I'm trying to obtain a raman spectra of chitosan (natural polymer with ammine groups) but the only thing that I get is the image I attached.
As you can see, there is the raylight (it is not perfectly filtered) and then a band from 100 to 1100 cm-1. The band is supposed to be the intrensic fluorescence of chitosan. Do you have any tricks to avoid the fluorescence because I'm afraid that it covers the raman signal?
(I'm already using a 780nm laser because I read that it can reduce the fluorescence but I still have it).
Hi,
Did you try to use a polarization filters on the scattered beam? That might help, I think.
Hi Davide,
The best way to obtain a good raman signal with chitosan is to performe FT-raman. Maybe at 780nm you will have small peaks in the fluorescence background, but you will need a good baseline correction in order to obtain your spectra.
Like suggested already, using even a longer wavelength of excitation (1064 nm) would be one option, but a couple of things listed below can be tried as well. However, let me tell you first that these two suggestions are based on our work with lignocelluloses which do not contain the amino group, so the sample response may be different. But since these are simple chemical treatments it's worth trying them. 1) treat the chitosan sample with 1/2 molar NaBH4 at RT for a couple of hours (with periodic shaking) and then wash the samples with excess water until the pH is neutral. 2) subject the sample to complexation by EDTA or DTPA.
Davide, I will be interested in knowing if any of the treatments worked.
Davide, it seems that there is no way to avoid fluorescence of chitosan when using 780-nm excitation. Also, the chitosan spectrum distortion seem does not originate from laser heating the sample – since (i) the spectrum baseline turns out to be equally distorted when using 5 or 100 mW of laser power, and (ii) dilution with KBr or KCl (1-3 % chitosan by weight) and subsequent pressing the disk does not help either. Thus, one ought to live with these peculiarities of chitosan! In fact, however, chitosan usually does give useful Raman spectra (please see attached PDF).
Judging by the spectrum you’ve posted, I would advice the following: (1) somehow press the sample to increase the material density (the number of molecules) under the laser beam; (2) increase the laser power – usually chitosan endures up to 100 mW of 785 nm laser without burning; (3) lower the resolution (increase the instrument slit) – yours spectrum (detector) seems to be starving of Raman photons. As you may see in my attachement, the 2, 4 and 8 cm-1 instrument resolutions gave basically the same spectra (it is a condenced phase, and chitosan intrinsically is a “morphological nightmare”). So there is no use to fight for 1 or 2 cm-1 resolution; (4) considerably raise the CCD exposure time – chitosan is a poor Raman scatterer; (5) try somehow to filter the Rayleigh line completely – may be your CCD is saturated (or somehow is blinded) by the Rayleigh. Actually, there is little (if any) structural info at low (below ca. 300 cm-1) Raman shifts in chitosan case – so there is no need to fight for low Raman shifts; (6) certainly, use several (9-16) accumulated and averaged separate scans (each with exposure time of 60 seconds or more) – chitosan is a tough case. Good luck!
In any case, you will have to correct the baseline. This cosmetics helps a lot. Good corrections are provided, e.g., by Bruker OPUS, ThermoGrams, Perkin-Elmer Spectrum, etc.
If all above does not help – seek for a machine with 1.064 um excitation line as Mark and Umesh have advised already. Again, Good luck!
There is a lot of new stuff out there. If you are in a position to try new instrumental methods to remove the effect of fluorescence such as modulated wavelength Raman (http://pubs.rsc.org/en/Content/ArticleLanding/2013/AY/C3AY40193F) or SSRS (http://pubs.rsc.org/en/Content/ArticleLanding/1998/AN/A802802H#!divAbstract). Alternatively, if you are not in a position to try new instrumental methods such as FT or the others, another option is the use of multivariate based methods. These can handle very large backgrounds with a high degree of reproducibility.
http://onlinelibrary.wiley.com/doi/10.1002/jrs.2896/abstract
Another paper on that
Article Estimation of signal backgrounds on multivariate loadings im...
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs