Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

# Spin pumping during the antiferromagnetic–ferromagnetic phase transition of iron–rhodium

## Abstract

FeRh attracts intensive interest in antiferromagnetic (AFM) spintronics due to its first-order phase transition between the AFM and ferromagnetic (FM) phase, which is unique for exploring spin dynamics in coexisting phases. Here, we report lateral spin pumping by which angular momentum is transferred from FM domains into the AFM matrix during the phase transition of ultrathin FeRh films. In addition, FeRh is verified to be both an efficient spin generator and an efficient spin sink, by electrically probing vertical spin pumping from FM-FeRh into Pt and from Py into FeRh, respectively. A dramatic enhancement of damping related to AFM-FeRh is observed during the phase transition, which we prove to be dominated by lateral spin pumping across the FM/AFM interface. The discovery of lateral spin pumping provides insight into the spin dynamics of magnetic thin films with mixed-phases, and the significantly modulated damping advances its potential applications, such as ultrafast spintronics.

## Introduction

The B2-ordered material FeRh undergoes a first-order phase transition from the antiferromagnetic (AFM) to the ferromagnetic (FM) phase in the vicinity of room temperature accompanied by a lattice expansion of about 1% (ref. 1). FeRh has stood out as a highly intriguing material for applications in heat-assisted magnetic recording2,3. Nowadays, with the rapid development of AFM spintronics aimed at low power and ultrafast logic devices4,5,6,7,8,9, FeRh attracts renewed interest as a unique AFM due to its particular physical properties during the phase transition. Benefiting from the ability to grow thin films of high quality, intensified functional devices based on the phase transition of FeRh have been proposed10,11,12,13,14,15. Fundamentally, a closer insight into spin-dependent transport in its AFM state and the related spin dynamics during the phase transition are crucial issues. Although X-ray magnetic circular dichroism (XMCD)13,16,17,18, ferromagnetic resonance (FMR)19,20,21, and time-resolved methods13,14 have been adopted to study magnetization dynamics, direct experimental observation of spin generation and detection in FeRh is highly desired.

AFM metals with large spin–orbit coupling (SOC) can act as spin current detectors via the inverse spin Hall effect (ISHE), where a detectable transverse charge current is converted from a pure spin current22, demonstrating that AFM metals act similarly as heavy metals such as Pt, Ta, or W6,7,8. Spin pumping from FM across/into AFM has been reported23,24,25,26, where enhanced spin pumping efficiency around the AFM magnetic phase transition temperature TNéel has been observed24. In comparison with FM/AFM heterostructures with explicit vertical spin pumping, FeRh is unique concerning its magnetic structure, since AFM and FM matrixes/domains embed in the bulk during the first-order phase transition3,17,27, making it an ideal candidate to investigate spin dynamics at the interface of FM and AFM domains in a single material. In this intriguing case, a landscape of lateral spin pumping accompanied by the transfer of angular momentum from FM domains into AFM surroundings is presented. In addition to the previously reported non-equilibrium non-local transfer of angular momentum by optically excited electrons within the ferrimagnetic GdFeCo alloy28, lateral spin pumping within the equilibrium magnetization precession regime29, e.g., excited by FMR, has remained unexplored.

In this work, besides vertical spin pumping from FeRh into the adjacent normal metal (Pt) which is detected by ISHE voltages, we mainly report the experimental observation of lateral spin pumping from FM- into AFM-FeRh across its phase transition, evidenced by an enhanced linewidth and damping. The crucial role of AFM-FeRh as a spin detector is further confirmed by vertical spin pumping experiments from Py into FeRh. In addition to the low damping of ~0.0023 in the FM state which makes it competitive for magnonic and spin-orbitronic applications30, the greatly enhanced damping related to the AFM phase during the phase transition opens prospects towards ultrafast logic devices. The observation of lateral spin pumping and modulated damping are fundamentally significant for understanding the spin dynamics in FeRh and might advance its practical application in AFM spintronics.

## Results

### Magnetization properties of FeRh during phase transition

The bilayers studied here consist of FeRh(t)/Pt and FeRh(t)/Al, where t = 5 or 10 nm. α′-FeRh thin films were grown epitaxially on MgO (001) substrates by magnetron sputtering15,31. As capping materials we use either a 4-nm-thick Pt layer with strong SOC or an Al layer with weak SOC. As shown in Fig. 1, bilayers are fabricated into 6 µm wide and 300 µm long stripes which are placed in the gap between the signal line and ground planes of a coplanar waveguide (CPW). This configuration leads to out-of-plane rf-field excitation of the FeRh stripes32. The devices are placed in an external magnetic field which can be rotated both in the in-plane (φH) and out-of-plane (θH) directions. To provide an overview of the magneto-structural transition in FeRh, Fig. 1c presents the magnetization of full films measured by SQUID and the four-point resistance of corresponding stripes (10 µm × 60 µm) etched out of the same films (see Supplementary Note 1 for FeRh/Al). The T-dependent magnetization loops clearly demonstrate the phase transition between AFM and FM-FeRh, where the transition temperatures locate between 300 and 380 K for different samples. Since the AFM and FM states of FeRh contribute differently to the transport properties, an abrupt change in resistance is observed during the phase transition30,33.

### Vertical spin pumping in FeRh/Pt across phase transition

In order to investigate the ISHE due to spin pumping and separate parasitic anisotropic magnetoresistance (AMR) from pure ISHE, spin pumping voltages as a function of in-plane rotation angle φH are studied. The device configuration with the stripe located in the gap of the CPW (out-of-plane excitation) has great advantages for easy separation of AMR and ISHE signals through analyzing the angle-dependent signals32,34. Note that finite inductive currents induced in the conductive Pt (Al) layers can lead to an additional in-plane magnetic driving field, which can be accurately determined by analyzing the dc voltages. A typical voltage trace (offset voltage Voffset is subtracted) for an excitation frequency of 11 GHz measured at 400 K and φH = 45° is shown in Fig. 2a. To fit the lineshape of the spectra, symmetric voltage components VsymLsym (red line) and antisymmetric ones VasymLasym (blue line) are introduced. The antisymmetric part is solely a consequence of AMR following Lasym = −4ΔH(HHR)/[4(HHR)2 + ΔH2], where HR is the resonance field and ΔH is the linewidth (full-width at half-maximum, FWHM). In turn, the symmetric part contains both AMR and ISHE signals, and the lineshape can be described by Lsym = ΔH2/[4(HHR)2 + ΔH2] (ref. 34). The antisymmetric and symmetric d.c. voltage amplitudes at FMR as a function of angle φH are shown in Fig. 2c, d, which can be fitted by (for a detailed derivation see Supplementary Note 2).

$$V_{{\mathrm{a}} {\mbox{-}} {\mathrm{sym}}} = V_{{\mathrm{AMR}}}^{\mathrm{I}} + V_{{\mathrm{AMR}}}^{\mathrm{O}} = C^{\mathrm{I}}\sin ^2\varphi \cos \varphi + C^{\mathrm{O}}\sin 2\varphi ,$$
(1)
$$V_{{\mathrm{sym}}} =\,\, V_{{\mathrm{AMR}}/{\mathrm{ISHE}}}^{\mathrm{I}} + V_{{\mathrm{AMR}}}^{\mathrm{O}} + V_{{\mathrm{ISHE}}}^{\mathrm{O}} = D^{\mathrm{I}}\sin ^2\varphi \cos \varphi \\ + D^{\mathrm{O}}\sin 2\varphi + E^{\mathrm{O}}\cos \varphi .$$
(2)

Here CI (CO) is the coefficient of Va-sym due to in-plane (out-of-plane) excitation, DI (DO) the coefficient of Vsym due to in-plane (out-of-plane) excitation, and EO the magnitude of VISHE. Note that φ is the angle of the magnetization with respect to the axis of FeRh [110], which is assumed to be equal to φH due to the fact that the magnitude of in-plane magnetic anisotropic fields are much smaller than HR. From the fitting, sizeable VISHE can be obtained, which shows opposite polarities at φH = 0° and 180° (Fig. 2b). This clearly verifies spin pumping from FeRh into Pt. In contrast to FeRh/Pt, no ISHE signals are detected for FeRh capped with Al (Supplementary Note 2) due to the much smaller spin Hall angle of Al (ref. 32).

We now turn towards vertical spin pumping during the AFM–FM phase transition. Figure 3a shows the ISHE voltages of FeRh (10)/Pt during heating and cooling at 11 GHz, φH = 0°. The voltages decrease continuously during the transition into the AFM state from 320 to 300 K. The traces of VISHE and linewidth (Fig. 3b) upon cooling and heating show an open window, which are consistent with the hysteretic magnetization/resistance traces shown in Fig. 1c. Similar tendencies have also been detected for thinner FeRh (5) films (see Supplementary Note 3). To have a better understanding of this phenomenon, a schematic diagram of the spin pumping scenario for FeRh/Pt during the phase transition is illustrated in Fig. 1a. Upon heating, FM domains nucleate in the AFM matrix and grow, until the whole sample is FM17,27. Conversely, cooling down from the FM state reverses the phase transition process and AFM domains nucleate in an FM matrix. It is known that, during the first-order phase transition of FeRh, the microscopic magnetization of separate FM-FeRh domains remain in the same magnetic state as the fully FM phase; it has been verified by XMCD that the orbital to spin moment ratio for Fe/Rh remains constant during the phase transition16. Thus, the reduction of the proportion of FM domains, together with the decreasing interfacial contact area between FM-FeRh pillars and the Pt film, will directly lead to a decrease of the vertically injected pure spin current and consequently the detectable voltages in Pt. Quantitatively, the measured data can be well reproduced by calculation results (solid lines in Fig. 3a) based on the dynamic magnetic susceptibility33, clarifying the process of the first-order phase transition in FeRh.

### Magnetic anisotropy

To explicate the change of magnetic susceptibility during the phase transition, which is the basis for exploring the spin dynamics of FeRh, we investigate the T dependence of the magnetic anisotropies. Figure 4a depicts the in-plane angular (φH) dependence of the resonance field μ0HR for FeRh (10)/Pt at 11 GHz and 400 K. The data can be well explained by considering the effective demagnetization field $$\mu _0M_{{\mathrm{eff}}} = \mu _0M_{\mathrm{S}}-\mu _0H_ \bot = \mu _0M_{\mathrm{S}}-2K_{\mathrm{U}}^ \bot /M_{\mathrm{S}}$$ of 1.37 T (with H the perpendicular magnetic anisotropy field and $$K_{\mathrm{U}}^ \bot$$ the corresponding uniaxial out-of-plane anisotropy energy constant), a cubic in-plane biaxial anisotropy field μ0HB (along [100] or [010] directions) of 6.03 mT and a uniaxial anisotropy field μ0HU (along [$$\bar 1$$10]) of 5.23 mT (see Supplementary Note 4). When T decreases to 310 K in the phase transition region (Fig. 4b), the values of μ0Meff, μ0HB, and μ0HU increase due to the increase of magnetization. The T dependences of μ0Meff(T) and $$K_{\mathrm{U}}^ \bot (T)$$ obtained from the frequency-dependent μ0HR along the in-plane easy axis (φH = 45°) are summarized in Fig. 4c. The increase of μ0Meff(T) during cooling is simply related to the increase of the saturation magnetization μ0MS(T) of FM-FeRh, which increases with decreasing temperature. Accordingly, the T dependence of $$K_{\mathrm{U}}^ \bot (T)$$ can be obtained from the experimental data for μ0Meff(T) in combination with μ0MS(T), which also exhibits an increase towards lower T. Below 320 K when FeRh goes into the phase transition region (as indicated by the arrows in Fig. 4c), no abrupt change of μ0Meff and $$K_{\mathrm{U}}^ \bot$$ is observed, indicating again that the magnetic anisotropy of FeRh pillars remain, basically, unaffected during the phase transition. Similar behavior is observed for the heating up process (Supplementary Note 4).

### Detection of spin pumping voltages in FeRh

To confirm the role of AFM-FeRh as an efficient spin sink and spin detector, we also investigate vertical spin pumping in an FeRh (10)/Py (15) bilayer capped with Al. FeRh in this bilayer also exhibits a reversible phase transition (Supplementary Note 5). As shown in Fig. 5a, b, clear VISHE features, i.e., symmetric spectra with opposite polarities at φH = 0°/180°, have been observed at 300 and 400 K. This unambiguously proves that the spin current is pumped from Py into FeRh, either into the FM or the AFM state, and transforms into a charge current due to the ISHE of FeRh. Moreover, a clear signal induced by spin pumping from FM-FeRh into Py has also been observed (400 K). Since this is not in the focus for the present discussion, we ignore it here. The crucial role of FeRh as a spin sink is further evidenced by the enhanced linewidth and damping compared to that of pure Py (see Supplementary Note 5). The spin Hall angle θSHE of FeRh can be quantified by

$$V_{{\mathrm{ISHE}}} = RI_{\mathrm{C}} = R\frac{{2{{e}}}}{\hbar }\theta _{{\mathrm{SHE}}}J_{\mathrm{S}}w\lambda \tanh \frac{d}{{2\lambda }},$$
(3)

where R is the resistance of the FeRh/Py bilayer, IC the charge current induced by the ISHE, e the electronic charge, ħ the Dirac constant, w the width of the stripe, λ the spin diffusion length of FeRh, d the thickness of FeRh, and JS the magnitude of spin current at the interface. Instead of θSHE, we use the product of θSHE and λ, θSHEλ in unit of nm, to determine the spin-to-charge efficiency35. This is based on the condition that λ of FeRh is assumed to be much smaller than d, which holds also for other AFMs23. Based on the temperature-dependent VISHE shown in Fig. 5c, the magnitude of θSHEλ of FeRh as a function of temperature is plotted in Fig. 5d (for details see Supplementary Note 5). Note that the data of FeRh around 320 K have been excluded due to the scattering of the linewidth, probably related to inhomogeneities at the FeRh/Py interface. The magnitude of θSHEλ is comparable to heavy metals22, due to the strong SOC arising from the 4d electrons of Rh in FeRh. A ~2 times larger θSHEλ is found in the AFM phase, indicating that ΑFM-FeRh can be an efficient spin sink. This result, in turn, supports the crucial idea of the lateral spin pumping in single FeRh layer during its phase transition.

### Lateral spin pumping from FM-FeRh domains into AFM matrix

Apart from vertical spin pumping from FeRh into the Pt capping layer, lateral spin pumping from FM-FeRh into the AFM-FeRh surroundings during the phase transition can be expected. As illustrated in Fig. 1a, the side walls of the pillars provide channels for lateral spin pumping from the FM domains into the AFM matrix. Even though in this configuration injection of a spin current can be expected, it cannot be detected through a voltage across the FeRh/Pt layer: for the case of in-plane magnetization, a voltage due to the ISHE can only be expected in the out-of-plane direction22,36, which will average out due to the symmetric distribution of the injected spin current. However, a pronounced increase of linewidth is observed during the phase transition (Fig. 3b), indicating possible lateral spin pumping from FM-FeRh to AFM-FeRh.

It is known that both intrinsic and extrinsic effects can contribute to the T dependence of the linewidth. In addition to spin pumping, other extrinsic factors, including two-magnon scattering, inhomogeneity, and mosaicity broadening, could also enhance the experimental linewidth. Particularly along the in-plane easy axis (φH = 45°), mosaicity broadening is negligible and two-magnon scattering is minimized (see Supplementary Note 6). However, before we can conclude that the increase in μ0ΔH mainly arises from lateral spin pumping, other extrinsic effects contributing to the FMR linewidth must be excluded, which will be discussed below by analyzing the frequency dependence of μ0ΔH.

Through investigating the frequency-dependent linewidth of the spin pumping voltage spectra at different T, we are able to quantitatively study spin dynamics—here the effective damping—of FeRh during the phase transition. In the case of uniform magnetization, the FMR linewidth as a function of microwave frequency f for Gilbert damping (α) is given by37,38

$$\mu _0\Delta H = \mu _0{\mathrm{\Delta }}H_0 + 2\alpha \frac{{2{\uppi}f}}{\gamma }.$$
(4)

The damping calculated by linear fitting of the frequency-dependent μ0ΔH is called effective damping (αeff, since two-magnon scattering is included here37,39), which gives the upper limit for intrinsic damping. μ0ΔH0 represents the inhomogeneous term which is independent of f (also called zero frequency intercept). As plotted in Fig. 6a, linear fitting of f vs. μ0ΔH is adopted to allow a comparison between different T. Figure 6c summarizes the value of the calculated αeff at different T during heating (solid spheres) and cooling (solid squares). It is interesting to note that αeff is greatly enhanced during the phase transition, indicating a different dynamic behavior for FM/AFM domains coexisting in FeRh compared to the pure FM-FeRh. For all of the five samples with different thicknesses of FeRh and different capping metals, dramatic enhancement of αeff is observed in the temperature region of the FM–AFM phase transition (see Supplementary Note 7).

To exclude the influence of two-magnon scattering which widely exists in FM films due to, e.g., misfit dislocations or defects, the out-of-plane FMR configuration is used where two-magnon scattering is switched off37,39. Since the spin pumping voltages cannot be detected when the external field is perpendicular to the interface, we conduct a control experiment by measuring standard full film-FMR. A wide stripe sized 18 µm × 400 µm directly beneath the signal line of the waveguide is measured. From linear fitting of the f-dependent μ0ΔH in the out-of-plane FMR measurements (φH = 45°, θH = 90°) at five typical temperatures (Fig. 6b), the calculated damping values α,Pt(T) during cooling are compared to those obtained in the in-plane configuration in Fig. 6c. A larger value of αeff is obtained for in-plane spin pumping compared with α,Pt(T) for the out-of-plane FMR, and this difference is attributed to two-magnon scattering37. This is further confirmed by FMR measurements during out-of-plane rotation where φH is fixed to 45° (see Supplementary Note 8). The linewidth for the out-of-plane direction (θH = 90°) is smaller than that for the in-plane direction (θH = 0°), which clearly verifies the contribution of two-magnon scattering in the in-plane configuration37,38,39. It is worth emphasizing that by ruling out two-magnon scattering in the out-of-plane FMR measurements, α,Pt(T) still increases from 0.0035 to 0.0147 by a factor of 4 (open rhombus in Fig. 6c) when cooling from the FM to the FM–AFM coexisting states. The enhanced damping indicates the experimental observation of lateral spin pumping, through careful separation of different contributions in the following discussion.

## Discussion

For FeRh/Pt bilayers, the total effective damping αeff(T) obtained from in-plane spin pumping measurements is composed of four parts:

$$\alpha _{{\mathrm{eff}}}\left( T \right) = \alpha _{{\mathrm{Gilbert}}}\left( T \right) + \alpha _{{\mathrm{Vertical}}\;{\mathrm{SP}}}\left( T \right) + \alpha _{{\mathrm{Lateral}}\;{\mathrm{SP}}}\left( T \right) + \alpha _{{\mathrm{2M}}}\left( T \right),$$
(5)

where αGilbert(T) is the intrinsic Gilbert damping, αVertical SP(T) the vertical spin pumping from FeRh into Pt, αLateral SP(T) the damping contribution by lateral spin pumping from FM-FeRh into AFM-FeRh, and α2M(T) the damping contribution by two-magnon scattering. In comparison, the obtained damping from out-of-plane FMR (α(T), open rhombus in Fig. 6c) excludes α2M(T): α,Pt(T) = αGilbert(T) + αVertical SP(T) + αLateral SP(T), but also increases dramatically during the phase transition. For FeRh/Al at 400 K, α,Al (0.0023) is smaller than α,Pt (0.0035) due to the lack of vertical spin pumping (αVertical SP). Thus the value of αVertical SP is estimated to be 0.0012 and is basically a constant during the whole temperature range (see Supplementary Note 8), as shown by the dashed line in Fig. 6d. This is as expected because the interface properties between FeRh and Pt remain essentially unchanged during the phase transition, in spite of the reduced contacting areas. The effective spin mixing conductance $$g_{{\mathrm{eff}}}^{ \uparrow \downarrow }$$ is estimated to be 9.20(±0.53) × 1018 m–2, which is comparable to the magnitude of other FM/heavy metal bilayer structures29,32,38.

To highlight the crucial role of lateral spin pumping for the enhancement of damping, the intrinsic Gilbert-like damping αGilbert(T) as a function of T is discussed first. Based on the microscopic theory of Gilbert damping which originates from spin–orbit scattering of band electrons in FM metals, αGilbert(T) is expected to be quantitatively related to the out-of-plane magnetic anisotropy, which is also a second-order effect of the spin–orbit interaction40,41. When the whole FeRh film is in the FM state where lateral spin pumping is absent, αGilbert(T= 350 ~ 400 K) is experimentally obtained according to α,Pt(T) = αGilbert(T) + αVertical SP(T). It scales linearly with the uniaxial out-of-plane anisotropy $$K_{\mathrm{U}}^ \bot (T)$$ (see Supplementary Note 4), as already demonstrated in many FM systems41. Then, αGilbert(T) in the whole temperature range across the transition temperature can be extrapolated based on the experimentally determined $$K_{\mathrm{U}}^ \bot (T)$$ (Fig. 4c), as shown by the empty squares in Fig. 6d. As calculated, the intrinsic damping αGilbert increases from 0.0024 to 0.0031 by ~29.2%, in good agreement with the values reported previously30.

It is worth emphasizing that the slight increase of αGilbert is consistent with the experimental resistance change during the phase transition. For a common FM, the damping will increase towards high (low) temperature as a consequence of interband (intraband) scattering, which is called resistivity-like (conductivity-like) behavior30. Specifically, for the case of FeRh where the phase transition occurs above 300 K, interband scattering is expected to dominate and the Gilbert damping αGilbert(T) is generally proportional to the electrical resistivity ρ (ref. 30). So αGilbert(T) is expected to increase by 28.9% during the phase transition, which is consistent with the derived result based on $$K_{\mathrm{U}}^ \bot$$. In addition to the different resistance contributions from FM- and AFM-FeRh, the resistance change during the phase transition also reflects scattering at the FM/AFM interface and lattice expansion induced electron–phonon scattering. Consequently, compared to the total increase of more than 400% for α,Pt(T) or αeff(T), the increase of αGilbert(T) during the phase transition can be ignored. Thus the T dependence of Gilbert damping and electron–phonon scattering cannot explain the sharp increase of the effective damping during the phase transition.

Importantly, the T-dependent αLateral SP(T) due to lateral spin pumping can be obtained by subtracting αGilbert(T) and αVertical SP(T) from α,Pt(T), as plotted by the red spheres in Fig. 6d. From the sharp increase of αLateral SP(T) below 320 K, it can be concluded that the lateral spin pumping plays a crucial role in enhancing the total effective damping. The lateral spin pumping in FM-FeRh/AFM-FeRh, where AFM-FeRh acts as a spin sink, is also supported by the electrically detected ISHE voltages in Py/FeRh bilayers caused by spin pumping from Py into AFM-FeRh (Fig. 5). In addition, we also demonstrate the contribution from two-magnon scattering, where α2M(T) is estimated by subtracting α,Pt(T) from αeff(T), according to Eq. (5). As expected, α2M(T) (blue triangles in Fig. 6d) exhibits a clear increase during the phase transition, due to the increased inhomogeneity and scattering at the interface between FM and AFM domains. Although two-magnon scattering adds to the value of the total effective damping, it does not dominate compared with the dramatic enhancement of αLateral SP(T) arising from lateral spin pumping. In analogy to FeRh/Pt, a dramatic increase of damping attributed to lateral spin pumping is also observed in FeRh/Al (see Supplementary Note 8), verifying that the detection of lateral spin pumping is independent of the capping material. In summary, during the phase transition of FeRh from FM into AFM, the FMR linewidth and effective damping are crucially enhanced, indicating distinct spin dynamics related to lateral spin pumping from the FM domains into the AFM surroundings, where the interface between FM and AFM-FeRh acts as the channel. In addition, other non-local contributions including two-magnon scattering, lattice expansion induced electron–phonon scattering, and interfacial exchange coupling are also verified to be not the dominant mechanisms for the greatly enhanced damping. Thus we conclude that lateral spin pumping is the main mechanism at work and significant angular momentum is transferred from the FM domains to AFM matrix. According to the greatly enhanced damping during the phase transition of FeRh and the high efficiency of spin sinking in AFM-FeRh, the lateral spin pumping between FM- and AFM-FeRh in this case could be more efficient than in traditional FM/NM bilayers. This study provides the scientific basis for understanding spin dynamics during a first-order phase transition.

## Methods

### Sample preparation

FeRh (t = 5 nm, 10 nm)/(Pt, Al) and FeRh (10)/Py (15)/Al are deposited on MgO (001) substrates by magnetron sputtering. α′-FeRh is grown at 300 °C and then annealed at 750 °C for 1.5 h. After cooling down to room temperature, either 4 nm Pt or 4 nm Al capping layers are deposited at a base pressure of 4 × 10−7 Pa.

### Device

The film stacks are patterned into wires of 6 µm width and 300 µm length and integrated into a CPW structure, using electron beam lithography combined with Ar ion milling and lift-off process. The CPW consists of a 50 µm wide signal line and 30 µm wide gap corresponding to an impedance of 50 Ω in the GHz frequency range. The bilayers are placed in the gap between the signal line and ground planes.

### Measurements

For spin pumping measurements, microwave currents with a frequency ranging from 6 to 30 GHz are used and the input microwave power is 25 dBm. The measurements are carried out in vacuum with a base pressure of 5 × 10−5 Pa and at varied temperatures from 300 to 400 K. For full film-FMR measurements, a modulation field of 1 mT together with lock-in amplification is used to increase the signal-to-noise ratio. Note that the four-point resistance measurement on the Hall bar and the spin pumping measurements are carried out simultaneously and on the same substrate to provide a precise temperature reference.

## Data availability

The data that support the findings of this study are available from the corresponding author on request.

## References

1. 1.

Kouvel, J. S. & Hartelius, C. C. Anomalous magnetic moments and transformations in the ordered alloy FeRh. J. Appl. Phys. 33, 1343–1344 (1962).

2. 2.

Thiele, J. U., Maat, S. & Fullerton, E. E. FeRh/FePt exchange spring films for thermally assisted magnetic recording media. Appl. Phys. Lett. 82, 2859 (2003).

3. 3.

Lewis, L. H., Marrows, C. H. & Langridge, S. Coupled magnetic, structural, and electronic phase transitions in FeRh. J. Phys. D Appl. Phys. 49, 323002 (2016).

4. 4.

Marti, X. et al. Room-temperature antiferromagnetic memory resistor. Nat. Mater. 13, 367–374 (2014).

5. 5.

Wang, Y. et al. Electrical control of the exchange spring in antiferromagnetic metals. Adv. Mater. 27, 3196–3201 (2015).

6. 6.

Jungwirth, T., Marti, X., Wadley, P. & Wunderlich, J. Antiferromagnetic spintronics. Nat. Nanotechnol. 11, 231–241 (2016).

7. 7.

Železný, J., Wadley, P., Olejník, K., Hoffmann, A. & Ohno, H. Spin transport and spin torque in antiferromagnetic devices. Nat. Phys. 14, 220–228 (2018).

8. 8.

Baltz, V. et al. Antiferromagnetic spintronics. Rev. Mod. Phys. 90, 015005 (2018).

9. 9.

Wang, Z. et al. High-density NAND-like spin transfer torque memory with spin orbit torque erase operation. IEEE Electron Dev. Lett. 39, 343 (2018).

10. 10.

Cherifi, R. O. et al. Electric-field control of magnetic order above room temperature. Nat. Mater. 13, 345–351 (2014).

11. 11.

Lee, Y. et al. Large resistivity modulation in mixed-phase metallic systems. Nat. Commun. 6, 5959 (2015).

12. 12.

Liu, Z. Q. et al. Full electroresistance modulation in a mixed-phase metallic alloy. Phys. Rev. Lett. 116, 097203 (2016).

13. 13.

Radu, I. et al. Laser-induced generation and quenching of magnetization on FeRh studied with time-resolved x-ray magnetic circular dichroism. Phys. Rev. B 81, 104415 (2010).

14. 14.

Mariager, S. O. et al. Structural and magnetic dynamics of a laser induced phase transition in FeRh. Phys. Rev. Lett. 108, 087201 (2012).

15. 15.

Chen, X. Z. et al. Tunneling anisotropic magnetoresistance driven by magnetic phase transition. Nat. Commun. 8, 449 (2017).

16. 16.

Stamm, C. et al. Antiferromagnetic-ferromagnetic phase transition in FeRh probed by x-ray magnetic circular dichroism. Phys. Rev. B 77, 184401 (2008).

17. 17.

Baldasseroni, C. et al. Temperature-driven nucleation of ferromagnetic domains in FeRh thin films. Appl. Phys. Lett. 100, 262401 (2012).

18. 18.

Baldasseroni, C. et al. Temperature-driven growth of antiferromagnetic domains in thin-film FeRh. J. Phys. Condens. Matter 27, 256001 (2015).

19. 19.

Mancini, E. et al. Magnetic phase transition in iron-rhodium thin films probed by ferromagnetic resonance. J. Phys. D Appl. Phys. 46, 245302 (2013).

20. 20.

Heidarian, A. et al. Ferromagnetic resonance of MBE-grown FeRh thin films through the metamagnetic phase transition. Phys. Status Solidi B 254, 1700145 (2017).

21. 21.

Tanaka, K., Moriyama, T., Usami, T., Taniyama, T. & Ono, T. Spin torque in FeRh alloy measured by spin-torque ferromagnetic resonance. Appl. Phys. Express 11, 013008 (2018).

22. 22.

Sinova, J., Valenzuela, S. O., Wunderlich, J., Back, C. H. & Jungwirth, T. Spin Hall effects. Rev. Mod. Phys. 87, 1213–1259 (2015).

23. 23.

Zhang, W. et al. Spin Hall effects in metallic antiferromagnets. Phys. Rev. Lett. 113, 196602 (2014).

24. 24.

Frangou, L. et al. Enhanced spin pumping efficiency in antiferromagnetic IrMn thin films around the magnetic phase transition. Phys. Rev. Lett. 116, 077203 (2016).

25. 25.

Saglam, H. et al. Spin transport through the metallic antiferromagnet FeMn. Phys. Rev. B 94, 140412(R) (2016).

26. 26.

Qiu, Z. et al. Spin-current probe for phase transition in an insulator. Nat. Commun. 7, 12670 (2016).

27. 27.

Almeida, T. P. et al. Quantitative TEM imaging of the magnetostrutural and phase transitions in FeRh thin film systems. Sci. Rep. 7, 17835 (2017).

28. 28.

Graves, C. E. et al. Nanoscale spin reversal by non-local angular momentum transfer following ultrafast laser excitation in ferrimagnetic GdFeCo. Nat. Mater. 12, 293 (2013).

29. 29.

Tserkovnyak, Y., Brataas, A. & Bauer, G. E. W. Spin pumping and magnetization dynamics in metallic multilayers. Phys. Rev. B 66, 224403 (2002).

30. 30.

Mankovsky, S. et al. Temperature-dependent transport properties of FeRh. Phys. Rev. B 95, 155139 (2017).

31. 31.

Jiang, M. et al. Influence of film composition on the transition temperature of FeRh films. J. Cryst. Growth 438, 19–24 (2016).

32. 32.

Obstbaum, M. et al. Inverse spin Hall effect in Ni81Fe19/normal-metal bilayers. Phys. Rev. B 89, 060407(R) (2014).

33. 33.

Uhlíř, V., Arregi, J. A. & Fullerton, E. E. Colossal magnetic phase transition asymmetry in mesoscale FeRh strpes. Nat. Commun. 7, 13113 (2016).

34. 34.

Chen, L., Matsukura, F. & Ohno, H. Direct-current voltages in (Ga,Mn)As structures induced by ferromagnetic resonance. Nat. Commun. 4, 2055 (2013).

35. 35.

Rojas-Sánchez, J.-C. et al. Spin pumping and inverse spin Hall effect in platinum: the essential role of spin-memory loss at metallic interfaces. Phys. Rev. Lett. 112, 106602 (2014).

36. 36.

Ando, K. et al. Angular dependence of inverse spin-Hall effect induced by spin pumping investigated in a Ni81Fe19/Pt thin film. Phys. Rev. B 78, 014413 (2008).

37. 37.

Zakeri, Kh. et al. Spin dynamics in ferromagnets: Gilbert damping and two-magnon scattering. Phys. Rev. B 76, 104416 (2007).

38. 38.

Chen, L. et al. Emergence of anisotropic Gilbert damping in ultrathin Fe layers on GaAs(001). Nat. Phys. 14, 490–494 (2018).

39. 39.

Woltersdorf, G. & Heinrich, B. Two-magnon scattering in a self-assembled nanoscale network of misfit dislocations. Phys. Rev. B 69, 184417 (2014).

40. 40.

Hickey, M. C. & Moodera, J. S. Origin of intrinsic Gilbert damping. Phys. Rev. Lett. 102, 137601 (2009).

41. 41.

Pal, S., Rana, B., Hellwig, O., Thomson, T. & Barman, A. Tunable magnonic frequency and damping in [Co/Pt]8 multilayers with variable Co layer thickness. Appl. Phys. Lett. 98, 082501 (2011).

## Acknowledgements

Y.W. is grateful for the support from the Alexander von Humboldt Foundation. This work is financially supported by the National Natural Science Foundation of China (Grant Nos. 51601006 and 51671110), and German Science Foundation (DFG) via SFB 1277.

## Author information

Authors

### Contributions

Y.W., L.C., and C.H.B. planned the study. Y.W. fabricated the devices, collected and analyzed the data. Y.W. and T.N.G.M. carried out the spin pumping measurements. L.C. and M.M.D. carried out the calculations. Y.W. and T.G. carried out the full film-FMR measurements. X.C., C.S., and J.Z. grew the samples and performed the magnetization measurements. Y.W. wrote the manuscript with input from L.C., C.H.B., and W.Z. All authors discussed the results.

### Corresponding authors

Correspondence to Yuyan Wang or Lin Chen.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

Peer review information Nature Communications thanks Anatolii Kravets and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions

Wang, Y., Decker, M.M., Meier, T.N.G. et al. Spin pumping during the antiferromagnetic–ferromagnetic phase transition of iron–rhodium. Nat Commun 11, 275 (2020). https://doi.org/10.1038/s41467-019-14061-w

• Accepted:

• Published:

• ### Buffer layer engineering of L10 FePd thin films with large perpendicular magnetic anisotropy

• Xinjun Wang
• , Sergiy Krylyuk
• , Daniel Josell
• , Delin Zhang
• , Deyuan Lyu
• , Jian-Ping Wang
•  & Daniel B. Gopman

• ### Identification of spin-dependent thermoelectric effects in metamagnetic FeRh/heavy-metal bilayers

• Sheng Zhang
• , Siyu Xia
• , Qian Li
• , Bin Yang
• , Jun Li
• , Qingqi Cao
• , Dunhui Wang
• , Ronghua Liu
•  & Youwei Du

Applied Physics Letters (2021)

• ### Temperature dependence of the effective Gilbert damping constant of FeRh thin films

• T. Usami
• , M. Itoh
•  & T. Taniyama

• ### Electric-field control of spin dynamics during magnetic phase transitions

• Tianxiang Nan
• , Yeonbae Lee
• , Shihao Zhuang
• , Zhongqiang Hu
• , James D. Clarkson
• , Xinjun Wang
• , Changhyun Ko
• , HwanSung Choe
• , Zuhuang Chen
• , David Budil
• , Junqiao Wu
• , Sayeef Salahuddin
• , Jiamian Hu
• , Ramamoorthy Ramesh
•  & Nian Sun

• ### Local manipulation of metamagnetism by strain nanopatterning

• Michael Foerster
• , Enric Menéndez
• , Emerson Coy
• , Alberto Quintana
• , Carles Gómez-Olivella
• , Daniel Esqué de los Ojos
• , Oriol Vallcorba
• , Carlos Frontera
• , Lucia Aballe
• , Josep Nogués
• , Jordi Sort
•  & Ignasi Fina

Materials Horizons (2020)