Mr. Petrone, If your IR-spectra are measured in transmission, that the correction of wavelenth dependence of penetration dept and refractive index dispersion (ATR correction algorithmm) is not need for determination of the components of any band by the curve-fitting algorithm. But if your IR spectra are under ATR experiments than you should apply, but I am not sure that this shall be novelty. Most probably it has available chemometrics studies. For curve-fitting approach including on Amide I band please find a reference:

Talanta, 69, 2006, 822-828, Validation of reducing-difference procedure for the interpretation of non-polarized infrared spectra of n-component solid mixtures, Ivanova, B.B. , Tsalev, D.L., Arnaudov, M.G.

Luigi,

Your statement "Usually ATR correction is required for quantitative analysis..." seems to be "too strong". Why? Usually it is not. ATR correction is not a dogma. Rather, an ATR correction of ATR spectra is justified, wise, reasonable, useful, etc. when one has to compare ATR-acquired spectra with those recorded differently - e.g., during library search with spectral libraries recorded in transmission. By their own, ATR spectra are not worse (and are not better) for quantitative work than others.

Yet, when talking about amide 1 band (proteins or what?), other experimental sample prep parameters such as, concentration (if solution), water content (if dried) and hence secondary structure, reproducibility of applied pressure, etc., have much more stronger effects on the band envelope (and hence on deconvolved and fitted component bands) than the ATR corretion. So, controlling these sample parameters deserves more attention than performing ATR correction.

Actually, these topics are well covered in literature. Please try to find, e.g, (1) QUANTITATIVE SPECTROSCOPY: THEORY AND PRACTICE BRIAN C. SMITH, Spectros Associates, Shrewsbury, Massachusetts © 2002 Elsevier Science (USA)

or (2) Fundamentals of Fourier Transform Infrared Spectroscopy // the second edition, Brian C. Smith, © 2011 by Taylor and Francis Group, LLC

Thanks Alexander!

Maybe I was not clear in my question :)

I am not saying ATR spectra are bad and therefore need to be corrected. In fact, I have never applied ATR correction in my papers. As you rightly say, the correction is necessary when comparing IR spectra recorded in different modes.

But I feel the ATR correction is needed for instance when fitting the Amide I band (obtained from ATR-IR analysis of a protein) into its peak components to get quantitative information on its the secondary structure (% of alpha helix, beta sheet and other structures).

I am assuming all other precautions to be perfectly carried out, i.e. water subtraction, baseline and so on…I am also assuming that the molar absorptivity of each components is the same (I feel they are all different, but we can assume they are the same in all analyses).

Without the ATR correction, the area under the band components at higher wavenumber will be underestimated. If I simply apply the ATR correction, the shape of the amide I band slightly changes, enhancing components at higher wavenumber.

The paper you sent me is excellent, but they do not mention this aspect.

Furthermore, we still need a criterion to set the width of the band components on an Amide I band. I am digging into the literature but I haven't found anything helpful :(

Thanks Alexander for the stimulating discussion!

Luigi

Hi Bojidarka,

thanks for your reply.

Of course I am doing analyses in ATR mode, thus my question on ATR correction.

I checked your paper that you suggested me for the fitting of the Amide I band in its structural components. However, besides Figure 3, in which you show the deconvolution and fitting of an Amide band, where can I find an explanation on how the fitting was carried out?

there is no explanation and assignment for the deconvolution of the Amide I band, not to mention the band width issue.

Can you please elucidate it better?

thanks

Luigi

Mr. Petrone, for deconvolution particularly please find a review by P.R. Griffiths, where the method is explained comprehensively:

Analytical Chemistry, 63, 1991, 2557-2563

Comparison of fourier self-deconvolution and maximum likelihood restoration for curve-fitting (Review), Jackson, R.S., Griffiths, P.R.

By the way it have more recent reports on application of deconvolution on ATR spectra. Furthermore in paper 1, the authors claimed that the study is quantitative, but I have not time to pay attention on the paper content. May be you shall find where is the novelty in your study comparing the data:

1. Polymer Science - Series A, 55, 2013, 91-97

Detection and quantitative determination of the crystalline phase in poly(vinyl alcohol) cryogels by ATR FTIR spectroscopy

Tretinnikov, O., Sushko, N., Zagorskaya, S.

2. Colloids and Surfaces B: Biointerfaces, 85, 2011, 97-102

Time-dependent adsorption behavior of β-lactoglobulin on ZnSe crystal surface studied by 2D correlation ATR/FTIR spectroscopy

Xie, J., Huang, H.

it depends on how thick your protein film is and which ATR crystal (Ge, Si, ZnSe) you are using. For a dry protein film of 100 nm (n2=1.5) on a Si ATR crystal (n1=3.5) you get at 1700 cm-1 a depth of penetration (dP) of 414 nm and at 1600 cm-1 a dP= 439 nm (the lower wavenumber the higher dP). (n3 = 1 (air) was considered, n1, n2, n3 are refractive indices of crystal, film and environment).

Due to Harrick (Internal Reflection Spectroscopy, Harrick + Mirabella, 1985) you can calculate the effective thickness (de) at these wavenumbers, which is proportional to the measured absorbances at these wavenumbers.

They are nearly equal and thus negligible.

Best regards,

Martin Müller

... so for the case I have described in previous answer you need no ATR correction.

Best regards,

Martin

Thanks Martin for your answer. It is a good approximation, plus the decay of the intensity of the IR evanescent wave dictates that at longer penetration depth the sensitivity decreases..so, we can omit ATR correction in this narrow region.

thanks!

I would perform a dispersion analysis of the reflectance spectrum. Since dispersion analysis is based on wave optics, the best possible ATR correction is automatically included.