Electrons carry more than charge
Ordinary electronics moves electrons around and reads out their charge. Every electron also carries spin — an intrinsic quantum angular momentum that points either "up" or "down" relative to some reference direction — and spintronics is the discipline of building devices that exploit spin as an additional information channel, not just charge. The founding, and still the most commercially important, spintronic effect is giant magnetoresistance (GMR), discovered independently in 1988 by Albert Fert and Peter Grünberg, work that won them the 2007 Nobel Prize in Physics.
The sandwich: ferromagnet / spacer / ferromagnet
A GMR device is a thin-film stack: a ferromagnetic layer (commonly cobalt or a cobalt-iron alloy), a nonmagnetic metallic spacer (commonly copper) only a few nanometres thick, and a second ferromagnetic layer. One layer's magnetisation is fixed (pinned to an adjacent antiferromagnet), the other is free to rotate with an external field — that structure is called a spin valve. Current flows perpendicular to the layers (or, in an earlier geometry, in-plane), and the whole point of the device is that its electrical resistance depends measurably on whether the two magnetic layers are pointing the same way (parallel) or opposite ways (antiparallel).
Spin-dependent scattering: the mechanism
Inside a ferromagnet, electrons whose spin is aligned with the local magnetisation scatter off impurities and grain boundaries much less than electrons whose spin is anti-aligned — a direct consequence of the exchange splitting between spin-up and spin-down conduction bands in a ferromagnetic metal. Because the spacer is thin enough that electrons cross it without flipping their spin, the two-current picture (spin-up and spin-down currents flowing largely independently, first proposed by Mott in 1936) explains the whole effect:
PARALLEL layers (same direction): spin-up electrons → low scattering in BOTH layers → low-resistance channel spin-down electrons → high scattering in BOTH layers → high-resistance channel → the low-resistance channel short-circuits the stack → LOW total resistance ANTIPARALLEL layers (opposite directions): every electron is "wrong-aligned" in one layer or the other → BOTH spin channels see one low-scattering and one high-scattering layer → no low-resistance shortcut exists → HIGH total resistance GMR ratio = (R_antiparallel - R_parallel) / R_parallel — typically 10-200%
In the parallel state, one spin channel gets an easy path through both layers and dominates the current, acting like a resistor short-circuited by a much lower-resistance path. In the antiparallel state, every electron, regardless of spin, is forced through at least one high-scattering layer, so there is no easy channel and the total resistance is significantly higher. Flip the free layer's magnetisation with a small external field and the resistance visibly jumps — that jump, reliably read out, is the working principle of every hard drive read head made since the late 1990s.
From physics to petabytes
A hard disk stores bits as tiny regions of magnetisation on a spinning platter. Reading them means detecting a very small, very local magnetic field as the head flies over each bit — and GMR sensors are exquisitely sensitive to exactly that. IBM shipped the first GMR read head in 1997, less than a decade after the effect's discovery, and the improved sensitivity let engineers shrink the magnetic bits dramatically while still reading them reliably, driving roughly a thousand-fold increase in storage density over the following decade — one of the fastest transitions from a fundamental physics discovery in a condensed-matter lab to a technology inside essentially every computer on the planet.
Beyond GMR: tunnelling and spin torque
GMR's direct successor, tunnelling magnetoresistance (TMR), replaces the metallic spacer with a thin insulating barrier (commonly magnesium oxide) that electrons cross by quantum tunnelling rather than diffusive transport; the tunnelling probability is even more strongly spin-dependent, giving TMR ratios that can exceed 200%, and TMR-based magnetic tunnel junctions are now the standard read-head technology and the storage element in MRAM (magnetoresistive RAM). A related and more recent effect, spin-transfer torque, runs the physics in reverse: a spin-polarised current can itself exert enough torque to flip a nanomagnet's orientation, which is how modern MRAM writes bits — using spin, not just reading it, to store information directly.
Frequently asked questions
What is the difference between GMR and ordinary magnetoresistance?
Ordinary magnetoresistance, present in any conductor, is a small effect from the Lorentz force bending electron paths in a magnetic field, typically changing resistance by well under 1%. GMR is a quantum spin-dependent scattering effect in layered magnetic thin films that can change resistance by 10-200% depending on whether neighbouring magnetic layers are aligned or anti-aligned, which is why it earned the name giant.
Why does spacer thickness matter so much in a GMR stack?
The spacer must be thin enough, typically a few nanometres, for electrons to cross it without losing their spin orientation through scattering, and it also mediates an interlayer coupling (RKKY interaction) between the two magnetic layers whose sign oscillates with spacer thickness. Get the thickness wrong and you either destroy the spin memory needed for the effect or accidentally lock the layers into parallel alignment instead of the antiparallel default state the device needs.
Why did GMR win the Nobel Prize?
Albert Fert and Peter Grünberg shared the 2007 Nobel Prize for discovering GMR independently in 1988, because it was rapidly and directly applied — IBM shipped the first GMR read head in 1997 — enabling roughly a thousand-fold increase in hard drive storage density within a decade. It is one of the fastest transitions from a fundamental condensed-matter discovery to a technology in every computer on the planet.
Try it live
Everything above runs in your browser — open Spintronics and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Spintronics simulation