| Author | Message | ||
     Annick De Coster (Adc) |
I'm trying to collect NIR-spectra by transmission mode om extracts from flax-fibres. These extracts are water-soluble. When take 10 spectra from the same extract, there are band-shifts, which make it very hard to interpret them. How do I solve this problem?? Suggestions are welcome !!! | ||
| hlmark |
Annick - you didn't say which peaks were changing, but my guess is that it is the water peaks. I'm also assuming that you are reading one extract 10 times (i.e., 10 aliquots). Water spectra are uniquely sensitive to all sorts of external effects. On the other hand, if it is a true solution, then we would expect the extract to be homogeneous, so that chemical differences should not be a factor, especially if the sample is well-mixed. The possiblity that comes to mind immediately is temperature changes. The spectrum of water is known to be very sensitive to temperature. If my assumption above about using aliquots of one well-mixed extract is not correct, than any other differences might also be affecting the water spectrum: pH, dissolved salts or even organic matter, etc. If any of the other assumptions are wrong, please let us know, so that we can try to think of other possibilities you can check. Howard | ||
| Annick De Coster (Adc) |
Thanks for the comments so far. Maybe some additional information can help to find other suggestions: I'm indead measuring one extract 10 times, 10 scans /measurement, so I can make an average on 100 scans. The peaks that are changes are most likely linked to the water peaks, however it is very difficult to say. I did the same measurements using a special cup in reflectance mode. There i had no problems with shifting peaks. Because of practical reasons i wanted to redo the tests using transmittance and now I have problems. Is there a difference in sensitivity between both methods?? | ||
| hlmark |
Hard to say. More likely the differences in the setup might leave the sample more exposed to external influences one way or the other. Measureable spectral changes can occur for very small temperature changes. Is your sample cell thermostatted? If so, how tightly can it control the temperature? You might need to get better control of that. Howard | ||
| W. Fred McClure (Mcclure) |
Annick, Three things to be concerned about: 1. Instrumentation - I assume you have check wavelength stability? 2. Sample cell - If the instrument is transmitting horizontally through a vertical cell, the sample may be "settling" (assuming the sample is cloudy). For cloudy sample it is always better to transmit vertically through the sample - hence, settling is less of a problem. 3. In drawing the subsamples, are you sure you are obtaining a representative sample? Fred | ||
| Tony Davies (Td) |
Annick, I think we need a "check-list" for people to fill in, to save having to ask so many questions! Are you able to dry your extracts and look at them in refelection? There is a technique known as DESIR in which you dry samples on a glass-fibre paper. This will remove the problems that you have with the varying water bands and make you much more sensitive to the compounds that you want to measure. Best wishes, | ||
| Fernando Morgado |
Annick : Do you intent change the pathlenght of the cuvette?. I don´t know your product but maybe you need increase or decrease the cuvette pathlenght, using diferent cuvettes. Fernando Morgado | ||
| David Hopkins (Hopkins) |
Annick, There is another possibility. You did not state whether the samples are clear, turbid, or possibly contain particulate matter. If you look at difference spectra from the first scan in a series, do you see smooth, evolving differences that look like first derivatives, suggesting temperature changes, or do you see distorted, angular bands that suggest that particles might be moving in/out of the measuring beam during scans? Do you remove the effect by filtering your samples? It is very difficult to advise you without seeing the spectra. I hope these questions help you. Best wishes, Dave Hopkins | ||
| Annick De Coster (Adc) |
Dear tony, any idea where I can get information on this DESIR technique?? | ||
| Ian Michael (Admin) |
Have a look at: http://www.impublications.com/discus/cgi-nirpubcom/nir_srch.pl?file=abs&Range=Any&Format=Standard&Terms=DESIR |