The path to the source

The rapid determination of contamination sources in cleanroom production environments is fundamental for the smooth running of clean production. Conclusions regarding the source of contamination can be drawn when exact information about the material composition of the particles is available. There are two ways to ensure that a product produced in a cleanroom is free of particles: either the air in the room or the product itself can be checked for particles. The second method is the more commonly chosen one when parenterals are being produced. It is required by law in order to protect the consumer that the contents of every ampoule are checked for particles. A cleanroom operator can quantitatively assess particles and classify them by size using classical monitoring systems. A qualitative assessment use not usually conducted for production lines. This is a mistake, because knowing a particle's composition can give insight into possible faults occurring during production. With the help of an exact chemical analysis of the particle it is possible to identify its source and thereby eliminate it in the future. APSys has introduced an analysis method that allows the rapid chemical identification of particles to be conducted: ParticleExplorer™ analyses airborne particles LiquIdent® analyses particles in highly purified fluids Both systems are based on the same principle: RAMAN-spectroscopy. Using this method, most compounds that are used during clean production can be easily identified. The two systems differ in two points which arise from their different uses: LiquIdent® registers and analyses particles of 10 µm and larger, while ParticleExplorer™ can already analyse particles as small as 0.5µm. For both systems the particle must first be applied to a sample carrier; this is achieved externally through filtration in the case of LiquIdent®. ParticleExplorer™ takes the particles automatically out of the air. The path from particle to source is best described using an example from real life; in this case, the use of LiquIdent® in the quality control of parenteral production. The same technology and the same procedures are applied when using ParticleExplorer™, except for the preparation of the sample.

Parenteralia production In the production of parenterals and ophthalmics, the exclusion of visible insoluble contaminants from solutions is paramount. Particles that are 50µm and larger are visible to the naked eye. Every single bottle, ampoule, or vial has to be taken and examined by eye or by a machine. Finished products are then sorted out and rejected. Rejection rates are in the range of 0.2 to 3%. Particle contamination can originate during various stages of production and filling processes. Determining the origin of contamination is often very time consuming. When increased particle concentrations are detected during parenteral quality control, the manufacturer usually attempts to determine the material composition of the particles in order to ascertain the origin of the contamination. No standardised technology for such analytical procedures exists to date.

Possibilities for characterisation The simplest method is the light microscope, which allows conclusions to be drawn on the shape and colour of a particle. In a short time it is possible to find out if a particle is e.g. a black fibre or a small, undefined spherical object. There are several possibilities open to the user for determining the nature of the material; one of these is EDX analysis, which reliably supplies information on the elements in the material and their proportions in the particle. However, all of these methods have something in common: they are often carried out at a considerable distance from the production site in an external lab, and the result comes several days later than the actual particle problem did. In addition, they require the employment of specialists who can carry out analysis and interpret the results.

Rapid analysis system LiquIdent® and ParticleExplorer™ are both integrated measuring systems consisting of an encapsulated and easily washable special steel housing. The software is operated through a touch-screen monitor; the samplings can be inserted very easily through a flap in the front of the device. The work needed to conduct a measurement is essentially limited to the deposition of the sampling on a membrane through filtration. The membrane can be directly inserted into the device without any further steps. After pressing the start button, each particle is illuminated by a laser beam. The scattered light reflected off the particle is then captured by a detector and is qualified using RAMAN-spectroscopy. The computer then compares the resulting spectrum with a spectrum saved in the database. The user then automatically receives the precise name of the compound. Even unskilled workers can easily identify a particle. The results are automatically recorded in a hypertext document that conforms to CFR 21 Part 11.

Different particle sources After one learns what the particles are composed of, there is a better chance of finding out where they came from. Possible sources are active ingredients or other additives, water, gases, i.e. anything that is used to make the product. However, foreign particles can also find their way into the product through the wear-and-tear of the production machinery. In this way, the product is contaminated by foreign particles created during the production or filling processes. For this reason, the environment must be closely monitored for airborne particles. The operators wear special protective clothing to guard against this possibility, but foreign particles can arise from various sources e.g. during the cleaning process. Most such pollutants come in with the containers and their closure systems. There is a whole range of possible sources that can be responsible for pollutants, which makes it quite a challenge to actually pinpoint the actual source. However, the pharmaceutical manufacturer usually keeps a complete list of all materials involved in the process, a fact that limits the number of possible sources to a reasonable range. The long-term goal is to develop control parameters of the relevant process by statistical evaluation of the particle composition. This makes variations in the known particle pattern immediately visible and enables the user to react quickly to unknown particles.

Process-optimising through analysis At the pharmaceutical manufacturer which first tested the LiquIdent®, several methods for identifying particle contamination are available. These measurements and the interpretation of their results require time and a staff of highly qualified personnel. Furthermore, it can take several days after collecting the sample to obtain the results of the analysis. Although these analytical procedures were not carried out on a regular basis, the manufacturer succeeded in optimising the processes in question and increasing their yield. The team members in quality control sought to further optimise the processes by more often and more regularly carrying out chemical analyses of the particles. Then the next step was the search for an analytical method, which would allow the quality control to carry out many analysis simultaneously and without great difficulty.

Continuous particle characterisation at the production site Using the aforementioned list and a determination of the particle, it is possible to assign the origin of this particle to a process step. Consequently, one can perhaps optimise this process step so that the particle source is eliminated in the future. Although until now these methods required a lot of time to be carried out, the quality control staff have followed this path in the past and have become quite experienced in identifying and eliminating particle sources. However, in order to improve knowledge about particle reduction, the employees at the lab needed more data with which a trend analysis could be made; data which could be collected without additional personnel. This work could be carried out by a machine able to analyse particles very quickly and with a high degree of precision. This was the reason to start preliminary tests with LiquIdent® in the hope that the system would be able to simplify particle identification. The results of previous single measurements should be made complete using a standardised material composition test. The long-term goal is to develop control parameters of the relevant process by statistical evaluation of the particle composition. This makes variations in this known particle pattern immediately visible and enables the user to react quickly to unknown particles.

Challenge of the first tests The LiquIdent® system was used in production quality control. A clear goal was set at the beginning of the study: LiquIdent® should be able to differentiate between drug substance- particles and other particles, and should be able to identify the unknown particles.

Summary of the first tests A training of one hour qualifies an employee to operate the device and to independently identify the particles. After two hours of operation, Liquident® was shown to deliver qualitatively satisfactory results. The drug substance was added by APSys to the database before the measurement was performed; LiquIdent® then unambiguously identified the drug substance-particles and was able to tell the difference between these and foreign particles. Within minutes the device delivered a spectrum for a foreign particle that revealed a correlation to a spectrum with a high proportion of carbon-hydrogen. The databank then delivered the corresponding result: polyethylene. That was the first indication that somewhere in the process polyethylene is used and a bit of it unintentionally comes into the product.

Future development of fine process profiling Standardising particle identification is the first step towards lasting process optimisation. With the help of LiquIdent®, the manufacturer now has the possibility of entering all constituents of a product necessary for its production into the database of an automatic analysis device. In this way, all materials utilised in production and filling processes can be retrieved from the databank. After the source has been recognised, it can be eliminated or reduced. It addition, it will be detected by the automatic device; the number of particles and particle source are constantly monitored. Furthermore, a process-specific historical particle profile is fast and simply prepared by standardised record generation as well as by statistical evaluation of particle pollution. Minimal variations in the profile are registered and allow the easy development of control parameters for a lasting optimisation of the process. The actual particle profile is constantly compared to the historical particle profile as well as to the control parameters in order to introduce relevant measurements. With this new knowledge quality assurance will then be able to trace defects and deviations in the production before the specification limits are exceeded.

For further information contact Annick Eimer at Apsys tel: +49 (0)30 5301219