Previous Page

Next Page

phase scales in Figure 11, a higher phase corresponds to a later signal.  The magnitude and phase spectra were calculated using Equations (14) and (21) from the same pair (in-phase and quadrature) of interferograms.  The PMMA magnitude spectrum in the bottom panel of the figure (dotted line), shows no interferences larger than the small (about 1 unit tall) CO2 doublet at 2360 and 2335 cm-1 and no water-vapor structure.  Its companion phase spectrum in the top panel, on the other hand, has a prominent (about 10°) CO2 doublet and smaller (about 2°) water-vapor structure between 1500 and 1850 cm-1.  The PMMA itself absorbs sufficiently strongly at most mid-infrared wavelengths that the phase effects from the background or surface-film signal are modest.  The largest interference from the background signal is the broad, 9°-tall feature between 3100 and 3550 cm-1.  This is a common phase-spectrum interference probably caused by water or other hydroxyl species adhering to surfaces.  This broad feature demonstrates the apparent phase-band reversal that can result from background-signal interference.  PMMA has a small absorption peak at 3440 cm-1, which shows up in the phase spectrum as a gap in the interfering hydroxyl feature, so it looks as if the 3440 cm-1 band were pointing toward higher phase instead of lower phase like the other PMMA bands.  This is the characteristic appearance of phase interference from the background signal – a broad region of faster than expected phase in which small absorption bands of the sample appear to be partly or fully reversed in direction.
          Figure 11 also shows a phase spectrum and companion magnitude spectrum (solid line) for polyethylene.  Polyethylene has a generally lower absorption away from its main peaks than PMMA does.  As a result, its phase spectrum is much more susceptible to background-signal interference.  At first glance, the polyethylene phase spectrum appears noisy because there are so many features that apparently have little or no corresponding structure in the magnitude spectrum, but many of the features are real, not random noise.  The water interference band between 3100 and 3550 cm-1 is again present, and the dip in phase centered at 1690 cm-1 is also probably from superficial water.  Many small absorption bands appear reversed in the phase spectrum because of the background interference.  The upward-pointing features in the phase spectrum at 2350, 2150, 2020, 1900, 1815, 1360, 1300 and 910 cm-1 all correspond to small absorption peaks in the magnitude spectrum.  Phase spectra contain much useful information complementary to magnitude spectra, but the background interferences can make phase spectra more difficult to interpret in weakly absorbing spectral regions, especially at high modulation frequencies.