A few tips on running spectra

Before we consider the thorny issue of how best to present the sample to the infrared radiation, it is worth considering briefly the general approach to use when recording spectra. 

Traditionally, infrared spectra are displayed with abscissa units of wavenumber and ordinate units of percent transmission. The infrared active bands of the sample are consequently revealed as the dips in the 

Typically, 20 or 30 scans will do for background and sample single beams. Recording hundreds of scans should be avoided, except for specialist applications (such as very weakly absorbing samples, e.g. monolayers, or in areas where high signal-to-noise is essential, deconvolution, for instance). 

Another way of boosting signal-to-noise is to reduce the resolution. Although 4 cm is largely the standard in FTIR analysis, bands are never that narrow, except for gases. Indeed, a peak with an intrinsic FWHM 

---