The little metal cans inside your phone are not heat sinks. Every circuit is an accidental radio, and they are the answer

Open a phone or a laptop and you find small rectangular metal cans soldered flat onto the circuit board, with more foil and metallised tape around the edges. They are often mistaken for heat sinks. They are not — they are usually poor at moving heat and are sometimes in the way of it.

They are there because of an awkward fact about electronics: any circuit fast enough to be useful is also a radio transmitter, and a radio receiver, whether or not anybody wanted it to be. A clock line switching sharply radiates. A power converter switching a few amperes radiates rather more. And the same traces that radiate will also pick up whatever else is radiating nearby, including the phone's own transmitter sitting centimetres away.

Shielding is the business of containing that. And its most useful lesson is the same one that turned up in our article on masks: the performance of the system is decided by the leaks, not by the quality of the material.

Why circuits radiate in the first place

A steady current does not radiate. A changing current does, and modern electronics changes current very abruptly on purpose, because sharp switching edges are what make a digital signal unambiguous and a power converter efficient.

The consequence is that the frequencies present in a circuit go far beyond the frequency it appears to run at. A square wave contains its fundamental plus a long series of harmonics, and the sharper the edges, the more energy is carried in the high ones. A board with a 100 MHz clock and fast edges has real energy at several gigahertz — which is precisely the band where the same device's Wi-Fi and cellular receivers are trying to hear signals a billion times weaker.

Any conductor of the right length then behaves as an antenna: a trace, a cable, a heat sink, a poorly grounded shield. Cables are usually the worst offenders, because they are long, and length is what an antenna wants.

This produces two separate obligations. Emissions — not interfering with everyone else — is a regulatory requirement, and products are tested against limits before they can be sold. Immunity — not being disrupted by others — is what stops a device misbehaving next to a motor, a lift or a transmitter. Together they are electromagnetic compatibility, and a product that fails either does not ship.

Shielding is reflection first, absorption second

The intuitive picture of a shield is a wall that blocks. The physics is a little different and considerably more useful.

When an electromagnetic wave travelling in air meets a conductive surface, it encounters an abrupt change in impedance — free space is about 377 ohms, a metal is a tiny fraction of that. Most of the wave is reflected, for the same reason light reflects off a mirror and a rope wave bounces back from a knot. For good conductors at most frequencies, reflection does the majority of the work.

Whatever does get in is then absorbed, decaying exponentially as it penetrates. The distance over which the field falls to about a third of its value is the skin depth, and it shrinks as frequency rises. In copper it is a couple of millimetres at mains-adjacent frequencies, tens of micrometres around a megahertz, and roughly two micrometres at a gigahertz.

That single number explains a great deal. At gigahertz frequencies, a shield only a few micrometres thick is already many skin depths and attenuates enormously — which is why metallised plastic, thin foil, and vacuum-deposited coatings work perfectly well for high-frequency shielding, and why the metal can over a chip does not need to be thick. It also explains the opposite case: at low frequencies the skin depth is large, so a thin conductive shield does almost nothing, and low-frequency shielding is a genuinely harder problem.

Shielding effectiveness is quoted in decibels, and as with every logarithmic scale on this site it compresses badly at the top: 20 dB blocks 99% of the power, 40 dB blocks 99.99%, 60 dB blocks 99.9999%. Most consumer equipment needs somewhere in the 30 to 60 dB range; going further costs disproportionately more for benefits nobody needs.

The holes decide everything

Here is the part that matters most in practice and surprises people who think about shields as walls.

A shield's effectiveness is almost never limited by its material. It is limited by its apertures — seams, ventilation holes, connector openings, display cutouts, the gap where a lid meets a wall.

A slot in a conductive surface behaves as a slot antenna. It radiates efficiently when its longest dimension approaches half a wavelength, and its leakage depends overwhelmingly on that longest dimension rather than on its area. This produces a rule that sounds wrong and is right: many small holes leak far less than one large hole of the same total open area. A ventilation panel of a hundred small perforations can be nearly transparent to air and nearly opaque to a gigahertz field, while a single long slot of identical open area leaks badly.

It also means the failure in a real product is usually a seam. Two shield halves that touch only at their screw points leave long unbonded gaps between them, and those gaps are slots. This is why EMC design is full of things that look like overkill — conductive gaskets, dense rows of stitching vias along a board edge, finger stock around a lid, tight screw spacing. All of it exists to shorten the longest unbonded dimension.

A shield is only as good as its worst gap, and a gap that is electrically long is an antenna regardless of how good the metal around it is. The material is the easy part; the joints are the engineering.

Grounding is a related trap. A shield that is not properly bonded is a floating conductor, which can couple energy from one place to another and make matters worse than no shield at all. And a cable leaving a shielded enclosure carries the interference straight out through the wall unless it is filtered or its screen is bonded at the point of entry — the most common way a well-shielded box still fails its emissions test.

Low frequency is the hard case

Everything above concerns electric fields and plane waves, where conductivity is what matters. Low-frequency magnetic fields are a different problem and a much more stubborn one.

Magnetic fields at, say, mains frequency pass through copper and aluminium as though they were not there, because the skin depth is enormous and there is no useful reflection. Shielding them requires diverting the flux rather than reflecting it, which means a material of high magnetic permeability — mu-metal and similar nickel-iron alloys — that offers the field an easier path and guides it around the protected volume.

There is a lovely practical detail here. Mu-metal's permeability is destroyed by mechanical stress: bend or machine it and its performance collapses. Formed parts must be annealed afterwards in a controlled atmosphere to restore the property, which makes mu-metal shields expensive and rules out treating them as ordinary sheet metal. It is a good reminder that a material property can be a property of the material's history rather than of its composition.

This is why a transformer humming into an audio circuit is a harder problem than a gigahertz emission, and why sensitive instruments are sometimes surrounded by both a conductive shield and a permeable one.

Where nanomaterials genuinely help

A stamped metal can is cheap, effective and thoroughly solved. So the case for a new material has to come from somewhere the can does not work: weight, thickness, flexibility, conformality, or the need to absorb rather than reflect.

That last one is worth explaining, because it is the strongest technical argument in this area. Reflection does not destroy energy; it sends it somewhere else. Inside a dense assembly, a reflective shield can bounce interference onto a neighbouring circuit, and in an enclosure it can set up standing waves that make some spots worse. What is often wanted is a shield that absorbs — converting the field into heat inside the material rather than returning it.

Nanostructured materials are genuinely good at this, and for a structural reason. A conductive foam, aerogel or layered composite gives a wave many internal surfaces to scatter from, so it makes multiple passes through lossy material rather than one. Carbon nanotube and graphene composites and foams are used precisely this way, with the porosity doing as much work as the conductivity.

MXenes are the standout result. Solution-processed films of titanium carbide MXene combine metallic conductivity with a layered structure, and thin films of them deliver shielding effectiveness that is exceptional for their thickness — this is comfortably the strongest real application the material has, and it is why our earlier article on MXenes listed EMI shielding first among their uses. Films a few micrometres thick, spray-coated or dip-coated onto flexible substrates, do a job that would otherwise need a rigid metal enclosure.

The honest caveats are the ones this site keeps arriving at. Much of the literature reports shielding effectiveness normalised by thickness or density, which flatters very thin films and is a materials comparison rather than a product specification — the same trap as farads per gram in supercapacitor research. MXenes oxidise in humid air unless stabilised, which is a real and documented durability problem for a component expected to last a decade. Adhesion, abrasion and flex-cycling performance are less reported than shielding numbers. And the benchmark is not another laboratory material; it is a stamped steel can that costs a few paise.

The applications where this genuinely wins are the ones where a can cannot go: wearables and flexible circuits, conformal coatings on irregular shapes, aerospace where mass is expensive, and thin absorptive layers inside crowded assemblies.

Why it matters for students and researchers

Electromagnetic compatibility is a discipline most electronics graduates first encounter when a finished product fails certification, at the point where fixing it is most expensive. Learning it early changes design decisions that cost nothing at the schematic stage and a fortune afterwards.

The transferable lesson is the aperture rule, and it generalises well beyond electronics. In a shield, a mask, a cleanroom or a firewall, the performance of the system is set by the largest gap rather than by the average quality of the barrier — and effort spent improving an already-adequate material while ignoring a seam is effort wasted. Anyone who has internalised that will look for the gap first, which is usually where the answer is.

The open problems are practical rather than fundamental. Absorption-dominant shields that are thin, light and manufacturable would solve real problems that reflective shields cannot. Environmental stability of solution-processed conductive films — oxidation, humidity, flex fatigue — is the barrier between impressive laboratory numbers and shipped components. Shielding for millimetre-wave bands brings tighter tolerances, because as wavelength shrinks, so does the size of gap that leaks. And standardised, device-level reporting of shielding performance would make the literature far easier to translate into engineering.

Frequently asked questions

Does a phone case block the signal?

An ordinary plastic or leather case does not, because those materials are transparent to radio at these frequencies. A case with a metal back, or a metallised layer, genuinely does attenuate the signal, and the phone responds by transmitting at higher power to maintain the link — so the practical effect is worse battery life rather than a dropped call. Wallet cases holding cards against the phone can also interfere with the near-field antenna used for payments.

Do EMF-blocking pendants, stickers and phone shields work?

A properly constructed conductive enclosure genuinely does attenuate radio fields — that part is ordinary physics, and it is why a microwave oven has a mesh in its door. A small sticker or pendant does not enclose anything and cannot attenuate a field arriving from every direction; there is no mechanism by which it could. Worth noting too that anything which does partially block a phone's signal makes the phone transmit harder, which is the opposite of the stated intention.

Why does a microwave oven door have a metal mesh with visible holes?

Because the holes are far smaller than the wavelength being contained. Microwave ovens operate near 2.45 GHz, a wavelength of about 12 centimetres, and holes of a few millimetres are so far below the half-wavelength that leaks efficiently that they are effectively opaque to it. Visible light has a wavelength of a fraction of a micrometre, so it passes through the same holes easily — which is why you can see in while the microwaves stay put.

Why did my speaker buzz just before my phone rang?

That is a classic case of insufficient immunity rather than emissions. Older cellular protocols transmitted in rapid bursts, and an audio input with inadequate filtering rectifies that burst envelope into an audible buzz at the burst repetition rate. The phone was operating normally; the speaker's input stage was acting as an unintended radio detector.

Can I use aluminium foil as a shield?

For a quick diagnostic experiment, yes — wrapping a suspect cable or module in foil and seeing whether a problem changes is a legitimate troubleshooting step. As a permanent fix it usually disappoints, because a shield must be bonded and grounded properly rather than merely present, unbonded foil can make coupling worse, and the seams and the exit path of any cable are where the performance actually goes.