Public Lab Research note

This is an attempt to replicate an activity.

Foldable Spectrometer Workshop for Cities

by nicholas | February 28, 2017 15:51 28 Feb 15:51 | #13969 | #13969

The foldable spectrometer workshop was held with ten students as a part of the Remote Sensing Class at NYU's Center for Urban Science and Progress. An initial overview of spectrometry was provided describing the fundamentals of using a diffraction gradient to split light into its component spectra.

The workshop goals were to:

  1. Build a foldable spectrometer (largely following this research note)
  2. Collect sample spectra from light sources in the room (CFL bulbs, LEDs and sunlight)
  3. Create a profile and upload images to Spectral Workbench
  4. Calibrate the device using Spectral Workbench

The workshop last approximately one hour which was just enough time to achieve goals 1-3. Some students were able to start calibrating but did not have much time to thoroughly understand the process via spectral workbench.

General Observations and Challenges

We used Gaffers tape to build the foldable spectrometers which worked really well. The tape is strong but can easily be removed without destroying the material underneath.

One challenge that I encounter was understanding the fundamental analysis techniques of capture spectra. For example, @dhaffner describes in this note dividing a sample spectra with a reference spectra in order to obtain transmittance which can then be used to calculate absorption. However, @straylight describes analyzing concentrations in this excellent post, by subtracting a sample spectrum from a reference spectrum in order to obtain absorption. I would like to be able to capture an image and perform the fundamental analytical techniques use Python though I'm not clear on the overarching methods for doing this.

I also noticed students were timid to attach the spectrometer to their phones. Over and over I observed students holding the spectrometer in front of the phone's camera while led to light leakage and stray light entering into the device. As a moderator, I continued to explain that they would observe better results if they attached the device to their phone but very few of them did. I think this might be from a concern of applying an adhesive to their mobile device.

We also found in two of the spectrometers, the slit allowing light to pass into the spectrometer at one end was not entirely 'open'. We manually used a pair of scissors to further create a larger slit to allow the appropriate amount of light to enter. I'm not sure if this resulted from a 'rugged' folder who mistakenly applied too much tap and thus warped the inlet slit or if the material was supplied like this.

Most students had difficult in determining the proper placement of the DVD-R diffraction gradient. This was partly the fault of the instructor (myself) but reflects the need for special attention in the splitting and cutting of the diffraction gradient. Most students immediately started cutting and quickly lost track of how the DVD-R should be placed. In most situations we had to find the side with the strongest rainbow.

Several students were entirely unsuccessful in splitting the DVD-R into two 'clean' sections. The aluminum part peeld off and stayed with the clear layer leaving only a small usable section. I'm curious if there are better methods for splitting the DVD-R and removing the dye. This post by @MrBumper is excellent but I wasnt able to remove the purple dye after washing in warm water with soap. I'll try using other material to clean in the future.

in the calibration process, students immediately used their mobile devices to access spectral workbench. Several people were having difficulty navigating the interface and noted a few 'bugs' or lack of response from the site via the mobile interface. I advised them that the desktop version would probably be a better experience

What's next:

This post by @warren about Fraunhofer absorption lines in sunlight is very interesting and I'd like to explore this more. One thought is to use a raspberry pi to capture spectra images periodically (every hour?) as well as basic weather information temp/humidity etc. I'm curious 1) what is the sensitivity of the spectrometer and what can be measured and 2) what can be learned from repeated measurements of the Fraunhofer absorption lines and 3) what other gases (if any) can be detected. Does this require further modification to the camera (remove ir filter) and/or the diffraction gradient?


Hi Nicholas,

That sounds like it was an exciting hour.

If a live video connection to Spectral Workbench is an obstacle, it is always possible to take a still photo of the diffraction pattern and upload that to Spectral Workbench. There is no difference in analysis, and the quality of the image might even be better.

The linked wiki page on absorption does not seem to be a helpful place to start. Too bad it can't be labelled as such. A critical procedure in measuring spectral absorption is usually missing from discussions at Public Lab. The goal of these measurements is to see that a substance in the sample blocks some wavelengths compared to the control. So both spectra (sample and control) must be made with the same exposure settings. This requires that you use a phone camera app (or webcam app) that allows you to lock the exposure for two or more photos or video frames. Phone cameras (and webcams) usually use automatic exposure which can compensate for brightness differences between the two photos and compromise your results.

Fraunhofer lines are easy to see in spectra of a clear or cloudy sky. The lines are thin, so to see the fainter lines you want good resolution on the wavelength scale. That means a narrow entrance slit, and/or a slit farther from the grating. It also requires good alignment and focus so a clear photo can be captured. These lines will not change much over time, so a timelapse might not reveal much. The overall shape of the sky spectra will change as the sky becomes bluer, grayer, or reddish at different times.

Another fun urban project is to compare the spectra of "neon" signs. All red ones do not contain neon, and with a spectrometer it is really easy to learn which ones do. Also, street lamps can be mercury lamps, sodium lamps, or LEDs. It's easy to tell which is which with a spectrometer.


Reply to this comment...

The Raspberry Pi mod would be very cool, especially as it presents opportunities to do un-corrected images (re: chris's above comments); several folks have posted about it:

and a few have been built:

I agree that this wiki page has some problems:

But I do think it can be labelled as such. Chris, if you think there are key errors with it, you should feel free to edit it or add comments/warnings. I'm also working hard on the #rich-wikis project, which should allow us to do more interactive editing/suggesting/revising of wiki pages.

A critical procedure in measuring spectral absorption is usually missing from discussions at Public Lab.

Do you mean a discussion of how to do comparisons without exposure compensation? I agree -- there are a few easier ways I think this could be done, and it'd be great to see an activity posted for them:

  • non-exposure-adjusted cameras (as Chris outlines), either still cameras or via a Raspberry Pi
  • two spectra in the same image (side by side comparison) so the exposure settings are the same

However, @straylight has posted an absorption activity that's listed on the spectrometry wiki page:

Thanks @nicholas and @cfastie!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Login to comment.