NIR and Crystallisation Log Out | Topics | Search
Moderators | Register | Edit Profile

NIR Discussion Forum » Bruce Campbell's List » I need help » NIR and Crystallisation « Previous Next »

Author Message
Top of pagePrevious messageNext messageBottom of page Link to this message

Solomon Abebe
Posted on Sunday, October 31, 2004 - 7:56 am:   

Hi every one, i have a question about particle size determination during cooling crystallisation process. Am using SIMCA software to determin the particle size and distibution. Using SIMCA the scatter plot after applying PCA gives a certain profile how the particle grows. I am asking, is there any plot in this software which gives a clear understanding about the particle size and distribution?
Regards
Solomon
Top of pagePrevious messageNext messageBottom of page Link to this message

spectrokid
Posted on Monday, November 01, 2004 - 12:54 am:   

Fot crystalline products (sugar), the NIR spectrum will "bend" upwards for bigger particle sizes. If you draw a second degree polynomial through all your spectra, the coefficients of those polynomials would already be a good indication for particle size. I am puzzled you use a classification algorithm (SIMCA) to measure a quantitative value (particle size). Could you not try to use PLS instead? Of course we are talking about a certain distribution of part.sizes, but you could at least try to predict the average particle size. Getting a histogram of particle sizes will be pretty difficult I believe, but maybe you could use PLS to predict a standard deviation? Never done this myself though...
Top of pagePrevious messageNext messageBottom of page Link to this message

Solomon
Posted on Monday, November 01, 2004 - 4:55 am:   

Hi spectrokid, Can you explain how the polynomial coefficients can be a good indication of the particle size. The mean particle size is it in what unit micro or nano? I used SIMCA score plot just to see the spectra similarity. there is quite a good distribution as the crystal grow. according to the particle size. I have not tried PLS yet becuase i need a reference size.
Top of pagePrevious messageNext messageBottom of page Link to this message

spectrokid
Posted on Monday, November 01, 2004 - 6:02 am:   

In the good old days, I studied pure sugars and benzoic acids and tried to compensate for particle size differences. A quadratic compensation removed > 99 % of differences between spectra, much better than MSC which is a linear method. Although I didn't have reference values either, it was clear that the coefficients of a second degree polynomial were the "only" difference between the spectra and were therefore related to particle seize. As usual, a reference method is your biggest problem. If you can accurately describe what you want to measure, you will be halfway. If you can't get a good reference method, you could think in the direction of a "golden batch", or mapping a value against past batches.
Top of pagePrevious messageNext messageBottom of page Link to this message

hlmark
Posted on Monday, November 08, 2004 - 2:01 pm:   

Have you tried simply plotting the absorbance at a couple of wavelengths as a function of particle size? Emil Ciurczak has shown good and pretty linear results by sieving ground powders to obtain the various mesh sizes, then plotting absorbance versus mean particle size (or maybe it was mesh size).

Howard

\o/
/_\
Top of pagePrevious messageNext messageBottom of page Link to this message

Erik Skibsted
Posted on Wednesday, November 10, 2004 - 3:35 pm:   

@ spectrokid

how do you aplly the quadratic funtion to the spectra ? As a kind of baseline, what criteria??
do you have a reference?
Erik
Top of pagePrevious messageNext messageBottom of page Link to this message

geraldine
Posted on Wednesday, November 17, 2004 - 9:18 am:   

Hello to everyone, I am very interested in NIR Spectroscopy so I would like to know how the IR radiation is generated. that would be dangerous for the health or environment? The high intensity source probe from Labspecpro (ASDI) generates the nir radiation? How it works?. Thank you
Top of pagePrevious messageNext messageBottom of page Link to this message

hlmark
Posted on Wednesday, November 17, 2004 - 7:26 pm:   

Geraldine - the vast majority of NIR instruments use a tungsten-halogen lamp as the source of the NIR radiation. These are similar in principle to the ones used in automobile headlights; basically they are incandescent lamps (the type commonly found in homes) but with a halogen (I believe bromine is the one most commonly used) in the lamp's housing that allows them to run at a somewhat higher temperature than ordinary incandescent lamps, which, in turn, allows them to put out more light, and also to change the color of the light to maximize the output in the NIR region.

The main hazard of the lamps is that they get very hot in operation, so you should not touch it or you'll get a burn. It would also not be a good idea to stare at it while it's on because it's very bright and might damage your eyes.

Howard

\o/
/_\

Add Your Message Here
Posting is currently disabled in this topic. Contact your discussion moderator for more information.