Lyophilization LYOPLUS™ PAT for Pharma Freeze Dryers

LYOPLUS™ mass spectrometer is a multi-purpose measurement device for pharma freeze dryers helps save money and time. The system is able to work alongside any existing PLC / SCADA system as a stand alone unit or fully integrated within the control system. It can also be operated independently as a monitoring system that does not interfere with any qualified processes.

The LYOPLUS™ mass spectrometer provides essential data from within the freeze dryer itself, including: the ability to detect very small silicone oil leaks inside the dryer; the ability to monitor the moisture content within the freeze drying chamber; identifying the endpoint of primary and secondary drying; and a much easier and faster leak detection.

Silicone Oil Detection

Top of the list of system benefits is its unique ability to detect very small traces of silicone oil within the dryer. Silicone oil is used to transfer the required heat energy to the product. After years of operation and many cycles the arduous conditions within the dryer can lead to small leaks. As a consequence oil can contaminate the end product.

After the leak in the circulation system has reached a certain size this contamination will be detected during end product testing in quality control. That particular batch is of course lost but additionally suspicion immediately rises about the purity of previous batches.

Using LYOPLUS™ however, it is possible to detect even small traces of silicone oil leaking within the dryer during the operating cycle. No additional product is at risk as the leak would have been detected immediately.

Moisture Content

The LYOPLUS™ system can also monitor moisture levels within the drying chamber during the drying cycle. At the beginning of the drying process the chamber will be filled with a saturated water vapor as water is removed rapidly from the product. During the later stages, however, the moisture level in the chamber reduces significantly. LYOPLUS™ measures this drop very accurately and correlates the measurement with the actual average product moisture inside the vials. This information can be used to refine drying recipes and avoid any unnecessary drying time. One of our customers has reported a three-hour saving per cycle providing improved productivity and a reduction in power consumption.

An additional benefit comes from being able to predict the drying curve of the product through the moisture content in the chamber. Drying behaviour for each batch can be compared with a ‘qualified batch’ to confirm that the process is progressing normally.

 

Chamber Leak Tests

To prevent any contamination through a leak in the system, it is mandatory to perform a leak test after each critical process, i.e. after sterilization. This can take up to four hours, depending on the type of test performed, causing loss of productivity and consuming additional energy to create the necessary test conditions. Using LYOPLUS™ however it is possible to reduce the time for this standard procedure down to approximately one hour owing to its very high sensitivity.

If during the test a leak is detected it is necessary to perform a helium leak test to find the source of the leakage. Using standard external equipment for helium leak testing sometimes takes several hours to set up; LYOPLUS™, however, is permanently connected to the system so the helium leak test can begin instantly, detecting leaks much faster thereby saving valuable time.

Technical Data

LYOPLUS™Type 1Type 2Type 3
Contamination monitoringYes
Drying endpoint verificationYes
Leak CheckYes
OperationStandaloneIntegrated
PLC update requiredNoNo-
SCADAProcess EyeWinCCFALCO*
MultilingualNoYes
21 CFR Part 11NoYes
Audit trail, User logNoYes
Batch reportNoYes
Data exportYes
ConnectionTri Clamp or customer demand (<3d)
IQ/OQYes
PC HardwareLaptopDesktopDesktop or Server
Power supply240V / 115V ; 50/60Hz
PortableRestrictedNoNo

Detection of silicone oil leakages in freeze dryers with LYOPLUS™

As a small leakage from the silicone oil tubes in pharmaceutical freeze dryers may go undetected for several cycles it is a major threat both to product safety as well as to the economic performance of such an installation. A properly mounted mass spectrometer offers a non-invasive way to detect amounts of silicone oil down to 1 ppm. Additionally the system allows for improved leak testing procedures and also the use as a PAT tool for improved process control.

Introduction

Freeze drying has been used for many years in the production of parenterals to stabilize thermo labile molecules. With the continually rising number of large molecules the use of freeze drying will further increase as it offers the gentlest way of converting a liquid formulation into a more stable solid.

From an engineering point of view the typical process is very rough and demanding on the hardware. After washing every part of the entire dryer potentially in contact with the product, it is exposed to saturated steam at 127°C which relates to an overpressure of ~1,5barg. After loading the vials the dryer is evacuated down to 0,05mbar while the shelves and ice condenser are cooled down to temperatures typically in the range of -85°C. Various types of silicone oils are used to transfer the required heat energy in and out of the dryer. As a result of these rough conditions, and after many cycles, small leaks can occur in the silicone oil circulation system. This article identifies the most critical areas for the occurrence of leaks and introduces a system for the detection of small amounts of silicone oil in the system. It will also introduce new and improved state-of-the-art designs that minimize the risk of leakage.

Silicone oil is brought to the required temperature in the heat exchangers in the technical zone and then pumped into the system. Most important is the even heating/cooling of the shelves. From an engineering point of view, building the shelves to be as light weight as possible while ensuring an even surface temperature and a plan surface of all shelves is a challenge. The shelves are moved up and down during the loading, unloading and when closing of the vials at the end of the drying cycle. This means that the tubes supplying the silicone oil to the shelves need to be constructed and guided in a way that reduces bending that could increase the risk of cracks in the tubes.

Freeze dryers of older design tend to get small cracks generated through numerous cycles. It is through these cracks that silicone oil can escape into the drying chamber. To begin with these leaks are relatively small and will not cause malfunctions of the system. This means there is the risk that several batches could be produced before, by chance, QC may pick up the problem analyzing vials produced. That batch would most likely have to be destroyed but there is little guarantee that there hadn’t already been unacceptable amounts of silicone oil present in previous batches. For these reasons it would be helpful to have a method allowing the detection of traces of silicone oil from the moment the first leak occurs.

Feasibility Trials: The Detection of Silicone Oil Traces

Mass spectroscopy was chosen as a method of detection because of its sensitivity. Quadrupole mass spectrometer systems are widely used for the monitoring of critical processes and the detection of residual contamination. The principle of the operation is based on electron impact ionization, the separation of the formed ions in an electromagnetic field, and their detection. Early trials proved the principle of the method and determined the sensitivity to be evaluated. An MKS Vision 2000-P mass spectrometer, as shown in picture 2, was connected via a standard flange with an LYOVAC™ FCM-2 freeze dryer installed in the test centre at GEA Lyophil in Hürth (Germany).

When silicone oil is introduced into the mass spectrometer the molecule is broken into several fragments. By detecting the characteristics of the fragments, traces of silicone oil can be found. During the tests different types of silicone oils were used. As KT5 (Bayer AG) has the lowest vapor pressure it was used for all measurements. After the first trials a characteristic peak at 73 AMU was found. Additionally other types of silicone oil which may be present in a freeze dryer were tested. While the oil typically used for the siliconization of stoppers gave a characteristic peak at 56 AMU, silicone oil-based maintenance spray- typically used for lubrication of moving parts in loading and unloading systems- gave a characteristic peak at 58AMU.

As the characteristic peak at the mass of 73 AMU represents the type of oils used in the temperature circles of a freeze dryer this signal was used for the further measurements.

As a next step an amount of 2 mg of silicone was introduced into a vial which was closed by a stopper and held by the upper shelf in position. After the desired vacuum level was reached the upper shelf was lifted. As a result the stopper popped off and the oil evaporated into the freeze dryer. After just seconds a signal was detected.

During a further simulation 100mg KT5 was introduced over a small buffer volume into the drying chamber with a shelf size of 40m². This test confirmed again the characteristic signal pattern of silicon oil.

 

Measurements at Production Units

LYOVAC™ Freeze Dryer with Automatic Loading- and Unloading System ALUS™
The detection system was mounted onto production size freeze driers.
UnitShelfarea m³ChamberVolume m³mg
LYOVAC™ FCM 20,10,080,4
LYOVAC™ GT 100,80,211
LYOVAC™ GT 300-D24,89,145
LYOVAC™ FCM 400-D206,130
LYOVAC™ FCM 500-D4412,260

While two of these LYOVAC™ freeze dryers were installed at GEA the other two are manufacturing units installed at European production sites of two multi-national pharmaceutical companies.

Dynamic effects have to be taken into consideration as a result of the larger dimensions of production scale driers. To achieve an enhanced sensitivity the measurement is done when the mushroom valve between dryer and condenser is closed.

For the determination of sensitivity it was decided that the signal would be measured after the chamber was contaminated with a known amount of silicone oil that should be at least 5 times higher than the average background noise level detected. After 12 hours of evacuation the noise level was slightly below 0,2 ppm resulting in the ability to safely detect silicone oil contamination down to 1 ppm.

Evaluation of further possibilities

Performing a leak test is a standard procedure when assuring the integrity of a freeze drying chamber. The test requires the chamber to be evacuated to a pre-defined pressure, all valves closed and the rise in pressure monitored. To assure a meaningful measurement this is often done for between two and four hours. A mounted mass spectrometer offers a much faster way. Before evacuation the chamber is flushed with nitrogen. After the desired vacuum level has been reached all valves are closed and the composition of the remaining gas (> 99% Nitrogen) is monitored. As a result of small leaks- present in every industrial scale freeze dryer- air flows into the chamber which can be detected by monitoring the rise in concentration of oxygen.

Outlook

A model for the quantitative analysis of the oxygen signals has been developed. Due to the sensitivity and fast response of the instrument it can additionally be used to determine the location of leaks in the chamber.

In addition the spectrometer can also detect the concentration of water in the freeze drying chamber. This means that it can be used as a PAT tool for monitoring the drying cycle and the safe and non- invasive detection of the end points of primary and secondary drying.

The system is currently retrofitted to two production scale freeze driers owned by multinational pharmaceutical companies with more companies evaluating the potential of the method.

References

1. Handbook of Vaccuum Technology, Karl Jousten, Wiley-VCH Verlag GmbH & Co. KGaA; Etition: 1. (17. September 2008)   
2. NIST Chemistry WebBook, NIST (National Institute of Standards and Technology) Standard Reference Database Number 69, http://webbook.nist.gov/chemistry by Uwe Meissner, MKS, Munich, Germany, Dr. Harald Stahl, Senior Pharmaceutical Technologist for GEA and Daniel Steinkellner, Process Engineer, GEA Group.