# Fileset

[NT10414.pdf](https://mdr.nims.go.jp/filesets/807fad67-6176-469d-ae59-af13c235e1c8/download)

## Creator

[Daisuke Ogawa](https://orcid.org/0009-0007-5388-6916), [Anton Bolyachkin](https://orcid.org/0000-0003-0420-1806), [Angayarkanni R. Dilipan](https://orcid.org/0000-0003-1859-1008), [Nikita Kulesh](https://orcid.org/0000-0001-7046-2671), [Hossein Sepehri-Amin](https://orcid.org/0000-0002-7856-7897), [Yukiko K. Takahashi](https://orcid.org/0000-0001-9197-7236)

## Rights

©2024 American Physical Society[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Exchange-coupled <math display="inline">  <mi>Fe</mi></math>-<math display="inline">  <mi>Pt</mi></math>/<math display="inline">  <mi>Ru</mi></math>/<math display="inline">  <mi>Fe</mi></math>-<math display="inline">  <mi>Pt</mi></math> nanogranular films as potential heat-assisted-magnetic-recording media with reduced writing temperature](https://mdr.nims.go.jp/datasets/02c9db18-b811-453f-bc9b-d58ecd12cb3e)

## Fulltext

manuscript_FePt_Ru_FePt.dviExchange-coupled Fe-Pt/Ru/Fe-Pt nanogranular films aspotential HAMR media with reduced writing temperatureDaisuke Ogawa,1, ∗ Anton Bolyachkin,1 Angayarkanni R. Dilipan,1, 2Nikita Kulesh,1 Hossein Sepehri-Amin,1, 2 and Yukiko K. Takahashi1, †1National Institute for Materials Science, Tsukuba 305-0047, Japan2Graduate School of Science and Technology, University of Tsukuba, Tsukuba 305-8577, JapanHeat-assisted magnetic recording (HAMR) is a promising high-density data storage technologywhose intensive practical deployment requires a perfect balance of many factors, including devicereliability, which is challenged by severe writing conditions. In this work, we propose HAMR mediacomposed of Fe-Pt/Ru/Fe-Pt trilayer nanograins embedded into a carbon segregating matrix. Whenthe Ru nonmagnetic spacer is thinner than 2 nm, the Fe-Pt layers were found to be exchange-coupledand exhibit different Curie temperatures, which is beneficial for lowering the HAMR writing tem-perature. Given the grain size of 17 nm, sufficient thermal stability was preserved according to thetime dependence of coercivity analyzed using the Sharrock equation. The granular microstructureof Fe-Pt/Ru/Fe-Pt films with different Ru thicknesses was examined in detail, while their magneticproperties were interpreted by micromagnetic simulations distinguishing the properties of individualFe-Pt layers. Grain size reduction, suppression of in-plane variants, improvement of the degree ofL10 order in the top Fe-Pt layer - all these are further steps to optimize the developed exchange-coupled Fe-Pt/Ru/Fe-Pt nanogranular films, that would open a new avenue for tuning the HAMRmedia toward its ultimate performance.Keywords: heat-assisted magnetic recording, exchange-coupled media, Fe-Pt, Curie temperatureI. INTRODUCTIONIn the epoch of digital transformation, the amount ofdata being accumulated and handled globally is grow-ing exponentially. As a result, increasing the storage ca-pacity of data centers has emerged as a critical issue.Data centers mainly utilize hard disk drives (HDDs) ow-ing to their substantial capacity, non-volatility, and lowbit cost. The current recording density in commercialHDDs is about 1.5 Tbit/in2. This is achieved with aconventional perpendicular magnetic recording media – agranular media with ferromagnetic Co-Cr-Pt grains uni-formly dispersed in a SiO2-based nonmagnetic matrix.The next generation ultra-high-density HDDs are underdevelopment and are expected to exceed 4 Tbit/in2 inresponse to societal demands [1, 2].Achieving a recording density over 4 Tbit/in2 requiresreducing the grain size to 5 nm or less. The con-ventional Co-Cr-Pt recording media cannot withstandthat without violating the thermal stability condition:KuV/kBT > 60, where Ku is the uniaxial magnetocrys-talline anisotropy constant, V is the volume of a grain, kBis the Boltzmann constant, and T is the temperature [3].L10-ordered FePt can overcome this problem due to itshigh Ku of 7 MJ/m3 [4] that theoretically ensures ther-mal stability down to 4 nm grain size [5]. However, sucha strong magnetic anisotropy leads to another challenge- an external magnetic field larger than 3 T is neces-sary to switch FePt grains at room temperature, exceed-ing the 1.5 T capability of standard recording heads [6].∗ ogawa.daisuke@nims.go.jp† takahashi.yukiko@nims.go.jpAn energy-assisted magnetic recording (EAMR) has beenproposed, which essential idea is to stimulate the magne-tization reversal by applying external energy such as heatand/or microwaves. The former, heat-assisted magneticrecording (HAMR) [6], appears to be the most feasibleoption with first HAMR HDDs deployed in 2020. How-ever, there are still several issues on the way to the de-sired ultra-high recording density and mass production.One is the low reliability of the HAMR HDD, which isattributed to writing at elevated temperatures approach-ing the Curie temperature (Tc) of FePt around 760 K[7, 8]. Even a few nanoseconds of such intense heatingmay cause deformations in both the recording media andwriting head [9, 10]. Furthermore, lubricants protectingthe surface of the medium are also prone to degrade withheat exposure [11].To address these issues, reducing the Tc of FePt me-dia can be considered to lower the writing tempera-ture. This can be achieved by doping FePt with Ni[12], Mn [13], or Ru [14, 15]. However, such dop-ing has inherent limitations as it significantly reducesthe magnetic anisotropy and saturation magnetization,Ms, thus affecting KuV/kBT and the signal-to-noise ra-tio. Exchange-coupled composite (ECC) can inspire to amore adjustable solution which could be less detrimentalto Ms and effective Ku. For the ECC media, originallyproposed as granular media with stacked high anisotropyand soft magnetic layers, a well-tunable switching fieldwith maintained thermal stability was demonstrated [16–22]. This concept can be adapted to HAMR media:nanograins are split into exchange-coupled high- and low-Tc FePt layers to adjust the writing temperature of themedia, while preserving sufficient KuV/kBT (see Fig. 1for a schematic explanation). Other potential benefits2Thermal stabilitythreshold60Coercivity,(T)μ0HcTECCAmbientTemperatureHeating to write1.5TbTtWriting fieldthresholdTemperatureK Vuk TBBottom layer solelyTop layer solelyExchange-coupled trilayerMgO(100)C segregantFe-Pt:Fe-Pt:Ru (0-2 nm)low &T Hc chigh &T Hc cFIG. 1. The schematic temperature dependencies of coerciv-ity (Hc) and normalized energy barrier (KuV/kBT ) for theindividual Fe-Pt layers and the entire Fe-Pt/Ru/Fe-Pt granu-lar stack (inset). Although KuV/kBT is expected to decreasemore steeply than Hc with increasing temperature, both arerepresented by one curve for simplicity.of ECC-HAMR in different realizations include reducedthermal write errors [23–25], improved switching proba-bility [26], increased near-field transducer lifespan [27],and suppressed transition jitter noise [28].In this work, exchange-coupled HAMR granular me-dia was developed by depositing Fe-Pt/Ru/Fe-Pt tri-layer structure with C as a segregant (Fig. 1, inset).The measured hysteresis loops of the films were inter-preted with micromagnetic simulations, which clearly in-dicated exchange coupling of the Fe-Pt layers when theRu spacer was thin enough, while temperature depen-dencies of magnetization revealed the difference in Curietemperatures. The appropriate thermal stability of thedeveloped films was confirmed by the time dependenceof the coercivity. Thus, the practical feasibility of theECC-HAMR media has been demonstrated. With fur-ther microstructural optimization, it will contribute tothe realization of robust high-density HAMR media.II. EXPERIMENTAL DETAILSThe nanogranular films were fabricated using an ul-trahigh vacuum magnetron sputtering system with abase pressure around 10−7 Pa. All films were depositedon a single crystal MgO(100) substrate to promote the(001) textured growth of Fe-Pt with L10 chemical or-der [15, 29–31]. The nominal film stack was composedof Fe-Pt(4.5)/Ru(0-2.0)/Fe-Pt(4.5) with 20 vol.% of Cas a segregant in each layer. Both the top and bottomFe-Pt layers of 4.5 nm thickness were deposited by co-sputtering Fe-Pt and C targets at 500 ◦C under Ar pres-sure of 0.478 Pa. The Ru layer, also called the spacer,was deposited under the same conditions using Ru andC targets; its thickness was varied from 0 (no spacer) to2 nm. To prevent from oxidation, a 5 nm thick C cappinglayer was deposited at room temperature. For simplicity,the described films are referred to as Fe-Pt/Ru(tsp)/Fe-Pt hereinafter with tsp representing the thickness of theRu spacer in nm.X-ray diffraction (XRD) was used to evaluate the crys-tal structure and degree of the L10 order (S) using aRigaku SmartLab with a Cu-Kα X-ray source. S wasevaluated following the equation [32–34]:S =√I001I002×(|F |2LPAD)002(|F |2LPAD)001= α√I001I002, (1)where I001 and I002 are experimental integrated intensi-ties of the L10 (001) superlattice and (002) fundamentalpeaks, respectively. F , L, P , A, and D are the structurefactor, Lorentz factor, polarization factor, absorption fac-tor, and temperature factor, respectively – all are usedto calculate theoretical (001) and (002) integrated peakintensities assuming perfect out-of-plane (001) texture ofthe Fe-Pt grains. The coefficient α stands for the pro-cessed ratio of theoretical intensities and depends on theexperimental setup, in our case α = 0.85 [35–37].The top view and cross-sectional transmission elec-tron microscopy (TEM) of the films was performed us-ing a Titan G2 80-200 with a probe aberration correc-tor. Electron-transparent samples for these TEM ob-servations were prepared using chemical etching and alift-out technique using a focused ion beam (FEI HeliosNanolab 650), respectively. To prevent damage duringion beam milling, the films were coated with Ni. Energy-dispersive X-ray spectroscopy (EDS) was carried out us-ing a FEI Super-X EDX detector, while the results wereanalyzed using the Bruker Esprit v1.9 software. The localcrystal structure was examined via selected area electrondiffraction (SAED) using TEM. Statistical analysis of thegranular microstructure in the top view TEM images wasperformed using deep learning models [38].Out-of-plane (OOP) hysteresis loops were measuredusing a superconducting quantum interference device(SQUID) magnetometer (Quantum Design) in magneticfields up to 7 T over the temperature range from 100to 700 K. Measurements of the in-plane (IP) hysteresisloops were carried out using a Dynacool PPMS (Quan-tum Design) with 14 T maximum magnetic field, whichwas equipped with a large bore coil set. Curie temper-atures (Tc) were determined by fitting the M -T curvesusing the Kuz’min formula [39].To analyze the OOP demagnetization curves, micro-magnetic simulations were performed by solving the Lan-dau–Lifshitz–Gilbert equation with unit damping con-stant and sweeping rate of 0.4 T/ns via the Fastmag soft-ware [40]. Finite element models of nanogranular Fe-Pt3films with the trilayer structure were developed followinga similar approach to that in Refs. [15, 41].III. RESULTS AND DISCUSSIONA. MicrostructureThe reference Fe-Pt film devoid of a Ru spacer (tsp= 0 nm) exhibited a fine granular microstructure withthe mean grain size (D) of 17.5 nm and the numberdensity of 1.43 Tgrain/in2 (Fig. 2(a)). The introduc-tion of a thin Ru(0.5) spacer did not alter the top viewmicrostructure much, preserving the grain size and num-ber density at 17.6 nm and 1.48 Tgrain/in2, respectively(Fig. 2(b)). These parameters deteriorated to 19.3 nmand 1.23 Tgrain/in2 as the Ru thickness increased to1.0 nm (Fig. 2(c)). This was accompanied by a tendencyto grain coalescence, which finally resulted in a network-like structure when the Ru spacer approached 2 nm inthickness, as shown in Fig. 2(d). For all films, the SAEDpatterns revealed (110) superlattice spots indicative ofL10 Fe-Pt, along with (020) and (022) fundamental spots.The (020) diffraction spots in films with Ru spacers of 1.0and 2.0 nm had some satellites, suggesting the possibleformation of twins.Cross-sectional high-angle annular dark-field scanningtransmission electron microscopy (HAADF-STEM) im-ages of the Fe-Pt/Ru(0.5)/Fe-Pt and Fe-Pt/Ru(1.0)/Fe-Pt films are shown in Fig. 3 along with the correspond-ing EDS maps and composition line profiles evaluatedfrom the highlighted regions of interest. The grains inboth films demonstrated the desired trilayer structure,which was featured with a curved interface. The bottomFe-Pt tended to grow into nanograins with pronouncedsphericity, which were then well wetted by Ru, formingsuch an unplanarity. According to EDS (Fig. 3, middle),there was a possibility of nonuniform chemical distribu-tion and thickness of Ru along the curved interface with adecrease toward its edges. Across the center of the spacer(Fig. 3, bottom), Ru approached 35 and 55 at.% in thefilms with tsp of 0.5 and 1.0 nm, respectively, while Feand Pt occupied the rest content equally. However, theRu content and thickness are prone to under- and over-estimation, respectively, because of the spacer curvature.Therefore, additional experiments were performed withcontinuous Fe-Pt(30)/Ru(tsp)/Fe-Pt(30) films depositedunder the same conditions but without C segregant. Insuch an ideal case with a planar interface, the interdiffu-sion of layers was confirmed with Ru 55 at.%, Fe 20 at.%at tsp = 0.5 nm and Ru 60 at.%, Fe 15 at.% at tsp =1.0 nm (Fig. S1 in Ref. [42]).The alternating atomic layers of Fe and Pt were clearlyvisible in the bottom Fe-Pt (Fig. 3, top), indicating theformation of the L10 structure with a nice (001) texturein the out-of-plane direction. A similar feature was ob-served in the top Fe-Pt, but with some regions of deterio-rated L10 order or texture (in-plane variants). Also notethat despite similar nominal thicknesses designed for thetop and bottom Fe-Pt layers, the top layer tended to bethinner due to wetting over the Ru spacer.Fig. 4 shows the XRD patterns of Fe-Pt/Ru/Fe-Ptgranular films with the inset depicting a change in thedegree of L10 order (S) with increasing Ru thickness. Inaddition to the Fe-Pt (002) peak at 2θ ≈ 47.5◦, the (001)superlattice peak was observed at 2θ ≈ 23.5◦, indicatingthe L10 order in all films. The reference Fe-Pt film with-out Ru demonstrated a high degree of order, S = 0.86.When Ru was introduced, S started to decrease to 0.72at tsp = 2 nm. Note that here S is an effective degree oforder that means an averaged value over the entire film,which may have different S in the top and bottom Fe-Ptlayers following our previous study [15]. An insufficientmisfit between the L10 Fe-Pt and fcc-Ru (1.3%) as wellas the curved interface can result in a reduced S in thetop Fe-Pt. A larger misfit is required to promote L10ordering, e.g., 10.3% in the case of MgO(100) substrate[43, 44].B. Magnetic propertiesThe hysteresis loops of the Fe-Pt/Ru(tsp)/Fe-Pt filmsare shown in Fig. 5(a) for the OOP and IP directions.The reference Fe-Pt film (tsp = 0 nm) had a typicalsquare OOP loop with a high coercivity of 4.0 T, demon-strating a strong perpendicular magnetic anisotropy.When a thin Ru(0.5) spacer was introduced, the coerciv-ity abruptly decreased to 1.7 T while the squareness ofthe OOP loop was preserved. The corresponding IP hys-teresis loop was S-shaped near zero magnetic field, whichcould be attributed to some soft magnetic regions. Withincreasing Ru thickness, a step-like inflection appearedin the OOP loops close to the coercivity – the case ofRu(2.0) was the most illustrative. There were trends forthe deterioration of OOP remanence and slight increasein coercivity shown in Figs. 5 (b) and (c), respectively.The S-shape of the IP hysteresis loops became more pro-nounced that can be quantified by the increasing IP re-manence (Fig. 5(b)). The described transformations ofthe hysteresis loops could be caused by exchange couplingbetween the top and bottom Fe-Pt layers, the intensityof which was suppressed by increasing the Ru thickness.To verify this hypothesis and distinguish the magneticproperties of individual Fe-Pt layers, micromagnetic sim-ulation was performed.There are two different scenarios for the exchange cou-pling between Fe-Pt layers through the nonmagnetic Ruspacer. First, it can be a weakened direct exchange in-teraction enabled by a network of Fe atoms diffused intothe Ru spacer (as in Ref. [45]). Another alternative isan indirect RKKY exchange interaction through metal-lic Ru [46, 47]. Although interdiffusion of the layers wasobserved, favoring the first scenario (Fig. 3 and Fig. S1),it is difficult to justify with certainty whether the con-centration of Fe atoms is sufficient to realize such a cou-4(a) (d)(b) (c)Fe-Pt Fe-Pt/ /Fe-PtRu( .0)2D = 17.5 ( ) nm6.850 nmFe-Pt/ /Fe-PtRu(0.5) Fe-Pt/ /Fe-PtRu(1.0)1.43 Tgrain/in250 nm 50 nm 50 nmD N/AD = 1 ( ) nm7.6 5.2 D = 1 ( ) nm9.3 5.71.48 Tgrain/in21.23 Tgrain/in2Density N/A(020) (022)(110)FIG. 2. Top view TEM images of the Fe-Pt/Ru/Fe-Pt films with different Ru thickness. The SAED patterns are shown in theinsets. Where applicable, the average grain size (D) and number density are given at the bottom.5 nm 5 nmDistance (nm)0 4 8Content (at.%)020(a) (b)Fe-Pt/ /Fe-PtRu(0.5) Fe-Pt/ /Fe-PtRu(1.0)Fe Pt RuFe Pt Ru Fe Pt Ru10 0 4 8020406080100Fe Pt Ru1012 12Distance (nm)FIG. 3. Cross-sectional HAADF-STEM images, correspond-ing EDS maps, and composition line profiles of the Fe-Pt/Ru/Fe-Pt granular films with Ru thickness of (a) 0.5 nmand (b) 1.0 nm.pling or not. An additional complexity arises from thecurvature and potential nonuniformity of the Ru spacerdescribed in the previous section. The up-to-date DFTcalculations can clarify the origin of exchange couplingin the Fe-Pt/Ru/Fe-Pt granular media, either it is theRKKY interaction [49? ] or direct exchange interaction2 (deg.)θ0.5 nm1.0 nm2.0 nm20 30 40 6050tsp = 0. nm0Fe-Pt(002)Intensity(arb.u)nitsMgO substrateFe-Pt(001)0 1.00.51.00.90.80.7tsp (nm)L1order,0SFIG. 4. XRD patterns of the Fe-Pt/Ru/Fe-Pt films with dif-ferent Ru thickness (tsp) and the effective degree of L10 order(S) as a function of tsp (inset). Dashed line is a guide to theeye.via diffused Fe atoms [50].However, both scenarios of exchange coupling wouldaffect the hysteresis loops in a similar way. Therefore,we proceeded with a simplified micromagnetic model inwhich the Ru spacer is plain and weakly magnetic due todiffused Fe atoms (Fig. 6). The latter means that somesaturation magnetization and exchange stiffness Asp wereprescribed to the spacer. They were varied in a gridsearch manner until the best correspondence between thesimulated OOP demagnetization curves and experimen-tal ones. Other varied parameters were the magneticanisotropy constants K and volume fractions of in-planevariants V in the top and bottom Fe-Pt layers. The ini-5Magnetic field, (T)μ0HNormalizedmagnetization,/MMs0.0-0.5-1.00.51.00.0-0.5-1.00.51.00 10-10 -5 5 0 10-10 -5 5OOPIP0.5 nm1.0 nm 2.0 nmRemanence,/MMrs0.10.00.91.0OOPIPtsp= 0. nm0(a)(b)Coercivity,(T)μ0Hc042OOP(c)0.0 1.0 2.0Ru thickness, (nm)tsp0.0 1.0 2.0Ru thickness, (nm)tspFIG. 5. (a) Out-of-plane (OOP) and in-plane (IP) hystere-sis loops of Fe-Pt/Ru/Fe-Pt films with different Ru thickness(tsp) measured at room temperature. (b) The normalized re-manent magnetization and (c) coercivity as functions of Ruthickness.tial parameters for the grid search were adopted fromRef. [15] in which the Fe-Pt(5)/Ru(3)/Fe-Pt(5) film witha thick spacer was studied. The saturation magnetizationand exchange stiffness of both Fe-Pt layers were fixed tobe 1.43 T and 10 pJ/m, respectively [15]. Despite thesame nominal thicknesses of the Fe-Pt layers, the top Fe-Pt layer in the model was assumed to be 20% thinnerthan the bottom one. This was in agreement with TEMobservations (Fig. 3) and with another feature - shiftedup inflection points in the OOP loops with respect to thecoercivities, as can be seen in Fig. 5(a).The simulated OOP demagnetizations curves after thegrid search are shown in Fig. 6(d). They describedwell the experimental demagnetization curves of Fe-Pt/Ru(tsp)/Fe-Pt films with tsp of 0.5, 1.0, and 2.0 nm.The characteristics obtained for each layer are summa-rized in Table I. The top Fe-Pt layer had a significantlyreduced magnetic anisotropy constant (lower L10 order)and a higher content of in-plane variants compared to theTABLE I. Characteristics of each layer in Fe-Pt/Ru/Fe-Ptgranular films, evaluated from the micromagnetic approxima-tion in Fig. 6(d). Ku and V are the magnetic anisotropyconstant (mean and standard deviation in parentheses) andvolume fraction of in-plane variants in Fe-Pt layers, respec-tively. Ms and Asp are the saturation magnetization andexchange stiffness of the Ru spacer with diffused Fe atoms.Spacer Bottom Fe-Pt Top Fe-Pttsp µ0Ms Asp Ku V Ku V(nm) (T) (pJ/m) (MJ/m3) (vol.%) (MJ/m3) (vol.%)0.5 > 0.6 > 1.8 3.5 (0.45) 2 0.7 (0.1) 151.0 0.4 0.8 3.2 (0.45) 2 0.7 (0.1) 152.0 < 0.1 < 0.05 2.8 (0.45) 2 0.7 (0.1) 15bottom Fe-Pt layer. The deteriorated quality of the topFe-Pt layer was also evidenced in TEM (Fig. 3) and indi-rectly in XRD data (inset in Fig. 4). This was a detrimen-tal consequence of depositing on the curved Ru spacerwith insufficient interfacial stress for nice L10 (001) pref-erential growth.The parameters of the top Fe-Pt layer did not varywith changing Ru thickness. The bottom Fe-Pt layer hada much higher magnetic anisotropy constant and only aslight trace of the in-plane variants. We observed thatthe magnetic anisotropy constant of the bottom Fe-Ptlayer deteriorated with increasing Ru thickness, i.e., itdecreased from 3.5 to 2.7 MJ/m3 when tsp increased from0.5 to 2.0 nm. This could be attributed to a deeper diffu-sion of Ru into Fe-Pt considering that Ru is detrimentalto the magnetic anisotropy of Fe-Pt [16]. According toour simulations, the stack of magnetically soft/hard Fe-Pt layers was almost exchange-decoupled with the thickRu spacer of 2 nm (Table I), resulting in such a pro-nounced step-like inflection in the OOP hysteresis loop(Fig. 6(d)). With decreasing Ru thickness, the exchangecoupling between Fe-Pt layers became stronger, whichwas quantified by Asp in Table I – Asp of 1.8 pJ/m inFe-Pt/Ru(0.5)/Fe-Pt was enough for a strong couplingwith the eliminated M(H) inflection.The exchange coupling at room temperature was con-firmed in Fe-Pt/Ru/Fe-Pt films with tsp < 2.0 nm. Asthe next step, the magnetic properties of the films wereexamined with temperature. Fig. 7 shows the magnetiza-tion of Fe-Pt/Ru(tsp)/Fe-Pt films vs. temperature. Thereference film without Ru spacer had theM -T curve typi-cal for a uniform ferromagnetic material with Tc of 665 K.When the Ru layer was introduced, slight inflections wereobserved in the M -T curves of Fe-Pt/Ru(0.5)/Fe-Pt andFe-Pt/Ru(1.0)/Fe-Pt films, which required the superpo-sition of two Kuz’min equations with different Curie tem-peratures Tc1 and Tc2 for a better description. The lowerTc1 was attributed to the top Fe-Pt layer due to its re-vealed defectiveness with lowered L10 order, while Tc2corresponded to the bottom Fe-Pt layer. Although it washard to distinguish Tc1 for films with tsp of 0.5 and 1.0 nmdue to experimental uncertainty, a distinct Tc1 of the topFe-Pt layer was observed for the exchange-decoupled Fe-6Grain size, (nm)D4Nofgrainso.2086 10 12 14 1816025Mean = 12.5 nmStandard deviation = 2.0 nm20 nm( )a (b)(c)4 nm5 nmtspSide viewTop view-6Magnetic field, (T)μ H0Normalizedmagnetization,/M Ms-4 -2 0 2-1.0-0.50.00.51.0 (d)Experimental dataSimulationstsp = 0.5 nm 1.0 nm2.0 nmFePtFePtRuFIG. 6. (a) Top view fragment of the 240 × 240 nm2 micromagnetic model of Fe-Pt/Ru/Fe-Pt granular film and (b) corre-sponding grain size distribution. (c) Side view on a meshed grain with highlighted Ru spacer which thickness (tsp) was variedfrom 0.5 to 2.0 nm. (d) Experimental and simulated out-of-plane demagnetization curves of the films with different tsp.Pt/Ru(2.0)/Fe-Pt film with a value of about 400 K. Thisvalue can be considered as a reasonable landmark for Tc1in films with a thinner Ru spacer, assuming similar mi-crostructural state of the top Fe-Pt layers. The Tc2 ofthe bottom Fe-Pt layer in the exchange-decoupled Fe-Pt/Ru(2.0)/Fe-Pt film was of 655 K, that was close toTc of the reference Fe-Pt film. Note that the interlayerexchange coupling should have no effect on Tc2 given thesame other factors, while Tc1 could be enhanced to someextent by a strong coupling.There are two factors that affect the Curie tempera-ture of Fe-Pt and can result in low/high Tc layers. First,doping with Ru decreases the Tc of Fe-Pt [14, 15]. Ifthere is a difference in the diffusion of Ru in the topand bottom Fe-Pt layers, distinct Tc1 and Tc2 can ap-pear [15]. However, similar Ru diffusion profiles wereobserved in the top and bottom layers of the continuousFe-Pt/Ru/Fe-Pt films (Fig. S1 in Ref. [42]). In addition,Tc2 of the bottom layer in the Fe-Pt/Ru(2.0)/Fe-Pt filmwas close to Tc of the reference Fe-Pt film without Ru,indicating that Ru did not have a pronounced effect. Thesecond factor is the degree of L10 order [51], which wasdistinct in the top and bottom Fe-Pt layers according toour experimental insights and simulations (indirectly viaestimated Ku). This factor is supposed to be dominantfor the low/high Tc observed in our study.Fig. 8 shows the temperature dependencies of coer-civity for Fe-Pt/Ru/Fe-Pt films with different Ru thick-nesses. The reference Fe-Pt film without Ru exhibiteda monotonous decrease in coercivity approaching 1.5 T,the current writing field threshold, at 565 K that canbe considered as a writing temperature. The exchange-decoupled Fe-Pt/Ru(2.0)/Fe-Pt film possessed a lower0 400 600 800200Temperature (K)0 400 600 800200Normalizedmagnetization,/MM@300K0.60.40.81.00.20.00.60.40.81.00.20.0Tc = 6 K65tsp= 0. nm0 0.5 nm1.0 nm 2.0 nmTc = K4001Tc = K6552FIG. 7. Temperature dependence of magnetization for Fe-Pt/Ru/Fe-Pt films with different Ru thickness (tsp). Dashlines correspond to the approximation by a superposition oftwo Kuz’min equations (one for tsp = 0 nm) with distinctCurie temperatures (Tc).writing temperature, while its coercivity dependence hada maximum at Tc1 = 400 K. The magnetically softenedtop Fe-Pt layer magnetized oppositely to the bottom Fe-Pt layer in the vicinity of coercivity. Thus, the stray7500 600 700400Temperature (K)01234Coercivity,(T)μ0cH5Writing fieldthreshold0.5 nm2.0 nm0. nm0Fe-Pt/ /Fe-PtRu( )tsp300FIG. 8. Temperature dependence of coercivity for the Fe-Pt/Ru/Fe-Pt granular films with different Ru thickness (tsp).The blue area indicates the magnetic field currently consid-ered for the HAMR HDD write head [52].field from the top layer reduces the coercivity of the bot-tom layer. Near Tc1, the magnetization of the top Fe-Ptlayer abruptly decreases, as well as the correspondingstray field, thus increasing the coercivity of the bottomFe-Pt layer and resulting in a local maximum. The coer-civity dependence of the Fe-Pt/Ru(0.5)/Fe-Pt film withthe strongest exchange coupling decreased without anyanomaly, demonstrating a significantly reduced writingtemperature of about 380 K. At the same time, its coer-civity at room temperature was also notably decreased,so the next question was whether sufficient thermal sta-bility was preserved or not.The thermal stability of the exchange-coupled Fe-Pt/Ru/Fe-Pt films was depicted from the dynamic coer-civity, whose time dependence follows the Sharrock equa-tion [53]:Hr(t) = Hc0[1−[kBTEbln(f0tln2)]1n]. (2)Such dynamic coercivity Hr can be interpreted as thereversed magnetic field at which half of the grains areswitched during time t. Here Hc0 is a coercivity inthe absence of thermal agitations, Eb is an effective en-ergy barrier for magnetization switching at zero mag-netic field, and f0 is an attempt frequency [53]. Theexponent n may vary depending on microstructural fea-tures, e.g., grain misalignment, thicknesses of the ECCstack, etc. [53, 54]. For recording media including ECC,n of 1.4 − 1.5 is usually considered [22, 55, 56]. Theprotocol for measuring Hr(t) is described in Fig. 9(a)and (b) in details. Obtained Hr(t) dependencies for theexchange-coupled Fe-Pt/Ru/Fe-Pt films and the refer-ence Fe-Pt film are shown in Fig. 9(c). They were fittedusing Eq. (2), providing values of the thermal stabilitycoefficient summarized in Fig. 9(d). The reference Fe-Ptwithout Ru spacer had KuV/kBT of about 260. The Fe-Pt/Ru(0.5)/Fe-Pt film with the strongest exchange cou-pling exhibited a lower KuV/kBT of about 120, althoughit was still above the thermal stability threshold of 60.A bit better KuV/kBT value of 160 was observed for theFe-Pt/Ru(1.0)/Fe-Pt film.Some critical notes should be emphasized in theend. First, the writing temperature derived from Hc(T )(Fig. 8) may differ from that in a practical HAMRHDD with local heating and switching [57–59]. In ad-dition, local switching is a dynamic process, so the inter-play of exchange coupling and damping behavior in Fe-Pt/Ru/Fe-Pt is worth of further investigation via simula-tions [60, 61] and experimentally including elevated tem-peratures. The latter could be done with the all-opticaltime-resolved magneto-optical Kerr effect (TR-MOKE)[62], contributing to a discussion on how microstructuralfeatures affect the damping constant [61, 63, 64]. Second,although sufficient thermal stability was demonstrated inthe case of exchange-coupled Fe-Pt/Ru(0.5)/Fe-Pt film,it was achieved for a relatively large grain size of 17.6 nm.To be of interest for practical application, the grain size ofthe developed media should be reduced to at least 7 nm.This should be feasible with further optimization of thesputtering conditions and exploration of other materialsto be used as the spacer layer and segregant. For thelatter, h-BN [65–68] and CrOx [69, 70] can be potentialcandidates. The grain size reduction should be accompa-nied by an improvement in the microstructure of the topFe-Pt layer. In particular, the Ku of this layer should beincreased to maintain the thermal stability, while the ap-pearance of in-plane variants should be suppressed. Therealization of a flat interface between the top and bot-tom Fe-Pt layers would contribute to resolving the aboveproblem.IV. CONCLUSIONIn this work, a series of Fe-Pt(4.5)/Ru(tsp)/Fe-Pt(4.5)films with tsp = 0, 0.5, 1.0, 2.0 nm and 20 vol.% C seg-regant were fabricated by ultrahigh vacuum magnetronsputtering. The nanogranular microstructure was real-ized with the Fe-Pt/Ru/Fe-Pt trilayer interface in eachgrain. This interface had a pronounced curvature. Thethin Ru spacer of 0.5 nm thickness did not affect themean grain size, maintaining it at 17.6 nm, while fur-ther Ru thickening gradually deteriorated it, provokinggrain coalescence. The nice L10 chemical ordering with(001) out-of-plane texture was confirmed for the bottomFe-Pt layer. In contrast, the top Fe-Pt layer was moredefective, showing some in-plane variants and reduced8H iH2H1waitingM Hd i( )Magnetic field, (T)μ0H0 6-2 2 4-6 -4M(T)agnetization,μ0M0.0-0.5-1.00.51.00.0-0.1-0.2Waiting time (s)30501203201020-4.2 -4.0 -3.8μ0Hi (T)μ0Md(T)tsp= 0. nm0 tsp= 0. nm010 100 1000Waiting time, (s)t3.02.01.0μ0Hr(T)tsp= 0. nm00.5 nm1.0 nmK Vuk TBRu thickness, (nm)tsp0.0 1.50.5 1.015010050200Stability threshold(a) (b) (c) (d)H tr( )FIG. 9. (a) The protocol of measuring demagnetization remanence (Md) as a function of increasing inverse applied magneticfield (Hi) performed for the reference Fe-Pt granular film (tsp = 0 nm). After getting each Md in the shown sequence, thefilm was exposed to magnetic saturation, then a higher Hi was applied with certain waiting time at this field. Finally, (b) theMd(Hi) plots at different waiting times allow to extract dynamic coercivities (Hr: Hi at which Md = 0). (c) Hr as functionsof waiting time for Fe-Pt/Ru/Fe-Pt granular films with different Ru thickness (tsp) were analyzed using Eq. (2), giving (d)effective KuV/kBT .L10 order, which resulted in softened magnetic proper-ties of this layer. In addition, different Curie tempera-tures (Tc) were observed for the Fe-Pt/Ru/Fe-Pt films,which were attributed to the top (low Tc) and bottom(high Tc) Fe-Pt layers. The Fe-Pt layers were foundto be exchange coupled when tsp < 2 nm. In particu-lar, the Fe-Pt/Ru(0.5)/Fe-Pt film with the strongest ex-change coupling exhibited an inflectionless out-of-planehysteresis loop, as well as nearly monotonic temperaturedependencies of both magnetization and coercivity. Ad-justing both the effective magnetic anisotropy and Tc ofthe Fe-Pt/Ru(0.5)/Fe-Pt film, we decreased its writingtemperature to 380 K with a sufficient thermal stabilitycoefficient KuV/kBT = 120. Thus, the feasibility of de-veloping exchange-coupled HAMR media with a uniqueFe-Pt/Ru/Fe-Pt trilayer nanogranular structure and themerit of controlling the writing temperature was demon-strated. The actual mechanism of exchange coupling andthe effect of the curved interface on it are the subject offurther studies, as well as the thorough microstructuraloptimization of such HAMR media toward a less defec-tive the top Fe-Pt layer, smaller grain size, etc.ACKNOWLEDGMENTSThe authors thank T. Hiroto, K. Yamaura, Y. Toy-ooka, T. Anzaki, Y. Mori, and A. Nakama (NIMS) fortheir contribution in conducting experiments. This workwas supported in part by the JST-CREST (JPMJC22C3)and MEXT program: Data Creation and Utilization-Type Material Research and Development Project (JP-MXP1122715503).[1] B. Marchon, T. Pitchford, Y.-T. Hsia, and S. Gangopad-hyay, The head-disk interface roadmap to an areal den-sity of 4 Tbit/in2, Adv. Tribol. 2013, 1–8 (2013).[2] D. Weller, G. Parker, O. Mosendz, E. Champion, B.Stipe, X. Wang, T. Klemmer, G. Ju, and A. Ajan, AHAMR media technology roadmap to an areal density of4 Tb/in2, IEEE Trans. Magn. 50, 1–8 (2014).[3] D. Weller, A. Moser, L. Folks, M. Best, W. Lee, M.Toney, M. Schwickert, J.-U. Thiele, and M. Doerner,High Ku materials approach to 100 Gbits/in2, IEEETrans. Magn. 36, 10–15 (2000).[4] O. Ivanov, L. Solina, V. Demshina, and L. Magat, De-termination of the anisotropy constant and saturationmagnetization, and magnetic properties of powders of aniron-platinum alloy, Phys. Met. Metallog. 35, 81 (1973).[5] G. Varvaro and F. Casoli, Ultra-High-Density MagneticRecording: Storage Materials and Media Designs (JennyStanford Publishing, 2016).[6] J. J. M. Ruigrok, Limits of conventional and thermallyas-sisted recording, J. Magn. Soc. Jpn. 25, 313–321 (2001).[7] Y. K. Takahashi, Microstructure control for magneticthin films with high functionality, J. Magn. Soc. Jpn.46, 76–84 (2022).[8] T. D. Trinh, S. Rajauria, R. Smith, E. Schreck, Q. Dai,and F. E. Talke, Temperature-induced near-field trans-ducer failure in heat-assisted magnetic recording, IEEETrans. Magn. 56, 1–4 (2020).[9] B. X. Xu, Z. J. Liu, R. Ji, Y. T. Toh, J. F. Hu, J. M. Li,J. Zhang, K. D. Ye, and C. W. Chia, Thermal issues andtheir effects on heat-assisted magnetic recording system(invited), J. Appl. Phys. 111, 07B701 (2012).[10] S. Bhargava and E. Yablonovitch, Lowering HAMRnearfield transducer temperature via inverse electromag-netic design, IEEE Trans. Magn. 51, 1–7 (2015).9[11] J. Zhang, R. Ji, J. Xu, J. Ng, B. Xu, S. Hu, H. Yuan, andS. Piramanayagam, Lubrication for heat-assisted mag-netic recording media, IEEE Trans. Magn. 42, 2546–2548(2006).[12] M. Kaneko, K. Sueki, and Y. Kitamoto, Magnetic prop-erties of L10 Fe-Ni-Pt films for heat assisted magneticrecording, J. Magn. Soc. Jpn. 30, 588–591 (2006).[13] D. B. Xu, J. S. Chen, T. J. Zhou, and G. M. Chow, Ef-fects of Mn doping on temperature-dependent magneticproperties of L10 FeMnPt, J. Appl. Phys. 109, 07B747(2011).[14] T. Ono, H. Nakata, T. Moriya, N. Kikuchi, S. Okamoto,O. Kitakami, and T. Shimatsu, Addition of Ru to L10-FePt thin film to lower Curie temperature, Appl. Phys.Exp. 9, 123002 (2016).[15] P. Tozman, S. Isogami, I. Suzuki, A. Bolyachkin, H.Sepehri-Amin, S. Greaves, H. Suto, Y. Sasaki, T. Chang,Y. Kubota, P. Steiner, P. Huang, K. Hono, and Y.Takahashi, Dual-layer FePt-C granular media for multi-level heat-assisted magnetic recording, Acta Mater. 271,119869 (2024).[16] D. Suess, Multilayer exchange spring media for magneticrecording, Appl. Phys. Lett. 89, 113105 (2006).[17] D. Suess, T. Schrefl, S. F¨ahler, M. Kirschner, G. Hrkac,F. Dorfbauer, and J. Fidler, Exchange spring media forperpendicular recording, Appl. Phys. Lett. 87, 012504(2005).[18] R. Victora and X. Shen, Composite media for perpendic-ular magnetic recording, IEEE Trans. Magn. 41, 537–542(2005).[19] Y. Sonobe, K. Tham, T. Umezawa, C. Takasu, J. Du-maya, and P. Leo, Effect of continuous layer in CGCperpendicular recording media, J. Magn. Magn. Mater.303, 292–295 (2006).[20] Y. K. Takahashi, K. Hono, S. Okamoto, and O. Ki-takami, Magnetization reversal of FePt hard/soft stackednanocomposite particle assembly, J. Appl. Phys. 100,074305 (2006).[21] T. Shimatsu, N. Asakura, Y. Inaba, K. Kudo, A. Sato, H.Muraoka, H. Aoi, S. Okamoto, and O. Kitakami, Thermalstability and switching field of hard/soft-stacked perpen-dicular media, J. Magn. Magn. Mater. 320, 3088–3091(2008).[22] J. Wang, H. Sepehri-Amin, Y. Takahashi, S. Okamoto,S. Kasai, J. Kim, T. Schrefl, and K. Hono, Magnetiza-tion reversal of FePt based exchange coupled compositemedia, Acta Mater. 111, 47–55 (2016).[23] D. Suess and T. Schrefl, Breaking the thermally inducedwrite error in heat assisted recording by using low andhigh Tc materials, Appl. Phys. Lett. 102, 162405 (2013).[24] O. Muthsam, F. Slanovc, C. Vogler, and D. Suess, Im-proving the signal-to-noise ratio for heat-assisted mag-netic recording by optimizing a high/low Tc bilayer struc-ture, J. Appl. Phys. 126, 123907 (2019).[25] K. Eason, H. T. Wang, M. R. Elidrissi, B. Xu, Z. Yuan,and K. S. Chan, Recording performance and compari-son of graded-Tc and -Ku HAMR systems, IEEE Trans.Magn. 50, 107–113 (2014).[26] N. Natekar, W. Tipcharoen, and R. Victora, Compositemedia with reduced write temperature for heat assistedmagnetic recording, J. Magn. Magn. Mater. 486, 165253(2019).[27] N. Zhou, X. Xu, A. T. Hammack, B. C. Stipe, K. Gao, W.Scholz, and E. C. Gage, Plasmonic near-field transducerfor heat-assisted magnetic recording, Nanophotonics 3,141–155 (2014).[28] O. Muthsam, C. Vogler, and D. Suess, Noise reduction inheat-assisted magnetic recording of bit-patterned mediaby optimizing a high/low Tc bilayer structure, J. Appl.Phys. 122, 213903 (2017).[29] B. M. Lairson and B. M. Clemens, Enhanced magnetoop-tic Kerr rotation in epitaxial PtFe(001) and PtCo(001)thin films, Appl. Phys. Lett. 63, 1438–1440 (1993).[30] J.-U. Thiele, L. Folks, M. F. Toney, and D. K. Weller,Perpendicular magnetic anisotropy and magnetic domainstructure in sputtered epitaxial FePt (001) L10 films, J.Appl. Phys. 84, 5686–5692 (1998).[31] I. Suzuki, J. Wang, Y. K. Takahashi, and K. Hono, Con-trol of grain density in FePt-C granular thin films dur-ing initial growth, J. Magn. Magn. Mater. 500, 166418(2020).[32] C.-b. Rong, N. Poudyal, G. S. Chaubey, V. Nandwana, R.Skomski, Y. Q. Wu, M. J. Kramer, and J. P. Liu, Struc-tural phase transition and ferromagnetism in monodis-perse 3 nm FePt particles, J. Appl. Phys. 102, 043913(2007).[33] S. D. Granz and M. H. Kryder, Granular L10 FePt (001)thin films for heat assisted magnetic recording, J. Magn.Magn. Mater. 324, 287–294 (2012).[34] E. Yang, D. E. Laughlin, and J.-G. Zhu, Correction oforder parameter calculations for FePt perpendicular thinfilms, IEEE Trans. Magn. 48, 7–12 (2012).[35] A. Perumal, Y. K. Takahashi, and K. Hono, L10 FePt–Cnanogranular perpendicular anisotropy films with narrowsize distribution, Appl. Phys. Exp. 1, 101301 (2008).[36] E. Yang, D. E. Laughlin, and J.-G. Zhu, Buffer layers forhighly ordered L10 FePt-oxide thin film granular mediaat reduced processing temperature, IEEE Trans. Magn.46, 2446–2449 (2010).[37] J. Christodoulides, P. Farber, M. Dannl, H. Okumura,G. Hadjipanaysi, V. Skumryev, A. Simopoulos, and D.Weller, Magnetic, structural and microstructural proper-ties of FePt/M (M = C, BN) granular films, IEEE Trans.Magn. 37, 1292–1294 (2001).[38] N. Kulesh, A. Bolyachkin, I. Suzuki, Y. Takahashi, H.Sepehri-Amin, and K. Hono, Data-driven optimization ofFePt heat-assisted magnetic recording media acceleratedby deep learning TEM image segmentation, Acta Mater.255, 119039 (2023).[39] M. D. Kuz’min, Shape of temperature dependence ofspontaneous magnetization of ferromagnets: Quantita-tive analysis, Phys. Rev. Lett. 94, 107204 (2005).[40] R. Chang, S. Li, M. V. Lubarda, B. Livshitz, and V. Lo-makin, FastMag: Fast micromagnetic simulator for com-plex magnetic structures (invited), J. Appl. Phys. 109,07D358 (2011).[41] A. Bolyachkin, H. Sepehri-Amin, I. Suzuki, H. Tajiri, Y.Takahashi, K. Srinivasan, H. Ho, H. Yuan, T. Seki, A.Ajan, and K. Hono, Transmission electron microscopyimage based micromagnetic simulations for optimizingnanostructure of FePt-X heat-assisted magnetic record-ing media, Acta Mater. 227, 117744 (2022).[42] See Supplemental Material at [URL will be inserted bypublisher], for details about the cross-sectional HAADF-STEM images of Fe-Pt/Ru/Fe-Pt continuous films andfor out-of-plane hysteresis loops of these films with vari-ous temperatures.10[43] M. Futamoto, M. Nakamura, T. Shimizu, M. Ohtake, andN. Inaba, Influence of stress and strain on L10-orderedphase formation in FePt thin film, IEEE Trans. Magn.54, 1–4 (2018).[44] I. Suzuki, S. Kubo, H. Sepehri-Amin, and Y. K. Taka-hashi, Dependence of the growth mode in epitaxial FePtL10 films on surface free energy, ACS Appl. Mater. In-terfaces 13, 16620–16627 (2021).[45] R. F. L. Evans, Q. Coopman, S. Devos, W. J. Fan, O.Hovorka, and R. W. Chantrell, Atomistic calculation ofthe thickness and temperature dependence of exchangecoupling through a dilute magnetic oxide, J. Phys. D:Appl. Phys. 47, 502001 (2014).[46] J. Jiang, N. Tezuka, and K. Inomata, Exchange couplingbetween FePt and Fe through Ru interlayer, J. Appl.Phys. 98, 063902 (2005).[47] J. Jiang, N. Tezuka, and K. Inomata, Indirect exchangespring between FePt and Fe with a Ru interlayer, J.Magn. Magn. Mater. 302, 40–46 (2006).[48] T. Nomoto, T. Koretsune, and R. Arita, Formationmechanism of the helical Q structure in Gd-basedskyrmion materials, Phys. Rev. Lett. 125, 117204 (2020).[49] S. Liang, R. Chen, Q. Cui, Y. Zhou, F. Pan, H. Yang,and C. Song, Ruderman–Kittel–Kasuya–Yosida-type in-terlayer Dzyaloshinskii–Moriya interaction in syntheticmagnets, Nano Letters 23, 8690 (2023).[50] T. Fukushima, H. Akai, T. Chikyow, and H. Kino, Au-tomatic exhaustive calculations of large material spaceby Korringa-Kohn-Rostoker coherent potential approx-imation method applied to equiatomic quaternary highentropy alloys, Phys. Rev. Mater. 6, 023802 (2022).[51] D. Isurugi, T. Saito, S. Kaneko, K. K. Tham, T. Ogawa,and S. Saito, Evaluation of blocking temperature and itsdistribution for L10-type FePt granular films, J. J. Appl.Phys. 62, 045503 (2023).[52] Z. Li, D. Wei, and F. Wei, Micromagnetic modeling forheat-assisted magnetic recording, J. Magn. Magn. Mater.320, 3108–3112 (2008).[53] M. Sharrock, Recent advances in metal particulaterecording media: toward the ultimate particle, IEEETrans. Magn. 36, 2420–2425 (2000).[54] D. Suess, S. Eder, J. Lee, R. Dittrich, J. Fidler, J. W.Harrell, T. Schrefl, G. Hrkac, M. Schabes, N. Supper, andA. Berger, Reliability of Sharrocks equation for exchangespring bilayers, Phys. Rev. B 75, 174430 (2007).[55] J. Harrell, Orientation dependence of the dynamic coer-civity of Stoner-Wohlfarth particles, IEEE Trans. Magn.37, 533–537 (2001).[56] S. Okamoto, Experimental approaches for micromagneticcoercivity analysis of advanced permanent magnet mate-rials, Sci. Technol. Adv. 22, 124–134 (2021).[57] P. K. Venuthurumilli, Z. Zeng, and X. Xu, Inverse designof near-field transducer for heat-assisted magnetic record-ing using topology optimization, IEEE Trans. Magn. 57,1–6 (2021).[58] N. A. Natekar, E. Roddick, and R. M. Brockie, Inter-play of the thermal and magnetic fields in HAMR, IEEE.Trans. Magn. 58, 1–8 (2022).[59] F. Akagi and N. Matsushima, Relationship between tem-perature rise and thermal conductivity in a magneticmedium during heated dot magnetic recording, J. J.Appl. Phys. 62, SB1001 (2022).[60] R.-V. Ababei, M. O. A. Ellis, R. F. L. Evans, and R.W. Chantrell, Anomalous damping dependence of theswitching time in Fe/FePt bilayer recording media, Phys.Rev. B 99, 024427 (2019).[61] C. Liu, K. Srinivasan, A. Ajan, E. McCollum, A.Kalitsov, V. Kalappattil, and M. Wu, Ferromagnetic res-onance in FePt thin films at elevated temperatures, J.Magn. Magn. Mater. 563, 169988 (2022).[62] Y. Sasaki, I. Suzuki, R. Mandal, S. Kasai, and Y. K.Takahashi, Thermal modulation of magnetization dy-namics in nanometer-thick L10-FePt nanogranular andcontinuous films for high-density magnetic recording me-dia, ACS Appl. Nano Mater. 6, 5901–5908 (2023).[63] D. Richardson, S. Katz, J. Wang, Y. K. Takahashi, K.Srinivasan, A. Kalitsov, K. Hono, A. Ajan, and M. Wu,Near-Tc ferromagnetic resonance and damping in FePt-based heat-assisted magnetic recording media, Rhys.Rev. Appl. 10, 054046 (2018).[64] I. Kurniawan, Y. Miura, G. Xing, T. Tadano, and K.Hono, Theoretical study of the effect of lattice dynamicson the damping constant of FePt at finite temperature,Phys. Rev. B 108, 094426 (2023).[65] C. Xu, B. Zhou, T. Du, B. S. D. C. S. Varaprasad, D.E. Laughlin, and J.-G. J. Zhu, Understanding the growthof high-aspect-ratio grains in granular L10-FePt thin-filmmagnetic media, APL Mater 10, 051105 (2022).[66] C. Xu, B. S. D. C. S. Varaprasad, D. E. Laughlin, andJ.-G. Zhu, Fabrication of 16 nm Thick Granular L10FePt-hBN Thin Film Media, IEEE Trans. Magn. 59, 1–5(2023).[67] B. S. D. C. S. Varaprasad, C. Xu, M.-H. Huang, D. E.Laughlin, and J.-G. Zhu, FePt–BN granular HAMR me-dia with high grain aspect ratio and high L10 orderingon corning LotusTM NXT glass, AIP Adv 13, 035002(2023).[68] C. Xu, B. S. D. C. S. Varaprasad, D. E. Laughlin, and J.-G. Zhu, Bias sputtering of granular L10-FePt films withhexagonal boron nitride grain boundaries, Sci. Rep 13,11087 (2023).[69] T. Shiroyama, B. S. D. C. S. Varaprasad, Y. K. Taka-hashi, and K. Hono, Microstructure and Magnetic Prop-erties of FePt–Cr2O3 Films, IEEE Trans. Magn. 50, 1–4(2014).[70] I. Suzuki, T. Abe, H. Sepehri-Amin, K. Hono, andY. Takahashi, Microstructure evolution in FePt-Cr2O3granular thin films, J. Magn. Magn. Mater 579, 170874(2023).