Physics / mechanism
Heat-assisted magnetic recording (HAMR) addresses the “magnetic recording trilemma”: as bit size is reduced, signal-to-noise ratio, thermal stability and writability cannot all be optimised at once. High-anisotropy media give the thermal stability needed for small bits, but their coercivity exceeds what a write head can supply. HAMR resolves this by transiently heating the medium during the write cycle, lowering the switching field long enough for the head to reverse the bit, after which the medium cools back into a thermally stable state.
In the mainstream implementation, the heat is delivered optically: a laser and a plasmonic transducer concentrate energy into a sub-diffraction spot on the disk surface. This optical delivery chain is what confines HAMR largely to hard-disk drives and makes it difficult to integrate with on-chip or embedded memory architectures.
Two research directions attempt to change the heating and switching physics. The first replaces laser heating with Joule heating in an intercalated two-dimensional magnet, giving an electrically controlled variant of HAMR that does not require optics or a plasmonic near-field transducer. The second targets the medium itself: iron-rhodium (FeRh) has a first-order phase transition near room temperature between antiferromagnetic and ferromagnetic phases, which makes the write temperature lower and the transition sharper than in alternative materials.
The relevant design parameters follow from this: the write temperature and the sharpness of the thermal transition set the thermal engineering budget of the head and medium, and lower write temperatures are argued to extend write/read head lifetime.
Competitive landscape
Within magnetic recording, HAMR is described as the leading solution to the trilemma relative to conventional bit-size scaling. The competition is therefore mostly internal to HAMR: optical/plasmonic heating, which is the deployed form but is tied to hard-disk drives, versus electrically driven Joule heating, which is being explored specifically to make heat-assisted writing compatible with on-chip and embedded memory. On the media side, FeRh competes with other high-anisotropy HAMR media on write temperature and transition sharpness rather than on areal density directly, but antiferromagnet-based magnetic memory using FeRh has not yet been realised despite considerable effort.
Evidence base
- HAMR is identified as the leading answer to the magnetic recording trilemma, enabling high-density storage by transiently heating the medium during the write cycle (8 May 2026).
- Reliance on laser optics and plasmonic transducers restricts HAMR primarily to hard-disk drives and limits integration with on-chip or embedded architectures (8 May 2026).
- An electronic variant of HAMR based on Joule heating in intercalated 2D magnets was demonstrated (8 May 2026).
- FeRh has a first-order antiferromagnetic-to-ferromagnetic transition near room temperature, with a comparatively sharper transition and lower writing temperature than alternative materials, implying fewer thermal engineering constraints and longer write/read head lifetime (3 June 2026).
- Antiferromagnet-based magnetic memory using FeRh had not been realised as of this work (3 June 2026).
- Wide-field and scanning nanoscale quantum diamond microscopes were used to image directly the magnetic field of patterned FeRh structures (3 June 2026).
Frontier (open questions)
- What write energy, switching time and cycle endurance does Joule-heated HAMR in intercalated 2D magnets achieve relative to laser-plus-plasmonic-transducer HAMR in shipping drives?
- Can electrically heated HAMR bits be scaled to the areal densities and bit pitches required by hard-disk or embedded memory roadmaps, and at what thermal crosstalk penalty to neighbouring bits?
- Does the quantum diamond microscope imaging of patterned FeRh demonstrate reversible, addressable AFM-to-FM bit writing, or only static contrast between phases?
- How much lower, quantitatively, is the FeRh write temperature than incumbent HAMR media, and does that translate into a measured increase in head lifetime?
Synthesised 2026-08-31 from 2 KB sources by the resynth pipeline; citations are KB source slugs.