NIR Calibration USP and Ph.Eur. Log Out | Topics | Search
Moderators | Register | Edit Profile

NIR Discussion Forum » Bruce Campbell's List » I need help » NIR Calibration USP and Ph.Eur. « Previous Next »

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

Schaurt Toby
Posted on Wednesday, February 23, 2005 - 3:53 am:   

Hello all,
we have 2 Nir´s, and we started with the qualification.
Now we have a problem mit the calibaration against USP 1119 and Ph.Eur 2.2.40 ( Sup. 5.0)
Does anybody know, where i can get good information, if they are similar? Or datds, how too start the calibration?
Does anybody has experience with this calibration?
Top of pagePrevious messageNext messageBottom of page Link to this message

Nuno Matos (Nmatos)
Posted on Wednesday, February 23, 2005 - 4:11 am:   

Hi Toby

First of all, what kind of NIR equipment are those?

Nuno
Top of pagePrevious messageNext messageBottom of page Link to this message

schaurt toby
Posted on Wednesday, February 23, 2005 - 4:15 am:   

Hi Nuno,
two Bruker MPA´s,
one only with Sample wheel and a sphere,
the other with External Transmission, solid fibre optics and transmission optics

Toby
Top of pagePrevious messageNext messageBottom of page Link to this message

Nuno Matos (Nmatos)
Posted on Wednesday, February 23, 2005 - 4:18 am:   

In what sense didn't you meet the criteria from USP and EP?
Top of pagePrevious messageNext messageBottom of page Link to this message

schaurt toby
Posted on Wednesday, February 23, 2005 - 4:29 am:   

The problem is, that i don´t know,
how too start the Ph.Eur. Test.
I think they are the same like the USP test, but I´m quite not sure.
There is one more test in Ph. Eur. ,which is called Wavelength repeatability, do you know, how too manage that?
There´s a plug in from Bruker, which makes all test´s for the USP, i think that´s o.k.
But I find no comments in the Medicla books, about calibration in Transmission area.
Can I fell free for this part?
Top of pagePrevious messageNext messageBottom of page Link to this message

Nuno Matos (Nmatos)
Posted on Wednesday, February 23, 2005 - 4:39 am:   

Bruker should have a software (complied with 21CFR11) that would do all the calibration sets.

By memory:
- Frequency validation: Collect an air spectra and analyse a determine peak wavenumber;
- Spectral quality: Collect two air spectra, one next to the other. The first one will be the background while the second one will be obtained in absorbance units.
- Photometric noise: The same above but using a 0% transmitance standard (low flux noise) and a 99% transmittance standard (high flux noise).
- Photometric precision and linearity: This test is done with several standards with different transmittance grades.
- Wavelength accuracy: the same as frequency validation but with a reare earth oxide standard in severeal peaks.

If I'm wrong with something please warn.

Nuno
Top of pagePrevious messageNext messageBottom of page Link to this message

schaurt toby
Posted on Wednesday, February 23, 2005 - 4:48 am:   

Nuno,
that´s correct,
there´s a plug in from Bruker,
all test were done with this Software,
so it´s o.k. to do these test and than the MPA is calibrated for USP requirement?
But there is nothing for Ph.Eur. Test,
thát´s my problem,
how could i manage this

Thank´s a lot
Top of pagePrevious messageNext messageBottom of page Link to this message

Nuno Matos (Nmatos)
Posted on Wednesday, February 23, 2005 - 6:53 am:   

Toby,

If the software is well constructed then yes.

What's missing for EP?
Top of pagePrevious messageNext messageBottom of page Link to this message

schaurt
Posted on Wednesday, February 23, 2005 - 8:39 am:   

The Test Verification of the wavelength,
it´s only a part from EP, but not from USP.
So how could I manage this?
Top of pagePrevious messageNext messageBottom of page Link to this message

Nuno Matos (Nmatos)
Posted on Wednesday, February 23, 2005 - 9:27 am:   

Both, USP and EP, do mention the wavelength verification/accuracy/uncertainty. USP refers in the wavelenght uncertainty the following:
"In transmittance measurements, NIST SRM 2035 rare earth oxide in glass standard, or Holmium oxide solution NIST SRM 2034 3 are available. Alternative standards may be used with appropriate justification."

While EP refers in wavelength verification:
"Measurement in transmission mode: Methylene chloride R may be used at an optical pathlength of 1.0 mm. Methylene chloride has characteristic sharp bands at 1155 nm, 1366 nm,1417 nm, 1690 nm, 1838 nm, 1894 nm, 2068 nm and 2245 nm. The bands at 1155 nm, 1417 nm, 1690 nm and 2245 nm are used for calibration. Other suitable standards may also be used."

Nuno
Top of pagePrevious messageNext messageBottom of page Link to this message

Steve Choquette (Choquette)
Posted on Wednesday, February 23, 2005 - 12:33 pm:   

SRMs 2035, 2065 and 2036 are used for validation/calibration of wavenumber/wavelength scales for FT and dispersive instruments in the NIR (975 nm to 1945 nm). In addition, SRM 2065 is certified for 13 additional bands spanning 334 nm to 805 nm. SRM 2034 is a holmium oxide perchloric acid solution in a cuvette that is certified for calibration of UV-Visible spectrometers only. It is not intended, nor useful for calibration of NIR spectrometers. Not sure how that got into the USP chapter, but we're working with them to correct this.

steve choquette, NIST
Top of pagePrevious messageNext messageBottom of page Link to this message

venkynir
Posted on Wednesday, February 23, 2005 - 11:13 pm:   

Dear NIR-friends,
Any body have worked on Self Organization Mapping ( a good clustering technique based on ANN) of NIR spectral of consumer plastics?.

How to select output neurons and no.of epochs ?

If any publicatiion reg. is highly helpful
Top of pagePrevious messageNext messageBottom of page Link to this message

VENKYNIR
Posted on Wednesday, February 23, 2005 - 11:25 pm:   

Dear NIR-friends
we have collected more than 200 sample(Indian) of consumer plastics.

our success rate in PVC with PP is not appreciable.

We follow standard pre-processing techniques.

Can any body suggest a suitable compact algorithms .
Top of pagePrevious messageNext messageBottom of page Link to this message

Nuno Matos (Nmatos)
Posted on Thursday, February 24, 2005 - 2:20 am:   

Steve,

Another question. Wouldn't be suffice the water vapor bands to evaluate the wavelenght accuracy in FT-NIR? The EP refers it but the USP doesn't.

Other question. By evaluating the linearity, precision and photometric noise with the reflectance mode (solid apparatus), can't I expand those results for the transmittance mode (liquid apparatus)?

Nuno
Top of pagePrevious messageNext messageBottom of page Link to this message

schaurt toby
Posted on Thursday, February 24, 2005 - 2:34 am:   

Hello,
I think, that there are no specification in the USP/EP for the transmission part.
Only the Wavelength, with a BRM 2065 Std.
But for the rest ther is no describtion with other details.

We check the transmission with internal Stds.

The only Problem is the Wavelength repeatability,
is it enough, that I check my system daily with a System check?
Top of pagePrevious messageNext messageBottom of page Link to this message

hlmark
Posted on Thursday, February 24, 2005 - 9:47 am:   

VENKYNIR - The spectra of those two plastics are considerably different (see, for example, Spectroscopy, 9(1), p.27-32 (1994) for partial spectra of those, along with others) and therefore should be easily distinguishable. But there is much that your message didn't say about the problem, for example, how many different plastics are included among the 200 samples? Something about the sample set must be masking the differences. Or maybe the data pretreatment you are using is making them look too much alike.

From the spectral differences, I think that Mahalanobis Distances, even without any data pretreatments, should work for this problem, although you'd have to be careful to choose the correct wavelengths. You will also need a good representation of both the PVC and the PP, I'd say at least 20-30 samples of each.

Howard

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

Steve Choquette (Choquette)
Posted on Thursday, February 24, 2005 - 12:22 pm:   

Hi Nuno,
Regarding h20 vapor...only if you have sufficient resolution to measure, i.e. 2 cm-1 or better. It is my understanding that many process FT-NIR's do not have sufficient resolution to measure h20 vapor. Also many are hermitically sealed, making it difficult to get a useable spectra. H20 vapor should work fine, given its limitations (i.e. only 2 bands are really useable at 2 cm-1 resolution) , however we do the primary calibration of our spectrometers with SRM 2517, an acetylene vapor cell.
Not sure I understand your 2nd question, but if you're using a FT and the transmission accessory and the reflectance accessory are two unique channel's, then no, one can't assume that calibration of one channel is the same as the second. Maybe I'm misunderstanding your question?

steve
Top of pagePrevious messageNext messageBottom of page Link to this message

Nuno Matos (Nmatos)
Posted on Friday, February 25, 2005 - 2:17 am:   

Steve,

In what concerns to h2o vapour, the equipmetn I use goes to a resolution of 1 cm-1. Nevertheless, and hit if I'm wrong, for FT-NIR spectrometers, is suffice to examine one peak. If this peak is correct than all the others are correct due to the modus operandis of the interferometer.

For the 2nd question, it's answered. I totally agree with that. One channel will be enough to garantee the correct functioning of the equipment.

nuno

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