Introduction

Metal particles contamination is a recurring issue in pharmaceutical manufacturing, especially when dealing with powders, as associated manufacturing equipment often involves mills or other mechanical equipment that may get damaged in time, and magnetic filtration is a simple solution that can be applied cost effectively to most processes.

Magnetism is a phenomenon known since the antiquities (natural magnetite ore) and magnets (including electromagnets) have been studied by modern science since the XIX century. Ferromagnetic materials (such as iron, nickel, cobalt and some rare earth metals) are in fact strongly affected by magnetic fields while paramagnetic materials (such as aluminum, tin magnesium) show a much weaker attraction.

Permanent magnets (AlNiCo) have been manufactured industrially since the 1930s, but common industrial magnets manufactured at that times were able to attract just ferromagnetic materials, while most common metal contaminants in pharma are paramagnetic materials, so they were mostly unsuitable for applications in pharma.

Samarium Cobalt magnets, that were developed in the 1960s, were too expensive and delicate for many industrial applications. Finally, in the 1980s, Neodymium Iron Boron magnets (a.k.a. ‘high intensity’ magnets) were developed for industrial applications, effectively opening the door for this technology in many industrial pharmaceutical applications.

The principles of magnetic filtration

Magnetic separation is a physical separation of discrete particles with different susceptibility to a magnetic field, such as minutes pieces of (paramagnetic or ferromagnetic) metal in a (non magnetic) bulk powder. This kind of contamination can be difficult to detect early in the process but can result in whole batches being recalled. For this reason magnetic filtration can be considered a general, ‘end of pipe’ good practice, especially when dealing with bulk powders.

Magnets are in fact used in almost every processing industry to remove ferromagnetic and paramagnetic particles, as they can attract and remove particles continuously without having to stop the process.

The strength of a magnet is usually measured in two ways (that can of course be correlated by calculation): the pull test, that measures the force required to remove a piece of metal from the magnet, and the use of a Gauss meter allowing to read the value of the magnet field.

A ferrite magnet produces about 3,000 Gauss (typically adequate to attract ferromagnetic materials), while a high intensity magnet can produce about 13,500 Gauss (suitable for paramagnetic materials as well). Therefore, every metal can be measured.

Magnetic properties of stainless steel in process conditions

Some stainless steels are magnetic and some are not. Ferritic and martensitic stainless steels are magnetic due to their iron composition and molecular structure. In its natural state, austenitic stainless steel (such as stainless steel 316, widely used in pharmaceutical productions) is non magnetic.

This is due to the presence of nickel which alters the physical structure of stainless steel and removes or inhibits any magnetic qualities. This may be considered an indication that magnetic filtration is not suitable for this kind of metal contamination.

However, when a piece of non magnetic stainless steel breaks off from the line, such as a blade, valve or line wear, its atomic lattice strains, forming martensite, that is magnetic and can be removed by magnetic filtration. For this reason, magnetic filters are considered effective to remove stainless steel contamination as well.

Evaluating ‘unknown’ alloys

Most of the times, the metal contamination that can be realistically found in pharmaceutical powders is easily identifiable and the number of gauss required to deal with it can be easily found in the literature. However, if a specific alloy is not yet in the literature, the best course of action is simply to send a coupon of that metal to the magnetic filter supplier, that will grind part of it and will test the shards against various magnets, until a suitable strength is found.

For example, Hastelloy is known as a paramagnetic metal, but there are various grades of Hastelloy and literature doesn’t necessarily reports them all, so it may be difficult to identify exactly which magnet is suitable for removing, for example, Hastelloy C22. In this case, a coupon of Hastelloy C22 can be ground and shards of the sample are tested for paramagnetism using various high intensity rod magnets.

It is found that the shards possess enough paramagnetism to be firmly caught and held by a 9,000 Gauss rod. See Image 2.

These results prove that Hastelloy C22 has enough paramagnetism to be used with a high intensity rod and could therefore be removed during magnetic filtration if a magnetic field greater than 9,000 gauss is present.

Of course, the same procedure can be followed for any other material.

Corrosion

The Neodymium Iron Boron magnet itself has a poor resistance to corrosion and therefore it is better not to keep it in direct contact with the process. This is also true for Clean In Place treatments that typically require alkaline or acid detergents.

For this reason, the equipment which houses the magnet and that is in contact with the process is often stainless steel, but it may be coated with an organic coating like PTFE in specific corrosion conditions. Of course, distancing the magnet from the process material decreases the magnetic flux density, so it is recommended having the minimum distance and protection required.

It is critical that the arrangements to separate the magnets from direct contact with the process material do not cause isolated conductors within the separator.  If an arrangement with a ‘magnetic door’ is to be used it must be bonded by means of a conducting wire or strip to eliminate any difference in potential between the door and the housing unit. The housing unit itself must be bonded to ensure static can flow between all conductive parts.

Temperature Effects

Pierre Curie, between the XIX and XX centuries, was one of the most important scientists that studied magnetism and he noticed that temperatures can have a strong effect on the magnets. When magnets pass some time at certain temperatures, their performances may change and may decrease.

The change is in general reversible up to a certain temperature and beyond it the change becomes irreversible until an even higher temperature is reached which removes permanently all magnetic properties (Curie temperature, which is dependent on the specific magnet composition, particle size, pressure and orbital ordering).

Under ambient temperatures, in fact, the atoms in a magnet “align” between the two poles, fostering magnetism. When exposed to heat, however, these atoms begin to fluctuate locally faster and more irregularly.

This movement “misaligns” the atoms, causing magnetism to be lost. Below certain temperatures, instead, the atoms in magnets move progressively slower and less randomly, “aligning” the atoms in a more stable way, at first strengthening the magnetism, then, at even lower temperatures, spin reorientation once again decreases the magnet’s strength.

For most pharmaceutical applications with Neodymium Iron Boron magnets, the effect of low temperature is unlikely to be relevant, as partial, reversible demagnetisation begins below -138°C, which is an uncommon processing condition in mainstream pharma.

Partial, permanent demagnetisation due to high temperature instead can begin above 80°C, that is more likely to happen as part of pharmaceutical manufacturing, while the Curie temperature (above which all magnetisation is lost) is around 320°C.

As previously mentioned, the temperatures that affect magnetic materials are not exactly fixed and depend on various fine parameters that can be controlled during magnets manufacturing, so the exact temperature thresholds have to be double checked with the supplier.

It is important that the normal operating range lies comfortably below the maximum operating temperature to ensure that any partial demagnetisation is reversible. Above the maximum operating temperature partial demagnetisation will be irreversible and above the Curie temperature the demagnetisation will be total and irreversible.

Cleaning

The unit must be regularly cleaned to remove the metal which has built up on the surface of the magnets. Cleaning can be carried out manually or, more elegantly, using a fully-automated system, that moves the magnets outwards to remove the magnetic field from the housing; the captured contamination is then released and can be discharged away from the product flow (see Image 3).

Magnetic filter arrangements

Identifying the right position and the right design for a specific magnetic filter requires knowledge of the process and experience, as there are many possible arrangements for many possible process conditions. It is in fact critical identifying the right situation where the pull of the magnetic force will be prevalent on other forces. For example, if we have a long chute, it will be better having a magnetic filter near the top, when the speed of the particles is usually lower, rather than near the bottom, where the acceleration of the fall may prevail over magnetic attraction.

If the product is not too corrosive and abrasive, grids are often preferred, typically on two layers, to maximize the contact with the powder (see Image 4).

Another common arrangement, particularly useful for thinner pipes and pressure / vacuum driven transfers, is “bullet” or in line magnet.

In worse corrosion and abrasion conditions it is more common having chute arrangements, that offer better protection to the magnets (see Image 6).

Conclusion

Magnetic filters are a suitable and flexible solution for cost-effectively removing metal particles from the process in many pharma applications, especially when bulk powders are handled, as some of the machines involved in powders production are at risk of causing metal particles contamination.

There are many different possible arrangements and design for deploying this technology and, in order to identify the most suitable magnetic filtration design, it is necessary having good knowledge of the process and good experience with magnetic filtration.