Magnetic shielding and the distortion of magnetic fields Magnetic Shielding When electromagnetic fields affect sensitive equipment, shielding is needed in order to lessen or negate the effect. Magnetic shielding works by redirecting the propagation of magnetic field lines so that magnetic flux flows through the walls of the shield itself, resulting in the by-passing of the enclosed area or object needing protection. In order for the shield to successfully provide this alternative pathway, there must be a route through the shield that offers less resistance than travelling through at an angle perpendicular to the surface. This is achieved through two approaches; using a high permeability material and using an appropriate geometric design [1,2]. Permeability is a measure of how easily a magnetic field is established in a material. This means a higher permeability material is more conductive to a magnetic field. Mumetal and Supra50 are commonly used for shielding due to this property. Fig.1: A diagram depicting field distortion. The chosen material must also have a suitable saturation point. The saturation point is the applied field strength after which the material cannot be magnetised further, regardless of a continued increase in external field (i.e. there is no further alignment of electrons within the material). After saturation, no further magnetic flux can be absorbed by the shield, rendering it ineffective. Fig. 2: Two Pylab intensity maps produced from results taken at MSL. The first depicts a Mumetal can distorting the field of a 0.05T magnet. The second shows the same can become saturated by the field of a stronger 0.1T magnet. As Mumetal has a relatively low saturation point, it is often combined with other materials in multi-layer shields if a stronger magnetic field is in effect. Geometry also plays a large part in how effective a shield is. Whilst hollow spheres provide the best shielding properties, they are impractical for most uses, as well as being very difficult to make. Closed cylinders are the next best design, as they limit the flux leakage observed at corners, sharp edges and joins. For a cuboidal shield, the losses are due to providing a difficult path for flux to flow along; any wall that is perpendicular to the external field will contribute a small amount to the overall effectiveness of the shield, with most of the attenuation occurring along the parallel sides. Fig.3: A diagram depicting the field distortion observed for a shielding cube. If the shield is of high permeability, it is observed that flux that has leaked through a perpendicular side may attenuate again within the shield.
The thickness and length of the shield is also an important factor in optimising shielding properties. Upon testing, the following results on shielding a 0.1 Tesla neodymium magnet were obtained describing the effect of shield dimensions on field attenuation (shielding factor): Magnetic Field Theory Magnetic fields are produced by electric currents [3]. At any given point, a magnetic field has a vector denoting the direction and strength of the field in that location. This field will exert a force on a charged object within range, defined by the charge Q producing the field, its velocity, and the size of the magnetic field (B): F Q( v B) This force is the cause of the observed attraction or repulsion between two magnets. An important feature of magnetic force is that it does no work; this can be proven by considering the vector identity (a x b) a = 0 and that work W=force distance moved. If charge Q moves a distance l: W F l Q.( v B) l Q.( v B) v. t 0 Fig. 4: Graph showing the effect of shield dimensions on shielding factor. An increase in both length and thickness is shown to improve the performance of the shield; this could be predicted from the approximate theory for the shielding factor of a cylindrical shield: d 4N D S 1 D 1 2L S=shielding factor d=thickness N=demagnetisation factor L=length of cylinder D=inner diameter Increasing both d and L results in a greater S. It has been suggested from theory that for good shielding properties the optimum dimensions for a cylindrical shield is a ratio of 1:10 for D:L; other ratios have proven as effective in practise. Magnetic shielding is best suited to low frequency external fields, as it is the only form of attenuation for fields of frequency <100kHz. As frequency increases, eddy currents induced by the external field take over as the primary shielding effect, producing a magnetic counter field that opposes the applied field. At these frequencies, permeability plays a small role; a conductive metal such as copper can then be used instead for suitable electromagnetic shielding. As magnetic force does no work, energy is conserved within the magnetic field and cannot be extracted as a form of free energy. Phenomenon such as a magnet lifting a paper clip is instead a result of the magnetic force redirecting other forces; in this example, the magnetic field converts kinetic energy at a subatomic level into potential energy that results in the attraction. Another important aspect of magnetic theory is magnetic flux. This is defined as the component of the field B passing through an area A. After rigorous testing for hundreds of years, it has been observed many times that there is no net flux surrounding a magnetic object. This means that all magnetic objects are dipoles, creating a closed loop of flux around itself departing from one pole and returning to the other. This phenomenon can be observed by placing a bar magnet near iron fillings: Fig.5: Iron filings following the magnetic field lines of a bar
This is also defined by Gauss s law of magnetism, stating that as there is no net flux through a closed surface, there are no magnetic monopoles: 3-axil Hall probe. These values were taken at regular intervals using 2D gridded paper as a reference. B d A 0 Recently monopole-like phenomena has been observed in spin ice experiments [4]. This is due to the fractionalization of particles resulting in quasiparticles that are monopole of the H field, an auxiliary field that describes an aspect of B. As this is still not a monopole in B nor an example of an elementary monopole existing in a vacuum, Gauss s Law still stands. Isolating magnetic poles Beyond shielding sensitive equipment from the effects of external magnetic fields, there are several devices or ideas that make use of the ability of magnetic shields to manipulate the external field into a new shape by redirecting magnetic flux. A popular example is the idea of shielding one magnetic pole from another or keeping only one pole shielded, thereby effectively separating the two poles from each other. Such a creation could be used as a form of magnetic switch, or could even shed light on the Grand Unified Theory. As described previously, it is currently impossible to completely separate two poles as there have been no observed magnetic monopoles beyond the level of quasiparticles. However, research into the extent of attenuation in such cases has been conducted in order to depict how useful this field redirection can be in such an application. Several open cylindrical shields of different dimensions were first used in order to investigate the extent of shielding either side of a long magnet and the difference in field strength observed at either end. The graph, produced in Python, depicts the intensity of the magnetic field in a 2D plane of the shield that displayed the strongest shielding properties (one of dimensions 40x60x1.5mm). The field strength is portrayed as a logarithmic scale of the actual values for the sake of clearer colour differentiation, and the field values were measured using a Fig. 6: Colour map showing the field shape of a shielded 0.1T bar magnet. It is worth noting that the slight leakage shown on the bottom of the shield could be explained by damage possibly being obtained during its creation; the asymmetry suggests it is not an intrinsic property of the shield. However, despite this, a reasonable amount of leakage is shown regardless of any damage, with a large radial field being observed at either end of the shield. This describes only a limited field channeling effect from the shield, and certainly does not prevent field lines flowing from one pole to the other. After this setup was measured, brief tests with the Hall probe also showed an increase in the field of the shield itself at an average of ~5%. This suggests that the shield was slightly magnetized by the close exposure in the experiment. Whilst experimentation suggested completely shielding each pole from the other was impossible, practical application could still be found if it were possible to shield the effects of just one magnetic pole whilst leaving the other unchanged. This would involve creating a shield resembling an open box in which only one end of a bar magnet would reside, allowing the other to be relatively unaffected by the shield. To reduce the possibility of a strong field affecting the shields again, the magnet strength used in this part of experimentation was reduced to ~0.05T.
The investigation undertaken used a cylindrical Mumetal shield of dimensions 35x26x1.5mm, closed at one end, which again required the Hall probe to measure the field strength. Observations are similarly displayed in Python: basic function of a switch designed to alternate between shielding and exposing two magnets). This would require an energy input into the system that could use mechanical energy to turn the shield. Further investigation was put into the idea of increasing the spacing between magnets as a way of creating a more effective shielding effect. The space was increased from 5.6cm to 9.8cm to observe the effect: Fig. 7: Map of the field of two unshielded 0.05T magnets 5.6cm apart. Fig. 9: Map of the field of two unshielded 0.05T magnets 9.8cm apart. Fig. 8: Map of the field of a pole-shielded 0.05T magnet 5.6cm from an unshielded magnet. The results displayed show that the obvious attraction between the shield itself and the magnet is almost as great as between the unshielded magnets. Like any other ferrous material, Mumetal is always attracted to the source of the field. This makes shielding unhelpful for devices in which no interaction from the shielded pole is desired. The force between the magnet and the shield must also be overcome if the exposed side is to be directed towards the magnet (i.e. the Fig. 10: Map of the field of a pole-shielded 0.05T magnet 9.8cm from an unshielded magnet. As predicted, there is less of an attraction between the shield and magnet at a greater distance. However, this drop in field strength closely
mimics that of the separated, unshielded magnets. This suggests that the reduced potential between the shield and magnet is mainly down to the way the magnetic field strength falls away with distance, rather than a better shielding effect; again, the potential energy between the shield and the magnet is not much less than when both magnets are exposed. From results obtained at Magnetic Shields Ltd., it can be said that a magnetic shield is still not a suitable method of completely shielding one pole of a magnet from the other. It also stands that magnetic monopoles are non-existent, and that shielding is again not a solution. Despite such limitations, the useful role of Mumetal shielding for applications in which field distortion or confinement is necessary cannot be stressed enough. Figure 4 depicts some typical shielding factors obtained by smaller shields containing strong magnets, whilst some larger shields that contain weaker fields reach factors many times this magnitude. References [1] T. Rikitake, Magnetic and Electromagnetic Shielding, Springer, 1987. [2] K. L. Kaiser, Electromagnetic Shielding, CRC Press, 2005. [3] D. J. Griffiths, Introduction to Electrodynamics (3rd Edition), Addison Wesley, 1998. [4] D. J. P. Morris, D. A. Tennant, S. A. Grigera, et al, 2009, Science, vol. 326 no. 5951.