Author Topic: Measuring light spectrum intensity  (Read 1870 times)

0 Members and 1 Guest are viewing this topic.

Offline scatterandfocusTopic starter

  • Regular Contributor
  • *
  • Posts: 165
  • Country: us
Measuring light spectrum intensity
« on: October 06, 2019, 04:05:17 pm »
I was planning to experiment with some photodiodes with an oscilloscope for measuring light intensity of various sources at wavelengths from uv to infrared.  I'm noticing that photodiodes have a peak response given in datasheets by the wavelength specification in nm (nanometers).  And it seems that cost goes up drastically when looking at photodiodes with a peak response in the blue to uv spectrum, and cost gets really high at around 400 nm and higher.  And phototransistors seem to only go up to around the 500nm wavelength, which would make them not useful for the blue to to uv spectrum.  Maybe someone knows of another means for measuring light spectrum intensity, or has other ideas for getting around the cost of expensive photodiodes.
 

Offline StuartA

  • Regular Contributor
  • *
  • Posts: 103
  • Country: gb
Re: Measuring light spectrum intensity
« Reply #1 on: October 06, 2019, 04:26:50 pm »
Chemical analysts have been using IR, visible and UV spectroscopy since the middle of the last century, so this requirement was being met long before photo-transistors came on the scene. I cannot recall now what devices were used in these systems, but I'm fairly certain that there would have been separate optics for each range. If you Google 'Infrared Spectroscopy' (etc), I suspect that Wiki will come up with an overview of how the different machines evolved.
 

Offline KaneTW

  • Frequent Contributor
  • **
  • Posts: 811
  • Country: de
Re: Measuring light spectrum intensity
« Reply #2 on: October 06, 2019, 05:01:38 pm »
Photospectrometers use a photon detector (single tube/diode or diode array) with monochromatic light. Monochromacy is achieved either via an adjustable filter (single detector) or a diffraction grating before the diode array.

There might be some more ways that I'm not aware of, but that's what I've worked with in the past.
 


Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5571
  • Country: us
Re: Measuring light spectrum intensity
« Reply #4 on: October 06, 2019, 06:29:08 pm »
Historically the detector was a photomultiplier tube (PMT).  Spectral response was determined by the photoemissive properties of the target material and the transmission of the tube envelope.  Quartz windows were used to achieve the UV wavelengths.  Since response varied with wavelength, just as it does with current semiconductor diodes some method of calibrating the response was needed.  Now we tend to use calibration tables to provide a gain adjustment.  Then many instruments used a dual optical path to provide a ratio correction.

PMTs are still available, are still very good for performance, are far from cheap and require high voltages not readily available in today's hobby labs.  The links from previous responders give good options for current use.
 

Offline KaneTW

  • Frequent Contributor
  • **
  • Posts: 811
  • Country: de
Re: Measuring light spectrum intensity
« Reply #5 on: October 06, 2019, 07:33:57 pm »
You can probably get a PMT device for pretty cheap. My Varian Cary 1 cost maybe 100E on ebay and is fully functional with a barely used deuterium lamp.
 

Offline eliocor

  • Supporter
  • ****
  • Posts: 531
  • Country: it
    • rhodiatoce
 

Offline scatterandfocusTopic starter

  • Regular Contributor
  • *
  • Posts: 165
  • Country: us
Re: Measuring light spectrum intensity
« Reply #7 on: October 07, 2019, 05:23:49 pm »
Many thanks for all the suggestions.  I probably never would have found this stuff by searching. 

This is the first I have heard of a photomultiplier tube.  Those things seem to have a wide detection spectrum, which is very interesting, where using photodiode arrays or cameras seems limited by comparison.  Using a photomultiplier seems like it would essentially boil down to having a suitable power supply (up to maybe 1500 volts), utilizing it in some sort of spectrometer setup, and bringing the output voltage down to something usable for an oscilloscope or other device.  Probably digitizing the output from the photomultiplier for having some means of graphing and measurement would be best here.

But very naively, I hadn't even considered the frequency spectrum of light and bandwidth of the measuring device, where light sits in the THz to PHz range.  So how are any of these devices capable of measuring frequencies in the light spectrum? 
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1530
  • Country: gb
Re: Measuring light spectrum intensity
« Reply #8 on: October 07, 2019, 06:19:31 pm »
Quote
So how are any of these devices capable of measuring frequencies in the light spectrum?
I do not think the devices measure the freq.
They respond to light/em waves between certain freqs.
They can measure amount of said light within the freqs.

Measuring system use some kind of filter for each specific freq and the device measures the amount of light that gets through.
 

Offline KaneTW

  • Frequent Contributor
  • **
  • Posts: 811
  • Country: de
Re: Measuring light spectrum intensity
« Reply #9 on: October 07, 2019, 07:16:37 pm »
Yeah, they don't measure the frequency. You filter the incoming light instead, and calibrate the gain (either via dual path or stored coefficients)
 

Offline ker2x

  • Regular Contributor
  • *
  • Posts: 83
  • Country: 00
Re: Measuring light spectrum intensity
« Reply #10 on: October 07, 2019, 07:35:25 pm »
Excuse me but... aren't you simply trying to build an optical spectrometer here ?
https://en.wikipedia.org/wiki/Optical_spectrometer

a grating plate (or a prism, pretty much the same thing) and a camera ?
You will have a problem with UV and IR and anything out of the humanly visible spectrum however.
You can buy astronomy camera (with IR/UV filter) but they tend to buy expensive because it's a niche market. And if you add an IR bandpass filter you'll end up with very VERY low sampling rate. (1hz or even less)
often times... there are other approaches which are kinda crappy until you put them in the context of parallel machines
 

Offline edpalmer42

  • Super Contributor
  • ***
  • Posts: 2351
  • Country: ca
Re: Measuring light spectrum intensity
« Reply #11 on: October 07, 2019, 08:15:57 pm »
I'm surprised nobody has mentioned https://spectralworkbench.org/ .  Depending on the level of precision required, any old DVD and a webcam plus an hour or two of work might be all you need.

Ed
 
The following users thanked this post: CatalinaWOW

Offline jbeng

  • Regular Contributor
  • *
  • Posts: 53
  • Country: us
Re: Measuring light spectrum intensity
« Reply #12 on: October 08, 2019, 12:01:33 am »
As far as PMTs go, it is rare to find inexpensive ones.
I previously worked for a company which built machines for medical research called flow cytometers (cell-sorting machines), which relied on PMTs for detection of the very weak high-speed optical signals.  The ones we used (mfg.= Hamamatsu) were modules which had thier own integral HV power supplies running on 15vdc and had an input for gain control.  IIRC they were approx. 1200 USD each.
"It's such a fine line between stupid and clever" - David St. Hubbins
 

Offline scatterandfocusTopic starter

  • Regular Contributor
  • *
  • Posts: 165
  • Country: us
Re: Measuring light spectrum intensity
« Reply #13 on: October 08, 2019, 12:21:36 am »
It seems that there are quite a few diy projects out there for spectrometers using cameras, but they all seem to drop off at the beginning of the UV spectrum.
 

Offline donlisms

  • Frequent Contributor
  • **
  • Posts: 308
  • Country: us
Re: Measuring light spectrum intensity
« Reply #14 on: October 08, 2019, 08:54:16 am »
If you want to experiment with photomultiplier tubes, an easy way to get started might be to pick up an old densitometer on an auction site. These were used a lot in darkroom work.  I have some Macbeth photographic-oriented devices, mostly tube-based.  They have gorgeous big mirrored panel meters, the power supply for the PM tube is already there, there's a log conversion circuit if you'd like to keep that.  Assuming you get one that works or only needs a little attention, you can just start right in.  I can't speak to the spectral response, but for photographic purposes, the goal was "flat through the visual spectrum," which was accomplished with three filters and some tuning.  I have a vague recollection that there was an IR filter in the path, but that could be my imagination.
 

Offline scatterandfocusTopic starter

  • Regular Contributor
  • *
  • Posts: 165
  • Country: us
Re: Measuring light spectrum intensity
« Reply #15 on: October 08, 2019, 02:02:22 pm »
Quote
So how are any of these devices capable of measuring frequencies in the light spectrum?
I do not think the devices measure the freq.
They respond to light/em waves between certain freqs.
They can measure amount of said light within the freqs.

Measuring system use some kind of filter for each specific freq and the device measures the amount of light that gets through.

So I have to wonder how it can be verified which range of the spectrum is being measured.

Also, I am now seeing that LED's can be used as light sensors, and UV led's aren't expensive, where UV photodiodes are.  But it isn't clear to me how to know which wavelengths a particular LED will be sensitive to when used as a light sensor.  But if that can be known, it seems that an LED could be used as an already filtered light sensor.
« Last Edit: October 08, 2019, 02:08:03 pm by scatterandfocus »
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5571
  • Country: us
Re: Measuring light spectrum intensity
« Reply #16 on: October 08, 2019, 06:43:24 pm »
There are probably other and better ways to know the frequency, but ones I have worked with use a hammer and tongs approach.  First identify one frequency.  Using a line emission from an ionized gas.  The sodium line is a common choice.  That sets the offset on the scale.  The slope of the scale is built into the instrument and is based purely on geometry.  Using the diffraction equation for units using a grating, or the dispersion equation for prism based instruments the relationship between angle and wavelength is known, and the gear train that moves the grating (prism) and the scale on the dial are chosen to match that.  That linear relationship between wavelength and angle is why spectrometers are marked in wavelength or wavenumbers.  Much easier to calibrate a linear scale than a reciprocal one. 

Recognize that due to the high frequencies involved even very narrow band filters are hundreds of MHz wide.  Its just a scaling problem.  A 1kHz bandwidth filter in the SW RF is roughly 0.1 percent bandwith.  The same 0.1 percent bandwidth at optical frequencies is about 500 GHz.  So your very good and accurate spectrometer might have three or four digits of wavelength accuracy but have dozens or hundreds of GHz frequency error. 

You might be able to use LEDs as your sensors, but it will be a tough road to follow.  They are optimized for emission, not the other way around and so are relatively insensitive.  In most cases in spectrometry there is relatively little light available and sensitivity is a prime requirement in the detection chain.  It is why photomultipliers were so dominant in the tube era.  Even with the extra noise of the multiplication process the gain provided was useful in getting above amplifier noise levels.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf