From Ultrafast Demagnetization to Ultrafast Spintronics : a 30 years story
Quentin Remy, Stéphane Mangin
Abstract
The discovery of femtosecond laser-induced ultrafast demagnetization in 1996 opened a new field, femtomagnetism, in which magnetic order can be quenched on timescales shorter than a picosecond. This seminal observation revealed that angular momentum can be transferred out of the spin system with unprecedented speed, launching intense efforts to disentangle the roles of electrons, phonons, and spins in the non-equilibrium regime. Soon it became evident that ultrafast demagnetization generates spin-flips, spin polarization, magnons and spin currents, providing new channels for angular-momentum flow. These insights laid the foundation for linking femtomagnetism with spintronics. An emblematic breakthrough in this evolution is the helicity-independent single-pulse all-optical switching (AOS) observed in rare-earth transition-metal (RE-TM) ferrimagnets such as GdFeCo. This mechanism, operating at femtojoule-scale energies and without external magnetic fields, establishes RE-TM alloys as benchmark systems for understanding and exploiting angular-momentum flow at the femtosecond timescale. Building on these concepts, the combination of ultrafast optical excitation with spintronic devices has demonstrated deterministic magnetization reversal driven by femtosecond pulses in spin valves and tunnel junctions, including rare-earth-free systems. Ultrafast spin injection, acting analogously to spin transfer torque but operating three orders of magnitude faster, allows reversal of both ferromagnetic and ferrimagnetic layers. By enabling ultrafast and energy-efficient switching, ultrafast spintronics promises scalable technologies for high-speed information processing while raising fundamental questions about angular momentum transfer in strongly out-of-equilibrium quantum materials.
AI Impact Assessments
(3 models)Scientific Impact Assessment
1. Core Contribution
This paper is a comprehensive review spanning 30 years of development from Beaurepaire et al.'s 1996 discovery of femtosecond laser-induced ultrafast demagnetization to the current state of ultrafast spintronics. The core contribution is not a single novel finding but rather a synthesis that articulates a coherent narrative connecting femtomagnetism to functional spintronic devices. The review identifies a unifying thread: angular momentum redistribution on femtosecond timescales—initially studied as a fundamental curiosity—has evolved into the physical basis for deterministic magnetization switching in technologically relevant architectures (spin valves, magnetic tunnel junctions, racetrack memories).
The paper makes several specific intellectual contributions beyond mere compilation: (i) it presents a detailed thermodynamic formalism (equations 1–4) connecting electron-magnon scattering, spin accumulation, and magnon chemical potentials to interfacial spin currents, providing a unified language for ultrafast spin transport; (ii) it distinguishes four distinct classes of all-optical switching (helicity-dependent, precessional, non-thermal phononic, and helicity-independent single-pulse), clarifying a taxonomy that has been muddled in the literature; and (iii) it explicitly connects recent THz emission spectroscopy results to a revised microscopic picture where electron-magnon scattering at ~10 fs timescales precedes and drives ultrafast demagnetization at ~100 fs, challenging the conventional three-temperature model hierarchy.
2. Methodological Rigor
As a review, the paper's rigor lies in its selection, organization, and critical evaluation of existing literature rather than in new experimental or theoretical results. The authors demonstrate strong command of the field, citing over 310 references and presenting experimental data from numerous groups alongside their own contributions. The theoretical framework presented in Section 2.3, building on Rouzegar et al.'s recent work (ref. 63), is presented with sufficient mathematical detail to be assessed critically. The distinction between spin accumulation-driven electronic spin currents and magnon chemical potential-driven magnonic spin currents (equation 4) represents a careful formalization.
However, there are limitations. The review is inevitably shaped by the authors' own research perspective—the emphasis on Lorraine-based work (GdFeCo switching, spin-valve architectures, hot-electron switching) is pronounced. While the authors acknowledge competing interpretations (e.g., the debate over exchange splitting collapse, the mechanism of P→AP switching in ferromagnetic spin valves), some controversies receive more balanced treatment than others. The claim that electron spin-flip scattering to the lattice is "negligible" compared to electron-magnon scattering (Section 2.3) is presented as relatively settled, though this remains an active area of debate.
3. Potential Impact
The review's impact potential is multifaceted:
For the research community: It provides an authoritative roadmap connecting fundamental ultrafast magnetism with device-oriented spintronics. The clear articulation of open questions—particularly regarding non-linear regime dynamics, quantitative role of interfaces, and the need for predictive multiscale models—should help guide future research priorities. The identification that thermal stability and write energy can be decoupled in spin-valve-based AOS architectures (Section 4.3) is highlighted as a potentially transformative insight for memory technology.
For technology: The review makes a compelling case that ultrafast spintronics is approaching technological viability, with demonstrations in MTJs showing >100% TMR, sub-100 fJ/bit switching, and compatibility with CMOS processing. The demonstration that picosecond electrical pulses (not just femtosecond optical pulses) can drive switching significantly lowers the barrier to integration.
For adjacent fields: The connections drawn to THz technology (spintronic THz emitters), photonic integrated circuits, and orbitronics expand the review's relevance beyond the core ultrafast magnetism community.
4. Timeliness & Relevance
The timing is excellent. The field is at an inflection point where fundamental discoveries are being translated into device demonstrations, yet no comprehensive review has previously connected the full arc from Beaurepaire's 1996 experiment through to 2025 MTJ and spin-valve demonstrations. The inclusion of very recent results (multiple 2025 references, including refs. 63, 64, 112, 133, 136, 137, 280, 290, 291, 317, 318) ensures currency. The review addresses current bottlenecks including: the need for rare-earth-free switching systems, scalability to nanometer dimensions, and integration with photonic platforms.
5. Strengths & Limitations
Key Strengths:
Notable Limitations:
Additional Observations:
The paper serves as both a historical document and a forward-looking perspective. Its value for the field lies in consolidating scattered results into a coherent framework and identifying the most promising research directions. However, the technological promise should be tempered by the substantial engineering challenges that remain between laboratory demonstrations and viable products.
Generated Apr 29, 2026
Comparison History (38)
Paper 1 is a comprehensive review covering 30 years of ultrafast spintronics, a field with transformative potential for high-speed, energy-efficient information processing. It synthesizes foundational discoveries (femtosecond demagnetization, all-optical switching) with cutting-edge spintronic device applications, bridging fundamental physics and technology. Its breadth of impact spans condensed matter physics, materials science, and computing technology. Paper 2, while valuable as a benchmarking study for machine-learned potentials applied to MOFs, addresses a narrower methodological question with more incremental contributions to the computational materials science community.
Paper 1 likely has higher impact potential: it proposes an intelligent, closed-loop, agent-based automation platform for high-throughput materials computation, addressing a broad, timely bottleneck in materials discovery and potentially transferable to other computational sciences. If rigorously validated, such infrastructure can scale research productivity and enable new discovery paradigms. Paper 2 is a valuable synthesis of a mature field (largely review/narrative), impactful for context and education, but offers less novelty and methodological contribution than a new automation framework with cross-domain applicability.
Paper 2 likely has higher impact due to greater novelty and timeliness: it clarifies the debated crystal/electronic ground state of an emerging room‑temperature d‑wave altermagnet candidate and identifies charge disproportionation as the key mechanism, with clear methodological lessons (symmetry-breaking channels in DFT+U). This advances a rapidly growing field (altermagnetism) and provides a concrete materials platform relevant to spintronics/magnonics and correlated-electron physics (coupled distortion/charge order). Paper 1 is a broad retrospective/review of a mature area, valuable but less likely to shift the frontier.
While Paper 1 offers a valuable historical review of ultrafast spintronics, Paper 2 presents a highly innovative, original machine learning-driven workflow for materials design. By combining effective-Hamiltonian molecular dynamics with a conditional autoencoder surrogate, Paper 2 enables rapid screening and multi-objective design of substituted ferroelectrics. Its methodological rigor and integration of AI into materials science represent a timely, highly impactful advancement with broad applications in computational physics and next-generation electromechanical technologies.
Paper 1 is a comprehensive review spanning 30 years of ultrafast spintronics, covering foundational discoveries (femtosecond demagnetization), all-optical switching in ferrimagnets, and emerging device applications. Its breadth connects fundamental physics to scalable technology for high-speed information processing, giving it enormous cross-disciplinary relevance and citation potential. Paper 2, while a solid first-principles study of PbRe2O6 revealing interesting quasi-1D and flat-band physics, addresses a narrower topic with more limited immediate impact. The review's timeliness, technological implications, and broad audience ensure substantially higher scientific impact.
Paper 1 presents highly novel primary research that bridges large language models, machine learning, and quantum physics for materials discovery. Its highly interdisciplinary approach and timely application of LLMs to physical sciences suggest a broader potential impact and methodological innovation compared to Paper 2, which serves as a historical review and perspective on the field of ultrafast spintronics.
Paper 2 is a comprehensive review covering 30 years of ultrafast spintronics, spanning from fundamental discoveries (ultrafast demagnetization) to technological applications (all-optical switching, spin valves, tunnel junctions). Its breadth of impact across condensed matter physics, materials science, and information technology, combined with its timeliness in summarizing a mature yet rapidly evolving field, gives it broader citation potential and impact. Paper 1, while methodologically rigorous, addresses a narrower topic—the recombination mechanism of a specific defect in silicon—with more limited cross-disciplinary reach.
Paper 2 has higher estimated impact due to its broad relevance across femtomagnetism, spintronics, and ultrafast optics, and its direct link to scalable, energy-efficient information technologies (ultrafast switching in devices). It addresses timely, high-profile questions (angular momentum transfer out of equilibrium) and summarizes a 30-year arc with clear real-world device implications. Paper 1 is novel and useful for diamond/NV characterization with potential cross-material extension, but its scope is narrower and more specialized, with more limited immediate cross-field and technological reach.
Paper 2 is a comprehensive review of a major field (ultrafast spintronics) over 30 years, synthesizing foundational breakthroughs and scalable technologies for next-generation computing. Review papers of this scale typically garner exceptionally high citations and broadly impact physics, materials science, and engineering. Paper 1, while providing a valuable experimental technique for diamond NV centers, is more narrowly focused on a specific methodological advancement and material system.
Paper 2 is a comprehensive 30-year review of ultrafast spintronics, covering the evolution from femtosecond demagnetization to all-optical switching and ultrafast spin-transfer torque. Reviews of mature, high-impact fields tend to garner significant citations by serving as authoritative references. It spans fundamental physics and technological applications across a broad community. Paper 1, while novel in proposing the SALL paradigm for altermagnetic semiconductors, is a more specialized theoretical proposal with first-principles calculations on a single proof-of-concept system, limiting its immediate breadth of impact.