Rare-Earth-Free Ultra-Strong Permanent Magnets: A Multiscale Framework Integrating Topological Spin Textures, Quantum Thermodynamics, and AI-Driven Inverse Design

The escalating demand for high-performance permanent magnets in decarbonization technologies, such as electric vehicles and wind turbines, has intensified the need for alternatives to rare-earth-element (REE)-dependent materials like Nd$_2$Fe$_{14}$B, which dominate over 90% of global REE consumption amid geopolitical and environmental challenges. This work presents a comprehensive multiscale framework for designing REE-free magnets that achieve a theoretical maximum energy product $(BH)_{\max} = 72 \pm 4.2$ MGOe and Curie temperature $T_c > 820$ K, surpassing current benchmarks. By synergistically integrating quantum thermodynamics, topological spin engineering, first-principles materials science, and AI-driven inverse design, the proposed Fe$_{65}$Co$_{20}$B$_{10}$N$_5$ system exploits interstitial nitrogen (5 at.%) to induce tetragonal distortion ($c/a = 1.08$) in a body-centered $\alpha$-Fe lattice (space group $I4/mmm$), generating giant uniaxial magnetocrystalline anisotropy ($K_u = 1.8$ MJ/m$^3$) via Fe 3d orbital spin-orbit coupling, with cobalt enhancing saturation magnetization ($M_s = 2.45$ T) and boron promoting 8–12 vol.% Fe$_2$B phases for domain-wall pinning ($\Delta G_f = -0.12$ eV/atom). Nanoscale Néel-type skyrmion lattices ($R \approx 8$ nm, $\rho_s = 1.2 \times 10^{12}$ cm$^{-2}$) are nucleated at sulfur-vacancy disclinations on MoS$_2$ substrates during atomic layer deposition, stabilized by Dzyaloshinskii–Moriya interactions ($D = 1.5$ mJ/m$^2$) within a generalized spin Hamiltonian incorporating Kitaev anisotropy, yielding a twofold coercivity increase ($H_c = 28$ kOe) through topological protection, as confirmed by micromagnetic simulations (MuMax3). A graph neural network (MagGen, $R^2 = 0.94$) optimizes the composition from 12,840 DFT+U configurations (VASP, HSE06), enforcing REE exclusion via penalized loss functions (MAE = 2.1 MGOe). Scalable fabrication bridges 2D films to 3D bulk via ALD growth at 300°C, tape delamination, 5 T field-assisted stacking, spark plasma sintering (>98% density at 850 K), and 10 T pulsed-field kinetic arrest, ensuring thermal stability ($\Delta E_f > 0.45$ eV/atom). A three-phase experimental roadmap targets phase confirmation ($M_s > 2.3$ T) in 0–6 months, skyrmion imaging in 6–18 months, and bulk validation ($(BH)_{\max} > 60$ MGOe) by 24 months. Discussion addresses defect uniformity via ion irradiation, skyrmion annihilation temperatures exceeding 473 K, and pinning dominance over nucleation, positioning this framework for >90% REE reduction and industrial adoption in sustainable magnetics.在脱碳技术（如电动汽车、风力发电机）领域，高性能永磁体的需求持续攀升，这使得人们亟需开发替代依赖稀土元素（rare-earth-element, REE）的永磁材料的方案——目前钕铁硼（Nd₂Fe₁₄B）占据了全球稀土元素消费的90%以上，同时面临地缘政治与环境层面的多重挑战。本研究提出了一套完整的多尺度设计框架，用于开发无稀土永磁体，该类永磁体可实现理论最大磁能积$(BH)_{max}=72pm4.2$兆高斯奥斯特（MGOe），居里温度$T_c>820$开尔文，性能超越当前行业基准。通过协同整合量子热力学、拓扑自旋工程、第一性原理材料科学与人工智能驱动的逆向设计方法，本研究提出的Fe₆₅Co₂₀B₁₀N₅体系利用间隙氮（5原子百分比）在体心立方$alpha$-Fe晶格（空间群$I4/mmm$）中诱导出四方畸变（$c/a=1.08$），通过Fe的3d轨道自旋轨道耦合产生超大单轴磁晶各向异性（$K_u=1.8$兆焦每立方米）；其中钴可提升饱和磁化强度（$M_s=2.45$特斯拉），硼则促进8%~12%体积占比的Fe₂B相形成，用于畴壁钉扎（$Delta G_f=-0.12$电子伏特每原子）。纳米尺度奈尔型斯格明子晶格（$Rapprox8$纳米，$ ho_s=1.2 imes10^{12}$厘米⁻²）可在原子层沉积（Atomic Layer Deposition, ALD）过程中于二硫化钼（MoS₂）衬底的硫空位向错处形核，在包含基泰耶各向异性的广义自旋哈密顿量框架下，通过德让-莫里亚相互作用（$D=1.5$毫焦每平方米）实现稳定，最终通过拓扑保护使矫顽力提升一倍（$H_c=28$千奥斯特），该结果已通过微磁学模拟（MuMax3）验证。一款图神经网络（MagGen，决定系数$R^2=0.94$）从12840个密度泛函理论+U修正（DFT+U）构型（VASP软件、HSE06泛函）中优化合金成分，并通过惩罚损失函数（平均绝对误差MAE=2.1 MGOe）确保体系不含稀土元素。可规模化制备工艺通过300摄氏度下的原子层沉积生长、胶带剥离、5特斯拉磁场辅助堆叠、放电等离子烧结（850开尔文下致密度>98%）以及10特斯拉脉冲场动力学阻滞，实现从二维薄膜到三维块体的制备，并确保体系热稳定性（$Delta E_f>0.45$电子伏特每原子）。该研究制定了三阶段实验路线图：0~6个月内完成物相确认（$M_s>2.3$特斯拉），6~18个月内实现斯格明子成像，至24个月时完成块体样品性能验证（$(BH)_{max}>60$ MGOe）。讨论部分涵盖了通过离子辐照实现缺陷均匀性、斯格明子湮灭温度高于473开尔文，以及畴壁钉扎优于形核过程等内容，该框架有望实现稀土元素用量降低90%以上，并推动其在可持续永磁领域实现工业化应用。