an appropriate baseline correction method depends on the baseline. if your baseline is independent of wavenumber you can just substract a constant from your intensity values. if it has a spectral shape, you may want to do a more sophisticated approach, e.g. fitting the baseline to a polynomial and subtracting it. there are many ways to do this.

in general, I'd recommend to arrange the experiment such that there is no significant baseline in the data. this can often be achieved by using appropriate laser sources, filters, signal collection arrangement, data acquisition  

134

1. Cut te cpectrum below  100 cm-1.

2. Use the standard software, which should be available on your spectometer. It is easy.

If you have not such a software, send me your spectrum in the ascii form (two columns)

You need to subtract the background first. Then you could use the software to do the baseline correction, the parameter you could use lines or polynom(the higher degree you choose, the baseline will closer to your data). BTW, if the baseline introduces more peaks, you could use Der2 instead of baseline. Good luck!

If you don't have any spectroscopic software you can:

1. If you are familar with programming in R - download freeware e.g. HyperSpec

http://hyperspec.r-forge.r-project.org/ - it can much, among others also baseline correction;

2. The Origin has also a "baseline correction" tool, although not particularly suitable for Raman spectra. I don know how this menu item called in english, i have german software. In my programm it is 7th menu button from left in the menu bar. I have Origin 6.0.It results normally an overfitted baseline in the first step but you can correct this line manually and then, it necessary fit the line once more e.g. with polynomial and use the resulting function to subtract.

Dear all,

Thanks for your kind reply and will try the same to do base line corrections.

Thanks alot.

Dear Hamza,

Thanks for your usefull information.

How to do deconvolution Raman spectrum ? what information we are getting from after deconvolution of raman spectrums ?  Can any one explain to me ?

What the influence of baseline correction on quantification analysis? 

I understand the question is a bit old by now but I will try to provide a better answer than the above in case someone is still looking for an answer.

I think a good polynomial fit should allow to reduce the signal without losing information of small peaks and changing the peak ratio which, in my opinion, hold important information.

I strongly recommend the use of the so-called ModPoly algorithm, which is arguable the most used in the community. It was developed in 2003. I give the original ref:

Lieber CA, Mahadevan-Jansen A, Automated method for subtraction of fluorescence from biological Raman spectra, Applied spectroscopy, 57 (2003) 11, 1363-1367

 I implemented the algorithm in Matlab to use it, although you may use R or python for example which are free. The default values are 200 iterations, 5th order polynomial, but note that the iteration number should be set to avoid any “overfitting” : i.e., the baseline fit should not “cut” through your original spectra or you will lose data.

The 5th order is good when you use the whole spectra (5OO to 1800 or to 3000) for example. If you consider individual peaks then a straight line or 2nd order is better.

 I compared this method with other such as recently published Goldindec method by Liu and colleagues (Appl Spectrosc. 2015 July ; 69(7): 834–842. doi:10.1366/14-07798).

The authors provided a Matlab algorithm and a sample test. The sample test runs well when using 5th order. However when I tried to use it on my data of E. coli Raman spectra, using same parameters, the algorithm performed very poorly. (I provide the figure here). In fact, it looked more like a simple polynomial fit of the 2nd order leaving wide spaces on the edge of the spectra, while I was expecting the algorithm to fit better each peak. In addition, the baseline fit “cut” through small peaks. In this regard, ModPoly performs better.

I am also interested by another recently published paper (He et al. 2014, Anal. Methods, 2014, 6, 4402 doi: 10.1039/c4ay00068d) unfortunately the authors do not provide any code so we can’t really test it… At present I don’t think there is a comprehensive study that compare various methods on different datasets. Also, the nature of the baseline itself is still unknown so it is difficult to find a consensus on which method should be applied.

I would recommend using R software, there are many packages for baseline correction about Raman data. ChemoSpec is a very popular package.

Me too I rather advice you to use Originlab

http://www.originlab.com/doc/Tutorials/Create-Subtract-Baseline

Hi, You can follow the steps detailed in the link above to make the baseline corrections using originlab software. 

http://www.originlab.com/doc/Tutorials/Create-Subtract-Baseline

The software Origin provides a function called "Create Baseline" that matches your need.

Is there any any to do it using excel???

In Origin2016 there is a good option to subtract the baseline of whatever shape: https://www.originlab.com/doc/Tutorials/Create-Subtract-Baseline

Previous versions don't provide this option.

Hi there!

I would recommend Peakfit either 4.0 or 4.12 (which is the lastest) for Raman. I'm also using it for baseline correction and peakfitting differential scanning calorimetry data. Keep in mind that it does not work properly on ultrahigh resolution screens (touch-sensitive), so I keep using my old WinXP for such a case (what a waste for my new gaming laptop!). Not sure if it is also due to some conflict with Win10 OS. Peakfit is a bit pricey, but not as Origin.

Please see this video

How to Correct Baseline in Origin

https://www.youtube.com/watch?v=sUi_0jWKDWs

Hi Venkatesan,

You can use XRDsingle.py to do the baseline correction of your Raman data automatically, you'd just have to change the x and y axes labels

Links below:

https://github.com/andrewrgarcia/xRD/blob/master/howto.ipynb

https://github.com/andrewrgarcia/XRD

It is very much easy in Origin software. Check out this tutorial link

https://www.youtube.com/watch?v=3ij7blBQ0SQ

This video shows how to perform a baseline correction for XPS, XRD, FTIR, Raman & UV-Vis in origin.

https://youtu.be/Uyby8vrox6U

Hmm, many suggest using Origin, no one suggested this one: why not use free optical spectroscopy software http://Spectragryph.com, which has "adaptive" baseline subtraction algorithm. It is single-pass, non-iterative, extremely fast and gives pleasant results...