| Author | Message | ||
     Gabi Levin |
Dear Nuno, Sorry for taking the time - that's what happens when you have to be in so many places and travel. Solvents in powders - I will make some assumtpions based on you being from a pharma company and try to get the answer you can possibly use. 1. I assume these are organic solvents, such as MEK, ethanol, toluene, Dimethyl Chloride, etc. 2. I assume the powder is of an organic ingredient that you are trying to dry in some process of manufacturing. 3. I assume you are trying to do it in the lab. If these assumptions are correct then with different degrees of success you shall be able to measure the desired range (0.3 to 2.5 weight %) of the different solvents. The difference in degrees of success depends on the difference between the spectrum of the solvent and the organic compound. For organic compounds you will find most interesting the C-H stretches in the ~1640 to 1750 region. With solvents like ethanol, the difference will be large because they have different spectrum in that region. For toluene the difference will be smaller, so to determine the degree of success you will need to perform experiments. If you write to me I can help you set up a set of tests that will help you to determine the feasibility for relevant combinations of compounds and solvents. After the feasibility is demonstrated, you will have to create a calibration based on substantial number of samples, the number will depend on how much variability exists in the composition of the powder from batch to batch. I hope you find this useful. Gabi Levin Brimrose | ||
| hlmark |
Nuno - as MPDC started to indicate, one of the key parameters is what level of residual solvent you want to detect. In all other respects it would seem to be an ideal application for the "classic" type of NIR measurement: a powdered sample which could be put into a closed cup and pressed up against the cup window to minimize the "particle size effect" as it's called. All manufacturer's standard closed cups of which I'm aware include provision to press the sample against the window in a more-or-less reproducible manner (using a spring or other device). The only other consideration that might mitigate against success would be if the liquid part was unusually caustic, volatile, flammable or had some other special characteristic, although in most cases something could be done about them, too. \o/ /_\ | ||
| MPDC |
Depends very much on how many other contaminants there are (including water!!!). Can you remove trace solvents on lab-scale yourself? (Vacuum oven?) Take a spectra "before" and "after". Plot them over each other. If you can't see disappearing peaks, your spectrometer will probably not be able to see them either. If it does work, you will have to evaluate the difference between those 2 spectra against differences caused by humidity, particle size and other contaminants. If you are talking big-bags of powder or bigger, the differences will have to lie outside of the 1400 nm and 1900 nm water peaks, as these will be wildly varying compared to your trace solvent. | ||
| Nuno Matos (Nmatos) |
Can anyone give me hints on the best way of measuring residual solvents in a powder using NIR? I'm trying to see from 3000 to 25000 ppm, with a FT-NIR equipped with a InAs detector. Am I searching for an Utopia? Thanks in advance Nuno |