Friday, September 24, 2010

Contamination Detection Basics

We are often asked about the relative importance of cleaning method qualification (aerospace, military, and some other groups like that term) or validation (for biomedical devices, pharmaceutical applications, etc.) and ongoing monitoring. Both are important and we suggest you monitor using a subset of the tests that are performed for qualification/validation. Even more important, as we recently explained,1 plan the validation/ qualification during development of the critical cleaning method.
Even if the process is running smoothly, keep an eagle eye out for cleaning process changes both inhouse and by your components suppliers, and for formulation changes by manufacturers of cleaning chemicals and process chemicals. Determine what might be a problem. Monitor appropriately; and, if there appears to be a problem, revalidate.
Identify, then qualify/validate and monitor the critical cleaning steps. The terms critical cleaning and precision cleaning are often used interchangeably. We have a preference for the term critical cleaning. Precision cleaning conjures up a vision of cleaning in a highlyrestricted cleanroom. Perhaps each individual component is cleaned separately by a highly-trained technician; perhaps there are wet benches with automated product handling; maybe there is a multi-chamber automated spray system that feeds directly into the cleanroom. This is a limited view of the important cleaning step. In our experience, the important cleaning step, the critical cleaning step, may occur in a machine shop or in a job shop (eg. a coating facility), in what looks at first glance like an automotive repair facility. If the soil (matter out of place) is not adequately removed at that step, subsequent cleaning steps may not resolve the problem.
Subsequent processing and cleaning may actually exacerbate contamination by inadvertent chemical reaction of the soil, drying of the soil, or by embedding the soil in the surface of product. Contamination happens long before the product enters the cleanroom. A cleanroom can minimize recontamination, but the most sophisticated cleanroom or controlled environment may not correct a contaminated product.
As with cleaning, more testing is not always better. Whether the issue is validation/qualification or monitoring, select the appropriate tests. This involves re-enacting that venerable game of “Animal, Vegetable, or Mineral.” Consider what might have to be tested.
What soils are you looking for? Where does the soil come from and why are you worried about it? In what form are the soils? A solid? A liquid? A vapor? Are you concerned about all soils, certain categories (such as silicones or large fluorinated compounds), or do you need to identify a single specific soil that might be toxic to the host? Are the soils particulate or film (as from Airborne Molecular Contamination)2 or both?
How much soil can you tolerate? The concept of “none” just isn’t going to happen. Besides, how would you know? Every test has limits of detection (See Figure 1).
Where are the soils of concern coming from? Some possibilities include air, water, process chemicals, cleaning chemicals, the surface of the product, and/or extract of the surface of the product. People also have to be monitored in terms of conformance to policies, gowning requirements, and adherence to cleaning and assembly processes.
Figure 1
The monitoring test that is free or low-cost may in fact be costly. There is a tendency to depend on a technique that is readily-available and familiar, such as Scanning Electron Microscopy/Energy Dispersive X-Ray (SEM/EDX) analysis. It is a useful, valid technique for surface elemental analysis; but it will not detect everything. If the contaminant of concern is a specific organic compound that forms a film on the surface, SEM/EDX is not your method of choice. Perhaps Fourier Transform Infra-Red (FTIR), X-ray Photoelectron Spectroscopy (XPS), Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS) or other techniques would give the information you need. SEM/EDX is not the choice for volatiles; they will evaporate during sample preparation. Volatiles that might outgas from the surface might be better detected by headspace gas chromatography. Volatiles in the air, such as ammonia, that “swoop and destroy” materials on the surface are better detected by AMCrelated methods.2 On the other hand, airborne particles and volatiles that alight on the surface are better picked up by witness samples.
On the third hand (or perhaps it’s the fourth hand, we have lost track), particles introduced by the cleaning process or other manufacturing process may have to be detected by examining the product. This can start with visual observation by experienced technicians and progress to more complex and specific methods. Depending on the product and the contaminant(s), both surface analysis and extractive analysis may be needed.
Surface analysis has the advantage of detecting the soil in-situ; it is time consuming and operator dependent. Analysis after extraction is more thorough in that theoretically soil is extracted from areas that cannot be detected by visual inspection or line-of-sight surface analysis. At the same time, extraction is also methoddependent and operator-dependent; and setting up an effective extraction process is analogous to setting up a good cleaning process. Also, extraction may tell you what was on the surface, but it will not specify where on the surface it came from.
The ultimate goal is to build a competitive device or component. That means achieving high quality and reliability while minimizing true costs. Omitting appropriate qualification and monitoring analyses may not save money if the process does not remain in control. On the other extreme, the most sophisticated and expensive tests may be unnecessary, or if it is not the right test, will not guarantee high process yield or component reliability.
  1. Kanegsberg & Kanegsberg, Validation Readiness, Parts 1 and 2, “Controlled Environments Magazine,” January and February, 2010.
  2. Kanegsberg & Kanegsberg

No comments: