Docstoc

VisioNIR_Lyophilization

Document Sample
VisioNIR_Lyophilization Powered By Docstoc
					Welcome to
PAT in pharmaceutical Production




100% Inspection of residual
Moisture on a
Lyophilization Line
with NIR




                                   Your final touch to pharma quality.
Introduction
Abstract

 Integration of a high speed quantitative analytical NIR-inspection system in a
  lyophilization production process in order to identify and to reject out of spec products
  with respect to the residual humidity in freeze dried cakes.


 The VisioNIR system is capable to handle up to 3000 vials per minute.


 The process uses Near Infrared Spectroscopy in order to analyze the water content
  in freeze dried product cakes in glass syringes, cartridges or vials.


 The NIR model contains specific algorithms and mathematical pretreatments to
  realize the high speed analysis.


 The spectroscopic model is based on a calibration of artificially humidified products in
  the concentration range between 0% and 25% residual water content.


2
VisioNIR Lyophilization
vs. 5.1
The goal
Process Integration


                          The VisioNIR inspection
                          system has been
                          integrated into a
                          WILCOMAT R 36
                          MC/LFC/V/HSA/NIR Leak
                          Tester for liquid and
                          lyophilisate glass vials with
                          integrated laser based
                          headspace analysis




3
VisioNIR Lyophilization
vs. 5.1
The goal
Measure the NIR spectrum of each lyo-product



                                          • The principal set up which to
                                            achieve the NIR spectra in
                                            reflection mode
                                          • The illumination is realized
                                            through a 40 Watt tungsten
                                            halogen lamp which
                                            illuminates the vial
                                          • The NIR reflection is collected
                                            in a fibre optic cable which is
                                            connected to the NIR PDA
                                            spectrometer with 256 diodes
                                            and a wavelength range of
                                            1000 nm to 2100 nm




4
VisioNIR Lyophilization
vs. 5.1
The realization
Probe integration into the machine




5
VisioNIR Lyophilization
vs. 5.1
Before production
Method Development



                                        The calibration set with different moisture levels
                                        from 1% to 23% rel. humidity




             The Off line calibration probe
          combined with the spectrometer
     system in the back . The same probe
                        can be used inline
6
VisioNIR Lyophilization
vs. 5.1
Before production
Method Development




                          Click here to play the video

7
VisioNIR Lyophilization
vs. 5.1
                               Before production
                                    Method Development – The moisture signal and the result

                        0.174
                                                                         Normalized Spectra
                                                                                                                                 At the end the model spectrum
                                    0.15
                                                                                                                                 is calculated by taking the
el 0  wav len_ norm_i ndex

el 1  wav len_ norm_i ndex
                                     0.1
                                                                                                                                 mean of all the spectral data of
el 2  wav len_ norm_i ndex

el 3  wav len_ norm_i ndex                                                                                                     all samples for each diode for

l 3  wav len_ norm_i ndex          0.05
l 2  wav len_ norm_i ndex
                                                                                                                                the spectral regions of interest

l 1  wav len_ norm_i ndex
                                      0

l 0  wav len_ norm_i ndex

       
pari son 0  wav len_ norm_i ndex

pari son 1  wav len_ norm_i ndex 0.05
       

       
pari son 2  wav len_ norm_i ndex

       
pari son 3  wav len_ norm_i ndex
                                     0.1
      
ari son 3  wav len_ norm_i ndex

      
ari son 2  wav len_ norm_i ndex
                                    0.15
      
ari son 1  wav len_ norm_i ndex

t ion wav len_ norm_i ndex

                                     0.2



                                    0.25


                       0.292
                                     0.3
                                        25      30   35   40   45   50            55          60   65   70   75   80        85
                                         28                               wavlen_norm_index                            82




                                              Building the weighted integral
                                                    value leads to the result
                                    8
                                    VisioNIR Lyophilization
                                    vs. 5.1
Before production
           Method Development – The calibration


                                                                 NIR vs KF

                     7,00
                                                                     y = 1,0112x - 0,0926
                     6,00                                                 2
                                                                         R = 0,9951
Inline Average [%]




                     5,00
                     4,00
                                                                                                Inline Average [%]
                     3,00                                                                       Linear (Inline Average [%])
                     2,00
                     1,00
                     0,00
                            0      1        2       3            4            5        6    7
                                                        KF [%]

                                Calibration: NIR predicted vs. Karl Fischer moisture content plot

         9
         VisioNIR Lyophilization
         vs. 5.1
Production examples
The effect of broken cakes




                                                              Separation of different moisture contents

                                         10


                                         9


                                         8


                                         7                                                                            15% H2O
                                                                                                                      3% H2O
                    moisture content %




                                         6                                                                            6.5% H2O
                                                                                                                      broken cake 1
                                         5                                                                            broken cake 2
                                                                                                                      broken cake 3
                                         4                                                                            good cake 1
                                                                                                                      good cake 2
                                         3                                                                            good cake 3


                                         2


                                         1


                                         0
                                              0   100   200    300         400          500         600   700   800
                                                                        no. sample



10
VisioNIR Lyophilization
vs. 5.1
Production examples
Inline measurement with production samples - reproducibility


                                                      Inline NIR Measurement vs Karl Fischer

                   8

                   7

                   6
    Moisture [%]




                   5
                                                                                                            Moisture [%] Inline
                   4
                                                                                                            Moisture [%] KF
                   3

                   2

                   1

                   0
                       0   2   4   6   8   10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50
                                                                 Vial #




                   5 repeat measurement of vials in comparison with reference analysis

11
VisioNIR Lyophilization
vs. 5.1
The method quality aspect
Stability of the method – ruggedness


                                                 12h run static

                                               Sdev= 0,049 %
                       2


                      1,8


                      1,6
                                                                                          The inline application
                      1,4
                                                                                          requires a long term
 moisture content %




                      1,2
                                                                                          stability test.
                       1


                      0,8
                                                                                          The graph shows a run of
                      0,6                                                                 good product for a
                      0,4                                                                 duration of 12 hours
                      0,2


                       0
                       18:28   20:28   22:28      00:28           02:28   04:28   06:28
                                                    time




12
VisioNIR Lyophilization
vs. 5.1
The method quality aspect
Influence of temperature



                                          Influence of temperature on moisture content prediction

                        6,00                                                                                                   Individual prediction of
                                                                                                         47                    product moisture with a
                        5,00
                                                                                                         45
                                                                                                                               nominal relative moisture
                                                                                                                               content of 4% and the
                        4,00                                                                             43
                                                                                                                               associated sensor
 moisture content %




                                                                                                              temperature °C
                                                                                                         41
                                                                                                                               temperature.
                        3,00


                                                                                                         39
                                                                                                                               A clear correlation of both
                        2,00
                                                                                           moisture
                                                                                                         37
                                                                                                                               is given.
                                                                                           temperature

                        1,00
                                                                                                         35
                                                                                                                               From those values a
                         0,00                                                                            33                    temperature correction
                      10.07.08 16:33:36        10.07.08 21:21:36
                                                                   date time
                                                                                11.07.08 02:09:36
                                                                                                                               can be derived.




13
VisioNIR Lyophilization
vs. 5.1
The method quality aspect
Influence of temperature


                                       temperature corrected moisture content

                      6


                                                                                                      In the graph shows the
                      5
                                                                                                      corrected values which
                      4
                                                                                                      have been calculated
 moisture content %




                                                                                                      from the temperature
                      3                                                                               calibration.

                      2                                                                               The residual variation is
                                                                                                      the same compared to the
                      1
                                                                                                      repeatability
                                                                                                      measurements
              0
       10.07.2008 16:33:36   10.07.2008 20:24:00         11.07.2008 00:14:24    11.07.2008 04:04:48
                                                        date time




14
VisioNIR Lyophilization
vs. 5.1
The application result
Conclusion

        The system is able to perform 100 percent product analysis within the required
         validation range, and the accuracy is comparable to the Karl Fischer analysis and
         shows about 2 times the variance of the Karl Fischer results.
        The project demonstrates the possibility of a 100% quantitative control of the
         residual water in freeze dried products, inline in the primary production process.
        The system worked with a production speed of 300 products/minute – 3000/min is
         possible.
        A mass production result of 2.5Mio measured products is shown to demonstrate the
         ruggedness and stability of the process.
        Measurements proof for specificity and accuracy of the method.
        The data proof that the scientific basics and the acquisition of the required process
         parameters leading to a fast and reliable non destructive analysis method for a
         potential real time release.
        The PAT based approach demonstrates a successful implementation and shows
         that direct process understanding using scientific methods to achieve a
         technological system which is able to control the complete process can be realized.
15
VisioNIR Lyophilization
vs. 5.1

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:13
posted:3/28/2011
language:English
pages:15