| Author | Message | ||
     Kathryn Lee |
I am interested in inexpensive moisture sensors for QC of thin moving films on backing plus all the tricks for making them work well. If anyone has information about these, could you please contact me? The rules of the forum prevent vendors from "selling" on the site, so I would appreciate any tips on what qualities to look for and how to make them work successfully and for pitfalls. Thanks! | ||
| MPDC |
There are a couple of companies out there who specialise in thin films. If you think about measuring on a moving web, these guys can help you with crossweb-frames etc. Look at http://www.ndcinfrared.com/ and try to find others. "Cheap" is however how you look at it. Some of the best and cheapest analysers work in mid-infrared, but in transmission. Problem is there that you need to keep source and detector perfectly aligned, while there is maybe a 6 meter wide web of plastic going in between. This requires an extremely strong and sturdy O frame which can cost a ton. www.applitek.com will sell you a reflection solution where the spectrometer costs both your arms AND your legs (www.brimrose.com) but any "el cheapo" traveller mechanism will do nicely. An other thing to take in account is batch to batch variation in film thickness. Specially if you measure transparent film, you will see interference fringes (if you use a scanning spectrometer of course) which vary with film thickness. The cheaper filter instruments solve this by letting you make a slope/bias adjustment at the beginning of every roll, but that requires extra operator work. The full scanning machines can use PLS; maybe somebody of the honourable chemometricians here could even come up with a double FT algorithm. Some pieces of advice: + What do you want to measure? Sure only water is interesting? How about CWT, color, film thickness? More parameters means more money! + Test your application in the lab. Find a place with a full scanning spectrometer. I know a guy who bought a 100000$ spectrometer and came up with a perfect calibration on TWO wavelengths. + You can try to measure on a shiny roll, reflection only off course. This can have a stabilising effect on the spectra. + Think about your reference measurement for moisture. Thin films take/lose moisture at lightning speed. You will need a setup you can put right beside the spectrometer, also when it goes "live". Good luck! | ||
| hlmark |
Kathryn - "MPDC"'s advice is good, if somewhat general. On the issue of interference fringes there are two other considerations: 1) Correcting for interference fringes by slope and bias by measuring at the beginning of a roll is only good if you're sure the thickness is constant throughout the roll. 2) Interference fringes in an FT instrument give rise to an interesting effect: since they are sinusoidal in the spectrum, they give rise to a distinct, distinctive, and usually fairly sharp "spike" in the interferogram. In fact, in the early days Myron Block sold interferometers as film thickness gauges based on this effect; the size of the spike is, in most cases, large enough to be detected automatically and is related to the film thickness through its location in the interferogram, and the magnitude of the spike is otherwise immaterial to the film thickness measurement. By virtue of this, the interference fringes can often be removed from the spectrum by simply excising the reading in the interferogram corresponding to the spike, with minimal effect on the rest of the spectrum. It's one of those things you'd have to try once you've got some data, but could be done automatically with some relatively simple programming if the data is suitable and it works out. Howard \o/ /_\ | ||
| Gabi Levin |
I thoutght this is not the room for commercial information, but when MPDC says that a given solution from a given source will cost both Arms AND Legs without knowing anything about our recent miniature spectrometers with extremely attractive pricing, I feel free to respond in a some what commercial way. The Brimrose solution will not cost arms nor legs, and I invite MPDC to contact me or Igor at any time to find out about our pricing so next time he can either avoid commercial information altogether and leave it to the interetsed people, or if he chooses to include such information he can use correct one. Many thanks, Gabi Levin | ||
| MPDC |
Sorry, sorry! Before I get the US navy down my back... I guess I should do a better job staying on top of current technology. Even if I have some reservations (is Snap! still Windows 3.1, Gaby?), it is difficult to deny that Brimrose seems to be one of the very few players who are still making BIG steps in R&D. It amazes me a little how slow evolution is going in the sector in general. Is NIR such a bad market to be in there is no money left for R&D? About the miniature spectrometer: Brimrose succeeded in folding all the electronics and optics of their "classic" FreeSpace machine into a very small box, reducing weight, losing the heavy anaconda and in the process (apparently!) also reducing the price. Remark however that the only way they could do this was by reducing the strength of the lamp. So if you are probably fine for clear films and light colours, you should test your darkest products on such a mini-model to see what S/N you still can get out of it. Before Gaby gets on his high horse again, this is actually something you should do whatever you consider buying, specially if you have thick black products. NIR is still sufficiently similar to visible light for it to dissapear in black ink, and I would be suspicious of any vendor who claims not to have any problems with that. | ||
| Gabi Levin |
Many thanks to MPDC for the candid response. About the black products - a word of agreement that any black looking product should be treated with suspicion when coming to work in NIR. My experience taught me one thing - don't let black product scare you away until you understand the reason why it is visibly black. Some organic compounds appear very black in the visible but are quite clear in the NIR, and you need to verify what you have by testing a simple test. Also, some molases in the processing of beet sugar appear very dark, almost pitch black, yet, they transfer more NIR light than the clear products because they are heavier in sugars and lower in water. If carbon black is there, or similar particulate material (and any ink contains particulate material) NIR light will disappear very quickly. I don't want to ride horses, but our software is long beyond the Windows 3.1, and if you are 21CFR11 sensitive, our 32 bit software will run on practicaly anything, and is fully compliant with 21CFR11. The reduction of lamp power has not reduced the signal, optics is used to compensate for this. Distance to the sample was sacrificed, evidently, but in many instances it is an acceptable sacrifice. Can we have some coffee somehwere, MPDC? once again, my thanks. | ||
| Gabi Levin |
Many thanks to MPDC for the candid response. About the black products - a word of agreement that any black looking product should be treated with suspicion when coming to work in NIR. My experience taught me one thing - don't let black product scare you away until you understand the reason why it is visibly black. Some organic compounds appear very black in the visible but are quite clear in the NIR, and you need to verify what you have by testing a simple test. Also, some molases in the processing of beet sugar appear very dark, almost pitch black, yet, they transfer more NIR light than the clear products because they are heavier in sugars and lower in water. If carbon black is there, or similar particulate material (and any ink contains particulate material) NIR light will disappear very quickly. I don't want to ride horses, but our software is long beyond the Windows 3.1, and if you are 21CFR11 sensitive, our 32 bit software will run on all new platforms, and is fully compliant with 21CFR11. The reduction of lamp power has not reduced the signal, optics is used to compensate for this. Distance to the sample was sacrificed, evidently, but in many instances it is an acceptable sacrifice. Can we have some coffee somehwere, MPDC? once again, my thanks. | ||
| Karl Norris |
MPDC is wrong about " NIR is still sufficiently similar to visible light for it to dissapear in black ink, and I would be suspicious of any vendor who claims not to have any problems with that." The black in the visible results from electronic transitions, and such absorptions in the NIR are quite rare. Black anodized aluminum reflects more than 80 % of the radiation for wavelengths from 900 t0 2500 nm. As a matter of interest NASA has spent a lot of research money searching for low reflectance coatings for the NIR. This does not mean that the user should assume that their "black" sample is not black in the NIR, but check to be sure. | ||
| W. Fred McClure (Mcclure) |
MPDC (I would rather know the man's name) probably was aware that the electronic transitions related to color are noticable all the way to 1400 nm. To see this all you have to do is run the reflectance spectrum of white and brown paper. Yet, one can never assume that darker materials do not reflect the NIR - nor can we assume that black materials have no unique absorption bands. So, one is always faced with recording preliminary spectra to make such determinations. Fred | ||
| Gabi Levin |
Quick remark - Let's not forget that the absorption coefficients of visible light stemming from color compounds are huge comapred to the NIR absorption coefficients, therefore, a product can look black with an extremely small concentration of the absorbing compound, and this will have zero influence on the NIR spectrum. Example is some colorants they use in pharma tablets. Less than 0.05% of the total weight is the colorant in many cases, so the effect on NIR is non existent. I worked with black looking hard gelatine capsules, they transmitted the NIR light just as well as the transparent ones. Again, the only valueable advice is test the stuff you are dealing with, unless you know it is black carbon, and then you can still find ways to perform some meaningful measurements, depending on what and the circumstances. Thanks, Gabi | ||
| Gabi Levin |
Please note - Let's not forget that the absorption coefficients of visible light stemming from color compounds are huge comapred to the NIR absorption coefficients, therefore, a product can look black with an extremely small concentration of the absorbing compound, and this will have zero influence on the NIR spectrum. Example is some colorants they use in pharma tablets. Less than 0.05% of the total weight is the colorant in many cases, so the effect on NIR is non existent. I worked with black looking hard gelatine capsules, they transmitted the NIR light just as well as the transparent ones. In fact, some white capsules that were made white because of added dispersed Titanium dioxide transmitted the light much worse than the black colored ones, and the reason is obvious, scattering by the TiO2. Again, the only valuable advice is test the stuff you are dealing with, unless you know it is black carbon, and then you can still find ways to perform some meaningful measurements, depending on what and the circumstances. Thanks, Gabi |