| Author | Message | ||
     Tony Davies (td) Moderator Username: td Post Number: 186 Registered: 1-2001 |
Hello David and Art et al., Yes Teflon is good. The risk with "permanent" standards are that they get contaminated. You can keep a spare but when do you know to use it? So you need alternatives. David, I have exhausted a fair amount of my limited brain-power in trying to find a solution to this problem. "Sphere" could have been a solution but the technology was not up to it in 1994. "Grey" and "Gold" standards where too limited. If sufficient people wanted it, I am sure that the "Sphere" idea could be implemented with mirco mirrors but unless the instrument people decide that they need something, it is not likely to happen. I am positive that: There is a solution to be discovered! The potential future of NIR spectroscopy requires it to be found! Best wishes, Tony | ||
| Art Springsteen (artspring) Member Username: artspring Post Number: 12 Registered: 2-2003 |
David says some nice things about us but the primary reason for using something like our sintered PTFE materials for a standard is consistency. It's always going to be the same. That said, I will let the cat out of the bag, so to speak, by saying WE used packed PTFE as a reference standard. But we can carefully control both the packing density and the environment in which we make the standards (we are a metrology lab, after all). I wouldn't want to have to pack standards in a grain elevator or sugar refinery. | ||
| David von Boisman (david_von_boisman) Member Username: david_von_boisman Post Number: 11 Registered: 2-2006 |
Art have a point here - I would much prefere a solid sample as a standard, as any difference in packing will create difference much bigger than the ones you're trying to capture. So you would have to apply a derivative or someting to the scan in order to filter out the packing effect if you use a powder. Actually, Art makes some great solid standards that can be used for the purpose - I once used a white teflon-style standard made by him for evaluating glass quality. | ||
| Art Springsteen (artspring) Member Username: artspring Post Number: 11 Registered: 2-2003 |
Hello all, A word of warning to all. While packed sulphur or PTFE powder are potentially good standards, the reflectance is highly dependent on packing density. You can really mess up your baseline if you don't pack these materials to the same density every time. And in a glass- or quartz-windowed cell, you do risk breaking the window if you press a little too hard- that's especially true of sulphur. | ||
| Nick Glover (milkyone) New member Username: milkyone Post Number: 4 Registered: 5-2007 |
Hi Everyone adding to the thread. Cells, cups, vials etc will get damaged, broken etc, whatever else happens. The impact on performance, like just about everything else will depend on the application. I have one going now where a completely split cell base has no apparent effect on performance. Regards use of reference materials eg sulphur or teflon, how can I readily interpret the spectra to look for day to day instrument checks as with a check cell. I have tried using product calibrations but because calibrations developed on different material results seem meaningless. Currently trying a PCA qualification method with deliberately spiked reference. Cheers Nick | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 222 Registered: 9-2001 |
Good point, Tony. I've never used sulphur as a standard, but finely powdered Teflon, also known as Halon, Algoflon and other trade names, is also highly reflective and Lambertian. I'd say that if used as a disposable, as you recommend for the sulphur, it should work as well as sulphur. \o/ /_\ | ||
| Tony Davies (td) Moderator Username: td Post Number: 185 Registered: 1-2001 |
Hi Russell, I would just like to tell you (and remind the others) that sulphur is a very useful high reflectance material. It can be obtained very pure and quite inexpensively. Use it once and discard. The idea is almost as old as NIR analysis; Kronstein (1963) and Tkachuk and Law (1978). Best wishes, Tony | ||
| David von Boisman (david_von_boisman) Junior Member Username: david_von_boisman Post Number: 10 Registered: 2-2006 |
Hi Russel, a gold reflector is not a bad choise, just make sure that its inserted in the same way from time to time (not rotaded between readings). The surface needs to be rather rough (sandblasted) to avoid direct reflectance (i.e. should not work as a mirror). Its unlikely that the glass itselv would change over time (I never seen that), more likelythat different batches of cups may differ. But takin a reference scan with a standard every now and then provides a way of troubleshooting if problems shoudl occouor - so it may be worth the effort even if you don't evaluate the data after each reading. | ||
| Russell Wilkie (rwilkie) New member Username: rwilkie Post Number: 3 Registered: 2-2009 |
Thanks Sheelagh, David and Howard for your comments on this issue. I agree that disposable vials are not the most consistent way of performing NIR testing, although we do currently use these for raw material identification as peak positions remain consistent regardless of vial thickness/surface variations. What does change however is the magnitude of the absorptions from vial to vial even using the same sample of milled powder; not helpful for minimising quantitative error. After having taken your comments on board, the fact that standards also change along with Howard's 'infinite regress' comment, QA agree that trying to develop a test to ensure the cup is suitable for use may be just too complicated, and perhaps we could alternatively address the issue by adopting a documented sample cup cleaning/handling methodology, adequate training and Risk Assessment. Because the likelihood of physical damage is low (but can be checked visually) and the likelihood of chemical damage even lower, we may just accept that one day damage may occur, but we took as many measures as possible to prevent it. Cheers, Russell | ||
| Sheelagh Halsey (shalsey) Member Username: shalsey Post Number: 15 Registered: 12-2006 |
Howard mentions the word change quite a lot in his last message and in my experience this can also relate to quartz cups. Unless you want to calibrate on one cup, very dangerous, there will also be variation from cup to cup. In my view, as long as you choose good quality vials with flat bases, the variation will be more than with quartz cups, but this can be comnpensated for in the calibration process. In addition, with active pharmaceutical materials, you don't want to have loose powder in the open. Disposable vials allow for more easy sample containment. There is a quick experiment you could try, analyse the calibration set both in cups and vials. In this way you can check whether the vials are reducing your accuracy or not. This will be application dependant, particularly with regard to the level of active you are trying to detect. Regards Sheelagh | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 220 Registered: 9-2001 |
Russell - the problem of things changing is not new, nor is it limited to spectroscopic analysis. I recall that when I was taking my first-year analytical chemistry courses, even when doing titrations, it was necessary to standardize the titration solutions to account for possible unknown changes between uses. That's the nature of the world: instruments change, chemicals change, everything changes. The solution to the problem back then, as it is still now, is to use something whose properties you can rely on to remain constant. Back then, it was pure chemical "standards" of some sort, depending on what the analysis was for. Nowadays, using NIR, we have to rely on the physical properties of what we can put into the instrument, that we have reason to believe is constant from one time to the next. If you keep arguing "I don't know that the xxx standard is constant and therefore I need something else to check it against", you'll wind up with an infinite regress, since there's no absolute guarantee about any of that. That's done even in the wider world of metrology. The standard for mass is still a platinum-iridium weight kept in Paris, that is defined as 1 gram, and since the whole world depends on the constancy of that, extraordinary means are used to make sure it doesn't change (although, to be sure, there are proposals under test for replacing that with natural constants, e.g., counting out a certain number of hydrogen atoms, for example). Similarly, even though the current definition for length (the meter) is based on a certain number of wavelengths of light of a specific atomic transition, in practice the two scratches on a Pt-Ir bar, also kept in Paris is used as the standard meter. So in reality, for NIR it's a matter of selecting something that is expected to not vary, and then taking care to not do anything to it that will cause changes, and to guard against natural changes. If you have, say, a quartz cell, then you have to keep it clean. If you're worried about scratching it during cleaning than you have to clean it in a way that will not cause scratches, then you have to just take it, almost on faith, that it's the same as it was before. This concept also hasn't changed over the years, when we had to take it on faith that the pure chemical "standard", kept in a tightly closed container, didn't change it's properties from one time to the next. Unfortunately, Sheelagh's proposal has an a priori serious problem: how do you know that the vials are the same to start with, especially when ordinary off-the-shelf vials are not specified for their optical properties? It's not easy to keep a traceable set of standards, but nobody ever said that doing good science, or even good NIR, was easy. You might read up on Metrology (the study of the science of measurement) for ideas on how to keep a set of standards that can be checked against each other for internal consistency of the set. That's one way to avoid having to rely on only a single standard, although it does add more effort to the solution of the problem. \0/ /_\ | ||
| Sheelagh Halsey (shalsey) Member Username: shalsey Post Number: 14 Registered: 12-2006 |
Hi Russell Have you coinsidered using good quality, disposable glass vials? This will eliminate the need for cleaning cups. Sheelagh | ||
| Russell Wilkie (rwilkie) New member Username: rwilkie Post Number: 2 Registered: 2-2009 |
Hi David, We were proceeding toward developing assay methods of pharmaceutical actives in milled powder, which should be relatively straight forward. Typical validation strategies require that all systems and methods be 'in control', and we think that periodic qualification of the empty cup may be important in validation of our NIR testing process. Due to the requirement to clean/dry the sample cup prior to every sample, how do we know that all of the cleaning/wiping has not made the vessel less and less suitable to gather data over time? If we cannot check that the vessel is suitable for use prior to gathering data through it then we have no control over this portion of the NIR testing process. We had considered acquiring a gold reference the exact size of the sample cup, and taking reference scans when the cups were new. But we are unsure as to what sort of test/model we could employ and what limits to adopt to check that the cup still remains in useable condition over time? Any further thoughts would be greatly appreciated! Cheers, Russell | ||
| David von Boisman (david_von_boisman) Junior Member Username: david_von_boisman Post Number: 9 Registered: 2-2006 |
What is the application? The sort of disturbances that make no difference for one application may have an inpact for another. Basically you need some sort of standard that fits well into the cup, that is very stable, don't absorb moisture and that don't attract dirt. Some sort of plastic cut to fit may work. The easiest test would be to use a PLS calibration, that's typical for the sort of work you do. If numbers remain stable over time then everything is fine. If they drift significantly, you need to figure out a metodology to check wheter your instrument, the cup, or the standard is changing. Thus you probably want to set up the same sort of standard in a sealed cup, using the same calibration, for instrument verification. | ||
| Russell Wilkie (rwilkie) New member Username: rwilkie Post Number: 1 Registered: 2-2009 |
Hi everyone, If the integrity of a reflectance measurement through a quartz sample cup is reliant on the cup window remaining unaltered physically and chemically during both calibration development and future testing, how does one ensure that a sample cup is qualified as being clean, undamaged and suitable for use and what types of tests or limits could flag a cup as needing replacement? Cheers, Russell |