Cosmic rays won't consistently appear in the same spot.  the ~600 cm-1 broad peak and the 1649 cm-1 peak both sound like features of residual water - although CaF2 shouldn't really absorb much water, you might be getting condensation on it for some reason.  If that's the cause, then you can try drying it, or washing the water away with a small amount of acetone. 

First,.. many thanks to Aurelio and Kaoru for your suggestions.

Let me clarify this problem further.

Earlier i had taken spectra of glass cover (on 532nm and 785nm lasers) and aluminium (on 785nm laser) substrates.

For the glass cover, I did obtain SAME broad signal around 700cm-1 and of course the much fluorescence around 1400cm-1 and again,..the SAME sharp peak at around 1650cm-1.

Similar observations are found on Raman signal from aluminium substrate (on 785nm laser).

See the attached signals for all substrates.

I can't see the files  (the file format isn't supported on my mobile device), but based on your description, I would suspect first that you are getting room light  (fluorescent lamp) that is contributing to this constant background spectrum.  You can test this by turning off all your lights to see if your spectrum still has the same background.  If it helps, and you still prefer to have the room light on, then you can figure out how to block room light from reaching your detector.  Remember that Raman scattering intensity is several orders of magnitude lower than the incident laser intensity, so even a light leakage from room light can be a problem.  (This is probably more common with confocal Raman benchtop equipment, as older Raman equipments tend to be large and enclosed in a windowless room, intended to be operated in darkness).

Would you also mind telling us the model of your equipment, if it is a benchtop equipment? 

Yes,.. It is a benchtop assembled / customized Raman system

Kaoru:

I can open and view your MATLAB figure files. I'm not seeing any fluorescence in the traditional sense for the CaF2 or the Al metal. As for the glass sample, I think you've got some organic contamination on the surface of the glass. Assuming it's a standard microscope slide (Boro-silicate glass), there shouldn't be any bulk fluorescence. Physisorbed water has no significant Raman signal in the 1200 to 2000 cm-1 band. If you have it, try soaking the glass slide in conc. H2SO4 overnight, then rinsing it in DI water followed by a dry N2 blow dry and a +100C bake. That should give you a pretty clean and dry surface. Prepare a couple of slides and compare results.

As for the very narrow peak occurring at about 1660 cm-1, that narrowness is typical of an atomic emission line. It doesn't look like a 'hot' pixel it's too broad for that. You weren't clear about what laser was used to take the spectra you've provided. But, if the laser wavelength was 532 nm, 1660 cm-1 would correspond to a wavelength of about 583 nm. Cesium has a super bright line at that wavelength. And, Cesium is pretty common in lighting systems.  The fact that it doesn't move from spectra to spectra would indicate that you took all three spectra with the same laser ... no? Otherwise, if it was room light, and you did used different laser wavelengths, that line would move a bit.

The broad peak(s) near 600 cm-1 are also not fluorescence. I notice that the peaks for the Al and the glass sample are very similar, while the peak for the CaF2 sample is red-shifted a bit. If you used the same wavelength for all three spectra (532nm) that peak shouldn't have moved if it was due to room light light leakage. It's possible you have a contaminant common to all three samples. Other than that, I can't say.

Having also made my own Raman spectrometer, I agree with all the others: Put your spectrometer in a dark box.

If you've got the ability to do depolarization measurements, you might want to check the VH, VV polarization combinations and compare. Light leakage won't show any sensitivity to the polarization state.

Oh, one last thing. If you have a fluorometer and/or a UV spectrometer, check your samples for any sensitivity to your laser wavelengths.

I hope this helps!

Thanks a lot  to all of you and your comments.

@P J Bisson: thank you for the posting, it clarified what was on the file (which I canot view either on my computer or my other devices).

Just for clarity, the mention of MATLAB figure refers to what J Githaiga attached (I was saying that I couldn't read the file on my device - so I did not post it).  Also, the reference I made to fluorescent lamps is regarding room light, and *not* a reference to any fluorescence that sometimes occurs when laser light hits a sample or substrate.

I certainly agree with your statement that the narrow peak probably corresponds to an atomic emission line.  The fact that the narrow peak doesn't move, despite J Githaiga's statement about having three spectra, one with a 532 nm laser and another with a 785 nm laser, is very odd, I agree.  If the narrow peaks are coming from fluorescent lamps in the room, then their absolute frequency does not change, so that when one switches to a different laser frequency, then the Raman shift of that atomic emission should also move.  So one would think all three then came from the same laser (likely the 785 nm)...

@Kaoru, P J Bisson,..well for clarity purpose, the few spectra files i attached were specifically based on 785nm laser.

Meanwhile,.. today i have been single stepping through our system, with aim of collecting the mentioned problem. I will attach some spectra findings once i am through with analysis.

Thanks and enjoy the remaining part of the day.

Based on your advice, I have taken measurements on a blank Calcium Fluoride substrate at exposure time: - 50 s, No of accumulations: -2, on 785nm laser at laser power – approx. 7mW, 50x magn. I have also attached four consolidated matlab graphs for your review.

Data1 - In the presence of ambient light I have obtained Raman peaks at 320cm-1, 364cm-1, 380cm-1, 411cm-1, 1264cm-1, and 1656cm-1.

Data2 - On switching off the lights, but epi-illumination still present, the Raman peaks are 320cm-1, 326cm-1, 364cm-1 and 1656cm-1. Compared to Data1 scenario, It is more likely that the now inactive Raman peaks at 380cm-1, 411cm-1, and 1264cm-1 may have been contributions from fluorescent lamps in our lab.

Data3 - In order to find out any contributions from epi-illumination originating from Tungsten-Halogen lamp, I have repeated measurements when all lights are switched off and epi-illumination completely switched off. I have obtained the Raman peaks at 320cm-1, 362cm-1, and 1657cm-1.

Data4 – But on repeating measurements with all lights  switched off and epi-illumination set at lowest minimum (at preset level in our Raman system), the peaks are 364cm-1, and 1656cm-1. For some reasons i have to find out, the expected 320cm-1 peak is inactive.

I tend to think the 362cm-1(or 364cm-1) may have been a likely contribution from Tungsten - Halogen lamp. I also note the two Raman peaks at 362 (or 364cm-1) and 1656cm-1 have the less intensities when measurements are taken with all lights switched off and epi-illumination light off. Again at these settings the 320cm-1 peak is prominent and is closer to 323cm-1 peak as Aurelio pointed out. This seems to be my best setting option for further measurements.

In spite of their relative less intensities, I haven’t managed to completely eliminate the 362cm-1 and the prominent 1656cm-1peak. Currently, I am finding out ways of bringing these peaks contributions to the lowest level possible.

Any comments please?,…..