In order to calculate effect of AgNPs in fluorescent dye we tried to measure SRES. Unfortunately, fluorescence signals are interfering the actual Raman signals. We deposited nanoparticles on glass slide add organic dye and we compared it with dye in same enviroment without AgNPs. We used LabRAM (Horiba) as Raman spectrometer. kindly, suggest me some protocol for removing Fluorescence from Raman. I also changed excitation laser but fluorescence is still there.
Almost like this type of experiment.....however they no encounter fluorescence problem
http://link.aps.org/doi/10.1103/PhysRevB.79.241404
Dear Debabrata,
You cannot arbitrarily "remove" fluorescence from already collected spectra, but you can minimize it; for example, by using cooling of the sample, minimizing the excitation power density (via total incident power or focal adjustments), and/or using FT-Raman with excitation in infrared range.
Good luck!
Slava
There are several ways in which you can minimize the effect of fluorescence in your sample. The simplest way is to simply change the laser excitation- which you have already tried. You could try collecting two spectra at different wavelengths and subtracting one from the other. This should remove the unchanged fluorescence background.
Another technique to try is photo-bleaching. This method involves irradiating the sample with intense light from the laser for a few minutes. If the fluorescence is caused by containment fluorescence species this method will decompose them. However, this method can also have the effect of damaging the sample through localized heating so you need to be very careful with it.
I hope this may be of some help to you!
Dear Debabrata,
As Vyacheslav and Alexander commented, it is much better to “fight” fluorescence before data actual analysis.
However, if you want to remove fluorescence from already collected spectra I would recommend the following algorithm: http://code.google.com/p/airpls/ . This algorithm doesn’t require (almost I would say) any user intervention and prior information. They have C++, Matlab, R and Python implementations. Try different levels of parameter “Lambda” which can be treated as a rough estimation of fluorescent peak width.
Most probably, you fluorescence background is very variable in shape and amplitude. Try relatively small lambda. Definitely, by this way you can lower intensities of some wider Raman peaks relative to their actual intensities. Nevertheless, if you will process all spectra using the same settings, you will be able to do comparative analysis.
Still not ideal algorithm, but worked very well with multiple data sets.