| Author | Message | ||
     Xuxin Lai (Laixuxin) New member Username: Laixuxin Post Number: 14 Registered: 6-2004 |
Dear all, Thanks a lot for all of your suggestions. I agree that laser diffraction is a good method. The instrument I used is from Malvern. However NIR is a versatile method, to measure particle size is only one of the purposes in my project. I think I have read the theory of time of flight. But it seems that it is mainly used for NIR imaging for clinical diagnostic by using a probe. I have also read the representative layer theory. But it is suitable for the particle size which is much bigger than the wavelength. Thanks for your discussion. I think I know how to write in my thesis now. Best regards, Xuxin | ||
| (Unregistered Guest) Unregistered guest |
Hi Xuxin, I am working with particle size (suspension and dry powder methods). I would say that the better method is Laser diffraction (laser scattering). Some companies in this branch are Sympatec, Horiba, Malvern. About theory, try to see the Fraunhofer diffraction. Best regards, Renato Guchardi | ||
| Gabi Levin (Unregistered Guest) Unregistered guest |
Hi, You can also try to talk to people who do particle size by a somewhat different method, called Time Of Flight, I am not very familiar with the technology, it uses a different principle from scattering, and I think that if you do a search for GALAY which is somewhere in the name of the company you will find them. I am not sure NIR is your best way to do aprticle size. Gabi Levin Brimrose Corp. | ||
| (Unregistered Guest) Unregistered guest |
Xuxin - I recommend you follow Lois' advice and read the manufacturer's description of how they implement the Mie theory. If they use a goniometric method, even if at only two angles, then you should be in pretty good shape as far as the practical application of the measurement is concerned, i.e., your calibration should give a good approximation of the "true" particle size. As for the theoretical description, you have two choices: do a full-fledged literature search (i.e., repeat all Don's work) or wait until his book comes out. Alternatively, you can do a simplified search (read relevant sections in Kortum and in Peter Griffith's chapter) and accept that fact that theory is not a good description of actuality - something that you'll have to accept even if you do a comprehensive literature search. Howard \o/ /_\ | ||
| Xuxin Lai (Laixuxin) New member Username: Laixuxin Post Number: 13 Registered: 6-2004 |
Howard, The samples I used is polystyrene latext standard. The two methods (NIR and LLS)agree quite well. I used PLS regression to construct the model. It seems that it can solve the nonlinearity quiet well. I am writing a paper about this method now, this is why I am looking for the theorectical support. Best regards, Xuxin | ||
| Xuxin Lai (Laixuxin) New member Username: Laixuxin Post Number: 12 Registered: 6-2004 |
Hi Both, Thanks for your reply. Howard: I am still interested in the theory of light scattering. Lois weyer: Actually I used the laser diffraction method as the reference method. But the problem is the sample used in laser diffraction need to be diluted many times to generate single particle light scattering, which is different from NIR method. The concentration I used is much higher than laser diffraction method. Thanks for your information. Best regards, Xuxin | ||
| hlmark (Unregistered Guest) Unregistered guest |
Lois - I considered recommending a light-scattering instrument to Xuxin, but decided not to because the problem of deciding whether it conforms to the theory becomes a circular argument: using one measure of light scatter to calibrate another. For a practical measure that's OK, and Xuxin seems to have solved it practically, but it leaves open the possibility of the two instruments agreeing with each other, but both giving the same wrong answers! By the way, Xuxin, can you tell us how you developed your calibration, and what you used to measure the particle sizes? It does occur to me that by diluting the sample sufficiently, so that the particles are well-separated, you could in fact use Mie theory to determine the particle sizes indpendently of the overall scatter, but you would have to take goniophotometric measurements. It's possible that the LLS instruments Lois mentioned work that way, in which case they would provide values not based on the sample scatter, and your calibration would then be against much better estimates of the particle size. Howard \o/ /_\ | ||
| hlmark (Unregistered Guest) Unregistered guest |
Xuxin - Yes, "representative layer theory" is what Don calls his approach. If you've read up on scattering theory, then you probably know by now that our problem is that there simply is no rigorous ab inito theory of diffuse reflectance. Those theories that attempt rigor, such as Mie theory, apply only to individual particles. As soon as you attempt to take into account the effect of neighboring particles and interactions between particles, especially particles of varying sizes and shapes, the equations get so complicated that they become almost impossible to write down, much less solve. Simplified theoretical approaches (such as Kubleka-Munk theory, or Representative Layer theory) abound, but because they are simplified, at best they "work" in only over a relatively limited regime of conditions. But it seems that you've at least solved your immediate practical problem of measuring particle sizes. If the general problem of describing diffuse reflectance theoretically is still of interest to you, I'm sure that Don would welcome assistance. He has made some major contributions to the field, but we all can help by cross-fertilizing each other thoughts. I know I've thrown plenty of fertilizer at him from time to time; I don't know if he's considered that as helping or not. A final note: "Don" is his name, he's not "the Don" of Near-infrared!! (Or maybe he is!!) Howard \o/ /_\ | ||
| Lois Weyer (Lois_weyer) New member Username: Lois_weyer Post Number: 15 Registered: 7-2002 |
Xuxin, You might want to talk with people who sell laser light scattering particle size analyzers such as Micromeritics, Malvern, Microtrac,etc. They use Mie theory for their instruments and they have some nice explanations of how the theory works. | ||
| Xuxin Lai (Laixuxin) New member Username: Laixuxin Post Number: 11 Registered: 6-2004 |
Howard, Thanks for your reply. Is the Don's approximation 'the representative layer theory for diffuse reflectance'? Actually the calibration model has finished. It seesm working. I supposed that the theory of describing the multiplicative light scattering in NIR transmission is available. Then I can use it to support my method. Now it looks like it is a difficult issue. Best regards, Xuxin | ||
| hlmark (Unregistered Guest) Unregistered guest |
Xuxin - First of all, I think that you are going to have to be able to carefully control the concentration of the latex, otherwise you'll have some difficulty distinguishing between the latex concentration differences and the particle size changes. If you can control the latex concentration, then you might be able to use an empirical approach to relating the particle size to the measured transmittance. This is similar to what is normally done in NIR calibration work. It does mean, however, that you'll have to have some alternate means of measuring the particle sizes. Using this approach, you will likely not need to use more than one wavelength, although you will probably have to do curve fitting instead of fitting a straight line. As Don indicated, in the size range you are dealing with, where the particles range in size from somewhat below to somewhat above the wavelengths you use, theory breaks down, and even the best ones (even his) are only approximations. You might also consider, however, working in a different wavelength region, so that the experimental conditions would conform more closely to the theoretical requirements. Going to longer wavelengths (i.e., mid-IR) would not work since the water would absorb all the radiation. You might consider trying to do the same thing in the UV, however, where a wavelength of, say 200-250 nm might be enough less than the particle size range of the latex for scattering theory to apply. Howard \o/ /_\ | ||
| Xuxin Lai (Laixuxin) New member Username: Laixuxin Post Number: 10 Registered: 6-2004 |
Dear all, Thanks a lot for your reply. I think I didnot describe my project properly. Actually I am using light scattering to determine the particle size in suspension. The size range is 0.5-1.5 micron. The material is latex, it has low adsorptivity in NIR range. The meadia is water. From the NIR spectra, I can only see water band with baseline shift due to light scattering. I can also see the light scattering is wavelength dependent, and it also relates to the latex concentration, particle size. My question is where can I found a equation states the relationship between Log(1/T) and the latex concentration, particle size, wavelength and the actual pathlength? Could you tell me about that? The particle size range I used is very close to the wavelength. I don't know which theory I should use. Reyleigh theory is only for the particle size which is much smaller than wavelength. Kubelka-Munk theory is used for the particle size much bigger than the wavelength. For the Mie theory is use for single particle scattering. I think for the light scattering I measured in NIR range should be multiple light scattering. I am not sure if the Mie theory can be still used. Dave, I do use EMSC for the moisture determination in freeze-dried products, but not for this project, since the light scattering is the one I want to keep this time. Looking forward to your reply. Best regards, Xuxin | ||
| David W. Hopkins (Dhopkins) New member Username: Dhopkins Post Number: 83 Registered: 10-2002 |
Hi Xuxin and Don, First, let me say I am glad to hear about your book publication date, Don. Second, I am interested in what you say does not work? I have used MSC or SNV with derivatives for analysis of transmission data with success. I know that you have also examined EMSC for your applications. Are these approaches not working? Can you tell us more? Good to hear from you. Best regards, Dave | ||
| djdahm (Unregistered Guest) Unregistered guest |
The reason it seems as if the theories are for "reflectance only" is that they results that are emphasized are for "infintely thick samples". As Howard has suggested, the theories are really for "scattering" rather than "difuse reflectance" or "diffuse transmission". There is a requirement that you have two measurements on a sample of fixed thickness to determine the absorption and scattering parameters. Furhtermore, all the theories are approximations. Obviously, I believe that my theory is a better approximation that most, or I wouldn't have published it. The problem is that the absorption becomes entangled with the scatter. In the absence of two pieces of data the best thing to do is to use a thin sample and a small area detector with the detector as far away from the sample as possible. All of these things will help make the log(1/T) more linear with sample absoprtion. Incidently, Peter Griffiths and I have just finished an article for the next edition of the "Handbook of NIR Spectroscopy". In my opinion it is a very good summary of the various theories. Peter's article in the second edition wasn't bad. My book will be published by NIR Publications this year. | ||
| hlmark (Unregistered Guest) Unregistered guest |
Xuxin - the guy who knows most about this topic these days is Don Dahm, but I don't know if hes following this thread. If he is he'll surely respond, if not I'll forward your question to him. My own reaction to your question is that most, if not all, theories of diffuse reflectance, of necessity take into account both the forward as well as the backward scattered light. Therefore while they may concentrate their attention on the backward-scattered part, the effect on the forward-scattered (transmitted) light is inherently present. Don is in the process of writing a book on his findings, but currently the best exposition is probably still Kortum's "Reflectance Spectroscopy" although that may be difficult to find. Don has published his work largely in the Journal of Near-Infrared Spectroscopy, and as a column in NIR News; there's a lot of information there but you'd have to dig it all out, until Don finishes getting it all together. Peter Griffith's chapter in "The Handbook of Near-Infrared Analysis" is probably the most accesible current discussion of the topic. Howard \o/ /_\ | ||
| Xuxin Lai (Laixuxin) New member Username: Laixuxin Post Number: 9 Registered: 6-2004 |
Dear all, Currently I used NIR transmission spectroscopy for the analysis of a suspension sample. I use 1cm cuvette as sample container, and use 700-1300nm as the spectra range. I have tried to search the light scattering theory, but most of them is for reflectance. It seems doesn't fit my result. Can anyone tell me about this light scattering theory in NIR transmission with a fix optical pathlength? Thanks in advance! Best regards, Xuxin |