Impact of Co doping on the magnetic and transport properties of FeRh

FeRh, known for its unique antiferromagnetic (AFM) to ferromagnetic (FM) phase transition,^1,2 has attracted significant attention due to its distinctive magnetic properties. This magnetic phase transition at ∼370 K,^3,4 exceeding room temperature, positions it as a compelling candidate for applications that extend beyond conventional magnetic materials. This transition is accompanied by anomalously high changes in resistance and entropy,^5–11 which has sparked interest for potential applications in spintronics and magnetic recording.^12–18 Notably, this AFM to FM transition is associated with notable physical changes, including a small volume expansion of around 1%.^3,4,19–22 In order to utilize FeRh in technological applications, it is important to maintain the unique magnetic properties in a thin film structure that is sensitive to volume expansion, i.e., strain effects.

To understand the unique properties of FeRh for potential technical applications, particularly in fields such as spintronics and magnetic memory devices, it is essential to tailor these properties to meet specific requirements, such as tunable transition temperature and high magnetoresistance, which are essential for achieving stable data storage, enhanced sensitivity in magnetic sensors, and improved energy efficiency in spintronic devices.^12–18 One previously explored approach is the doping of FeRh with different elements. By introducing other elements into the FeRh lattice, researchers aim to manipulate its electronic structure, magnetic ordering, and transition temperatures,^{6,7,9,23–31} thereby paving the way for novel functionalities. For instance, the transition temperature changes over a wide range; it increases with a small doping of Pt^28,29,31 and decreases with a small doping of Pd.^24–27,31 While many previous studies have predominantly focused on the doping effect in the bulk type of FeRh,^6,9,25,27 there have been few reports on the doping effect in the film type of FeRh.^{7,24,26,28–30} It is particularly noteworthy to investigate how different dopants influence the critical behavior for the magnetic phase transition in FeRh films, expanding the understanding of FeRh and its potential applications.

Another explored study is the effect of strain on FeRh thin films. Since the strain, especially at the interface, influences the phase transition, it is necessary to investigate the transition behavior in epitaxial thin films. In our earlier studies,³⁰ we identified key factors to influence the magnetic properties of FeRh thin films and determined the optimal conditions for high-quality epitaxial films. FeRh films grown under optimal conditions exhibit a pronounced transition behavior, characterized by maximal magnetization and resistance changes between the AFM and FM phases at 370 K. A crucial finding is that the crystal grains do not correlate with the formation of magnetic domains, which is contrary to previous reports exhibiting a strong correlation between surface topography and magnetic domains.³² 

Therefore, it would be interesting to examine how the magnetic transition behavior is influenced by the internal strain due to doping in epitaxial FeRh films. In this study, we present an in-depth investigation into the Co-doping effects on FeRh thin films, focusing on the AFM–FM transition temperature and its role in achieving high magnetoresistance. By employing controlled deposition and precise measurement techniques, we aim to uncover the underlying physics of the tunable transition temperature and magnetoresistance change in Co-doped FeRh, providing scientific insights that may guide future applications in spintronics and magnetic memory devices.

The pristine and Co-doped FeRh films were deposited on MgO (001) substrates using the DC magnetron sputtering technique. A co-sputtering method with FeRh and Co targets was employed to introduce Co atoms into the FeRh films. The deposition temperature and argon (Ar) pressure were maintained at 600 °C and 3 mTorr, respectively, during the deposition process. The sputtering gun power for the FeRh target was set to 100 W, and the sputtering gun current for the Co target was varied from 0, 5, 8 to 10 mA, while the sample rotation was fixed at 7 rpm/min. After deposition, the as-prepared films were subjected to in situ annealing at 800 °C for 10 h to achieve epitaxial crystallization. Upon cooling to room temperature, the samples were capped with a 2 nm-thick SiO₂ layer to prevent surface oxidation and environmental degradation of the FeRh films.

The Co doping rate was determined through energy-dispersive spectroscopy (EDS) measurements, yielding Co doping levels between 1% and 2%. In the case of the pristine FeRh sample, the composition of Rh is slightly lower than that of Fe, i.e., Fe:Rh = 52.2:47.8, as shown in Table I. As the Co sputtering gun current increases from 5 to 8 mA, the Co atomic percentage (at. %) increases by approximately 1%–2%, while the Rh at. % decreases by a similar amount. This observation suggests that the doped Co atoms predominantly replace the Rh atoms. However, with a further increase in the Co sputtering gun current to 10 mA, the Co at. % remains almost unchanged, but the Fe at. % decreases by ∼1%. This finding implies that the Co atoms begin to replace the Fe atoms for a further increment in the Co doping rate. The resulting compositions are approximately Fe₅₂Rh₄₈, Fe₅₂Rh₄₇Co₁, Fe₅₂Rh₄₆Co₂, and Fe₅₀Rh₄₈Co₂, corresponding to the Co sputtering gun currents of 0, 5, 8, and 10 mA.

The crystal structure of the films was characterized using x-ray diffraction (XRD) with Cu Kα radiation. The sample geometry comprises an area of ∼3 × 3 mm² and a thickness of 0.5 mm for the MgO substrate and 90 nm for Co-doped FeRh. Magnetic and transport measurements were performed using a superconducting quantum interference device-vibrating sample magnetometer (SQUID-VSM) with a maximum applied field of 70 kOe in the temperature range from 2 to 380 K. For transport measurements, we employed the van der Pauw method, which is commonly used for thin films. The current was applied using a Keithley 6221 current source, and the voltage was measured using a Keithley 2182A nanovoltmeter.

The magnetic phase transition of FeRh occurs within a narrow range of Rh compositions in Fe_1−xRh_x (0.47 ≤ x ≤ 0.50).^11,33 This transition is accompanied by a small volume expansion of ∼1% and is strongly influenced by the chemical composition.^11,33,34 Therefore, introducing Co doping can lead to a dramatic change in the magnetic transition behavior. For pristine and Co-doped FeRh films, we measured the temperature dependence of magnetization and electrical resistance, denoted as M(T) and R(T), respectively. The obtained results are presented in Figs. 1(a) and 1(b). The R(T) data are normalized to the resistance value at 380 K. All M(T) and R(T) curves, measured in an applied magnetic field of 1 T parallel to the film plane, show identical transition behaviors from high resistance in the antiferromagnetic (AFM) phase to low resistance in the ferromagnetic (FM) phase. The pristine FeRh film undergoes a phase transition from an AFM phase to a FM phase at 340 K, which is lower than the typical transition temperature of 370 K reported in FeRh.^3,4 This reduction in the phase transition temperature, T_PT, is attributed to the influence of the applied magnetic field, which tends to lower T_PT.

As the Co-doping concentration increases, it is clearly seen that T_PT decreases significantly, reaching 270 and 160 K for Fe₅₂Rh₄₇Co₁ and Fe₅₂Rh₄₆Co₂, respectively. Remarkably, only 1 and 2 at. % of Co atoms replacing Rh atoms cause a substantial change in T_PT, with a reduction of 70 K for Fe₅₂Rh₄₇Co₁ and 180 K for Fe₅₂Rh₄₆Co₂. Figure 1(c) shows a phase diagram, demonstrating that the transition temperature decreases with Co-doping concentration (x). This result has been observed with similar trends in bulk FeRh_1−xM_x samples, where the transition temperature decreases for M = Fe, Co, or Ni (magnetic elements) and increases for M = Ir or Pt (nonmagnetic elements).³¹ Thus, the reduced transition temperature with Co doping can be ascribed to the weakening of the AFM phase and/or the strengthening of the FM phase due to the Co replacement of Rh atoms. Meanwhile, for Fe₅₀Rh₄₈Co₂, the transition characteristics are smeared out, and an apparent FM moment is observed even in the AFM phase. This observation is also consistent with the EDS results, which indicate that Co atoms begin to occupy the Fe sites after reaching 2 at. % occupancy. Notably, this Co replacement of Fe atoms can destabilize the AFM state, leading to the loss of clear magnetic phase transition behavior.

In addition to the dramatic changes in the magnetic transition behavior, the Co doping exerts a notable impact on the magnetic phase. While the saturation magnetization value in the FM phase is relatively unchanged, the AFM phase undergoes significant changes by the Co doping. These findings not only emphasize the critical role of the chemical composition in determining the magnetic phase transition in FeRh films but also underscore the complex effects of doping on the magnetic phases of FeRh films. In previous theoretical studies, the appearance of Rh moment was described under the Stoner model by a strong hybridization between Rh and Fe states.³⁵ In the FM state of FeRh, Fe spins align parallel to each other, and the nonmagnetic Rh atoms become magnetized, leading to a non-zero magnetic moment of the Rh atoms, which align in the same direction as the Fe spins. In the AFM state, Fe spins align antiparallel to each other, leading to a frustration of the Rh spins, which ultimately results in a zero magnetic moment of the Rh atoms. When Co is doped into FeRh, in the FM state, Co atoms replacing Rh sites can align either parallel or antiparallel to the Fe spins. However, the increase in saturation magnetization observed in the M(H) results [Fig. 3(a)] suggests that the Co spins align parallel to the Fe spins. In the AFM state, while the net magnetization is originally zero, the M(T) results indicate that the Co doping induces a finite magnetization in the AFM state of FeRh. This can be understood as Co atoms substituting nonmagnetic Rh sites, which introduce an additional magnetic moment, interrupting the perfect antiferromagnetic ordering, and resulting in residual magnetization. Understanding such phenomena is crucial for designing advanced magnetic materials with tailored properties for various potential applications.

There have been several previous reports on the presence of FM moments within the AFM phase.^11,36,37 Regarding the origin of this residual moment, the proposed mechanisms include atomic-scale defects, compressive strain, and interface moments.^37–43 Atomic-scale defects, such as vacancies or impurities within the FeRh lattices, can serve as local magnetic moments that favor FM ordering.^37,43 As discussed earlier, it is evident that the doped Co atoms play an important role in enhancing the FM coupling, which will be explored in depth later. Compressive strain can be a crucial factor inducing a local distortion in FeRh lattices and, thereby, FM moments even in the AFM phase.^38,40,41 Interface moments, which arise at the interfaces between the FeRh sample and the substrates or capping layers, can also contribute to the observed FM behavior.^38,39,42 Therefore, one may anticipate possible crystal structural deformation as a result of the Co doping effect.

In order to gain the crystal structural information by doping, we conducted x-ray diffraction (XRD) measurements on Co-doped FeRh films. Figure 2 displays the XRD patterns obtained at room temperature. The [002] diffraction peak of the MgO substrate is retained at ∼43.1° as a reference peak, and two distinct diffraction peaks for Co-doped FeRh films are observed at ∼29.8° and ∼62.0°^22,44,45 for [001] and [002] reflections, respectively. The diffraction peaks of Co-doped FeRh films are slightly lower than those of the pristine FeRh film. As summarized in Table I, the estimated lattice parameters of Co-doped FeRh films are larger compared to those of the pristine FeRh film. The variation in lattice parameters is attributed to the different magnetic phases at room temperature, where the pristine FeRh film is in the AFM phase, whereas the Co-doped FeRh film is in the FM phase. Namely, the crystal lattice experiences an expansion of ∼1% during the magnetic phase transition from the AFM to FM phase, which is in good agreement with the previous results.^3,4,19–22 Within the Co-doped FeRh films, there is no change in diffraction peak positions, suggesting that the Co doping level of 1 to 2 at. % does not alter the crystal structure of FeRh.

Figure 3 depicts the magnetization and magnetoresistance data as a function of magnetic field M(H) and magnetoresistance (MR), respectively, measured at various temperatures. The M(H) data of pure FeRh samples were taken by subtracting the diamagnetic signal with a linear background originating from the MgO substrate, and the MR value is normalized by the low resistance at the FM phase. The overall features of M(H) and MR are identical, as shown in Fig. 3(a), where MR displays a concurrent change with the transition from the AFM to FM phase in M(H). The AFM phase exhibits high resistance, and the FM phase exhibits low resistance. The overall transition behaviors are similar between Fe₅₂Rh₄₈ and Fe₅₂Rh₄₇Co₁, but Fe₅₂Rh₄₆Co₂ shows a smoother transition behavior, and Fe₅₀Rh₄₈Co₂ does not show any significant transition behavior. The temperature-dependent characteristics can be described into three distinct temperature regimes: (i) At temperatures above T_PT, wherein the FM phase dominates, M(H) exhibits a typical FM hysteresis loop with a low coercivity, H_C. A closer examination of the strong correlation between M(H) and MR in the low field regime is depicted in Fig. 3(b), where the position of the MR peaks aligns with H_C in the M(H) curve, indicative of a characteristic anisotropic MR (AMR) response commonly observed in ferromagnetic materials. (ii) At temperatures below T_PT, wherein the AFM phase dominates, M(H) is close to zero and no notable change in MR is observed, while the residual magnetization increases with increasing the Co doping. (iii) At temperatures near T_PT, M(H) rapidly increases with a sharp transition from the AFM to FM phase, resulting in the concurrent change in MR. As the Co doping increases, the MR ratio reaches maximal values of ∼90%, 170%, and 190% for Fe₅₂Rh₄₈, Fe₅₂Rh₄₇Co₁, and Fe₅₂Rh₄₆Co₂, respectively, despite almost identical changes in M(H). However, for Fe₅₀Rh₄₈Co₂, the transition behavior nearly vanishes, leading to an MR ratio of only 40%. The MR change holds promising potential for the utilization of FeRh in magnetic memory and spintronic devices. Moreover, it would be even more useful because the MR value is further increased while T_TP is adjustable.

The observed MR ratio, reaching 190% for Fe₅₂Rh₄₆Co₂, is highly intriguing. This magnitude is too large to be interpreted as a spin-dependent mechanism, such as giant magnetoresistance (GMR), as frequently observed in metallic multilayers. In conventional GMR devices, resistance changes depend on the relative orientation of magnetic moments, and typical GMR values are in the range of a few tens of percent.^46,47 Notably, in the case of FeRh, where the FM to AFM transition is inherently present, such a large MR change is achieved even within a single material layer. The transition between the AFM and FM phases induces a change in the magnetic moment, thereby contributing to the observed resistance variation. This phenomenon serves as a distinctive example of GMR, especially in the absence of multilayer structures. Furthermore, it is noteworthy that the MR change remains evident even in Fe₅₀Rh₄₈Co₂, despite the nearly negligible change in magnetization. This result suggests that the MR change is not purely due to the change in magnetic moment between the AFM and FM phases.

To explore alternative contributing origin to the observed MR change, we performed Hall experiments to investigate the Co-doping effect on the band structure. Figure 4 illustrates the results of the Hall effect measurement. The pristine and Co-doped samples exhibit changes in Hall data during the phase transitions, by diving it into three temperature regimes: (i) at temperatures above T_PT, where only the FM phase is present, anomalous Hall and normal Hall appear in all samples without undergoing a magnetic phase transition; (ii) in the AFM phase below T_PT, the Hall data manifest as a straight line with only normal Hall behavior in the pristine sample, while the Co-doped FeRh in the AFM state shows a mixed form with an anomalous Hall signal due to residual moments; and (iii) at temperatures around T_PT, the Hall data rapidly increase with a sharp transition from the AFM to FM phase, consistent with the M(H) results. However, unlike M(H) data, the Hall resistance does not saturate into a single value at high magnetic fields, suggesting another source contributing to the MR change.

Table II displays the calculated carrier density from the Hall data. The p-type charge carrier density is approximately ∼10²³ cm⁻³ for the FM phase above T_PT, and the n-type carrier density is approximately ∼10²¹ cm⁻³ for the AFM phase below T_PT. The differences in carrier densities between the two magnetic phases are roughly 100 times the same for both samples, but Co-doped FeRh exhibits a slightly lower carrier density than the pristine sample. At temperatures around T_PT, where the magnetic phase transition occurs, the Hall slope changes before and after this transition, indicating a change in the carrier type. In the FM state, the major carrier is a hole, while in the AFM state, it is an electron, consistent with previous studies on FeRh.^48–50 This study holds significant importance as it demonstrates that the Co doping can significantly improve the MR ratio. Previous studies suggested that the conductivity change in FeRh is influenced not only by the intrinsic band structure but also by the extrinsic scattering mechanism.^51,52 However, our recent studies using terahertz time-domain spectroscopy have revealed that the conductivity change in FeRh is due to a reconfiguration of the band structure during the phase transition, specifically involving modifications in the charge density and effective mass.⁵³ This suggests that the phase transition in FeRh_1−xCo_x likely involves intrinsic changes in the electronic structure, rather than solely spin-dependent scattering.

Co-doped FeRh films were fabricated using a co-sputtering technique with FeRh and Co targets, with precise control of the Co doping level by the modulation of the Co sputtering gun current. As the sputtering gun current increases from 0 to 5 and 8 mA, the Co atoms mostly replace the Rh atoms at levels of 1 and 2 at. %, and then, the Co atoms begin to replace the Fe atoms for further increment in the Co sputtering gun current to 10 mA. We observed a dramatic change in the magnetic transition behavior of FeRh by Co doping, which is evidenced by the significant reduction in transition temperature (T_PT) even with small Co doping levels of 1%–2%. The most possible effect due to the Co substitution of Rh atoms is the strengthening of the ferromagnetic (FM) coupling. Moreover, we demonstrated a remarkable improvement in the magnetoresistance (MR) ratio by Co doping, reaching up to 190% for 2 at. % Co doping. This MR enhancement, even within a single layer, suggests another mechanism distinct from the conventional giant magnetoresistance effect. The Hall effect measurements revealed that the carrier density is slightly lowered by Co doping, while keeping the carrier-type change across the phase transition. These findings provide in-depth insights into the complex underlying physics beyond simple spin-dependent transport mechanisms, revealing intrinsic changes in the electronic structure. Overall, our study highlights the tunability of phase transitions to desirable temperatures and the significance of the AFM–FM transitions in achieving high magnetoresistance values.

This work was supported by the National Research Foundation of Korea (NRF) (Grant Nos. 2020R1A2C3008044 and 2022R1A4A1033562) and by Samsung Electronics Co., Ltd. (Grant No. 202470076.08).

The authors have no conflicts to disclose.

Sang-il Seo: Investigation (equal); Methodology (equal); Writing – original draft (equal). Min-Tae Park: Investigation (equal); Methodology (equal). Myung-Hwa Jung: Conceptualization (equal); Supervision (equal); Validation (equal); Writing – review & editing (equal).

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