Ferromagnetic CoPt: Properties and Applications

• Ferromagnetic CoPt is a cobalt–platinum alloy system exhibiting robust magnetism, strong spin–orbit coupling, and distinct ordered/disordered structures.
• Its magnetic anisotropy is highly sensitive to factors like crystal order, surface termination, defects, and finite-size effects, influencing spin–orbit torque and Hall effects.
• CoPt is pivotal in advanced applications ranging from field-free and current-induced switching to exchange-coupled architectures, interfacial charge control, and catalytic surface modifications.

Searching arXiv for the cited CoPt papers to ground the article in the specified literature. arxiv_search.query({"search_query":"id:(Yongming et al., 2023) OR id:(Brahimi et al., 2019) OR id:(Li et al., 2022) OR id:(Chen et al., 2021) OR id:(Zhu et al., 2020) OR id:(Karoui et al., 2012) OR id:(Al-Mahdawi et al., 2018) OR id:(Qu et al., 2019) OR id:(Miura et al., 2021) OR id:(Masuda et al., 2020) OR id:(Liu et al., 2024) OR id:(Belmoubarik et al., 2023) OR id:(Demidov et al., 2016) OR id:(Alexandrakis et al., 2016) OR id:(Rohart et al., 2011) OR id:(Cuadrado et al., 2012)","max_results":16}) to=arxiv_search.query code ൂട {"result":[{"arxiv_id":"(Liu et al., 2024)","title":"Efficient Terahertz Generation from CoPt-based Terahertz Emitters via\n Orbital-to-Charge Conversion","authors":"Yi Xing, Wenchao Hao, Xinwei Xu, Yifan Wei, Chuyi Ma, Huajin Wang, Aijun Chen, Xingfen Wang, Jinbo Yang, Weisheng Zhao, Xiaozhe Chen, Yong Jiang","abstract":"Orbitronics devices operate by manipulating orbitally-polarized currents.\n Recent studies have shown that these orbital currents can be excited by\n femtosecond laser pulses in ferromagnet as Ni and converted into ultrafast\n charge current via orbital-to-charge conversion. However, the terahertz\n emission from orbitronic terahertz emitter based on Ni is still much weaker\n than the typical spintronic terahertz emitter. Here, we report more efficient\n light-induced generation of orbital current from CoPt alloy and the orbitronic\n terahertz emission by CoPt/Cu/MgO shows terahertz radiation comparable to\n that of efficient spintronic terahertz emitters. By varying the concentration\n of CoPt alloy, the thickness of Cu, and the capping layer, we confirm that\n THz emission primarily originates from the orbital accumulation generated\n within CoPt, propagating through Cu and followed by the subsequent\n orbital-to-charge conversion from the inverse orbital Rashba-Edelstein effect\n at the Cu/MgO interface. This study provides strong evidence for the very\n efficient orbital current generation in CoPt alloy, paving the way to\n efficient orbital terahertz emitters.","categories":"cond-mat.mtrl-sci cond-mat.mes-hall cond-mat.str-el"}]} Ferromagnetic CoPt is a cobalt–platinum alloy system that appears in chemically ordered and chemically disordered forms, in bulk phases, thin films, nanoparticles, and heterostructures. Across these realizations, its defining features are robust ferromagnetism, strong spin–orbit coupling inherited from Pt, and unusually rich anisotropy and transport phenomenology. The literature spans tetragonal $L1_0$ CoPt with large uniaxial magnetocrystalline anisotropy, chemically disordered $A1$ CoPt single layers with perpendicular magnetic anisotropy and self-generated spin-orbit torque, and symmetry-engineered epitaxial CoPt films that support deterministic field-free switching, unconventional Hall responses, and efficient orbital-current generation (Karoui et al., 2012, Zhu et al., 2020, Yongming et al., 2023, Chen et al., 21 Nov 2025).

1. Crystal forms and ferromagnetic order

Ferromagnetic CoPt is not a single structural motif. First-principles and thin-film studies identify ordered tetragonal $L1_0$ CoPt, ordered $L1_1$-type CoPt in (111)-oriented junction concepts, chemically disordered fcc $A1$ CoPt, and interdiffused multilayer-derived (111)-oriented CoPt films (Karoui et al., 2012, Masuda et al., 2020, Chen et al., 2021, Yongming et al., 2023). In the ordered equiatomic $L1_0$ phase, DFT-GGA gives $a = 3.81\,\text{\AA}$, $c/a = 0.976$, and a magnetic moment of $1.14\,\mu_B/$atom, while Pt acquires a small induced moment of about $0.3\,\mu_B/$Pt atom through Co $A1$0–Pt $A1$1 hybridization (Karoui et al., 2012). The same study concludes that ferromagnetism strongly reinforces ordering tendencies: the formation enthalpy of ordered CoPt becomes roughly $A1$2 to $A1$3 eV/atom in the spin-polarized state, whereas the nonmagnetic or paramagnetic reference yields only about $A1$4 to $A1$5 eV/atom, with $A1$6 remaining the most stable equiatomic ordered phase (Karoui et al., 2012).

Chemically disordered $A1$7 CoPt is also ferromagnetic and technologically relevant. Sputtered $A1$8 Co$A1$9Pt$L1_0$0 single layers on MgO(111) were shown to retain perpendicular magnetic anisotropy over the composition window from Co$L1_0$1Pt$L1_0$2 to Co$L1_0$3Pt$L1_0$4, with the phase identified as chemically disordered fcc rather than ordered $L1_0$5 (Chen et al., 2021). In a separate single-layer study, chemically disordered fcc Co$L1_0$6Pt$L1_0$7 was described as polycrystalline, globally centrosymmetric, and homogeneous through thickness, yet capable of generating a strong bulk damping-like spin-orbit torque (Zhu et al., 2020). This combination is central to current CoPt research: strong ferromagnetism survives both long-range chemical order and chemical disorder, but the resulting anisotropy and transport differ sharply between these limits.

A recurrent misconception is that ferromagnetic CoPt should be equated only with hard-magnetic $L1_0$8 order. The published record does not support that restriction. Ordered $L1_0$9 CoPt is indeed a canonical hard ferromagnet, but chemically disordered $L1_1$0 CoPt can also be perpendicularly magnetized and can act as a single-layer spin-torque material (Karoui et al., 2012, Chen et al., 2021, Zhu et al., 2020).

2. Magnetic anisotropy from bulk crystals to nanoscale structures

Magnetic anisotropy is the most persistent organizing theme in ferromagnetic CoPt. In thin-film calculations and defect studies, the anisotropy energy is written as

$L1_1$1

with positive MAE denoting out-of-plane preference and negative MAE denoting in-plane preference (Brahimi et al., 2019). For pristine $L1_1$2 CoPt, large uniaxial anisotropy is a bulk characteristic rather than a small perturbation. A Néel-type nanoparticle analysis gives the bulk $L1_1$3 value $L1_1$4 J/m$L1_1$5 for CoPt (Rohart et al., 2011), while a first-principles study of $L1_1$6-ordered (111)-oriented junction electrodes reports $L1_1$7, tabulated as $L1_1$8 or $L1_1$9 depending on normalization (Masuda et al., 2020).

Surface termination and atomic-scale defects modify this anisotropy strongly. For tetragonal $A1$0 CoPt thin films, Pt-terminated surfaces show strong out-of-plane MAE, about $A1$1 meV per formula unit in the abstract and around $A1$2 meV in the supercell discussion, whereas Co-terminated surfaces are in-plane, about $A1$3 meV per formula unit in the abstract and around $A1$4 meV in the supercell picture (Brahimi et al., 2019). All investigated single-atom defects—Co and Fe adatoms on Pt-terminated surfaces, Pt adatoms on Co-terminated surfaces, and Co or Pt vacancies—favor an in-plane MAE contribution. Fe adatoms and Pt vacancies are strong enough to switch Pt-terminated CoPt from out-of-plane to in-plane, and the Pt vacancy produces the largest reported reduction, about $A1$5 meV (Brahimi et al., 2019). The same work interprets these trends through orbital moment anisotropy via Bruno’s relation,

$A1$6

emphasizing the special role of Pt because of its stronger spin–orbit coupling (Brahimi et al., 2019).

Finite size introduces a second layer of complexity. In CoPt nanoparticles, chemically disordered clusters can show MAE as high as $A1$7 J/m$A1$8 through a statistical finite-size effect, even though disordered bulk CoPt has essentially zero MCA by symmetry (Rohart et al., 2011). By contrast, $A1$9-ordered clusters inherit the large intrinsic uniaxial anisotropy of the ordered phase, but the surface reduces it. The fitted surface anisotropy for ordered CoPt nanoparticles is negative, with $L1_0$0 and $L1_0$1 mJ/m$L1_0$2, and for fewer than about 2000 atoms the MAE decrease exceeds 10% relative to bulk (Rohart et al., 2011). Fully relativistic DFT on $L1_0$3 cuboctahedral clusters with $L1_0$4 further shows that CoPt remains ferromagnetic after relaxation, that the total moment per atom stabilizes near $L1_0$5–$L1_0$6, and that the $L1_0$7 cluster exhibits a small in-plane MAE contribution rather than purely axial behavior (Cuadrado et al., 2012).

Temperature can rotate the easy axis as well. For CoPt/AlN multilayer thin films modeled with a third-order perturbed Heisenberg Hamiltonian, the spin reorientation temperatures are $L1_0$8 K, $L1_0$9 K, and $a = 3.81\,\text{\AA}$0 K for $a = 3.81\,\text{\AA}$1, $a = 3.81\,\text{\AA}$2, and $a = 3.81\,\text{\AA}$3, respectively, and the transition is reported to be highly sensitive to the second-order magnetic anisotropy constant $a = 3.81\,\text{\AA}$4 (Samarasekara et al., 2016). Taken together, these results show that ferromagnetic CoPt is intrinsically anisotropic, but the magnitude and even the sign of its anisotropy remain highly contingent on ordering, termination, defect chemistry, size, and temperature.

3. Self-switching CoPt and current-induced magnetization reversal

A major contemporary theme is that CoPt can function as a “self-switching” ferromagnet: the ferromagnetic layer itself generates the torque needed to reverse its magnetization, without an added heavy-metal spin source. The clearest early evidence came from chemically disordered fcc CoPt single layers, where a strong bulk damping-like spin-orbit torque was detected by harmonic Hall and ST-FMR methods. For $a = 3.81\,\text{\AA}$5 nm CoPt, the measured damping-like efficiency is approximately

$a = 3.81\,\text{\AA}$6

and the signal is nearly unchanged across MgO/CoPt/MgO, Hf/CoPt/Hf, Tb/CoPt/Tb, Ru/CoPt/Ru, and SiO$a = 3.81\,\text{\AA}$7/CoPt/MgO, while increasing monotonically with CoPt thickness above about $a = 3.81\,\text{\AA}$8 nm (Zhu et al., 2020). The same study therefore identifies the torque as bulk rather than interfacial and attributes it most likely to a strong spin Hall effect within ferromagnetic CoPt itself (Zhu et al., 2020).

Direct current-induced magnetization switching was then demonstrated in a chemically disordered $a = 3.81\,\text{\AA}$9 CoPt single layer with a composition gradient along the film normal (Chen et al., 2021). In that system, the composition-gradient film has $c/a = 0.976$0 Oe A$c/a = 0.976$1 m$c/a = 0.976$2 and $c/a = 0.976$3, compared with $c/a = 0.976$4 Oe A$c/a = 0.976$5 m$c/a = 0.976$6 and $c/a = 0.976$7 for uniform Co$c/a = 0.976$8Pt$c/a = 0.976$9, establishing that broken inversion symmetry from the gradient enhances the torque (Chen et al., 2021). The switching in that work requires an in-plane assist field, so it is not yet field-free (Chen et al., 2021).

Field-free reversal was subsequently advanced by substrate symmetry engineering. A simple alloyed multilayer stack, Pt(0.8)/Co(0.3)/Pt(0.6)/Co(0.5)/Pt(0.4)/Co(1) (nm), interdiffuses into an approximately $1.14\,\mu_B/$0 nm (111)-oriented CoPt layer with strong perpendicular magnetic anisotropy (Yongming et al., 2023). On flat Al$1.14\,\mu_B/$1O$1.14\,\mu_B/$2(0001) substrates, switching remains incomplete without an assist field; on vicinal substrates with miscut angles $1.14\,\mu_B/$3, especially $1.14\,\mu_B/$4, the switching becomes deterministic and 100% full-scale (Yongming et al., 2023). The decisive change is a tilt of the magnetic easy axis by about $1.14\,\mu_B/$5 in the $1.14\,\mu_B/$6 plane along the atomic-step direction. The uniaxial energy analysis is written as

$1.14\,\mu_B/$7

with equilibrium condition

$1.14\,\mu_B/$8

Micromagnetic simulations show that the field-free switching originates from this tilted anisotropy when the easy-axis tilt is perpendicular to the current, while DMI mainly assists by lowering the critical switching current density (Yongming et al., 2023). Experimentally, the zero-field Hall swing reaches $1.14\,\mu_B/$9, compared with the field-driven maximum $0.3\,\mu_B/$0, corresponding to about a 100% switching ratio; the effective DMI threshold $0.3\,\mu_B/$1 rises from about $0.3\,\mu_B/$2 Oe on a flat substrate to about $0.3\,\mu_B/$3 Oe at $0.3\,\mu_B/$4, and the damping-like efficiency increases by about 25% (Yongming et al., 2023).

A distinct symmetry route to field-free switching was later reported for epitaxial $0.3\,\mu_B/$5 nm CoPt single layers on MgO(111) with $0.3\,\mu_B/$6 symmetry (Chen et al., 21 Nov 2025). There, crystal spin-orbit torque (CSOT) produces nearly 100% field-free perpendicular switching at room temperature, with the optimized growth temperature $0.3\,\mu_B/$7 yielding a switching ratio of 98% and an out-of-plane effective field

$0.3\,\mu_B/$8

The current-angle dependence follows a $0.3\,\mu_B/$9 law, reflecting the threefold crystal symmetry (Chen et al., 21 Nov 2025). These results make clear that “field-free switching in CoPt” is not a single mechanism. Depending on the structural realization, the dominant ingredient can be bulk damping-like torque, inversion-symmetry breaking by a composition gradient, tilted magnetic anisotropy from vicinal steps, or a crystal-symmetry-allowed CSOT term (Zhu et al., 2020, Chen et al., 2021, Yongming et al., 2023, Chen et al., 21 Nov 2025).

4. Hall effects, Berry curvature, and symmetry-controlled transport

Ferromagnetic CoPt is also a model system for the relation between anomalous Hall, spin Hall, and crystal-symmetry-governed transport responses. In $A1$00-CoPt, first-principles calculations of Berry and spin Berry curvature show a symmetry mismatch: the Berry curvature preserves the tetragonal $A1$01 symmetry, whereas the spin Berry curvature is reduced to $A1$02 (Qu et al., 2019). Using the Kubo formalism, that work obtained

$A1$03

and concluded that intrinsic anomalous Hall conductivity does not simply scale with spin Hall conductivity through a Fermi-level spin-polarization factor when opposite-spin band crossings occur near the Fermi level (Qu et al., 2019). This directly challenges a common simplified two-current picture.

A related first-principles treatment of the spin anomalous Hall effect in $A1$04-type Pt-based ferromagnets found that CoPt occupies a regime of relatively large SAHE response (Miura et al., 2021). For CoPt at $A1$05, the calculated values are $A1$06, $A1$07, $A1$08, and $A1$09, with the table listing $A1$10 while the text emphasizes a large absolute Hall spin polarization for CoPt in comparison with FePt (Miura et al., 2021). The mechanism is traced to a negative spin-down–spin-down contribution to the anomalous Hall conductivity, originating from Pt-derived minority-spin anti-bonding states near $A1$11 (Miura et al., 2021).

In the $A1$12 single-layer geometry, CoPt supports two unconventional crystal-symmetry-driven effects simultaneously: CSOT and the crystal in-plane Hall effect (CIHE) (Chen et al., 21 Nov 2025). The effective-field term for the CSOT includes the $A1$13-specific contribution

$A1$14

which reduces under the switching condition to $A1$15, and the CIHE obeys the same threefold angular law, $A1$16 (Chen et al., 21 Nov 2025). The reported CIHE coefficient is

$A1$17

and both CSOT and CIHE peak near $A1$18, which the authors connect to a common origin in trigonal warping of the Fermi surface (Chen et al., 21 Nov 2025).

Ultrafast transport adds a further layer. In CoPt/Cu/MgO heterostructures, femtosecond excitation generates orbital accumulation within CoPt, which propagates through Cu and is converted into charge current at the Cu/MgO interface by the inverse orbital Rashba–Edelstein effect, yielding terahertz emission comparable to efficient spintronic emitters (Liu et al., 2024). The signal is strongest for Co$A1$19Pt$A1$20; it peaks at a Cu thickness of $A1$21 nm, shows a delay exceeding $A1$22 fs for Co$A1$23Pt$A1$24, gives an extracted orbital group velocity of about $A1$25 nm/fs, and fits an orbital decoherence length in Cu of about $A1$26 nm (Liu et al., 2024). The fact that a CoPt single layer gives no sizeable THz emission in that study rules out a simple anomalous-Hall-emission interpretation and supports an orbital-current mechanism (Liu et al., 2024).

5. Electrical, ionic, and interfacial control of CoPt magnetism

The magnetic state of ferromagnetic CoPt is unusually susceptible to interfacial charge control. At a ferroelectric ZnO interface, first-principles calculations show that reversing ZnO polarization changes the CoPt magnetic anisotropy from perpendicular to in-plane (Al-Mahdawi et al., 2018). The computed MAE values are $A1$27 meV/atom $A1$28 for the $A1$29 state and $A1$30 meV/atom $A1$31 for the $A1$32 state, corresponding to $A1$33 (Al-Mahdawi et al., 2018). Experimentally, CoPt/Mg-doped ZnO/Co junctions yield $A1$34 in $A1$35 and $A1$36 in $A1$37, so $A1$38, while the effective anisotropy field changes from $A1$39 kOe to $A1$40 kOe (Al-Mahdawi et al., 2018). The reported microscopic origin is a polarization-driven shift in Co/Pt hybridized interfacial states under strong Pt spin–orbit coupling, accompanied by a small interface charge change $A1$41 and a reversal of orbital moment anisotropy (Al-Mahdawi et al., 2018).

A related charge-control geometry uses MgZnO as a ferroelectric-like Schottky barrier. In epitaxial Co$A1$42Pt$A1$43/Mg$A1$44Zn$A1$45O/Cu heterojunctions, resistive switching is associated with a Schottky barrier-height change of $A1$46 meV at the CoPt/MgZnO interface (Belmoubarik et al., 2023). The reversed polarization is estimated as $A1$47 at $A1$48 K and $A1$49 at $A1$50 K, and stack-A gives a ferroelectric Curie temperature $A1$51 K (Belmoubarik et al., 2023). The current is modeled by thermionic emission through a back-to-back Schottky diode, and the authors explicitly argue against a filamentary mechanism (Belmoubarik et al., 2023). This suggests that CoPt can act simultaneously as a ferromagnetic electrode and as a spin-active, charge-sensitive interface.

Electrochemical control extends the tunability from anisotropy to multilevel magnetic plasticity. In a CoPt-based electrochemical transistor with a Ta (2 nm) / Co$A1$52Pt$A1$53 (9 nm) Hall-bar channel and DEME-TFSI ionic liquid, gate-driven ion insertion modulates both conductance and coercivity at room temperature (Li et al., 2022). The CoPt film shows an anomalous Hall loop under out-of-plane field, confirming perpendicular ferromagnetism. Negative gate bias increases coercivity and positive gate bias decreases it; the reported representative states are $A1$54 Oe and $A1$55 Oe, five levels can be distinguished in the multilevel $A1$56 modulation, and after sufficient training the AHE loops show no variation for at least $A1$57 s (Li et al., 2022). The same device also demonstrates paired-pulse facilitation, with a maximum PPF of $A1$58 at a $A1$59 s interval, linking ionic control of ferromagnetic CoPt to spin-based neuromorphic functionality (Li et al., 2022).

6. Exchange-coupled architectures and application domains

Ferromagnetic CoPt is a versatile component in exchange-coupled multilayers and device stacks. In CoPt/Pt/Co structures built from a periodic CoPt multilayer $A1$60 and a thick Co(10 nm) layer, the CoPt subsystem behaves as an easy-axis ferromagnet and the thick Co layer as an easy-plane ferromagnet (Demidov et al., 2016). MOKE and FMR show a non-collinear equilibrium state and strong exchange coupling for total Pt separation $A1$61 nm, with a fitted ferromagnetic interlayer exchange constant $A1$62; when the total Pt spacer is increased to $A1$63 nm, the coupling becomes much weaker (Demidov et al., 2016). The fitted anisotropy constants are $A1$64 for the CoPt multilayer Co layers and $A1$65 for the thick Co layer (Demidov et al., 2016).

In sputtered hard–soft CoPt/Co bilayers, the hard layer is L1$A1$66-ordered Co$A1$67Pt$A1$68 of thickness $A1$69 nm and the soft layer is Co of thickness $A1$70 nm (Alexandrakis et al., 2016). There, the exchange-bias field of the soft layer depends linearly on the hard-layer magnetization,

$A1$71

whereas the coercivity follows

$A1$72

Monte Carlo analysis attributes the coercivity primarily to domain-wall pinning by the tunable randomness of the hard CoPt layer’s domain state rather than to coherent rotation (Alexandrakis et al., 2016). This is a distinct use of ferromagnetic CoPt: not as a switching layer, but as a controllable random bias source for another ferromagnet.

The combination of strong anisotropy and favorable interfacial states also makes CoPt attractive in tunnel junctions. For (111)-oriented $A1$73-ordered CoPt/MgO/CoPt junctions, first-principles calculations give a TMR ratio of $A1$74 for a 7-monolayer MgO barrier (Masuda et al., 2020). The interfacial Co–O configuration is lowest in energy, and minority-spin interface resonant tunneling through Co–O antibonding states near $A1$75 is identified as the origin of the large TMR (Masuda et al., 2020). In the same framework, CoPt has the largest PMA among the studied $A1$76 alloys, with $A1$77 near $A1$78 and a dominant positive contribution from spin-conserving Pt-mediated spin–orbit processes near the Fermi level (Masuda et al., 2020).

Beyond memory and transport, ferromagnetism in CoPt can act as a chemical control parameter. In $A1$79-CoPt used as an oxygen-reduction catalyst model, self-consistent moments are about $A1$80 on Co and $A1$81 on Pt, with the easy axis along 001. The hollow site on CoPt(001) binds atomic oxygen most strongly at $A1$82 eV, and turning off Co spin polarization weakens O binding by about $A1$83 meV (Allen et al., 29 Aug 2025). For O$A1$84, Co magnetism strengthens adsorption by about $A1$85 eV per O, and thinner Pt overlayers increase the magnetic influence of buried Co on adsorption and dissociation energetics (Allen et al., 29 Aug 2025). The same study therefore presents ferromagnetic CoPt as a catalyst whose exchange-split surface electronic structure directly modifies oxygen chemistry (Allen et al., 29 Aug 2025).

Across these architectures, a consistent pattern emerges. Ferromagnetic CoPt can be a hard anisotropic storage medium, a bulk torque generator, a symmetry-sensitive Hall and orbital-current material, a ferroelectrically or electrochemically gated magnetic interface, a biasing layer in exchange-coupled stacks, and even a magnetism-assisted catalytic surface (Demidov et al., 2016, Masuda et al., 2020, Allen et al., 29 Aug 2025). A plausible implication is that the unusually broad utility of CoPt follows from the same microscopic ingredients recurring across all scales: strong Co–Pt hybridization, substantial Pt spin–orbit coupling, and a magnetic energy landscape that is exceptionally sensitive to symmetry, termination, and interfacial charge.