Analytics for Chemistry, Biology and Production:
Dear Sir, look, there are so many vineyards, and I have only two hands and one brain. P. Forster
Peter ForsterHome The author of this page is independent and has no commercial intention in mind ! Diese Seite in Deutsch.


System Suitability Test (SST) — IQ, OQ, PQ, and PV:
Are you spending a lot of time, solvent, glassware, washing powder, and manpower for the following tasks?:
  1. Doing the traditional tests of:
    •  Wavelength calibration and resolution
    •  Calibration of the Photometric Scale
    •  Test the Stability of your Spectrophotometer
    •  Determining the Stray Light of the Spectrophotometer
    •  or other test of such kind on your spectrophotometer ?
  2. Doing some work for Pharmacopoeias/Pharmacopeias (USP, EP, etc.) [P] ?
  3. Doing a lot of work for ISO 900x, GxP, etc. ?
  4. Doing IQ, OQ, PQ, and PV.                    Some weekly, - some monthly, - some every year !
  5. Doing daily a lot of manual Dilutions.    There is a special solution "Dynamic Dilutor" for you.
  6. Filling cuvettes manually like Mr. Beer and Lambert did.   There is a "Bar-Sipper" for you.
  7. Running a reference with your sample.  For every or most of your analysis ?

Here is a solution for you:
Do not "drop the hammer", when you are not absolutely familiar with:
'Photo Diode Array Spectrophotometers' !,   or with 'Inverted Optics' !
I'm sure, in time, not so far, you will have some difficulties to find a new spectrophotometer with scanning technology, at all !
Please let me give you for the first moment the cover-page' of our publication. When it will be published in the different spectroscopy journals, I will present you the whole contents and even a little more:

New System Suitability Test
for Diode Array Spectrophotometers


Authors:
Nicola J. Parvin, 135 Nanpantan Road, Loughborough, Leicestershire LE 11 3YB, UK.
Yvette Cornélis and Peter Forster*, Software Development & Consulting, CH-4002 Basle, Switzerland
Summary
Traditional System Suitability Tests (SST), as recommended by Pharmacopoeia, GLP, ISO, IUPAC and NIST, -- often related to the terms IQ, OQ, PQ, and PV --, are not satisfactory for photo diode array spectrophotometers (PDA's), especially when being used in conjunction with chemometric algorithms. They do not guarantee enough security and trustworthiness for daily and routine work. This report describes a new, more adequate and reliable system suitability test for an ultraviolet-visible photo diode array spectrophotometer, and compares it with the traditional tests. It is shown how easily the implementation for the routine and production control processes can be realised. Excellent experience of more than 500 spectrophotometer life years (over 35 instruments) with high reliability and reproducibility guarantees the success of the new SST. It's obvious, that this test with similar improvement can easily be applied to almost every other 'wavelengths range'.

Content:
Keywords
Introduction
Traditional SST
Standards for Wavelength Calibration in the UV/Vis
Standards for Wavelength Resolution in UV
Standards for Calibration of the Photometric Scale
Accuracy and Precision of the Wavelength and Absorbance
Stability
Instrument Stray Light (ISL)
New System Suitability Test
Instrumentation
Procedure
Interpretting the SST
Conclutions:
Appendix A: Multi Component Analysis (MCA)
Relative Fit Error (RFE)
Lagrage Interpolation
Appendix B:       Multi Component Analysis standards for the new SST
Appendix C:       Certificates and "marriage"                                    see: related Documents.
Appendix D:       "SOP" for the ARGUS system with a hp 8452A/8453 spectrometer
                            (As You can see, it's so easy!)

Keywords:
SST (SystemSuitabilityTest), DAD (Diode Array Detector), ARGUS, Wavelength accuracy, Wavelength resolution, Stray Light, IQ (Installation Qualification), OQ (Operation Qualification), PQ (Performance Qualification), PV (Performance Verification), FDA (Food and Drug Administration), Pharmacopoeia/Pharmacopeia (USP, EP, etc.) [P], GxP (GLP, GMP, etc.), ISO (International Organisation of Standardisation), IUPAC (International Union of Pure and Applied Chemistry), NIST (National Institute of Standards and Technology), ASTM (American Society for Testing and Materials), NNIR (Near Near Infra Red), MCA Multi Component Analysis), RFE (Relative Fit Error), OLS (Ordinary Least Square), ML/H (Maximum Likelihood), PLS (Partial Least Square analysis), PCA/PCR (Principal Component Analysis/Regression).


Introduction
Spectrophotometry is widely used in industrial quality control and is becoming a very important tool in production, including stoichiometric adjustment, kettle charging and end-point determination, for example. The reliability of the instrumentation and results must therefore be verified under the guidance of ISO, Pharmacopoeia, GLP, GMP and other organisations. The recommended procedure is called the System Suitability Test (SST).

Chemometric algorithms, such as MCA, PLS, PCA, PCR, combined with the latest diode array spectrophotometer generation, provide the required powerful tools for quality and production control tasks. At the same time, SST's are still performed in the traditional way. However, the requirements such a test has to fulfill today for a diode array spectrometer cannot be met, particularly where present day chemometrics is used.

For example, performing Multi Component Analysis (MCA), [1, 18], with reference and sample spectra measured at different times, while in the meantime the spectrophotometer wavelength has drifted by 3 nm (half the acceptable value in the visible range, [2]), will result in completely meaningless concentrations and quality parameters of the analysis. Also using the stricter limits of the US Pharmacopoeia, [3], does not even give more reliable results.
An important sentence from the paper:
By spending only ten seconds[Output] every 24 hours for carrying out the new SST, much more is achieved — apart from the paperwork — than what IQ, OQ, PQ, and PV can ever do together, for correct and reliable results in the daily routine work[AD].

Chemometrics[CHEM]:
We would have Publications of much higher quality and reliability, in science, if the authors would be able
to use the SST and the RFE at the correct moment, and in the correct manner! Peter Forster

With this we can summarise, that the only 'thing' the new SST is not able to garantee, is the expertise and reliability of the Method-Developer and his work, all other 'starts from:' / 'is in best care of:' — NIST! [7, 8, 9, AC]
This still required test must be left to other 'Test'/'Validation' - methodes!

So from now, at least for Spectrometers, with all this knowledge and the Conclusions(from below) in mind, You will need a special lot of 'considerateness' for argumentations like:
' Why System Suitability Tests Are Not a Substitute for Analytical Instrument Qualification'.

Conclutions:   (fom the paper!)
The new SST provides a daily means of verifying the accuracy and precision of a diode array spectrophotometer, and evaluating the ISL.  The use of MCA, which is usually included in the software delivered with the spectrophotometer, enables simultaneous verification accross the whole wavelength range, the traditional SSTinvolves only several seperate calibrations at specific wavelength. 
 It's important, to use only the basic OLS, or theML/H algorithm for carry out the MCA, as all sophisticated algorithms do not hold and are not in the right consistency with the required Beer-Lambert law. Algorithms like 'FSQ' can easily yield in 'dangerous' results special if they are used to 'eliminate' assumed noise 
 To carry out, the new SST is much less time consuming, and meets the much stricter limits, but as they are requried for modern, reliable chemometrics algorithms. 
 Verifying the same spectrophotometric parameters as the traditional SST, the new SST also evaluates the short and the long term reproducability, which is vital for reliable quality control or/and involving chemometrics. 
 The SST takes only a few seconds for measuring the Holmiumoxide filter and evaluating MCA. It can be placed in the start-up procedure by a macro, and automatically applaid every day [AD].
Correct implementation of the RFE[RFE] is, however, important, everything else is so easy to implement, that it can be done by anybody with absolutely no special knowledge, neither in computer science nor inmathematics. 
 The new SST and the instruments are of such robustness, that we had only three times to re- "marriage"[AC] the Holmiumoxide filter with the particular spectrometer(see appendix B):
         First,        the filter was broken("Nobody" knew why),
         Second,    the whole spectrograph(polychromator with diode array) had to be changed.
         Third,       the A/D-converter of the diode array was not working any more.
That means, that most of our customer are using the same Holmiumoxide glass filter and the installed reference spectrum for more than ten years, some even more than 15 years, as they are not able, to change (re-"mariage"[AC]) filter and instrument (reference spetrum) by themselves. 
 For all that, the ARGUS sytem has availability for routine analytics of higher then 99.98%. This is important for the production, where it is not possible/acceptable to wait for an analysis, whilst the reaction ius still going on. 
 By spending only ten seconds[SST] every 24 hours for carrying out the new SST much more is achieved — apart from the paperwork — than what IQ, OQ, PQ, and PV can ever do together, for correct and reliable results in the daily routine work. 
 In conclusion, the new SST is an easy, quick, thorough and robust means of assesing the performance of diode arry spectrophotometers. It meets the requirements of ISO 9001/2, GLP, GMP, and Pharmacopoeia in both production and quality control, with no loss of time. 
 Our "SOP" in the appendix D[AD] is documenting it. Note, that all spectrometer relevant tests are marked "equipment internal SST", meaning, that they are done by the system itself, and only the printout[SST] has to be signed and filed in the documentation folder for your system. 
 That this SST meets much more then only the real basic philosophy of the FDA, everybody would expect, is almost excrescent to be mentoined. 
 In other words, the new System Suitability Test guarantees, that the spectrometer is all time in the same good shape/conditions as it was, when the instrument was first certified and controlled, and in the identical good condition are met, as were presented, when the calibration of the corresponding method was worked out. 
 This means, that the measured method standard-spectra are valid and usable over several years, — not months —, without necessarily measuring all the time the reference against the sample over and over again. 
 That this procedure saves a lot of time, solvent, and other costs, and avoids a lot of mistakes, that normally occur during routine work, does not need , I think, any futher explanation.


Sincerely yours.



ReferencesReferences:
 

[P] "In correct Latin:  Pharmacopoeia"

[1] "Matrix Representations and Criteria for Selecting Analytical Wavelengths for Multicomponent Spectroscopic Analysis",
C. W. Brown, P. F. Lynch, R. J. Obremski & D. S. Lavery, Anal.Chem., 1982, 54, p 1472-1479

[2] "European Pharmacopoeia 2003", 2.2.25, Absorption spectrophotometry Ultraviolet and Visible. www.pheur.org

[3] US Pharmacopoeia XXVII, 2003, Physical Tests, 851 Spectrophotometry and Light Scattering,
United States Pharmacopoeial Convention, Rockville MD. www.usp.org

[4] "Hewlett-Packard HP 8452a UV-Visible Spectrophotometer (190-820 nm Version)”
Standard Operating Procedure, August 6, 1993. www.agilent.com

[5] "Hewlett-Packard HP 8453 UV-Visible Spectrophotometer Reference Manual”, May, 1995. www.agilent.com

[6] "Hewlett-Packard Operational Qualification / Performance Verification for HP 8453 UV-Visible Spectrophotometer”, September, 1996. www.agilent.com

[7] Holmiumoxide solution in perchloric acid with certificate, SRM 2034 from NIST, National Institut of Standards & Technology, www.nist.gov
(The certificate contains also some additional references)

[8] Holmiumoxide glass filter with certificate, SRM 930e from NIST, National Institute of Standards & Technology, www.nist.gov
(The certificate contains also some additional references)

[9] "Glass Filters as a Standard Reference Material for Spectrophotometry, Selection, Preparation, Certification, Use, SRM 930",

R. Mavrodineanu & J. R. Baldwin, NBS-Special Publication 260-51. www.nist.gov

[10] A.Savitzky, M.Golay, Analytical Chemistry Vol. 36, 8, 1964 1627ff, BUT YOU HAVE TO CORRECT THE FORMULA AND THE IMPLEMENTATION.

[11] "Liquid Absorbance Standards", R. W. Burke, E. R. Deardorff & O. Menis,
J.Res.Nat.Bur.Standards-A.Physics & Chemistry, 1972, 76A, No.5, p 469-482.

[12] "Standard Practice for Describing and Measuring the Performance of Ultraviolet, Visible, and Near Infrared Spectrophotometers", American Society for Testing and Materials, E 275-83, Annual Book of ASTM Standards, 1983. www.astm.org

[13] "A Wavelength Standard for Ultraviolet-Visible-Near Infrared Spectrophotometry", J. McNeirney & W. Slavin,
May Applied Optics, 1962, 1, No.3, p 365-367.

[14] "Standards for Checking the Calibration of Spectrophotometers (200 to 1000 nm)", NBS, Letter Circular LC-1017, Jan. 1955, re-issued 1967, in NBS Special Publication 260-51.

[15] "Standard Reference Materials: Metal-on-Quartz Filters as a Standard Reference Material for Spectrophotometry - SRM 2031", NBS Special Publication 260-68.

[16] "ASTM Standard Method of Estimating Stray Radiant Energy, E 387-72", Reprinted from the Annual Book of ASTM Standards, in NBS Special Publication 260-51. www.astm.org

[17] "Techniques in Visible and Ultraviolet Spectrometry Volume 1 : Standards in Absorption Spectrometry",
UV Spectrometry Group, ed. C. Burgess & A. Knowles, Chapman & Hall, London, 1981.

[18] "Near Near Infrared Spectroscopy for Quantitative and Qualitative Quality Control”, Tamzin A. Lafford,
Yvette Cornélis and Peter Forster*, ANALYST , Vol. 117, 1543ff 1992.

P. Forster Peter Forster: Software Development & Consulting
     Back Home
    Back to Index
Software Development & Consulting
Peter Forster
Neubadstrasse 88
CH-4054 Basle, Switzerland

Mail to: peter.forster@p-forster.com

Die selbe Seite in Deutsch