Monday, May 27, 2013

Intro to DIY Raman Spectroscopy

I've been working on a Raman spectroscopy setup in my shop for a while, and was finally able to collect some real, verifiable data this evening. Raman Spectroscopy is a technique where light is directed toward a target, and the reflected light is color-shifted by the size and type of the molecular bonds in the target. This is a non-destructive way to determine an object's molecular structure. The problem is that the color-shifted light is many, many times weaker than the non-color-shifted light. A Raman spectroscopy setup compensates for this, and allows meaningful data to be collected.

27 comments:

  1. Very interesting, too bad the filter lens is so expensive. It seems air the beam travels through doesn't change the result, but surely the ambient light might reflect into the DSLR. Getting rid of every light source (even IR) might be a bit difficult, how did you do it ?

    ReplyDelete
  2. Hats off to you sir!
    We're all looking forward to more. Eventually you'll see fit to toss in some of your x-ray tubes and delve into fluroscopy?

    ReplyDelete
  3. Hi Ben,
    very impressive (as always)! Commercial Raman machines sometimes use three diffraction gratings to block the excitation wavelength. The first one splits the light into a spectrum. Then the excitation wavelength is blocked by an absorber and the rest of the spectrum is recombined by the second grating. Finally the light is spit again by the third grating onto the detector (CCD or Photomultiplier). This method allows a very good suppression of the excitation wavelength. If this is not needed one can of course skip the first two gratings and just block the excitation wavelength.

    BTW these CCDs are sometimes cooled by a Peltier element or even liquid nitrogen to allow very long integration times. And if you really want precise measurements then go for a gas laser.

    Cheers
    Michael

    ReplyDelete
  4. Great intro to Raman scattering!!

    Could you write in a comment the name of the open source optical analysis software? I cannot understand its correct spelling from your video and google isn't infering it.. Thx in advance.

    ReplyDelete
  5. @Andre Esteves: Ben mentions Octave, which is an Open Source MATLAB analogue. You might also look at Scilab, which I believe is also open/free and in the same vein.

    ReplyDelete
  6. It seemed to me he was saying something like optiv. I know Octave and Scilab. Octabe is more matlab-like. But Scilab is more complete with inbuilt symbolic functions and XyCos, a simulink equivalent.

    ReplyDelete
  7. Hello
    Great stuff !!
    Would the camera be damaged if you would drop the expensive filter?
    Could you block some of the outgoing light between the spectrometer and the camera? At this point , the light is already split, blocking a certain frequency is just a matter of putting a paper strip at the right place between camera and spectrometer???? Would this work?

    Br
    jan

    ReplyDelete
  8. That's an idea. But i believe the light would be decreasing in intensity because it was not collimated (all rays running in parallel), you would have to put a few more lenses to collimate it , lengthening the optical route and lowering even more the beam's intensity (air disperses light). Using a filter is a way to minimize that and conserve the laser collimation.

    ReplyDelete
  9. hello Andre
    this is the way I saw it
    http://imageshack.us/a/img9/5005/naamloosor.jpg

    I did not try it, just an idea

    Jan Lietaer

    ReplyDelete
  10. As I said before: Some commercial Raman machines Do block the Rayleigh light by an obstacle in the path between the grating and the camera (or to be precise: They move the CCD in such manner that it can no longer be reached by the Rayleigh light. By doing so you can’t collect the anti-Stokes lines but they are less intense anyway).

    ReplyDelete
  11. Michael
    This must be far cheaper than the expensive filter? There must be a reason why this was not done?
    Maybe this would have ruined the camera if the absorber was in the wrong place?
    br
    jan lietaer

    ReplyDelete
  12. Cool and very interesting intro into that technique.
    I'm just curious, wouldn't it be possible to just don't use that notch filter at all and remove/ignore the HeNe laser wavelength in the final spectrum image? Or am I missing something obvious and if the HeNe light reaches the CCD it messes up everything else around?

    br, Kosi

    ReplyDelete
  13. Kosi, the light from the laser is so much more intense than the signal from the Raman scatter, stray reflections and scatter from the diffraction grating would completely saturate the image sensor.

    ReplyDelete
  14. It might be worth exploring optical disc drives. They use a laser diode and an interferometer. So you'd have laser, beam splitter and focusing lens in one tight little package. CD at 780nm, DVD at 650nm, and Blu-Ray at 405nm. Blu-Ray may cause fluorescence.

    MensaDropout

    ReplyDelete
  15. Ben, thanks for the clarification :)

    ReplyDelete
  16. A linear ccd might give you a higher dynamic range at the same price. Maybe you could use the "wasted" beam to cancel out the scattered excitation line by overlaying it (destructive interference)

    ReplyDelete
  17. Great job, Ben!! Have you considered using the sensor from a scanner? From my limited experience modifying them, they are pretty easy to pop open completely, and change out/remove everything between incoming light and the sensor. They have great resolution, albeit linear.

    You could tie the scanning mechanics to some kind of small movement, to adjust the focus or cause a slight panning movement with the sensor.
    ~~Wizzard

    ReplyDelete
  18. Awesome stuff. You might be running into a problem with the detector's response at the NIR range so that might be why you aren't seeing it. If you manage to get both sides try estimating the temperature from the ratio of the antistokes and stokes. It can really blow your socks off with how good it can be!

    ReplyDelete
  19. Like always, a great project!

    Another fun thing to build could be an ion mobility spectrometer.
    Then you can do full airport style security at home and check your guests for drugs and explosives.

    I've never seen a diy project, but it should be straightforward. You could try paperspray ionization, flame ionization, or a corona discharge instead of the Ni63, which is presumably impossible to get in our paranoid society.

    http://en.wikipedia.org/wiki/Ion-mobility_spectrometry

    ReplyDelete
  20. Hi,
    Can you provide the references you used in this project?

    ReplyDelete
  21. Mohammed, I don't have a list of references, but all of my information came from searching the internet (using only public sources). Good luck!

    ReplyDelete
  22. Did you ever figure out why you can't see the Stokes lines? I know you took off the IR filter on your camera, but it seems like you are using a color camera - so each pixel has a color filter to produce RGB images. Usually those filters don't allow much light at longer wavelengths than 650 nm (maybe 700 nm). So if you are using a color camera, I'm not totally surprised that you don't see the Stokes lines. However, it is awesome that you can see anti-Stokes lines!

    ReplyDelete
  23. Check the lines you get with the fluorescence lines from neon. HeNe lasers have significant borelight, light that comes from the discharge of the laser and makes it through the OC. The spectrum you get really looks like the neon spectrum. From what I know getting a Raman signal is very hard and you setup doesn't seem accurate enough. Getting only anti-stokes lines is even more suspicious as I think those were even harder to get. Try filtering the borelight out before using the laser beam, see if that removes the lines.

    But anyway I love your work I'm a physicist myself with a huge lab at home, I should start doing something with it.

    Regards,
    Johan

    ReplyDelete
    Replies
    1. Johan, it's possible that the signal is from helium or neon, but there is a strong green peak in the signal that I received (7:56). I don't think neon or helium has any emission peaks in the center of green. I haven't looked at the setup since I made this video, but I'll have to do an A/B comparison to make sure the signal that I'm getting is from Raman scattering.

      Delete
  24. Ben, I believe LEDs work in reverse (photoelectric effect?) as sensors. I have always wondered if the excitation range is as wide as the emission range, and if it's upshifted or downshifted. A ring of different colored LEDs in which one emitted and the others sensed could be an interesting and affordable way to collect ramen data... IF the the LED response could be amplified to a usable level and the differential response of the different LEDs used to create a synthetic notch filter.

    ReplyDelete
  25. Hello, Ben. This system is really cool. Thanks for sharing. The simplicity of the setup is what I like the most. However, I am a bit confused. I am no expert, but from all of the literature I've read, it seems that Raman shift typically has a red-shift (Stokes) scattering because, unless you have molecules which are already in an excited state, you cannot get more energetic photons than the ones you used for excitation. To get a blue-shifted signal (anti-Stokes) you need to have an already excited population of molecules, which can be done, but this is beyond the scope of this comment.
    The spectrum you showed had a negative shift (anti-Stokes), while the comparison spectrum has a positive shift. Have you figured out why you didn't get the expected infrared signal? Did it have anything to do with the stray light? Thanks again

    ReplyDelete
    Replies
    1. Anonymous, thanks. I'm pretty sure Stokes and anti-Stokes spectra will be collected from normal specimens under normal illumination. As long as the total amount of energy is conserved, it's OK to have some photons shifted into higher energy and some into lower energy. The Wikipedia page on Raman Spectroscopy is what I am using for my source on this. It's possible that my setup was actually capturing the "bore light" from the HeNe laser, which I didn't filter. I don't know why the IR part of the expected scattering was not visible. I haven't looked at this setup since posting the initial video.

      Delete