Sommaire
- Pourquoi et comment faire une caractérisation thermique d’un produit avant lyophilisation ?
- Les clés d'une qualification réussie
- Comment mener une analyse de risque contamination sur un procédé de lyophilisation ?
- CIP evolution in freeze dryers. Optimize turnaround time & water consumption
- Basiques en simulation de procédés aseptiques et lyophilisation
- Silicon oil detection in pharmaceutical freeze dryer by mass spectrometry
- Les spectromètres de masse. Retour d'expérience
- EN 17272. Nouvelle norme relative à la désinfection par voie aérienne automatisée
Silicon oil detection in pharmaceutical freeze dryer by mass spectometry.
Mass Spectrometry has a wide range of applications in the freeze drying process of pharmaceuticals. It can be used to monitor the gas composition inside the vacuum chamber, thereby aiding in keeping track of the drying progress, as well as possible external air leakages and internal silicon oil leakages.
In this paper, the use of mass spectrometry for silicone oil detection is discussed.
The shelves, which hold the pharmaceutical product vials, are supplied with silicon based oil for temperature control. If the oil leaks into the chamber, it could possibly contaminate a whole batch and if it goes undetected even multiple batches. Mass spectrometry can help in detecting such leakages and thereby enhance the safety of production especially for older freeze dryers where leakages from the shelves or silicone oil hoses could be more likely.
1. Basics of mass spectrometry
Mass spectrometry in any form is a method to analyse a sample’s composition. Therefore, the sample needs to be separated into its molecular components. This separation takes place through the different masses of the components.
In pharmaceutical environments quadrupole mass spectrometers are commonly used. Their principle is shown in Figure 1.
A quadrupole mass spectrometer functions through ionizing the sample, accelerating the then charged particles, filtering the particles depending on their charge and mass through an electromagnetic field and finally, detecting the filtered ions.
The gasses flow from the freeze dryer through a pipe into the mass spectrometer, where a filament inside the ion source is heated to a point where electrons are ejected. Those electrons ionize the gas and split up parts of the molecules in the gas. This means that the spectrum detected does not correlate directly with the sample, because only the fragments can be detected.
Silicone oil (polydimethylsiloxane) consists of long chains of silicon and oxygen atoms together with side-groups of hydrocarbons connected to the silicone centres. When subjected to the ion source of a mass- spectrometer the long chains break into several sub-chains and those sub-chains are detected. Also multiple ionization can occur, which further enhances the detection of further possible fragments.
The ionized particles are accelerated through a source slit and lead into four polarized rods (AC and DC on both pairs) which produce an electromagnetic field. The scan through the different mass by charge ratios (m/z) is done by increasing the voltages of the rods. Therefore different mass by charge ratios can pass into the detector.
Detection of the separated molecules takes place by means of a faraday cup, which simply detects any charge produced by incoming ions, or an additional secondary electron multiplier can enhance the detectable traces of ionized particles. Because of the low concentrations of the silicone oil vapours in the chamber the secondary electron multiplier is necessary to amplify the silicone-oil signal.
To simplify the presentation of the mass by charge ratio it is common to assume every particle is charged only once. Therefore, the mass by charge ratio can be expressed simply through its atomic mass. Calculating the concentration through the ion current, can simplified be evaluated by adding up all measured currents and setting the sum as 100%. Every single concentration is therefore a part of the 100% and can then be calculated.
2. Silicone oil detection
2.1/Silicone oil contammination
For the freeze drying process the shelves are heated and cooled in a wide temperature range by a silicone oil circuit and are moved during the stoppering process. The change in temperature and mechanical movement puts strain on the system.
Especially for older systems or after maintenance work in the chamber there is a risk of a leakage with the possibility of silicone oil evaporating into the chamber and thus contaminating the product.
Most likely, a contamination would occur when the sublimation of the product i.e. the vapour flow from the vial stops (at the end of the main drying phase). After the drying process the flow reverses because the chamber is vented before the vials are stoppered. This is when the contamination most likely would occur. Silicone oil contamination is not easy to detect, if a small leak stays undetected then the producer has to prove that earlier batches are not contaminated. The leakage would therefore lead to high costs.
2.2/Detection by mass spectrometry
A quadrupole mass spectrometer can easily detect even small leakages and therefore improves the safety of production. The silicone oils used in freeze drying differ in molecular structures, but they all have segments in common as shown in Figure 2.
After ionization the molecules fall apart into methyl groups bonded to a silicon atom, also oxygen can be part of the detectable fragments. This leads to detectable atom masses after their fragmentation by ionization of 71, 73 and 74. The fragments with these atomic masses are the easiest ones to detect because the peaks differ the most from their base signal. Therefore, they are the best indicator for silicon oil. The highest peak can be detected at 73 AMU.
3. Connection to the freeze dryer
Since mass spectrometry can be used for different purposes in pharmaceutical freeze dryers such as process monitoring, silicone oil and/ or general leak detection, the possibility of using the mass spectrometer at different process-stages and the mechanical connection to the chamber can differ accordingly.
For monitoring the water vapour during primary and secondary drying, method A i.e. continuous monitoring can be implemented to avoid any backflow into the sterile chamber. For service operations and silicone oil leak detection method B i.e. offline monitoring can be considered for implementation.
3.1/Method A : Continuous monitoring
While using mass spectromety in the freeze drying process, care must be taken to ensure no contamination of the product on the shelves. This can be done by implementing special
precautions in the control system (PLC / recipe) that minimize the risk of an unwanted backflow of unsterile gases from the mass spectrometer into the chamber.
Therefore, a fully integrated configuration is favoured, where the mass spectrometer is mounted directly on the chamber of the freeze dryer and the control system of the mass spectrometer is directly integrated into the one of the freeze dryer. Also a moveable mass spectrometer in a trolley can be used, if the relevant signals are exchanged.
The mass spectrometer is connected via a membrane valve to the chamber, which satisfies the 3d-rule. This valve separates the unsterile pipework of the mass spectrometer from the sterile chamber and must be closed during cleaning (CIP) and sterilisation (SIP). In principle, a mass spectrometer can also be connected to the freeze dryer at other points, e.g. in front of the vacuum pumps, at the condenser or at the main valve. For process monitoring it needs to be connected to the chamber. Also for general leak detection and silicone oil detection the chamber is the most preferable connection point since it is closest to the product.
During loading and freezing, the mass spectrometer starts up automatically to prepare for the measurement. During freeze drying, the drying progress can be monitored by measuring the gas composition inside the chamber with the mass spectrometer. To allow desublimation of water vapour on the condenser coils the main valve is open. However, since the cold condenser coils also condense possible silicone oil vapour, the detectability of silicone oil is reduced during the freeze drying process. In addition, the process of sublimation with high flow rates inside the chamber contributes to the worsening of the detection limit.
In order to detect silicone oil, it makes sense to use the mass spectrometer in a secondary process, e.g. before or after the leakage test of the system, where a low pressure is achieved in the chamber and the main valve is closed.
3.2/Method B : Offline monitoring
A standalone mass spectrometer can be used to check a freeze dryer in time intervals or after a service operation with high risk of silicone oil spill.
For that purpose a mobile mass-spectrometer, with or without signal exchange with the freeze dryer, can be used. The freeze dryer needs to be brought into a favourable condition i.e. low pressure, high shelf temperature and preferably a closed main valve or no condenser cooling to enhance the detection of silicone oil. The mass spectrometer is connected via a hose to a port on the chamber and is operated directly via the mass spectrometer unit. The biggest advantage of a mobile configuration is its mobility itself, which enables the operator to connect the mass spectrometer to different ports of the freeze dryer or to use the same mass spectrometer on several freeze dryers.
4. Tests
4.1/Freeze dryers and silicone oils
The silicone oil detection by mass spectrometry was tested on a variety of freeze dryers of different sizes:
Laboratory freeze dryer
The laboratory freeze dryer is a small unit with a chamber volume of 0,7 m3 and a shelve area of 0,6 m2. A stand-alone mass spectrometer was used which was connected by a flexible metal hose to a port on top of the chamber. For test purposes, liquid silicone oil was injected into the freeze dryer. A dosing system, which is capable to vary the injection rate from 40 mg/h to 1000 mg/h, was installed on top of the chamber and connected to a port via a 1/16″ tube. Figure 6 shows the installation of the silicone oil dosing system on top of the freeze dryer. The tests were performed with a silicone oil (TT3) with a viscosity of 3 cSt.
Production freeze dryer 1
Tests were done on two identical production freeze dryers. The freeze dryers are double floor design with the condenser below the chamber and have a shelf area of 6,77 m2 (8+1 shelves). The overall volume is about 8,7 m3, while the chamber holds about 5 m3 with the main valve closed. The mass spectrometer is fully integrated into the software of the freeze dryer and mounted on top of the chamber. The same dosing system used on the laboratory freeze dryer was installed on another port on the chamber top.
In these tests a different silicone oil (X-40 Fragoltherm) was used with a viscosity of 5 cSt.
Production freeze dryer 2
Tests were also performed on another production freeze dryer with a volume of 6,55 m3. Instead of liquid injection of silicone oil by the dosing system, the silicone oil was brought into the chamber via a dispense lock. The dispense lock was installed at a flange located on the freeze dryer chamber wall. The dispense lock, in which a vial with a certain amount of silicone oil was placed, was separated from the chamber by a hand valve. The silicone oil weight was measured before and after the test. The weight difference corresponds to the evaporated silicone oil mass during the test time. To further improve the detection with the mass spectrometer, a hot air blower was used on the dispense lock to increase evaporation of silicone oil. This test was performed with a silicone oil with a viscosity of 3 cSt but from a different manufacturer (KT3).
The mass spectrometers used for all experiments were either a QMG 250 PrismaPro or a QMG 220 PrismaPlus from Pfeiffer Vacuum GmbH. The secondary electron multiplier was always used to amplify the signal for silicone oil. The results are always shown in ppm (or %), and are calculated by the simplified formula mentioned before.
The mass spectrometer was configured to measure the following atomic masses: 18 (water), 28 (nitrogen), 32 (oxygen), 40 (argon), 44 (carbon dioxide) and 73 (silicone oil). For the test on production freeze dryer 1 instead of the mass of 40 (argon), 4 (helium) was used.
4.2/Basic silicon oil signal as a function of the chamber pressure
Since the mass spectrometer never shows zero signal for any mass, but always measures a certain but low ion current, it is important to measure the empty system. This is done to determine the silicone oil signal of a clean system before using the mass spectrometer for silicon oil leak detection.
A comparable test was done on the laboratory freeze dryer. During this test, the main valve was open, the condenser cooling was active and the pressure inside the freeze dryer was regulated to 500 μbar, 300 μbar, 200 μbar, 100 μbar, 50 μbar and 20 μbar. The shelves were controlled to a constant temperature of 20 °C during the whole test duration. No silicone oil was induced during this test. However, for the laboratory freeze dryer it must be taken into account that the system cannot be cleaned automatically or sterilised. In addition, it must be considered that tests with silicone oil dosing have already been carried out on the laboratory freeze dryer before this test. Therefore, there were most likely silicon oil residues inside the laboratory freeze dryer, which evaporated due to the low pressure inside the chamber.
Figure 7 shows the measured ion currents of those single gases divided by the total ion current.
Key learning : The test shows that a low chamber pressure enhances the detection of silicone oil. Even though these measurements are done with a very likely contaminated system, this test should also be performed on production freeze dryers before using the mass spectrometer to detect silicone oil within production runs for determining the basic signal of a clean system.
4.3/Influence of chamber pressure and shelf temperature on silicone oil detection
A test was performed on the laboratory freeze dryer to study the effect of different chamber pressures and shelf temperatures on silicone oil detectability. The silicone oil was injected into the freeze dryer via the dosing system with a flowrate of 40 mg/h for 45 min. The test was performed at four different chamber pressures i.e. 50 μbar, 100 μbar, 200 μbar and 300 μbar, and two different shelf temperatures, -20 °C and +20 °C. The main valve was kept open during the test and the condenser cooling was active.
To enhance the comparability of the different measurements, the change of concentration during the dosing of the silicone oil is used. This is due to the effect of a rising silicone oil signal caused by silicon oil residues from previous dosing tests.
In Figure 8 the rise of the measured silicone oil concentration (Δppm) is shown during the 45 min injection time at different shelf temperatures and chamber pressures.
Key learning : The test shows that a lower pressure inside the freeze dryer enhances the silicone oil detectability of the mass spectrometer. Since the substances must be present in gaseous form in order to be detected by the mass spectrometer, the liquid silicone oil brought into the system by dosing must first evaporate. This would also be the case in the event of a real silicone oil leakage from a hose or a shelf inside the freeze dryer. In order to detect the injected silicone oil, low pressures in the freeze dryer are therefore advantageous. The experiment also shows that higher shelf temperatures are advantageous due to the same effect, the necessary evaporation of the silicone oil.
It should also be noted that for this test the silicone oil signal, measured by the mass spectrometer, is still increasing after 45 minutes of injecting silicone oil.
4.4/Influence of sublimation on silicone oil detection
This test was additionally performed on the laboratory freeze dryer to study the effect of water sublimation on silicone oil detectability as it occurs during a regular drying process.
The silicone oil was injected into the freeze dryer via the dosing system with three different flow rates of 40 mg/h, 80 mg/h and 120 mg/h, each for 45 min. The shelves were maintained at a constant temperature of 5 °C for loading the shelves with vials filled with a water/ mannitol solution. After loading, the shelves were cooled down to -50 °C within one hour and held at that temperature for about 6,5 hours to freeze the water in the vials. After cooling down the condenser coils and opening the main valve, the system was evacuated and controlled at a pressure of 50 μbar. Then the shelf temperature was set to 20 °C to start the sublimation/ drying process. These conditions i.e. a shelf temperature of 20 °C and a chamber pressure of 50 μbar, were kept constant throughout the test.
In Figure 9 the rise of the measured concentration of silicone oil is shown during the 45 min injection time for different silicone oil injection flows during sublimation. As a reference, the rise of the measured concentration of silicone oil without sublimation is shown for a dosing flow of 40 mg/h silicone oil from the previous test.
Key learning : The test shows a decreasing detection capability of silicone oil during the sublimation process compared to an empty system. While the increase in measured concentration of silicone oil without sublimation was at a value of about 70 ppm after 45 min – and is still rising – the signal increase at the same injection rate with sublimation was below 5 ppm after 45 min. Even with higher dosing of 80 mg/h and 120 mg/h of silicone oil, the measured increase of silicone oil concentration by the mass spectrometer was much lower than that without sublimation.
4.5/Influence of cold trap and main valve on silicone oil detection
The test was performed on the production freeze dryer 1. For this test, the shelves were maintained at a temperature of 20 °C, the condenser was cooled down and the main valve was opened. At first, the system was evacuated to 50 μbar, the empty system was tested and the concentration of the silicone oil was noted (3,2 ppm / 1,2 ppm). Afterwards 48 mg/h of silicone oil were dosed into the freeze dryer. No significant rise of silicone oil was measured by the mass spectrometer. The main valve, which separates the cold condenser from the chamber, was closed. Almost immediately, a rise of the measured concentration of silicone oil was detected (19 ppm / 30 ppm). After opening the main valve again, the measured concentration by the mass spectrometer dropped again. The dosing was then subsequently increased up to 500 mg/h (780 mg/h for the second freeze dryer). An increase in measured silicone oil by the mass spectrometer could be detected (12 ppm / 4 ppm) but it was still lower than with the closed main valve and a dosing of only 48 mg/h.
Key learning : The closure of the main valve increased the detectability of silicone oil in the chamber at a pressure of 50 μbar and 20 °C shelf temperature for this freeze dryer. It could be advantageous to close the main valve for silicone oil detection to exclude the cold sink from the monitored system.
4.6/Influence of test conditions (vapour injection)
The test was performed on the production freeze dryer 2. For this test, the shelves were maintained at a temperature of 20 °C, the condenser was cooled down and the main valve was opened. At first, the system was evacuated to 10 μbar and the mass spectrometer was switched on to measure the clean, empty freeze dryer. A total of three times a vial filled with silicone oil was inserted into the dispense lock. The weight was measured before and after the test to determine the evaporated mass of silicone oil. The dispense lock was cleaned between the measurements in order to avoid having any residues of silicone oil inside. The manual valve for the injection of silicon oil to the dispense lock was opened for a duration of 5 minutes.
In Figure 10 the measured concentration of silicone oil is shown over the whole test time. The clearly recognisable three peaks are the times at which the manual valve for the dispense lock was opened and silicone oil was thus able to evaporate and flow into the chamber.
Key learning: With this “vapour injection” method, even very small doses of silicone oil (1 mg – 2 mg) were detectable by the mass spectrometer. In addition, the response time was very short (< 5min), and the detection signal was quite clear (about a tenfold increase in concentration) in this test.
Since the freeze dryer was not cleaned in between those three measurements while no silicone oil was added, it looks like the basic signal increases over the testing time. The measured concentration for silicone oil was around 7 ppm before the first test, which rose to about 12 ppm after the first test and to about 16 ppm after the second test.
Summary and further studies
With different tests on different sized systems, the general usability of mass spectrometry to detect silicone oil in a pharmaceutical freeze dryer was shown. The tests also show that the measurement result for the detection of silicone oil can be improved or worsened by certain parameters. In order to improve detection, and thus to be able to measure even small traces of silicone oil inside the chamber with the mass spectrometer, it seems to be advisable to:
– set the chamber pressure as low as possible,
– set the shelf temperature as high as possible,
– measure in an empty system without active sublimation,
– probably close the main valve between condenser and chamber.
For testing the limits of silicon oil detection with mass spectrometry, two different dosing systems were used, “liquid injection” and “vapour injection”. Since the silicone oil always has to evaporate before it can be detected by the mass spectrometer, vapour injection results in faster increase of detected silicon oil concentration. The used laboratory freeze dryer seemed to accumulate silicone oil over time, as it cannot be automatically cleaned or sterilized.
For this paper the simplified calculation method was used which does not accommodate for the different ionisation probabilities and fragmentation distribution. Fragmentation distribution of silicone oil is not known in detail and has to be determined in further studies to ensure a more precise calculation of gas concentrations.
For further studies, it would be interesting to investigate the degradation of the sensitivity of the mass spectrometer over time with and without exposing it to silicone oil. This degradation was seen on the used mass spectrometers after inducing silicon oil and has to be accommodated by increasing the gain of the secondary electron multiplier.
Another interesting area for further studies on mass spectrometry for freeze drying are experiments with a placebo product in a freeze dryer to which silicone oil is added during a normal freeze drying run to simulate a leakage. The aim is to investigate whether traces of silicone oil could be detected in the placebo product after freeze drying, and whether these could be detected beforehand i.e. during the freeze drying process or during a separate process by the mass spectrometer. The test can be used to tackle the question of which measured concentration of silicone oil in the freeze dryer might still be tolerable. The mass spectrometer as a PAT can also be used for other purposes than the detection of silicone oil. With the help of mass spectrometry, the water vapour concentration during freeze drying can be measured e.g. for end point detection. For leak detection from the technical or clean room, the gas composition in the freeze dryer can be analysed for air components such as oxygen, provided that the freeze dryer is operated with nitrogen.
Partager l’article
Paul CORNEL, Matthias KOPP, Jan-Malte MOSBLECH & Alexander TAMBOVZEV – OPTIMA
paul.cornel@optima-packaging.com
