| Author | Message | ||
     Niklas Warne (Niklas) |
Hi, I'm about to start a feasibility study to determine moisture in a confectionery product. It is a chocolate coated peanut covered with a sugar shell. The chocolate layer and the sugar shell are both a few mm thick. What I want to do is scan the whole product, but I'm only interested in determining moisture in the outer sugar layer. My main concern is that the NIR light will penetrate too deep, so that I'm actually measuring the chocolate layer. Is there a smart solution for this? For instance, would it be possible to use regions of near infrared with less energy, that will not penetrate as deep as the more energy-rich parts? Thanks for any suggestions. Regards Niklas | ||
| Gabi Levin |
To Niklas, Almost anyhting I will say will be commercial. However, When you say few mm thick, what is few, 2 or 3 or more? If it is more than 2 mm I think that several diffuse reflectance spectrometers that operate between 1200 to 2100nm will see little of the chocolate layer. If it is more than 3 most will probably do well. If it is less than 2mm, the critical question becomes the variability in moisture in the underlying chocolate layer. If this variability is small enough, you may get along nicely even if the spectrometer "sees" some of the chocolate layer. Anohter element is the variability of the sugar layer thickness. It appears that your model will have to learn a lot in order to be able to correlate to the water in the sugar only. Many wavelengths should help to get better results. Before proceeding with a feasibility, check different spectrometers to see what is the light inetensity of the light source, and this can help you to decide which one gives you a better chance of "staying" within the sugar layer. If you care to send an e-mail to [email protected] I will be glad to look into more specific help. Gabi. | ||
| Bengt Nordlund (Bengtn) |
To Niklas The higher wavelength range you use the shorter penetration deepth. So using the water peak at 1900 will probably give the penetration deepth that you are looking for. Even if there is some light that go deeper will it be most likely that the moisture content in the suger layer and in the chocolate is similar. | ||
| Bengt Nordlund (Bengtn) |
To Niklas The higher wavelength range you use the shorter penetration deepth. So using the water peak at 1900 will probably give the penetration deepth that you are looking for. Even if there is some light that go deeper will it be most likely that the moisture content in the suger layer and in the chocolate is similar. | ||
| Bengt Nordlund (Bengtn) |
To Niklas The higher wavelength range you use the shorter penetration deepth. So using the water peak at 1900 will probably give the penetration deepth that you are looking for. Even if there is some light that go deeper will it be most likely that the moisture content in the suger layer and in the chocolate is similar. | ||
| W. Fred McClure (Mcclure) |
I am a little confused about the statement that "the longer wavelengths give deeper penetration," especially for water-containing samples. Would someone explain why you get deeper penetration at 1940nm compare to 1450 nm, when the absorption coefficients at 1940 nm are much larger than at 1450nm. Have I missed something all these years? | ||
| DJDahm |
I believe that it is better not to think of depth of penetration as primarily a function of wavelength. It is a (inverse) function of the amount of absorption and the amount of scatter from the sample. With particles much larger than the wavelength, there is little change in the amount of scatter with wavelength. There is generally a large increase in NIR absorbtion as one moves to longer wavelengths. | ||
| Tony Davies (Td) |
I cannot find the statement that worried Fred. I do not think you get deeper penetration at longer wavelengths; the reverse is the normal case. If we forget scatter for the moment, then higher wavelength (say 1940nm) photons are more likely to be absorbed therefore they will not (on average) penetrate as far as lower wavelength (say 1450nm) photons. There is a theoretically interesting possiblity here. Water in the crystalline sugar is in a different environment to that in the chocolate phase. Thus the water spectrum in the two phases will be different. If you can isolate these [by some suitable PCA-like chemometric tool] then you would be able to chose which water you measured. There are of course many practical problems! If you can analyse a range of samples for H2O (sugar) and H2O (chocolate) then perhaps PLS would do something like this. Interesting problem, good luck! Best wishes, Tony | ||
| hlmark |
I think we're forgetting one point: the effects of absorption and scattering are opposite with wavelength. We all know that the absorption tends to increase at longer wavelengths, but we forget that scattering decreases at longer wavelengths. So from absorption considerations penetration would decrease at longer wavelengths, while from scattering considerations penetration would increase at longer wavelengths. So there is a competition between the two effects, and what will actually happen will depend on the particular material and its condition - and in general you have to find out - guess what: empirically. Howard \o/ /_\ | ||
| Erik Skibsted |
It is an interresting question and it surely highlights some of the extra benefits of NIR i.e. measuring at different dephts in a sample. I am myself looking on online granulation and drying in a fluidbed reactor and during the drying phase I would like to go into some investigations about where in the granules the water is located i.e. surface water vs. "inside granule bound water" - does anyone have some good idea how to do it? I was thinking about looking on the ratio between the absorbance of the water peaks in the first and second overtone region - is the water peak in the combinational regime usefull? I am working with FT-NIR and looking on the nir absorption table on my wall in front of me says water absorbs @ ~4570 cm-1, ~5200 cm-1, ~7143 cm-1, ~10300 cm-1 and ~13000 cm-1....does anyone have experience with using ratios between the different water peaks? I measure with a reflectance probe, online, InGaAs detector, with 3m fiber attached to a Bruker MPA NIR unit. And by looking on the spectra I think it is not a very good signal from 10000-12500 cm-1 (too noisy). Good ideas are welcomed ;o) Erik Skibsted University of Amsterdam | ||
| Marchel Snieder |
Information depth, also named penetration depth, has our attention for several years. Together with Akzo, DSM we had studied the information depth of PE films (Information depth of Vis-NIR light in Polyethylene films using transmission and reflectance measurements. W.G. Haanstra, W. Hansen, M. J. G. Huys, B.J. Kip, P. Palmen, J. Roumen, M. Snieder, T. van de Weerdhof, S. Wiedemann, V. A. L. Wortel. Applied Spectroscopy 52, 6, 1998 863-868). Recently we had studied the penetration depth of Vitro human skin ( Factors Influencin skin moisturization signal using Near-Infrared spectroscopy. J. W. Wiechers, M. Snieder, N. A. G. Dekker, W.G. Hansen. IFSCC Magazine 5 , 1, 1-8 2003). In both cases we had proven that the light penetrate more at shorter wavelength. This penetration depth depends on the matrix and the design of the equipment. Marchel Snieder Uniqema | ||
| Vitas Svedas |
To Eric, You can differentiate bound and "free" water by resolving band's "fine structure" components. For the 5200 cm-1 band, the 1904 nm peak attributes to free water, while the 1936 nm peak largely corresponds to bound water. Details are in the reference below Determination and Differentiation of Surface and Bound Water in Drug Substances by Near Infrared Spectroscopy GEORGE X. ZHOU, ZHIHONG GE, JASON DORWART, BILL IZZO, JOSEPH KUKURA, GARY BICKER, JEAN WYVRATT JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 5, 1058 (2003) Vitas Svedas PhD/Optics |