Clean on the surface

Those concerned with the manufacture of implantable medical devices, pharmaceuticals and other critical products justifiably expend enormous effort on monitoring particles in the cleanroom. Monitoring particles on the product is another critical part of contamination control. Viable and non-viable contaminants, both particulate and thin film, have the potential for negative interaction with the host. Implantable devices are expected to perform reliably for decades. Even very adherent, supposedly inert particles can compromise performance.

Minimising surface particulate contamination is likely to become more critical over the next few years. As medical devices become smaller, the ratio of surface to volume increases. As devices become more complex, the potential for entrapped particles is exacerbated. For combination devices, such as a drug-releasing stent, the performance of the active surface or delivery system depends on the cleanliness of the surface, on appropriate surface quality;1 this includes minimising particles that might migrate from surrounding surfaces to the active surface of the device.

The issue of surface particulate contamination control is not restricted to medical devices but is applicable to many critical applications, including vacuum components, analytical equipment, detectors and aerospace assemblies.

Surface particles

Cleanrooms are clean; manufacturing is not. Many manufacturing and assembly processes generate particles. Complex devices are often assembled in many steps, perhaps at remote locations, often in facilities that bear no resemblance to cleanrooms, with only the final assembly occurring in the cleanroom.

Cleanrooms cannot correct a contaminated product. Particles can be introduced during manufacturing; and they may be entrapped in a thin film residue. Particles can occur in process fluids; and they may be generated during necessary steps in fabrication. Water-soluble lubricants are often required, not only for environmental and worker safety considerations but also because they are considered to be simple to remove.

In my experience as the “cleaning lady” advisor for many critical applications, water-based metalworking fluids can be exceedingly difficult to remove, particularly with aqueous cleaning processes.

The lubricants contain particles as part of the formulation; or particles may be inadvertently introduced during the fabrication step. Drilling and polishing entraps particles in caramelised, dried soils. Such soils can be exceedingly adherent; and, in my experience, may not be completely removed during the cleaning process.

The cleaning process itself can introduce particles. A typical aqueous process for critical cleaning applications has a wash step, several rinse steps and a drying step. However, the initial processing may be performed in a non-critical manner. Many in the supply chain assume that the final assembler will ultimately clean the product. The manufacturer may simply wash without rinsing. Even if there is a rinse step, it may be with unfiltered tap water; and drying is likely to be air drying or blow-off with compressed air from the shop.

This sort of handling leads to residual, dried-on soils and residual cleaning agent residue. Such residue inevitably contains particles entrapped in a thin film mixture of organic and inorganic materials; and this sort of mixture is difficult to remove.

Cleanroom processing

Even with well-designed, carefully monitored cleanrooms, and where technicians are well-educated rather than simply trained by rote, particles can be generated after the product enters the cleanroom. This can occur during mechanical handling and even as the result of a non-optimal cleaning step. Drying is part of the cleaning process and must be controlled, even in cleanroom environments.

When we at BFK were troubleshooting a process, the lead technician pointed out product left to air dry within a very well-controlled area in the cleanroom. With a straight face, he referred to the process as “our particle generation step.” Particulate levels after extraction of the supposedly cleaned parts were consistent with his assessment. In this case and in similar situations, the recommendation is to add a drying step in a controlled chamber.

Surface particle monitoring

A “witness sample” can be a useful tool, as long as one is aware of the limitations. It can be difficult to design a witness sample programme that realistically emulates the entire manufacturing process, including the cleaning steps. A dry plate set out in an area of the cleanroom where the drying step occurs may not replicate conditions seen by larger, complex parts that entrap water or processing fluids.

Visual or microscopic examination of the part has its place. One is able to see the particle in situ, and if necessary, it can be further examined using scanning electron microscope (SEM) or Energy Dispersive X-Ray (EDX) technology. Where appropriate, further surface investigation using such techniques as X-ray Photoelectron Spectroscopy (XPS) or Auger electron spectroscopy (AUGER)2 may be used to determine molecular structure of the culprit particle. But obtaining a representative surface may be difficult: particles trapped in blind holes and crevices will not be readily observed.

Standards for observation, record keeping, and reporting must be instituted. Even with good employee education and training, differences among technicians can be significant. In addition, increased handling increases the possibility that the component may become recontaminated during the monitoring process.

For complex or large parts, particle monitoring after extraction with liquid is essential. Basically, the part is cleaned, dried and extracted using deionised water or a suitable solvent using an appropriate technique and with appropriate controls.

Particle monitoring after extraction can be performed in liquids. Typically, the particles are sized but not evaluated. Alternatively, the particles can be collected by passing the extraction solutions through an appropriately-sized filter. They can then be monitored visually or by automated methods. If necessary, the collected particles can be further analysed to determine chemical composition.

Often, particulate determination by filtration is paired with determination of non-volatile residue (NVR) to provide a more complete picture of extractable residue. The filtrate is taken to dryness and the NVR is determined gravimetrically. In this case, NVR includes thin-film residue along with any particulate that passes through the filter.

The operative phrases for reproducible, informative extractive particulate determin-ations are “suitable solvent” and “appropriate technique.” The design of the extraction process needs careful consideration.

Typically, the part is processed, washed, rinsed and dried according to the specified method. Extraction may consist of bathing the component in water or solvent, perhaps agitating the part, sometimes refluxing, and in some instances using ultrasonic cleaning. Some factors associated with designing an extraction process are listed in table 1.

Extraction is often assumed to be different from cleaning. Many process designers separate and compartmentalise the extraction process and associated analysis from the cleaning, to the extent that they are considered completely separate activities. While extraction and cleaning have different functions, operationally they are very similar processes. Soil is matter out of place. Critical cleaning is removal of matter out of place. Extractive techniques remove soil for further quantification or analysis. Therefore, one would want to avoid destroying or modifying the soil during extraction.

It is often assumed that water or hexane are the only appropriate extraction solvents. This is because water is very polar and hexane is very non-polar. However, using the principle “like dissolves like” is appropriate for both cleaning and extraction.

Since particles are often embedded in a thin-film matrix, it is often appropriate to select from among other solvents with different levels and ratios for the polar, dispersive (non-polar), and hydrogen bonding properties.3 Just as in developing cleaning processes, materials compatibility must also be considered. For example, methylene chloride is an extremely aggressive solvent that, if resorted to, must be used with consideration of employee safety and environmental quality issues.

Because of the broad solvency range, it will dissolve an array of residue and therefore release entrapped particulates. At the same time, it may dissolve materials of construction such as epoxies. If particles are embedded in the materials of construction, they will also be released, perhaps leading to falsely elevated particulate counts.

As with cleaning processes, the chemical, time, temperature and force must be specified, and it must be suitable to the component. One has to achieve a balance: at one extreme, it is important to extract a representative sample of the particles; at the other, the extraction process can damage the product and generate particles or break down particles to smaller sizes, resulting in artifactual findings. The problem can be exacerbated in extraction processes where delicate parts are subjected to excessive ultrasonic cleaning.4

It is difficult if not impossible to establish extraction of all possible particles from a complex part (or of all possible thin film residue, for that matter). In my experience, repeated extraction under a given set of conditions often continues to yield additional particles. Extracting with a variety of solvents and using a variety of conditions sequentially generates different populations of particles. After a certain point, a pragmatic judgment must be made, based on expected properties of the residue. Develop a reasonable, consistent, supportable, well-documented extraction process and do not deviate from it without very good justification.

The ideal way to minimise particles on the product would be to have all fabrication under well-controlled environments. However, this is not practical for many critical products. Optimising the critical and non-critical cleaning processes as well as the process chemicals and process flow will minimise surface particles. But surface particles can occur, even if the product is fabricated in very controlled environments using validated cleaning processes.

Certainly, monitor the air, the water, and the process chemistries as well as the product itself. Achieving appropriate surface quality involves particle monitoring; but it also involves an array of additional factors. Some, but by no means all, factors are indicated in table 2.

Monitoring is important; but no single monitoring step guarantees a superior product. Monitoring particulates in air and on product surfaces is crucial, but it is not sufficient. We have to carefully evaluate all steps in product fabrication; and then monitor appropriately.