A magnetic field changed a metal stack’s resistance
Send a small current through a metal film and measure the voltage along it. Now change the magnetic field. A resistance change is not surprising by itself: several forms of magnetoresistance were already known. What made the 1988 iron/chromium multilayer measurements striking was the scale of the change and its connection to the relative magnetization directions of neighboring magnetic layers.
Albert Fert’s group reported the effect in a stack containing many thin iron and chromium layers. Peter Grünberg’s group independently investigated related layered structures, with its paper published in 1989. The famous large low-temperature multilayer response and the smaller room-temperature trilayer response belong to different specimens and conditions. They should not be merged into a single universal performance number. The discovery was that a carefully made metallic stack could turn a change of magnetic arrangement into a strong electrical signal, and the electrical curve made that claim measurable.
The resistance must be inferred from voltage and the supplied current together. If the current changes during the field sweep, a voltage change alone does not isolate magnetoresistance. Repeating at a smaller measuring current also helps check whether self-heating is influencing the result.
Sources: [1][4][5]
The field reorganized magnetizations, not just individual trajectories
A ferromagnetic layer has a preferred collective magnetization direction. A thin nonmagnetic spacer separates it from another magnetic layer. In suitable iron/chromium structures, coupling through the spacer favors opposing directions at low applied field. A sufficiently strong external field brings the layer magnetizations toward parallel alignment. The resistance falls as the magnetic configuration changes.
The experiment therefore needs more than a resistance trace with a field axis. Layer thickness, temperature and current geometry identify the specimen being tested; magnetic measurements help establish the configuration associated with a resistance branch. The original celebrated multilayer measurements used current flowing in the plane of the layers. A sketch of electrons moving successively straight through every layer illustrates another geometry and cannot be substituted without explanation.
Fabrication was part of the physics. Rough or intermixed interfaces and changed spacer thickness can alter coupling and scattering. A stack that is almost the same in a low-resolution drawing need not have the same magnetic ground state or transport response. This sensitivity explains both the experimental challenge and the opportunity to design the effect.
Sources: [4][1]
Spin changes the probability of scattering
Spin is an electron’s intrinsic angular momentum, not the literal spinning of a small charged ball. In a ferromagnetic metal, the electronic states available to the two spin orientations differ. Electrons with spins defined relative to the layer magnetization can consequently encounter different scattering probabilities. Scattering removes directed drift, which is why different scattering rates can give different electrical resistivities.
When spin mixing is sufficiently weak, transport can be organized approximately into two conducting channels. The conductances add, as in the expression below; the channel resistances refer to effective paths through the structure in the chosen geometry. Parallel magnetic layers can leave one channel relatively favorable across the relevant structure. Antiparallel layers expose both channels to an unfavorable magnetic environment somewhere along their transport. The resulting total resistance is higher.
The cartoon is useful, but real electrons scatter within layers and at interfaces, and they need not preserve spin indefinitely. Temperature-dependent spin mixing reduces the independence of the channels. A quantitative calculation must describe these processes and the actual geometry rather than treating magnetic arrows as absolute one-way doors for selected electrons.
Sources: [4][2]
A percentage is meaningless without its reference resistance
The first Fe/Cr multilayer paper reported resistance reductions approaching a factor of two at low temperature. A drop from a high resistance to half that value is a fifty-percent decrease when normalized to the starting resistance. The same two endpoints give a one-hundred-percent increase when normalized to the lower resistance. The physics has not changed; only the reference has.
State both resistance endpoints and the chosen denominator when quoting the change. The parallel and antiparallel labels refer to magnetic configurations, not current directions. Temperature must accompany the quoted value, and a field-saturated state should be identified before it is equated with the parallel reference. Ordinary anisotropic magnetoresistance, which depends on magnetization relative to current direction, can contribute additional structure. Comparing angular rotations, layer arrangements and magnetic saturation helps identify which effect is responsible.
This care is essential when comparing a synthetic spin valve with a historical multilayer. It prevents a modest teaching example from inheriting the amplitude of a different low-temperature sample, and it makes experimental plots readable even when their authors use different percentage conventions.
Sources: [5][4][1]
An angle scan can turn a magnetic rotation into a voltage
A spin valve arranges one layer to be comparatively stable and another to respond more readily to a field. Their relative magnetization angle becomes a controllable variable. A simple angular interpolation connects known parallel and antiparallel resistances using a cosine. In this description, the angle is between the two magnetizations; its arguments are in radians. This interpolation describes an assumed angular response, not the dynamics by which an applied field rotates or reverses the magnets.
Take a synthetic device with a parallel resistance of one hundred ohms and an antiparallel resistance of one hundred and eighteen ohms. A one-milliampere sense current would produce one hundred to one hundred and eighteen millivolts across that range. At a right angle, the interpolation gives the midpoint. The example predicts an electrical signal from a specified magnetic change and makes the eighteen-percent convention explicit. Increasing sense current increases the voltage signal but also increases dissipated heat, so arbitrarily large electrical amplification of the readout is not free.
Sources: [4][1]
Reading magnetic bits required another engineering step
A practical magnetic read head must respond to fields from recorded bits with a stable, interpretable electrical output. A large resistance contrast in a laboratory field sweep is an encouraging starting point, but it does not directly specify sensitivity around the desired operating point, hysteresis, noise or fabrication tolerance. Spin-valve design and improved film production helped turn the material effect into a useful sensor technology.
This progression is part of the meaning of the 2007 Nobel Prize to Fert and Grünberg. Electron spin became a degree of freedom through which a magnetic configuration could control electrical transport. The result linked materials research to information readout without requiring the carriers to stop being ordinary charged electrons. It also encouraged researchers to distinguish metallic spin-dependent scattering from magnetic tunneling, whose barrier and microscopic transport mechanism are different. Similar-looking resistance changes can arise from different physical processes, and the specimen’s structure decides which description is appropriate.
Sources: [4][3]
Reading magnetization electrically
Spin-dependent scattering makes the relative arrangement of magnetic layers alter the conducting path. Holding one layer steady while another responds to a small field converts a magnetic change into resistance and voltage. That connection between thin-film structure and electrical readout drove the development of spin valves and magnetic information sensing.
Sources: [1][3]