The Pulsar Engineering Community Board

Hello,
I have been trying to redo the methanol concentration experiment described on the Pulsar Homepage. However I did not have any methanol and used propanol instead. I get a raman-signal on the camera from both alcohols however I found that the signal comming from the propanol is much stronger than for ethanol. This makes it difficult for me to compare the spectra when I mix both substances since the signal of the propanol will overshadow the signal comming from the ethanol.

When I tried to compute the concentrations the results were widely off with values that were above 1, meaning it is more than pure propanol, negative values etc. I suspect the propanol to be much more fluorecent due to the spectrum that I got.

I do not use the OpenRAMAN software and instead used a python algorythm that I programmed myself. It takes the mean value in each column of pixels as the intensity for this particular wavelength. This helps me avoid hotpixels and noise to some degree.

See my results below: The first image is the spectrum of ethanol, the second one of propanol and the third one compares the two intensities in one graph after background removal (yes I got negative values for some point).

I suspect the propanol to be more fluorecent than the ethanol due to the wide shape of the spectrum. If anyone else has encountred that problem and found a way to solve it I would love to hear about it and exchange ideas.

Based on the images I would rather orient to a mis-aligned system.

What integration time & gain did you use ? Even at 100 ms you should be able to recognize the Raman emission but here it looks like you're just collecting background noise.

Would you have the direct camera image to share with us ?

I wanted to add a note that the OpenRaman software also takes the average of each column of pixel data for the same reasons

I have the images taken of the isopropanol and ethanol in the zip-file attached to this post. I used 200002 µs exposure time on the camera and a gain of 47,99 dB gain. When I aligned my build up I first saw a signal with the isopropanol sample and tried ethanol next but when I did I saw nearly nothing. Therefore I tried to maximize the signal comming from the isopropanol by further ajusting the mirrors. Now those are the results of the adjustment. The ethanol sample is the dimmer one. I have in the meantime ajusted the height of the mirrors aswell to beter put the signal in the middle of the screen. I am not completely sure how misalignment could give me those results so if you have an explanation how this might have occured please let me know.

The second one is fluorescence for sure.

Can you try now to set the gain to zero and try and exposure time of ~10 sec ? If the image is still white, decrease the exposure time by 50% until you get something that does not saturate. We should see some lines then.

I have attempted to use your instructions to redo images for ethanol and isopropanol. I did not get any good image for the ethanol with an exposure time of 10 s and gain set to 0. I then tried if I get the lines for propanol using shorter and shorter exposure times. I tried 5 s, 2.5 s, 1.25 s and 0.625 s. I did not see anything really only that the spectrum got dimmer and dimmer. When I plotted the intensities I also did not see any improvements in the resolutions of more peeks. The intensity in creased but the shape stayed the same with expceptions of noise. For some reason the isopropanol is still just very fluorecent compared to the ethanol. Do you have any other suggestions as to what I could try? You mentioned misalignment earlier, but how can misalignment cause such a high degree of fluorecence and why would it only affect the isopropanol and not the ethanol? Anyway those are my results from this morning. Are there any books about raman spectroscopy you could recommend? Maybe I can search for them in our library and maybe find an answer there.

Alright I managed to redo the experiment. I used background removal to get to those results and wanted to share them here with you as a thanks for your help. I also fitted a linear function to it of the form y=a*x+b. The parameters a and b were compatible with their expected values (a=1 and b=0) within one standard deviation. The errors on the measured concentration were due to the resolution of my scale (about 200 mg) and the uncertainties on the computed values were calculated using the variance of the concentration parameter. I had to calculate the results slightly different than on the webpage in order to end up with one parameter alpha but it returned the same values for the computed concentration using the previous way. We now have our own idea on what experiment we would like to perform with our OpenRAMAN build. I will post the results and my progress here aswell since you guys strike me as the kind of people who would appreciate that.

Hope you have a great one.