colours back together, and sending a single colour through a prism again.

Safety first (read this before you start)

• Never look directly at the sun, and never look at it through the prism. In the sunlight version, keep your back to the sun and look only at the band on the paper.
• Do not use the prism or a mirror to bounce sunlight at anyone's face.
• In strong sun, put the prism away in its box as soon as you finish. Glassware left on a sunny windowsill can concentrate light enough to scorch a curtain or a tablecloth.
• A glass prism shatters if dropped. Work near the middle of the table with a cloth underneath.
• A laser pointer cannot produce dispersion at all — laser light is a single colour — and it risks eye injury. Do not use one here.

Materials

• A triangular prism. Failing that, a square clear plastic box filled with water, or a small mirror propped at an angle in a bowl of water, will produce a usable band;
• A white light source. Sunlight on a clear day is best, an incandescent torch second. Avoid LED torches and lasers — an LED spectrum has gaps and the band comes out missing colours;
• A piece of card with a 2 mm slit cut in it — the narrower the slit, the cleaner the band;
• White paper or a white wall as a screen;
• For the later steps: a second prism, and red cellophane or a red filter.

Procedure

1. Close the curtains except for a gap, or hold the slit card in front of the torch, to get a narrow, straight beam of white light.
2. Put the prism in the beam and rotate it slowly while watching a sheet of paper to the side. At some angle a coloured band appears on the paper.
3. Move the paper further away. The band widens and the colours separate more clearly. Note the order of colours from one end to the other.
4. Put it back together. Place the second prism in the path of the band, upside down relative to the first. Adjust until a white patch reappears on the paper behind it.
5. Try to split it again. Use the red filter to block everything but the red part of the band, then send that red light through the second prism. See whether it separates into anything else.

What you should see

The order of colours in the band is always the same:

The other two steps give:

• Second prism inverted: the band converges back into a white patch. If the alignment is slightly off you get a white patch with coloured fringes, which still counts — most of the colours have recombined.
• Red light only: after the second prism the red beam has changed direction (it still refracts) but its colour is unchanged and nothing separates out of it.

One detail worth noticing: look closely at the band and there are no dividing lines between the colours — one shades continuously into the next. The famous "seven colours" is Newton's division, chosen to match the seven notes of a musical scale. The real spectrum is continuous, and you could just as defensibly call it five colours, or ten.

The science: why the colours separate

White light is polychromatic — a mixture of light of many colours. What physically distinguishes those colours is frequency: red has the lowest, violet the highest.

The key fact is this: one and the same material has slightly different refractive indices for different frequencies. Glass has a somewhat higher index for violet than for red. Since the amount of bending at refraction is set by the index, violet gets bent hardest and red least. So a single beam of white light entering the prism at an angle comes out fanned into a band.

A prism works better than a flat sheet of glass because its two faces are not parallel: light refracts on the way in and again on the way out, and the two deflections point the same way and add together, magnifying the difference between colours. In a flat sheet the faces are parallel, the second refraction undoes the first, the beam merely shifts sideways, and the colours stay stacked and invisible.

Dispersion also makes one point about frequency very concrete. When light passes from air into glass its speed and wavelength both change, but its frequency does not — and neither does its colour. Red light is red in air, in water and in glass alike.

Dispersion around us

• Rainbows. Sunlight enters a raindrop and refracts, reflects once off the back, and refracts again on the way out; those three steps spread the colours. You have to stand with your back to the sun to see one, because the light is coming back out of the drops towards you.
• Fire in a diamond. Diamond has a very high refractive index and strong dispersion, and the cut bounces light around inside by total internal reflection, so the colours emerge widely separated.
• The coloured fringes thrown by the edge of a drinking glass or a crystal chandelier in sunlight are the same effect at small scale.
• A garden-hose rainbow. With your back to the sun, spray a fine mist ahead and above you and a band of colour appears — physically identical to the one in the sky.
• Purple fringing in photographs. Cheap lenses show violet or green outlines at high-contrast edges, because the colours focus at slightly different distances. Good lenses combine several glass types specifically to cancel it.

Five things students get wrong

• "The prism colours the light." This is exactly the pre-Newtonian explanation, and step 4 exists to demolish it: if the colour came from the prism, a second prism could only add more, never restore white.
• Calling the colours on a soap bubble or a CD dispersion. They are not. Soap-film colours come from thin-film interference, and the rainbow on the back of a disc from diffraction. All three produce colour and all three have different causes, and this is the trio students confuse most. The test to hold on to: was there a refraction — light passing at an angle through a transparent medium?
• "Red has the highest refractive index, which is why it's on the outside." The order is reversed. Violet has the highest index and bends most; red has the lowest and bends least. A rainbow is a useful mnemonic: red on the outside of the primary bow, violet on the inside.
• "White light is light with no colour." White is every colour added together, not the absence of colour. The real absence of colour is black, which is the absence of light. Conversely, red, green and blue light added together also make white, which is how every screen you own works.
• "There must be exactly seven bands." The spectrum is continuous and the colours blend smoothly; there are no seven boundaries anywhere in it. Seven is a convention, not a physical fact.

Teaching notes

• Time and grouping: 25 minutes. There are rarely enough prisms to go round, so demonstrate steps 1–3 from the front and let groups do steps 4 and 5.
• Say this before they start: never look at the sun, directly or through the prism.
• Prep that pays off: the light source decides whether this works. Incandescent, halogen or sunlight are all fine; an LED torch will not give a complete band, because its spectrum has a gap in it. Test yours before the lesson.

The step most often skipped. Recombining the colours is nearly always dropped for want of a second prism or a few more minutes. Without it, though, the lesson is just a look at a rainbow and proves nothing. Even one demonstration from the front is enough for students to see why that step carries the argument.

How to tell they have it. Ask: "If the prism were adding colour, what should the second prism do? What did it actually do?" A student who can answer both halves has understood the reasoning, not just the display.

Take it further

• In the simulator, switch the prism to a higher-index material such as diamond and compare how far the beam bends at the same angle. A higher index also means stronger dispersion.
• Increase the angle of the incoming beam step by step and find where the light stops leaving the far face altogether — total internal reflection. A diamond's sparkle is half dispersion and half this.
• Make a Newton disc. Divide a card circle into seven sectors, colour them, push a pencil through the centre and spin it fast. The colours blur into a greyish white — recombination again, this time performed by your own persistence of vision.
• Make a rainbow at home. With your back to the sun, spray a fine mist upwards and forwards and adjust until the band appears. Note where the sun, you and the bow are relative to each other: the rainbow always sits directly opposite your own shadow.
• Hold different coloured filters in the white beam before the prism and see which part of the band each one lands on. Red cellophane looks red because it absorbs the other colours and lets red through, not because it adds redness to the light.

When a student can say "the prism added nothing — it just separated colours that were already there", this experiment has done its job.