Waves in the Spectrum—A Possible Effect of Scatter on near Infrared Spectra?
George Bázár · NIR news · 2015
I have been working with NIRSystems instruments for many years and I am very familiar with these particular spectrophotometers. I am interested in looking at the spectra and trying to find out what the NIR signals mean. It is a kind of hobby, and this paper is a short explanation of some thoughts and doubts I have had regarding the results of recent experiments. I share these observations in the hope that if anyone else has similar questions and is interested in finding out the explanations for these phenomena, they will share their results with all of us also. One day, I was about to measure a very thin white ceramic plate in transmission mode. It ought to have been a dispersive transmittance reference for some measurements. As the plate was measured, some nicely shaped waves appeared instead of the expected more or less flat line. I did not pay very much attention to these at first. I thought the ceramic plate, the composition of which was unknown to me, had specific absorption bands at some regions. I became further interested in these spectra later on when I measured an open top quartz cuvette with fog on its surface (for some reason I simply breathed into the cuvette, so there were no droplets, only a thin deposit of vapour). Besides some characteristic water (vapour) signals at 1362, 1382, 1838, 1870 and 1902 nm, many bands appeared in the second derivative of the fogged surface which was something very similar to what I had previously observed months before with the ceramic. Figure 1 shows the second derivative transmittance spectra of the ceramic plate and the fogged window of the quartz cuvette. For some folks it might be obvious what these waves were, but it was not for me or for anyone who I asked. So I started playing with the spectra. According to Karl Norris, changing the size of the gap is of great help in understanding the nature of spectral signals. Small gap size enhances narrow band signals over wide band signals, while larger gaps enhance wider band signals over narrow band signals. If you go below the bandpass of the spectrometer and signals are still enhanced, than it is highly unlikely that the signal is coming from the absorbing sample or from the light beam, but from the electronics or some mechanical effects distorting the beam. My spectra were recorded with various spectrometers: Foss NIRSystems XDS and Foss NIRSystems 6500 models. Results were very similar, although some differences were apparent (Figure 2). Later, in another study based on Karl Norris’ idea of analysing the gap size, I have concluded that some parts of the spectrum might have problems in some instruments. In the 6500 models, the 770–790 nm region may show some systematic signal errors because of the changing bandpass filters in older monochromators, 1500 nm may contain some error due to the turning of the grating and some other parts of the