| Magnetoresistive random-access memory | |
|---|---|
| Name | Magnetoresistive random-access memory |
| Acronym | MRAM |
| Type | Non-volatile memory |
| Developer | Freescale / IBM / Hitachi (historical contributors) |
| Introduced | 1990s |
| Storage | Magnetic tunnel junctions |
| Access | Random-access |
| Technology | Spintronics |
| Operating voltage | varies (≈0.8–1.8 V) |
| Density | evolving |
Magnetoresistive random-access memory
Magnetoresistive random-access memory (MRAM) is a class of non-volatile memory that stores information using magnetic states of nanoscale elements rather than charge. MRAM is significant to Quantum Physics because its operation relies on quantum-mechanical spin phenomena, tunneling effects, and coherent electron transport central to spintronics and quantum device engineering.
MRAM bridges conventional semiconductor memory and quantum-enabled device physics by exploiting spin-dependent transport and quantum tunneling. Its study involves concepts from solid-state physics, quantum mechanics, and materials science including spin coherence, exchange interactions, and tunneling magnetoresistance. Research into MRAM has intersected with work at institutions such as IBM Research, Hitachi GST, Toshiba, and national laboratories that investigate spin-based quantum devices.
Fundamentally, MRAM utilizes magnetoresistance—the change in electrical resistance due to relative magnetization orientation. The field of spintronics studies electronic spin as an information carrier alongside charge; seminal contributions include the discovery of giant magnetoresistance by Albert Fert and Peter Grünberg, and later the theory of tunneling magnetoresistance (TMR). Quantum mechanics underpins spin polarization, exchange coupling, and spin-dependent scattering in multilayer structures. Phenomena such as spin transfer torque arise from angular momentum transfer between conduction electrons and localized magnetic moments, a quantum-process description linking microscopic scattering to macroscopic magnetization dynamics described by the Landau–Lifshitz–Gilbert equation.
The canonical MRAM bit uses a magnetic tunnel junction (MTJ): two ferromagnetic layers separated by an ultrathin insulating barrier, typically MgO or aluminum oxide. One layer is a fixed or "pinned" reference (often exchange-biased by an antiferromagnet such as IrMn), while the other is a free layer whose magnetization encodes logical states. TMR derives from spin-dependent tunneling across the barrier; high TMR ratios depend on coherent tunneling and band-structure matching between electrodes, a quantum transport problem treated with density functional theory and nonequilibrium Green's functions. Common ferromagnets include CoFeB and NiFe, with interface engineering critical to performance and thermal stability.
Writing MRAM states is achieved by manipulating magnetization via mechanisms tied to spin currents. Spin-transfer torque (STT) switching injects a spin-polarized current directly through the MTJ, transferring angular momentum to the free layer. Spin–orbit torque (SOT) switching uses in-plane currents in heavy-metal layers (e.g., Pt, Ta, W) to generate spin currents via the spin Hall effect or interfacial Rashba effect, exerting torques on adjacent ferromagnets. Reading is performed by measuring resistance across the MTJ, relying on TMR. Both STT and SOT are described by quantum scattering theory and spin-dependent transport models; device optimization often involves micromagnetic simulations and experiments at facilities like Argonne National Laboratory or university spintronics groups.
Key metrics include switching speed, write energy, endurance (write/erase cycles), and data retention time. MRAM offers low-latency reads comparable to SRAM and high endurance surpassing many flash memories. Scaling limits arise from thermal stability of nanoscale magnets (superparamagnetism), current-driven reliability, and barrier breakdown in MTJs. Quantum considerations affect scaling: reduced volumes increase susceptibility to thermal activation described by Arrhenius-type models and quantum tunneling of magnetization can impose limits at cryogenic sizes. Materials innovation—e.g., perpendicular magnetic anisotropy (PMA) stacks—seeks to maintain stability while reducing switching currents.
At the nanoscale, MRAM operation encounters quantum-level effects such as shot noise, spin noise, and quantum tunneling of electrons and, in extreme regimes, of magnetization states. Coherence of spin currents and phase-coherent tunneling influence TMR amplitude and switching dynamics; decoherence mechanisms include electron–phonon scattering and magnon excitation. Quantum fluctuations and stochastic spin-transfer switching introduce switching variability modeled with stochastic versions of the Landau–Lifshitz–Gilbert equation and full quantum master-equation approaches, relevant to device reliability and read/write error rates. Experimental probes often use ferromagnetic resonance and spin pumping techniques to measure damping and coherence.
MRAM targets applications from embedded memory in system on chip designs to cache and storage-class memory, with companies like Samsung Electronics, Micron Technology, and GlobalFoundries investing in production. Integration with CMOS processes involves back-end-of-line compatibility, thermal budgets, and voltage/current drivers. In quantum technology contexts, MRAM and spintronic elements are explored as classical control/readout elements for quantum computing architectures, as cryogenic non-volatile memory for superconducting qubits, and as building blocks for hybrid spin–qubit systems. Collaboration between semiconductor fabs, academic research groups (e.g., University of California, Berkeley, University of Cambridge) and national labs advances both practical MRAM and fundamental quantum spintronics research.
Category:Computer memory Category:Spintronics Category:Solid-state physics