IBM's GMR read head let hard disks shrink into laptops and pocket players

By the late 1980s the hard disk industry was squeezing against a wall of its own success: the smaller the magnetic domain, the weaker its field, and the standard read-out sensor — anisotropic magnetoresistance, known since Lord Kelvin measured it in 1857 — was running out of sensitivity. In 1988 the wall cracked twice: Albert Fert in Orsay and Peter Grünberg in Jülich independently discovered giant magnetoresistance, where resistance changes in layered iron-chromium structures far more than any known effect under a weak field.

Turning the effect into a product took two more moves, both driven from IBM. First the spin valve: one magnetic layer pinned, one free, so scanning a disk flips the free layer between parallel and antiparallel and converts each domain into a clean high or low current — binary data read directly off resistance. Second, manufacturing: epitaxy, the method both discoverers used, was too slow and costly for scale. Stuart Parkin's group at IBM showed the GMR effect survives layers made by sputtering, a cheap industrial process — perfection was never required.

The first commercial GMR read head shipped in 1997 and, as Fert's Nobel lecture records, GMR heads replaced AMR sensors that same year; the Nobel information page notes the new technology became standard soon after. The payoff was the continuation of the disk's exponential run: terabyte drives for home use, disks small enough for laptops and pocket music players. The successor technology, tunnelling magnetoresistance, now sits in the newest heads — and Fert and Grünberg shared the 2007 Nobel Prize in Physics for the effect underneath it all.

Read bits with electron spin, not induced current

Read sensitivity was the binding constraint on areal density, and GMR multiplied it over AMR by an order of magnitude in the lab — the physics matched the bottleneck exactly.

The spin-valve structure turns a field's direction into a clean two-level signal, no analog fine positioning required — the effect is digital-friendly by construction.

Sputtering removed the manufacturing wall: the effect did not depend on perfect epitaxial layers, so heads could be produced at industrial scale.

The payoff compounds through the disk's own economics: every gain in areal density cuts cost per gigabyte, which is what created the portable-device storage market.

A discovery becomes a product only once someone finds the cheap way to make it. Sputtering showed the effect survived imperfect layers — the manufacturing shortcut decided what shipped.

GMR read heads became the industry standard within a couple of years of 1997 and carried the hard disk's density climb through the terabyte era, a development the Nobel Foundation credits with the shrinking of hard disks into laptops and pocket devices. TMR heads — the same physics with an insulating spacer — displaced GMR in the newest generation. Fert and Grünberg shared the 2007 Nobel Prize in Physics, and the IBM materials work that industrialized the effect seeded the wider spintronics field, including magnetic tunnel junctions and MRAM.