What is NFC and how does it work?

What NFC is

Near field communication is a set of standards for exchanging data between two devices held within a few centimetres of each other, using a magnetic field oscillating at 13.56 MHz. The NFC Forum, the industry body that maintains the specifications, defines both the radio layer and the data format that runs on top of it.

The name is precise and worth taking literally. "Near field" is not a marketing way of saying "short range": it describes the region close to an antenna where the magnetic component of the field dominates and energy can be transferred by induction rather than radiated as a wave. Everything characteristic about NFC follows from operating in that region.

Two things are usually involved. One is powered and does the talking, typically a phone or a dedicated reader. The other is a tag, a card, or a second powered device.

How a phone powers a tag with no battery

An NFC tag is a chip connected to a flat coil of wire, usually printed as a few turns of aluminium or copper on a plastic film. There is no battery, no capacitor holding a charge, and nothing running when the tag sits on a table.

When a reader is brought close, its own coil is driven with an alternating current at 13.56 MHz. That produces an alternating magnetic field. The tag's coil sits inside this field and a voltage is induced across it, exactly as in a transformer whose two windings happen to be separated by air instead of an iron core. Rectified, that voltage powers the chip.

The reply direction is the part most descriptions skip. The tag does not generate a signal. It changes how much energy it absorbs from the reader's field by switching a resistance across its coil, and the reader senses that as a fluctuation in its own antenna current. This technique is called load modulation, and it is why the tag can answer while consuming almost nothing.

Data rates defined at this layer are 106, 212 and 424 kbit/s. That is enough for a URL, a Wi-Fi credential or a payment exchange, and hopeless for anything resembling a file.

Why the range is only a few centimetres

Inductive coupling in the near field weakens extremely fast as the two coils separate. Doubling the distance does not halve the coupling, it collapses it. Past a certain point the tag simply never receives enough energy to boot its chip, and the failure is abrupt rather than gradual: the tag either answers or it does not.

Antenna size matters more than the phone. A credit card sized tag has a large coil that intercepts much more of the field than a 25 mm sticker, so it reads from noticeably further away with the same phone. If a tag reads poorly, the first thing to change is the tag, not the handset.

The second variable is where the phone's antenna actually is. It occupies a specific area on the back of the handset, and it is not in the same place across manufacturers. A phone held flat against a tag but offset by three centimetres will fail as reliably as one held ten centimetres away.

The three operating modes

The standards define three ways two NFC devices can relate to each other. Only the first two are what most people encounter.

Reader and writer mode is the one that concerns anyone buying tags. Card emulation is handled by the operating system and the secure element, and applications reach it through a host-based interface on Android that Google documents under host card emulation.

Peer to peer is the odd one out. It is part of the standard, but phone to phone transfer over NFC is not how either mobile platform shares files today.

Android deprecated Android Beam in Android 10 and documents it as no longer in active development, remaining optionally available to device makers.

The standards behind NFC

NFC sits on top of contactless card standards that predate it. ISO/IEC 14443 defines the proximity card interface that most tags and payment cards use, ISO/IEC 18092 defines the NFC interface and protocol itself, and ISO/IEC 15693 covers the longer range vicinity cards that the NFC Forum later adopted as one of its tag types.

On top of that radio layer the NFC Forum defines two things that matter in daily use. The first is a set of tag types, each describing the command set a compliant reader must support. The second is NDEF, the NFC Data Exchange Format, a small container format that lets any compliant device understand a tag written by any other.

The current specification texts are published by the NFC Forum.

What NFC is bad at

Metal is the main one. A tag placed flat on a metal surface has its field distorted and its resonance detuned, and it usually stops working entirely rather than degrading. Tags sold as suitable for metal include a ferrite layer that separates the antenna from the surface, and they cost more for that reason. An ordinary sticker will not work on a laptop lid, a machine housing or the back of a phone.

Distance is the second. If the use case requires reading a tag from more than a hand's width away, NFC is the wrong technology and no amount of antenna tuning will change that. UHF RFID at around 900 MHz reads from several metres, which is precisely why warehouses use it and phones cannot talk to it.

Capacity is the third. Tag memory is measured in hundreds of bytes for the common consumer references, which comfortably holds a URL and nothing resembling a document or an image. The usual answer is to store a link and put the payload on a server.

There is also a failure mode that catches people out for a different reason. Most tags can be permanently locked, and a locked tag cannot be rewritten by any means. It is a one-way operation performed in software, with no physical trace, and it is easy to trigger by accident when a writing application offers it as a checkbox next to the write button.