Hi Terry,

how have done the manual base line correction? In Omnic, when you perform the manual base line correction, you can choose between polynomial, linear interpolation and spline. I think in your case the spline might work best.

Terry Franklin Have you got heating of the sample? Carbon black, graphene and similar materials all absorb the incident radiation and converting it to heat which interferes with the detectors (which also respond to heat).

Thank you Kevin and Alan.

Kevin Hellemans, I have tried the manual baseline correction tool, but there seems to be no difference between using linear, polynomial, and spline. All of them produce linear lines between the points. Is there something I am missing that allows for a nonlinear fit?

Alan F Rawle, I am not sure if my sample is heating up since it is a very small amount that is put on the FTIR-ATR. It would make sense that the carbon adsorbing and releasing heat could be the mechanism behind the sloping. If that is the case, there would be no way to correct for this except baseline correction.

Terry Franklin We see a similar effect for Raman spectroscopy for this type of material. Indeed with too much incident energy one can get the particles to explode! The amount of sample is irrelevant as it's the spectrum of the material you're measuring.

A very beautiful spectrum presented by Terry. Yes, such spectra are typical for substances with very high refractive indices.I agree with Kevin: spline baseline correction is most appropriate for such sort of spectra.

Hi,

As was mentioned by Alan, carbon black (and most black samples) will absorb in the mid-IR range, making them difficult to analyze using FTIR and resulting in low SNR ratios for absorption bands trying to come out of the backgournd.

One thing I might suggest is using a "pure" carbon black for background correction (if such may be obtained from your application) : acquire background on this "pure" sample and run your sample afterwards, using the background correction from your "pure" sample (and not that from the atmosphere or the window you're using). Maybe it will work better than software baseline correction.

We use this technique to avoid introducing "software" related artefacts in some sloping spectra, but I've never done it on carbon.

Pierre Caulet , that is a great idea. I can see how using a baseline correction tool can introduce false peaks. Although, it is my only option at the moment since I do not have a "pure" carbon black sample. I will closely verify corrected spectra with the original to qualitatively check that I am not introducing false peaks. Here is that example spectra corrected with the Omnic manual baseline correction using spline.

Thank you for all your help.

What is your ATR crystal and your angle of incidence? I suspect you use a diamond crystal and 45°, then you do not perform a ATR experiment, because the refractive index is larger than 1.7 and the critical angle would be higher. You have basically three options, either use a higher angle of incidence or a ATR crystal with higher refractive index (Ge!) or you correct your spectrum by applying dispersion analysis. Just removing the baseline will not correct for the errors you introduce concerning peak shapes, intensities and positions...

The best way to do baseline correction regardless the sample type is OriginPro software. have you tried it?

Ok, to be very clear, your (experimental!!!) baseline should be close to a reflectance of 1, otherwise your spectra will definitely not even remotely resemble absorbance if you take -log (R/R0).

I tried to correct it. But not sure whether it's what you want or not.

Thomas Mayerhöfer , it is a ZnSe crystal with an angle of 45. If I do not have another crystal or a way to change the angle, do you have a source that could show “dispersion analysis”? Would using a DRIFTS accessory with KBr mixture instead of ATR negate this issue?

ZnSe has about the same index of refraction as diamond, which means that the same limits apply. Concerning dispersion analysis, you can analyze and correct your spectra using RefFIT by Alexey Kuzmenko (https://reffit.ch/) or CODE by Wolfgang Theiss (https://www.wtheiss.com/). Both have good manuals. Also SpIRIT by Konstantin Shportko (http://www.shportko.com/spirit.html) may be able to help you. Concerning DRIFT, you have to use Kubelka-Munk theory to extract the absorption coefficient, which is highly approximative (I would not go for it).

Alan F Rawle and Pierre Caulet , would pure graphite powder be an appropriate background for biochar? I am assuming it would have the similar refractive index or heating effect as biochar without all the functional groups?

Terry Franklin You can just try it and see. These were just suggestions and one good practical experiment will defeat any number of Greek letters (more than an Athens post office) in a theoretical paper.

Terry Franklin, just the same as Alan with me : just give it a try.

I assume biochar are actually ashes from non-oxidative calcination experiments ?

For correcting the sloping baseline in FTIR-ATR, it is necessary to click on "Auto baseline"in the menu bar of the software (OMNIC) as soon as the test is finished.

Mina Karimi-Avargani In my opinion it is necessary to first find out what causes the baseline...

Of course the shape of your chart is a result of very high refractive index of biochar. BTW, ZnSe has a rather low refractive index, e.g. at 1000cm-1 appr. 2.4

- very close to KRS-5. It follows however from my experience - for IR spectra measurements of common organic compounds with refractive indices n = 0.8---1.6 KRS-5 gives much more accurate information than ZnSe. Usually if the ZnSe ATR-crystal was employed - the IR-peaks broaden and the highs of the relatively strong peaks are “cut off”, while KRS-5 gives good quality spectra This was always a puzzle for me (?).

Speaking about your question - my advice is to use a polynomial or even manual baseline correction.

If you have a baseline very different from 100 % in a (seemingly) ATR experiment, you should not correct the baseline, you should correct the experiment!

I've seen this behavior before when exciting shallow-trapped electrons into the conduction band for other materials. Maybe that is what you are actually seeing and not an instrumental or sample preparation problem. More specifically I've seen this for TiO2 that had lots of oxygen vacancy defects. Maybe this data gives you insight into the bonding network of your carbon system. Hope that helps! Good luck.

Robert B Balow Surely, semiconductor have plasmons that lead to contributions which increase with decreasing wavenumber (Drude's model!). But this also leads to an increase of the refractive index and a corresponding violation of the total reflection condition. Certainly this is neither an instrumental nor a sample preparation problem, but a problem of proper design of the experiment.