Infrared spectroscopy is a popular method for estimating protein structure, but it has its limitations. First, the region of interest (amide I ~1640/cm) is characterized by strong water absorption bands, making studies in aqueous solutions difficult. To overcome this problem, deuterium oxide can be used as a solvent. Even then the information obtained is only qualitative because absorption bands in the amide I region can not be resolved(1).
In order to overcome this problem, D. Michael Byler and Heino Susi applied Fourier self deconvolution to the spectra for 21 globular proteins they were analyzing. They were then able to characterize eleven frequencies for secondary structure and make estimates on the percentage of each type of secondary structure content. These estimates were in agreement with values derived from X-ray data, proving that infrared can be a useful technique for secondary structure analysis(2).