How Many Scans Should Be Used

The improvement of SNR with the number of measurements added together is not isolated to FTIR. All measurements of any quantity have a signal, a noise level, and 

FIGURE 2.11 Top The noise level in an infrared spectrum between 2200 and 2000 cm after 1 scan. Bottom The noise level in an infrared spectrum between 2200 and 2000 cm after 100 scans. The y-axis scales in the two spectra are identical. Note the reduction in noise for 100 scans versus 1 scan. 

The number of scans is perhaps the most important user-adjustable scanning parameter in FTIR, which begs the question, How many scans should I use Ultimately, it depends on your sample. For routine samples that are analyzed in the FTIR sample compartment using sample preparation techniques such as ATR, KBr pellets, and capillary thin films (more on these in Chapter 4), anywhere between 10 and 100 scans will typically yield a spectrum with a good SNR. For example, if 16 scans give a poor spectrum, by all means more scans should be tried. On the other 

The Noise Reduction Expected for a Given Number of Scans Compared to the Noise Level in a Spectrum Measured with One Scan 

FIGURE 2.12 A plot of the signal-to-noise ratio for a spectrum versus the square root of the number of scans used to measure that spectrum. Note that as the square root of the number of scans increases the signal-to-noise ratio increases. 

---