0
\$\begingroup\$

What components are sensitive to the wavelength (energy) of impinging light?

I am wondering if there are any that have a reliable and detectable response to wavelength instead of just intensity. For example, if some hypothetical component responded to light in the range of 400-1000nm, then when exposed to say 10 photons at 400nm perhaps its voltage (or impedance) is a significant multiple of the voltage when exposed to 10 photons of 800nm?

Clarification: I know that it's possible to do spectral imaging using prisms and/or filters so that a sensor can say, "I was just hit by this many photons," and based on the filters we know what those photons' wavelengths could be. What I'm wondering is whether there is any component that can say something like, "A photon hit me this hard," where this is a relatively precise measure of the photon's energy.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Have you seen the Pink Floyd logo? \$\endgroup\$ Mar 18 '17 at 18:45
  • \$\begingroup\$ @MarkoBuršič yes, I know prisms are used for old-fashioned spectrometers. I'm wondering if there is a way to measure the energy without spatially sorting or filtering the photons. E.g., a component that says something like, "Yes, not only did I just get hit by a photon, but it hit me this hard." \$\endgroup\$
    – feetwet
    Mar 18 '17 at 19:20
1
\$\begingroup\$

In order to measure the wavelength of a photon, you need to measure its energy. This is common for example with x-rays or gamma rays because they are ionizing radition -- they have enough energy to knock electrons free from atoms, and we can measure the energy of the electron. It is not normally possible with visible light because the energy it too small, but there are detectors, called transition edge sensors (TES) that are energy resolving photon counters. They measure the energy of a photon by the change in temperature of a thin metal film right at the superconducting transition. Unfortunately, they typically work at 0.1 K and require complicated electronics and optics to use.

\$\endgroup\$
2
  • \$\begingroup\$ Interesting. Reading on TES tends to categorize them as a special class of bolometer, and it appears that commercial uncooled bolometers already do exactly what I'm describing for the LWIR spectrum (which I assume means they could do the same for shorter wavelengths). \$\endgroup\$
    – feetwet
    Mar 19 '17 at 13:12
  • 1
    \$\begingroup\$ A bolometer by itself isn't enough, they tell you total power/energy. To get the wavelength, you need to know that and the number of photons. TES devices are special because they can measure the energy of a single photon. That said, you could split the beam 50/50, send half to a photodiode and half to a standard bolometer and work out the average photon energy. \$\endgroup\$
    – Evan
    Mar 20 '17 at 4:11
3
\$\begingroup\$

Basically all bandgap devices (read: semiconductor devices) are specifically sensitive to the wavelength for which photons have exactly the energy needed bring an electron from lower to higher energy band. So, for example, LEDs (light emitting diodes) work pretty well as selective photodetectors.

So, a blue LED will be significantly more sensitive to blue light than to red light.

Technically, it's much harder to do the opposite of what you want: have an output that depends only on intensity, not on wavelength of the incident light.


This is really simple to test: Get a monocolor LED. Take Blue.

Now, there's different ways to use an LED as photodiode, but I think the easiest for you will be to use the LED, and connect it in reverse to a (stable) voltage source (e.g. 4.5V coming from three AA batteries). Use a sensitive multimeter in series to measure how much current passes through the LED in reverse direction.

First, cover your LED in darkness, measure current, then expose it to bright blue light – another LED (operated in forward to actually emit light) of the same kind, and compare the current readings. Then, compare current readings when subjected to other colors of light – from a RED LED, for example.

If your ammeter isn't sensitive enough for µA to mA, you might want to feed the current to the base of a transistor pair in Darlington configuration.

\$\endgroup\$
4
  • \$\begingroup\$ Interesting. I was thinking more along the lines of CCD and CMOS sensors, which, although not uniformly sensitive, are responsive to photons across the visible and NIR spectrum. (And for all I know, maybe they do have a strong response to frequency, if the intensity can be held constant.) \$\endgroup\$
    – feetwet
    Mar 18 '17 at 16:57
  • 3
    \$\begingroup\$ @feetwet What Marcus has outlined is inherent wavelength sensitivity of the underlying semiconductor - CCDs and CMOS generally have a bias towards infrared. Additional optical treatment is added to their surfaces to make them more blue or green or red sensitive. And these "filters" are more narrow-bandwidth than the inherent infrared bias. \$\endgroup\$
    – glen_geek
    Mar 18 '17 at 17:13
  • 1
    \$\begingroup\$ A red LED is likely to be nearly as sensitive (on a photon flux basis) to blue light as to red. But a blue LED won't be sensitive to red light. \$\endgroup\$
    – The Photon
    Mar 18 '17 at 20:19
  • \$\begingroup\$ @ThePhoton whilst there's certainly energy levels that the electron could take, it'd usually take additional impulse to make it go there – aaaaaand this is where I realize we're talking about indirect semiconductors anyway, so you're right. \$\endgroup\$ Mar 18 '17 at 20:55
2
\$\begingroup\$

All opto-detection devices have a light frequency sensitivity profile.

However, they also all produce a single variable, e.g. voltage, resistance, or current depending on how it is configured. As such, it is impossible to tell what changed, the impinging light intensity, or the light's frequency.

What you are asking for is a chromatic light sensor. These do exist as equipment and rely on using prisms or lenses to "split" the light onto an array or line of light sensitive semiconductor devices or to focus a particular wavelength onto a single sensor.

However, the technology called "Foveon X3" uses a newer method to make a single spot on a silicon substrate sensitive to different wavelengths. However, unless someone else can find one, I know of no single "diode like" product that uses this technology. If you have a project in mind though, you may be able to use a Foveon X3 camera sensor.

\$\endgroup\$
1
\$\begingroup\$

For example, if some hypothetical component responded to light in the range of 400-1000nm, then when exposed to say 10 photons at 400nm perhaps its voltage (or impedance) is a significant multiple of the voltage when exposed to 10 photons of 800nm?

You can do this two ways:

  1. Choose a semiconductor material for your sensor with a bandgap energy somewhere between the photon energy associated with 400 nm wavelength and the photon energy associated with 800 nm wavelength.

    Unfortuntely I'm not familiar with the materials with bandgaps in this wavelength range so I can't suggest any particular material off the top of my head. As Marko's answer suggests, a red LED might be a good candidate.

    Also, this approach won't get you a detector that's sensitive at 800 nm but not at 400 nm.

  2. Start with a detector that's sensitive to both wavelengths and put a reflective coating on its surface (or on a piece of glass in front of it) that is highly reflective at 800 nm, but highly transmissive at 400 nm. Using multiple thin film layers, it's possible to design a coating with essentially any profile you like of reflectivity vs wavelength. This way you could also make a detector sensitive at 800 nm but not at 400 nm. Cost of course goes down if you choose a profile that can be achieved with a small number of layers.

    You could, of course, also think of using things like absorptive filters in front of your sensor to vary its response to different wavelengths.

\$\endgroup\$
2
  • \$\begingroup\$ I just added a clarification to my question: I already know we can use filters to ensure that if a photon hits a sensor it has to have defined characteristics. What I'm wondering is whether there is any sensor or component that can discriminate wavelength precisely, over some meaningfully wide range, for any photons that hit it. \$\endgroup\$
    – feetwet
    Mar 18 '17 at 22:18
  • \$\begingroup\$ @feetwet, no there aren't. That's why we're telling you about the nearest alternatives. \$\endgroup\$
    – The Photon
    Mar 19 '17 at 1:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.