Want to talk about wastewater? Join us for Open Call on this topic every Tuesday through September. Click here for call-in details!

Public Lab Research note

Grading oils with UV fluorescence

by mathew | September 19, 2014 17:44 19 Sep 17:44 | #11163 | #11163

What I want to do

I want to be able to grade oils from light to heavy. I'm going to use a means of UV Fluorescence with a 405nm laser pointer. I'll look at the emission spectra and see if I get a shift of the spectra's peak from green to blue as I move from light to heavy oils. My goal is to replicate a UV Fluorescence technique used in UV LIDAR research.



  • Can I replicate the location of emission spectra from a 337nm Nitrogen laser?
  • Can I see peak transition in oil grades from light to heavy oils?
  • what about motor oils?
  • can I distinguish diesel from crude?

    My attempt and results

the setup

The target oils were things I could find nearby-- a few grades of motor oil, some gear oil, crude oil from North Dakota, and deisel fuel. The idea is lighter oils should have a greener peak, and heavier ones a blue peak.

I started with a rectangular cuvette that @warren sent me and filled it with 1ml of mineral oil.


I determined concentration by adding single drops of the target oils.



do the different solutions have the same peaks?

These different concentration spectra seem to have similar peaks between 1-3nm apart.

the higher concentrations gave me a spectra with a clear cutoff, as if I saturated the image. Is this correct? I want to trust the lower intensity spectra more. should I?

Do I see a shift from green to blue as the oil gets lighter?
Motor Oils

Will the heavier 80W-90 have a more blue peak than the 20W-50?

  • 5W-30 has a peak at 444nm
  • 20W-50 at 436-8nm
  • 80W-90 at 432-435nm

Well, they seem to roughly follow the blue shift pattern

2-cycle is unlabeled, and so its grade isn't known. but it has a peak at 445

Crude Oil & Diesel

I see 10nm between North Dakota crude and Diesel. The North Dakota crude is dissolved 1 droplet in 2ml of mineral oil, the diesel is undiluted, I am comparing them at similar fluorescences, not concentrations.

Is that good? I'm not really sure. it seems like less of a gap than between different grades of crude.

Questions and next steps

I need people to replicate this experiment.

Hypotheses: 1) by varying the concentration of the target oil to get a lower, less blown out spectrum. true or not?

Why I'm interested

To distinguish crude oil from fuel oil, and ultimately from organic naturally occuring substances that also fluoresce, such as decaying organic matter (DOM) or humic acid. IMG_20140918_123107.jpg


Just a quick followup on area normalization:


We really need smoothing that we can save too.

Reply to this comment...

Diesel, gear oil, and N. Dakota crude:


Reply to this comment...

OK, added "equalize area" and scripting to the sets page: https://spectralworkbench.org/sets/show/1461

It still doesn't save, but makes that part of the comparison easier. I'll see if I can add savable smoothing over the weekend.

Reply to this comment...


Reply to this comment...

ok, trying to see if I can get good results from the Sanm camera, since all these spectra are from the discontinued Syba camera.... nope. the Sanm camera has a "blowout" at the spot where I'm expecting a crude peak. There is something that looks like a peak at lower brightnesses, but its a pretty inconclusive spectra:

Reply to this comment...

Login to comment.