Demasius et al. (Nat. Commun. 7, 10644 (2016)) reported that incorporating oxygen into sputtered tungsten films yields spin Hall angles up to approximately -0.5 in W(O)/CoFeB bilayers, remarkably insensitive to the incorporated oxygen content (12–44%), and argued that the spin–orbit torque is largely interfacial in origin — but the microscopic mechanism was not established. The open problem is the quantitative microscopic identification of the origin of this large antidamping-like torque: how it is partitioned, as a function of oxygen content, between interfacial spin–orbit coupling at the W(O)/CoFeB boundary and any residual bulk spin Hall effect, together with the supporting quantities needed to make this determination — the spin diffusion length in W(O), the magnitude and oxygen dependence of spin memory loss at the W(O)/CoFeB interface, the interfacial spin–orbit parameters, and a spin-torque-ferromagnetic-resonance methodology whose line-shape and line-width analyses are quantitatively reconciled. No quantitative interfacial-versus-bulk partition for W(O)/CoFeB, no oxygen-dependent spin-memory-loss quantification for this interface, and no independent spin diffusion length for the oxygenated films have yet been established.

Background

Spin–orbit torques (SOTs) are current-induced torques on a ferromagnet's magnetization that arise when a charge current in an adjacent non-magnetic layer is converted into a transverse spin current. Two mechanisms are commonly invoked. In the bulk spin Hall effect (SHE), spin–orbit coupling in the volume of a heavy metal deflects opposite spins in opposite directions, generating a spin current whose charge-to-spin conversion efficiency is expressed as the spin Hall angle (SHA, the ratio of spin current to charge current); the torque reaches the ferromagnet only if the layer is thicker than the spin diffusion length, the characteristic decay length λs\lambda_{s} of spin current in that metal. In interfacial mechanisms, inversion symmetry breaking at a heavy-metal/ferromagnet interface combined with spin–orbit coupling produces a Rashba-type spin-split surface state, whose current-driven spin polarization (the Edelstein or inverse spin-galvanic effect) acts directly on the adjacent magnetic moments. Telling these apart in a bilayer is difficult because both generate the same symmetry of torque (an antidamping-like torque plus a field-like torque) and both scale similarly with layer thickness when λs\lambda_{s} is short.

The standard experimental probe is spin-torque ferromagnetic resonance (STFMR): a radio-frequency (RF) current through a heavy-metal/ferromagnet bilayer drives magnetization precession, and the rectified DC mixing voltage is analyzed either from the Lorentzian line shape (symmetric part attributed to the antidamping-like spin torque, asymmetric part to the Oersted field) or from the resonance line width: a precessing magnetization pumps spin angular momentum into the adjacent heavy metal (spin pumping), which enhances its Gilbert damping, and this emitted spin current is reconverted into a measurable charge voltage in the heavy metal by the inverse spin Hall effect (the reciprocal of the SHE). The two analyses can disagree, and a quantitative reconciliation requires an accurate determination of the RF current actually flowing through the device. Interfacial spin transport introduces further entangled quantities: the spin mixing conductance (the interface's capacity to absorb spin angular momentum, extracted from spin-pumping-enhanced Gilbert damping) and spin memory loss (SML, the loss of spin angular momentum in the first few atomic layers at an interface, which reduces the spin current reaching or leaving the ferromagnet and biases the apparent conversion efficiency).

Tungsten is a flagship SOT source: the metastable, resistive A15 (β\beta) phase shows SHAs up to roughly -0.3 to -0.35, whereas the equilibrium bcc (α\alpha) phase is nearly inert; oxygen is known to stabilize the β\beta phase during sputtering. Demasius et al. (2016, IBM) sputtered W in an oxygen atmosphere, producing nanocrystalline to amorphous W(O) films with incorporated oxygen contents from 12% to 44%, and measured STFMR SHAs of up to about -0.5 in W(O)/CoFeB — at the time the largest reported for a conventional metal — that were insensitive to the oxygen content and to large changes in resistivity and microstructure, and essentially did not vary with W(O) thickness. Since a bulk SHE origin would require λs\lambda_{s} significantly smaller than 4.4 nm, and since oxidized W/CoFeB interfaces plausibly host an oxygen-enhanced Rashba interaction (by analogy with oxidation-enhanced Rashba splitting at the Gd(0001) surface), the authors concluded the torque is likely dominantly interfacial, while explicitly leaving open: the quantitative partition between interfacial and bulk contributions; the magnitude and oxygen dependence of SML at the W(O)/CoFeB interface; whether amorphization of highly oxygenated W(O) could instead enhance the SHA extrinsically (in analogy with the large anomalous Hall effects of amorphous magnets); the quantitative reconciliation of their line-shape and line-width analyses (limited by knowledge of the RF current distribution); and an independent determination of λs\lambda_{s} in the W(O) films. Resolving these would convert a phenomenological observation into a predictive, microscopic framework for engineering oxygen-tuned SOT materials, which matters because tungsten-based SOT sources are actively pursued for magnetic random-access memory.

Problem Statement

Establish the microscopic origin of the large antidamping-like spin–orbit torque in oxygen-incorporated tungsten W(O)/CoFeB bilayers, quantitatively and as a function of the incorporated oxygen content, in order to convert the phenomenological observation of Demasius et al. (2016) — spin Hall angles up to about -0.5 that are insensitive to oxygen content (12–44%), resistivity and microstructure — into a predictive, microscopic engineering framework. Concretely, determine: (i) the quantitative partition of the torque between interfacial spin–orbit coupling at the W(O)/CoFeB boundary, possibly oxygen-enhanced, and any residual bulk spin Hall contribution; (ii) the spin diffusion length λs\lambda_{s} of the W(O) films, independently of the growth entanglement between film thickness and incorporated oxygen content; (iii) the magnitude and oxygen dependence of spin memory loss at the W(O)/CoFeB interface, so that the extracted intrinsic charge-to-spin conversion efficiencies are not biased; and (iv) whether amorphization or extrinsic scattering in the highly oxygenated, nanocrystalline-to-amorphous W(O) films can instead produce an apparent enhancement of the spin Hall angle (in analogy with large anomalous Hall effects in amorphous magnets). The partition and the extracted interfacial spin–orbit parameters must be quantitative rather than indicative, which requires the line-shape and line-width analyses — including an accurate account of the RF current distribution in the device and of spin-pumping contributions — to be quantitatively reconciled. All components serve one objective: a microscopic, predictive account of how oxygen incorporation in tungsten produces very large spin–orbit torques; a complete resolution requires the partition (i) together with the supporting determinations (ii)–(iv), and no proper subset of these resolves the task.

References

  1. Kai-Uwe Demasius, Timothy Phung, Weifeng Zhang, Brian P. Hughes, See-Hun Yang, Andrew Kellock, Wei Han, Aakash Pushp, Stuart S. P. Parkin, 'Enhanced spin–orbit torques by oxygen incorporation in tungsten films', Nature Communications 7, 10644 (2016), DOI: 10.1038/ncomms10644, https://www.nature.com/articles/ncomms10644
  2. Jie Zhang, Timothy Phung, Brian P. Hughes, See-Hun Yang, Chirag Garg, Yong Jiang, Stuart S. P. Parkin, 'Effect of interfacial insertion layers on the spin–orbit torque in W(O)|CoFeB heterostructures', Applied Physics Express 12, 033001 (2019), DOI: 10.7567/1882-0786/aaf7f9, https://doi.org/10.7567/1882-0786/aaf7f9
  3. Yuito Kageyama, Yuya Tazaki, Hongyu An, Takashi Harumoto, Tenghua Gao, Ji Shi, Kazuya Ando, 'Spin-orbit torque manipulated by fine-tuning of oxygen-induced orbital hybridization', Science Advances 5 (11), eaax4278 (2019), DOI: 10.1126/sciadv.aax4278, https://www.science.org/doi/10.1126/sciadv.aax4278
  4. Satoshi Haku, Akira Musha, Tenghua Gao, Kazuya Ando, 'Role of interfacial oxidation in the generation of spin-orbit torques', Physical Review B 102, 024405 (2020), DOI: 10.1103/PhysRevB.102.024405, https://journals.aps.org/prb/abstract/10.1103/PhysRevB.102.024405
  5. Atsushi Okada, Yutaro Takeuchi, Kaito Furuya, Chaoliang Zhang, Hideo Sato, Shunsuke Fukami, Hideo Ohno, 'Spin-Pumping-Free Determination of Spin-Orbit Torque Efficiency from Spin-Torque Ferromagnetic Resonance', Physical Review Applied 12, 014040 (2019), DOI: 10.1103/PhysRevApplied.12.014040, https://journals.aps.org/prapplied/abstract/10.1103/PhysRevApplied.12.014040
  6. Ranen Ben-Shalom, Nirel Bernstein, Stuart S. P. Parkin, See-Hun Yang, Amir Capua, 'Determination of the spin Hall angle by the inverse spin Hall effect, device level ferromagnetic resonance, and spin torque ferromagnetic resonance: A comparison of methods', Applied Physics Letters 119, 042401 (2021), DOI: 10.1063/5.0057192, https://doi.org/10.1063/5.0057192
  7. Surya Narayan Panda, Sudip Majumder, Arpan Bhattacharyya, Soma Dutta, Samiran Choudhury, Anjan Barman, 'Structural Phase-Dependent Giant Interfacial Spin Transparency in W/CoFeB Thin-Film Heterostructures', ACS Applied Materials & Interfaces 13 (17), 20875–20884 (2021), DOI: 10.1021/acsami.1c03776, https://doi.org/10.1021/acsami.1c03776
  8. Deepika Jhajhria, Nilamani Behera, Dinesh K. Pandya, Sujeet Chaudhary, 'Dependence of spin pumping in W/CoFeB heterostructures on the structural phase of tungsten', Physical Review B 99, 014430 (2019), DOI: 10.1103/PhysRevB.99.014430, https://journals.aps.org/prb/abstract/10.1103/PhysRevB.99.014430
  9. W. Cao, J. Liu, A. Zangiabadi, K. Barmak, W. E. Bailey, 'Measurement of spin mixing conductance in Ni81Fe19/alpha-W and Ni81Fe19/beta-W heterostructures via ferromagnetic resonance', Journal of Applied Physics 126, 043902 (2019), DOI: 10.1063/1.5099913, https://doi.org/10.1063/1.5099913
  10. James Lourembam, Hong Jing Chung, Lisen Huang, Khoong Hong Khoo, Jinjun Qiu, Huiqing Xie, D.V. Maheswar Repaka, Sherry Lee Koon Yap, Hang Khume Tan, Bingjin Chen, Seng Kai Wong, Abhijit Ghosh, Haidong Liang, Sabpreet Bhatti, T.S. Suraj, Andrew Anthony Bettiol, Anjan Soumyanarayanan, S.N. Piramanayagam, Sze Ter Lim, 'Spin-orbit torques in laminated tungsten multilayers with nominally doped oxygen', Physical Review Applied 23, 014009 (2025), DOI: 10.1103/PhysRevApplied.23.014009, https://journals.aps.org/prapplied/abstract/10.1103/PhysRevApplied.23.014009
  11. L. Neumann, D. Meier, J. Schmalhorst, K. Rott, G. Reiss, M. Meinert, 'Temperature dependence of the spin Hall angle and switching current in the nc-W(O)/CoFeB/MgO system with perpendicular magnetic anisotropy', Applied Physics Letters 109, 142405 (2016), DOI: 10.1063/1.4964415, https://doi.org/10.1063/1.4964415
  12. Ananya Chattaraj, Mohammad Balal, Ashok Kumar Yadav, Sudipta Roy Barman, Anil Kumar Sinha, Shambhu Nath Jha, Sebastien Joulie, Virginie Serin, Alain Claverie, Vijay Kumar, Aloke Kanjilal, 'Unravelling oxygen driven alpha to beta phase transformation in tungsten', Scientific Reports 10, 14718 (2020), DOI: 10.1038/s41598-020-71650-2, https://www.nature.com/articles/s41598-020-71650-2
  13. Xueyao Hou, Keita Ito, Varun K. Kushwaha, Takumi Yamazaki, Junwei Gu, Zhenchao Wen, Keisuke Masuda, Jun Okabayashi, Shoya Sakamoto, Takeshi Seki, 'Amorphous Tungsten Driving Extraordinary Long-Range Interlayer Exchange Coupling and Large Spin-Orbit Torque', Advanced Electronic Materials, e70545 (2026), DOI: 10.1002/aelm.70545, https://doi.org/10.1002/aelm.70545