Prism Rainbow Experiment: Make a Rainbow Indoors

Most prism rainbow experiment photos online are rendered, not photographed. The real thing is narrower, dimmer and much more interesting. Hold a glass prism in a shaft of sunlight through a gap in the curtains and the spectrum it throws on the opposite wall is only about two and a half centimetres wide from red to violet, which is exactly why the room has to be dark and the beam has to be narrow. Get those two things right and you will see seven centimetres of separation nobody in the family forgets. Get them wrong and you will see a white smudge with coloured edges. The difference is physics you can calculate in advance.

Key Takeaways

• You pass a narrow beam through a prism in a darkened room and project the spectrum onto a white wall or card, never looking at the light source through the prism.
• The prism does not add colour. White light already contains every wavelength, and the glass bends each one by a slightly different amount because its refractive index depends on wavelength.
• The spread is small: for ordinary crown glass at minimum deviation, red and violet emerge about 0.7° apart, giving roughly 2.5 cm of separation at 2 m.
• Best for ages 6 to 12, though the quantitative version suits 12 and up.
• Never look at the sun through a prism, a slit, a camera or anything else. You project the beam onto a surface and look at the surface.

What You’ll Need

Item	How much	Where to get it	If you don’t have it
Glass or acrylic triangular prism	1, 25–50 mm faces	School supplier, hobby shop, optics seller	A thick clear glass paperweight, the corner of a clear acrylic box, or a chandelier crystal
Bright white LED torch	1	Any hardware shop	A phone torch works but gives a poorer spectrum, explained below
Stiff black card	A4 sheet	Stationery	Cereal box painted black, or two pieces of card held with a 1 mm gap
Craft knife or scalpel	1	n/a	Adult only. Or make the slit by taping two card edges 1 mm apart
White card or white wall	A4, or a bare wall	n/a	A sheet of printer paper taped up
Tape measure	1	n/a	n/a
Second prism	1, optional	n/a	For the recombination experiment; two cheap acrylic prisms are enough
Darkened room	1	Curtains or night-time	A large cardboard box on its side, with the slit in one end

Time: prep 15 min, experiment 20 min, cleanup 2 min Cost: a small glass prism costs about the same as a paperback; everything else is in the house Mess level: low

Do the Experiment: The Prism Rainbow Experiment

Step 1 - Cut a 1 mm slit, not a hole

Cut a straight slit 1 mm wide and 3 cm long in the black card. This is the step people skip, and it is the step that decides whether the experiment works. If you cannot cut a clean slit, tape two pieces of card down with a 1 mm gap between their straight edges, using a coin as a spacer and then removing it.

Step 2 - Darken the room and set up a long throw

Close the curtains. Stand the slit card in front of the torch so the whole beam has to pass through the slit. Aim it at a white wall at least 2 m away. The longer the throw, the wider the spectrum, in direct proportion.

Step 3 - Check the beam alone first

Look at the wall. You should see a sharp bright line about 1 mm wide at the wall if your torch is well collimated, or a few millimetres if it is not. If the line is already 2 cm wide, the torch is too diffuse. Move it closer to the slit or use a torch with a tighter beam.

Step 4 - Put the prism in the beam and rotate it slowly

Hold the prism in the beam about 30 cm from the slit, with the slit vertical and the prism’s long edges vertical too. Rotate it slowly about its vertical axis while watching the wall, not the prism.

What you should see: The bright line swings across the wall and, at a certain orientation, stops swinging and reverses direction. That turning point is the minimum-deviation position, and it is also where the spectrum is sharpest. At that point you should see a band of colour running red, orange, yellow, green, blue, violet, roughly 2.5 cm wide at 2 m with a glass prism.

Step 5 - Measure the spread

Mark the outer edge of the red and the outer edge of the violet in pencil on a card taped to the wall. Measure the distance between them, and the distance from prism to wall. The angle in degrees is roughly 57.3 × (separation ÷ distance). For 2.5 cm at 2.0 m that gives 0.72°.

Step 6 - Recombine it with a second prism

Place the second prism in the coloured beam, inverted relative to the first, about 10 cm after it. Adjust until the colours overlap again.

What you should see: White. Not grey, not muddy: a recognisably white line. This is Newton’s crucial experiment from 1672, and it is the single most convincing demonstration that the first prism separated something that was already there rather than creating colour.

Step 7 - Isolate one colour and try to split it again

Make a second slit card and put it in the spectrum so only the green band passes through. Send that green beam into the second prism.

What you should see: It bends, but it does not split. Green stays green. A prism cannot decompose a single wavelength, which tells you that the colours in the spectrum are elementary and white is the composite.

What’s Actually Going On

Refraction happens because light travels more slowly inside glass than in air. The ratio is the refractive index, n = c/v. Snell’s law relates the angles at the surface: n₁ sin θ₁ = n₂ sin θ₂. None of that, by itself, produces a rainbow.

Dispersion is the extra fact that makes it work: n is not a single number for a material, it depends on wavelength. For Schott BK7, a very ordinary borosilicate crown glass, the published index is 1.5143 at 656.3 nm in the red, 1.5168 at 587.6 nm in the yellow-green, and 1.5224 at 486.1 nm in the blue, as tabulated in the refractive index data for BK7. Blue light sees a slightly “slower”, optically denser glass than red light does, so it bends slightly more at each surface.

A prism has two non-parallel surfaces, so the two bends add instead of cancelling. For an equilateral prism at minimum deviation, the deviation is δ = 2·arcsin(n·sin 30°) − 60°. Put the three indices in and you get 38.40° for red, 38.62° for yellow-green and 39.14° for blue. The entire visible spectrum is squeezed into about three quarters of a degree.

That number explains every practical detail of the procedure. At a 2 m throw, 0.74° is about 26 mm of spread on the wall. If your slit projects an image 5 mm wide, each colour is a 5 mm-wide smear and neighbouring colours overlap across most of their width, which is why a broad beam gives white with coloured fringes. A 1 mm slit gives 1 mm bands inside a 26 mm spectrum, and you see them cleanly. Double the distance to the wall and you double the spread while the slit image stays nearly the same width, which is why throw matters.

Two corrections to common explanations. First, the prism does not “add” or “make” colour, and it does not “turn white into colours.” Sunlight already contains all of them; the glass only sorts them by angle. Newton established this in his 1672 letter to the Royal Society and specifically used the recombination experiment to prove it. Second, the colours you see from a sky rainbow come from the same dispersion but a different geometry: refraction into a spherical raindrop, one internal reflection, refraction out, which concentrates light at about 42° from the antisolar point for red and about 40° for violet.

A note on light sources, because this is where home attempts fail. A white LED is not a continuous spectrum. It is a blue chip at around 450 nm plus a yellow phosphor, with a real dip in the cyan region near 480 nm. Disperse it and you get a strong blue band and a broad yellow-orange band with a weak gap between them, which looks wrong and is correctly reported. Direct sunlight and an old-fashioned incandescent or halogen bulb are continuous and give the full smooth spectrum. If your LED rainbow looks like it is missing a colour, your LED really is.

How to Explain It to Your Kid

Ages 5–8: White light is all the colours stacked up

Tell them white light is every colour travelling together, in a bundle, and the glass is like a crowd of people walking into deep water: the blue ones slow down a bit more, so they turn a bit more sharply. Then do the second-prism step and let them watch the colours go back to white. That moment is the lesson. Afterwards, let them hunt for rainbows around the house: the edge of a mirror, a CD, a glass of water on a sunny sill. Do not let them look toward the sun.

Ages 9–12: Different colours bend by different amounts

Introduce the words refraction and dispersion as two separate ideas, because collapsing them is the usual mistake. Refraction is bending. Dispersion is bending by different amounts depending on colour. Have them measure the red-to-violet separation at 1 m, 2 m and 3 m and see that it doubles and triples. That is a proportional relationship they can plot, and it is the same geometry that governs how a pencil appears to break at the water’s surface.

Ages 13+: Calculate the deviation and check it

Give them the three BK7 indices and the minimum-deviation formula, δ = 2·arcsin(n·sin(A/2)) − A for an apex angle A. Have them compute the red and blue deviations, predict the separation at their measured throw distance, and compare with their ruler. Then ask the harder question: why does an acrylic prism, with n ≈ 1.49, give a spectrum of a similar width even though its index is lower? The answer is that width depends on how much n changes across the visible range, not on n itself, which is what the Abbe number measures.

The question to ask: “If we filled a hollow triangular plastic prism with water instead of using glass, would the rainbow be wider, narrower, or the same? Say which, and say why, before we try it.”

Why Teaching This Early Pays Off

Dispersion is the first place in a child’s science education where a property turns out to be a function rather than a constant. The refractive index of glass is not a number, it is a curve. That shift, from “materials have properties” to “properties depend on conditions”, shows up again in resistance versus temperature, solubility versus temperature and reaction rate versus concentration. Meeting it in a context you can see is a genuine head start.

It is also the foundation of spectroscopy, which is how we know what stars are made of. Once a child accepts that a material sorts light by wavelength, the step to reading a spectrum for information is short, and it leads directly into building a spectroscope from a CD and a cereal box.

Galili and Hazan’s 2000 study of learners’ knowledge in optics, published in the International Journal of Science Education, documented how persistently students hold onto a “colours are in the object” model rather than a “light carries the colours” model. Their work is a useful corrective to the assumption that one demonstration fixes a misconception. The recombination step in this experiment is targeted at exactly that belief, and it is the step most home versions leave out.

Before You Start: What Parents Should Know

Best age	6 to 12 for the demonstration, 12 and up for the calculation
Supervision	Adult required for cutting the slit with a craft knife, and adult present whenever sunlight is the source
Real hazards	Looking at the sun. Solar retinopathy can occur in seconds and is often painless as it happens. A prism does not make the sun safe to look at, and neither do sunglasses, stacked sunglasses, exposed photographic film, smoked glass or a welding filter below shade 12, as the American Academy of Ophthalmology states plainly. Keep your back to the sun and look at the wall. Glass prisms also have sharp edges and chip
Don’t	Don’t substitute a laser pointer for the torch. A laser is a single wavelength, so it produces no spectrum at all, and it is an eye hazard in a child’s hand

When It Doesn’t Work

• You get white with coloured edges. The beam is too wide. This is the dominant failure. Narrow the slit to 1 mm, push the torch right up against it, and increase the distance to the wall.
• The spectrum is faint and washes out. Too much ambient light. Close every curtain, turn off every lamp, or build a box: cut the slit in one end of a long cardboard box, put the prism inside, and view the far interior wall.
• A colour seems to be missing. Your source is a white LED, which genuinely has a dip around 480 nm. Switch to direct sunlight through a gap in the curtains, or an incandescent or halogen bulb if you still have one.
• The colours only appear at one angle and vanish when you move the prism. That is correct behaviour. The sharpest spectrum is at minimum deviation. Rotate slowly and find the point where the beam on the wall stops moving and turns back.
• The second prism gives grey, not white. It is not aligned or it is not inverted relative to the first. The colours must overlap completely. Moving it closer to the first prism makes alignment easier.

Take It Further

• Measure the Abbe number by proxy. Compare the red-to-violet spread for a glass prism and an acrylic prism at the same throw distance, with the same slit. The ratio of spreads tells you about the relative dispersion of the two materials, which is a measurement nobody expects a kid to make at home.
• Build a water prism and test the prediction. Tape three microscope slides or three strips of acrylic into a hollow triangle, seal the base, and fill it with water. The spread will be noticeably narrower than glass. Record it and work out why.

What to Watch For Over the Next 3 Months

• Week 4: They spot dispersion in the wild without prompting: the edge of a bevelled mirror, a crystal glass, the rim of a swimming pool.
• Month 2: They distinguish “the glass bends light” from “the glass bends different colours differently” in their own words. That is the concept landing.
• Month 3: If they ask why a rainbow is always a circle arc at the same angle, or why there is sometimes a second fainter bow outside the first, they have outgrown the prism and are ready for raindrop geometry. The second bow comes from two internal reflections and its colours run the other way.

Frequently Asked Questions

Can I just use a phone torch?

You can, and you will get colours, but the spectrum will look wrong because phone torches are white LEDs with a spectral dip near 480 nm. For a smooth, full spectrum you want sunlight or an incandescent bulb. Using both and comparing is actually the better lesson.

Why is my rainbow so small?

Because that is the real size. The whole visible spectrum from a glass prism spans about 0.7° at minimum deviation, which is 2.5 cm at 2 m and 5 cm at 4 m. Move the wall further away or accept a narrow band; the pictures you have seen are usually composites or illustrations.

Does the shape of the prism matter?

The apex angle does. A 60° equilateral prism is the standard because it gives good separation without the beam being totally internally reflected inside the glass. A very thin wedge barely separates anything; a very blunt one reflects the beam out of a side face instead of transmitting it.

Is it safe to leave a prism on a sunny windowsill?

Treat it with the same care as a magnifying glass. A prism can concentrate sunlight enough to mark paper or fabric over time, and it will throw moving patches of bright light. Keep it out of the sun when you are not using it, and never leave it where a child might look through it toward the window.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Newton, I. (1672). “A letter of Mr. Isaac Newton… containing his new theory about light and colors.” Philosophical Transactions of the Royal Society of London, 6, 3075–3087. https://doi.org/10.1098/rstl.1671.0072
2. Polyanskiy, M. N. “Refractive index of BK7 (Schott).” RefractiveIndex.INFO database. https://refractiveindex.info/?shelf=glass&book=BK7&page=SCHOTT
3. The Physics Classroom. “Dispersion of Light by Prisms.” https://www.physicsclassroom.com/class/refrn/Lesson-4/Dispersion-of-Light-by-Prisms
4. NASA Science. “Visible Light.” Tour of the Electromagnetic Spectrum. https://science.nasa.gov/ems/09_visiblelight/
5. Galili, I., & Hazan, A. (2000). “Learners’ knowledge in optics: interpretation, structure and analysis.” International Journal of Science Education, 22(1), 57–88. https://doi.org/10.1080/095006900290000
6. American Academy of Ophthalmology. “Solar Eclipse Eye Safety.” https://www.aao.org/eye-health/tips-prevention/solar-eclipse-eye-safety
7. National Research Council. (2012). A Framework for K-12 Science Education. National Academies Press. https://doi.org/10.17226/13165