# Can prism separate different colour of light?

Can prism separate colours of light? For example, I have a purple colour. purple colour is by mixing of red and blue. I that possible when purple colour light is passed through a prism it separate the red and blue colours. Why it cannot split purple into red and blue?

• Yes, definitely! Isaac Newton is rather famous for doing such things!
– Ed V
Mar 5, 2023 at 18:17
• But also note: "Purple" is special because you can only get purple by mixing red and blue. On the opposite side of the color wheel you've got yellow. If you see a light source that looks yellow, it either could be a mix of red and green, or it could be a single, pure wavelength. The prism will split red from green, but it will not split the single, pure wavelength into red and green. Mar 5, 2023 at 19:08
• @SolomonSlow Indeed! I even have an old bipolar LED that emitted red light when powered with one electrical polarity and it emitted green light when powered with the opposite polarity. So yellow was just a matter of using appropriate alternating polarity power for the LED.
– Ed V
Mar 5, 2023 at 19:23
• red blue and green are not the only fundamental frequencies of colour, you would not decompose a purple into its components colours since it is assumed we have one frequency of light incident the prism, purple. Mar 6, 2023 at 14:49
• If you make purple light by adding together red and blue light, then a prism can separate them. If you think purple has an assigned wavelength, or is a very narrow spectral band around some assigned wavelength, then a prism will do little or nothing. Personally, I think purple is a mixture of colors, so a prism will disperse the constituent colors. This is a very easy experiment to do.
– Ed V
Mar 6, 2023 at 15:27

Getting new colors by mixing different colors of light is a trick played on your brain by your retina, called metamerism. If you put a prism in the path of a beam of light that is mixture of two colors that appear to be a third color to your eye, the prism will separate them into the two original colors. If you put a prism in the path of a beam which consists of pure purple (produced, say, by another prism or a monochromatic source), it will not split that beam into more and different colors. You'll still get just purple.

• Pure purple? What wavelength would that be?
– Ed V
Mar 5, 2023 at 19:31
• @EdV, 380 nanometers. -NN Mar 5, 2023 at 19:40
• @edv, Yowza! it would be cool if you could post the spectra here for us to look at! -NN Mar 5, 2023 at 20:00
• There's a whole line of purples. en.wikipedia.org/wiki/Line_of_purples Unfortunately, they lie entirely outside the sRGB gamut, so they can't be correctly represented on most monitors, or stored in typical image file formats. Mar 7, 2023 at 7:46

It appears that there are two schools of thought in the matter of the color purple. One school, of which I am a member, holds that purple is a mixture of red and either blue or violet. It is not a color with a canonically or customarily assigned wavelength. It is not a narrow continuous spectral range of wavelengths. The other school considers purple to be a color, either pure or of narrow spectral width.

So the answer to the OP’s question depends on which school aligns with their notion or perception of purple.

To illustrate my view, consider the light produced by an old GE Glow lamp. These were formerly used as inexpensively operated night lights. They produced what I see as purple light:

But my visual perception of purple light from this lamp tells me nothing at all about the light’s relevant spectral properties. However, I have two homemade echelle spectrographs, so I can collect light from the lamp, using a lens and fiber optic collimator:

The other end of the 50 micron fiber optic cable is connected to one of my echelle spectrographs. This allows me to obtain a spectrum, called an echellogram, of the light from the lamp.

This next photograph is the resulting echellogram of the lamp’s light. It is a 30s exposure at ISO 6400.

A longer exposure, of 121s, produces this echellogram:

Both of these echellograms show that red and violet spectral lines are prominent and green spectral lines are relatively less prominent. This is consistent with me seeing the light as purple.

To begin, light from various light sources, such as LEDs, tungsten filament light bulbs, sunlight, hollow cathode lamps, etc., typically consists of spectral ranges of constituent ‘colors’. For LEDs, the spectral ranges are continuums, whereas for hollow cathode lamps (the typical atomic absorption spectroscopy light sources), the spectral output consists of many narrow spectral lines.

Using the echelle spectrograph shown here, it is easy to spectrally disperse the light from a desired light source. The spectra, called echellograms, are two dimensional, unlike the typically one dimensional spectra produced by single gratings or prisms. The echellograms are somewhat quirky in that the grating orders, which look like curved arcs, can have some redundancy.

Using various colored LEDs I happen to have in my junk drawer, here are their echellograms:

The UV LED has a nominal center wavelength of 395 nm and a measured center wavelength of 398 nm.

In contrast, here are echellograms for a strontium hollow cathode lamp with neon fill gas:

and an iron hollow cathode lamp with neon fill gas:

For both of the hollow cathode lamp echellograms, a blue filter was used to attenuate the intense red region neon spectral lines.

• Pretty! And I expect they look even better live. :) Modern colour displays are pretty good, but their gamuts are still noticeably smaller than what humans (without colour-blindness) can typically see. Mar 7, 2023 at 7:43
• @PM2Ring They do, indeed, look better live! In fact, I built the second spectrograph, in a shoe box sized salvaged metal box, specifically so I could position an eye as close to the cross dispersion prism as possible. The field of view then is stunning: the echelle grating arcs extend roughly 150 degrees.
– Ed V
Mar 7, 2023 at 12:20

According to the question, can prism split it separate different colours. Due to the conic nature of our retina, we process an extremely large amalgamation of light reflections, the ratio of which is to be seen as a perfect spectral color. So, in reality, you are seeing violet; red and blue with a little more blue, whereas when you see purple, it's more of a red and blue 1:1 situation As prism has not such biology in it.hence it will split purple colour light into red and blue colour light And only VIBGYOR has a wavelength in all colours.If any colour of light is passed through than only from these colours we will get through prism.