# Fileset

[126-Appl. Phys. Lett. 122, 112404 (2023)-Scheike.pdf](https://mdr.nims.go.jp/filesets/5fe9e0c1-d242-42c7-a12d-101f45f7c8d5/download)

## Creator

[Thomas Scheike](https://orcid.org/0000-0002-9163-5524), [Zhenchao Wen](https://orcid.org/0000-0001-7496-1339), [Hiroaki Sukegawa](https://orcid.org/0000-0002-4034-7848), [Seiji Mitani](https://orcid.org/0000-0002-1348-0774)

## Rights

©2023 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http:// creativecommons.org/licenses/by/4.0/).[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[631% room temperature tunnel magnetoresistance with large oscillation effect in CoFe/MgO/CoFe(001) junctions](https://mdr.nims.go.jp/datasets/4a5676f0-98dc-4a10-bb8e-ebaaf831f7da)

## Fulltext

631% room temperature tunnel magnetoresistance with large oscillation effect in CoFe/MgO/CoFe(001) junctionsAppl. Phys. Lett. 122, 112404 (2023); https://doi.org/10.1063/5.0145873 122, 112404© 2023 Author(s).631% room temperature tunnelmagnetoresistance with large oscillationeffect in CoFe/MgO/CoFe(001) junctions Cite as: Appl. Phys. Lett. 122, 112404 (2023); https://doi.org/10.1063/5.0145873Submitted: 08 February 2023 • Accepted: 28 February 2023 • Published Online: 15 March 2023 Thomas Scheike,  Zhenchao Wen,  Hiroaki Sukegawa, et al.COLLECTIONS This paper was selected as Featuredhttps://images.scitation.org/redirect.spark?MID=176720&plid=2023706&setID=378288&channelID=0&CID=740896&banID=520944490&PID=0&textadID=0&tc=1&type=tclick&mt=1&hc=227cb5a11ec71abd444d22629d95bb0bf1d563cc&location=https://doi.org/10.1063/5.0145873https://aip.scitation.org/topic/collections/featured?SeriesKey=aplhttps://doi.org/10.1063/5.0145873https://orcid.org/0000-0002-9163-5524https://aip.scitation.org/author/Scheike%2C+Thomashttps://orcid.org/0000-0001-7496-1339https://aip.scitation.org/author/Wen%2C+Zhenchaohttps://orcid.org/0000-0002-4034-7848https://aip.scitation.org/author/Sukegawa%2C+Hiroakihttps://aip.scitation.org/topic/collections/featured?SeriesKey=aplhttps://doi.org/10.1063/5.0145873https://aip.scitation.org/action/showCitFormats?type=show&doi=10.1063/5.0145873http://crossmark.crossref.org/dialog/?doi=10.1063%2F5.0145873&domain=aip.scitation.org&date_stamp=2023-03-15631% room temperature tunnel magnetoresistancewith large oscillation effect in CoFe/MgO/CoFe(001)junctionsCite as: Appl. Phys. Lett. 122, 112404 (2023); doi: 10.1063/5.0145873Submitted: 8 February 2023 . Accepted: 28 February 2023 .Published Online: 15 March 2023Thomas Scheike, Zhenchao Wen, Hiroaki Sukegawa,a) and Seiji MitaniAFFILIATIONSNational Institute for Materials Science (NIMS), Tsukuba 305-0047, Japana)Author to whom correspondence should be addressed: SUKEGAWA.Hiroaki@nims.go.jpABSTRACTWe demonstrate tunnel magnetoresistance (TMR) ratios of up to 631% at room temperature (RT) using CoFe/MgO/CoFe(001) epitaxialmagnetic tunnel junctions (MTJs). The TMR ratio increased up to 1143% at 10 K. The large TMR ratios resulted from fine-tuning of atomic-scale structures of the MTJs, such as crystallographic orientations and MgO interface oxidation by interface insertion of ultrathin CoFe andMg layers, which are expected to enhance the well-known D1 coherent tunneling transport. Interestingly, the TMR oscillation effect, which isnot covered by the standard coherent tunneling theory, also became significant. A 0.32-nm period TMR oscillation with increasing MgOthickness dominates the transport in a wide range of MgO thicknesses; the peak-to-valley difference of the TMR oscillation exceeds 140% atRT, which is attributed to the appearance of large oscillatory components in the resistance area product.VC 2023 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/5.0145873Tunnel magnetoresistance (TMR) effect, which is based on spin-dependent tunneling between ferromagnets (FMs), has been studiedin applied physics for over 40 years.1 It has been extensively investi-gated from both theoretical and experimental perspectives and hasbeen implemented in practical device applications such as hard diskdrives (HDDs) and nonvolatile magnetoresistive random accessmemories (MRAMs).2 At present, almost all applications use MgObarrier-based magnetic tunnel junctions (MTJs) due to the largeobtainable TMR.3 The largest experimental TMR ratio at room tem-perature (RT) is 604%, which was reported in 2008 by Ikeda et al.4using a pseudo-spin-valve CoFeB/MgO/CoFeB MTJ. However, theprogress in the RT-TMR ratio has stagnated over the recent decade.Notably, practical MTJ-based devices can achieve only 100%–200%TMR ratios at best to simultaneously satisfy several requirements fortheir operation such as circuit impedance matching, high-speed opera-tion, good thermal stability, low power consumption for operation,and low 1/f noise at a low frequency range.5–12A major drawback of MTJs as an integrated circuit element istheir small current on/off ratio (i.e., small TMR ratio). If much largerRT-TMR ratios can be achieved, the application ranges of TMR-baseddevices will significantly expand to very high-density MRAMs basedon a three-dimensional architecture, nonvolatile magnetic logics,13brain-morphic devices,14 etc. Some of the authors recently investigatedsimple Fe/MgO/Fe(001) MTJs with a significantly improved TMR ratioup to 417% and 914% at RT and 3K, respectively, by fine-tuning thecrystallinity near the MgO barrier interfaces.15 A �80% peak-to-valley(PV) difference in the RT-TMR ratio oscillation was shown,15 indicatingthe dominance of the transport properties by the oscillation effect.Furthermore, a MTJ with a spinel-based Mg4Al-Ox barrier, i.e., Fe/Mg4Al-Ox/Fe(001), showed a similar �0.3nm period oscillation withan enhanced PV difference of �125% and a TMR ratio of 429%at RT.16In this Letter, we report on giant TMR effect using CoFe/MgO/CoFe(001) epitaxial MTJs by improving the nanostructures of a MTJstack. The RT-TMR ratio reached a maximum of 631%, which is thelargest RT value reported in MTJs. The significant TMR enhancementwas due to the introduction of sub-nm to a few nm CoFe layers atboth the top- and bottom-sides of the barrier interface, together withthickness optimization of the Mg insertion at the bottom Fe/MgOinterface of the Fe/MgO/Fe(001) framework structure. In addition, wedemonstrated a significantly large TMR oscillation with a maximumPV difference of 141% in the CoFe/MgO/CoFe(001) MTJs at RT. Ourdemonstrations of the large RT-TMR ratio and its oscillatory behaviorare crucial steps for developing spintronic applications in the future.Appl. Phys. Lett. 122, 112404 (2023); doi: 10.1063/5.0145873 122, 112404-1VC Author(s) 2023Applied Physics Letters ARTICLE scitation.org/journal/aplhttps://doi.org/10.1063/5.0145873https://doi.org/10.1063/5.0145873https://doi.org/10.1063/5.0145873https://www.scitation.org/action/showCitFormats?type=show&doi=10.1063/5.0145873http://crossmark.crossref.org/dialog/?doi=10.1063/5.0145873&domain=pdf&date_stamp=2023-03-15https://orcid.org/0000-0002-9163-5524https://orcid.org/0000-0001-7496-1339https://orcid.org/0000-0002-4034-7848https://orcid.org/0000-0002-1348-0774mailto:SUKEGAWA.Hiroaki@nims.go.jphttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/https://doi.org/10.1063/5.0145873https://scitation.org/journal/aplThe MTJ stacks were deposited using an ultrahigh-vacuummultichamber deposition apparatus (ULVAC, Inc., base pressure:4� 10�7Pa). Typical stack structures were MgO(001) single crystalsubstrate//Cr (60)/Fe (50)/Co50Fe50 (CoFe) (dbot-CoFe)/Mg (dMg)/wedge-shaped MgO (dMgO¼ 1.0� 3.0)/natural oxidation/CoFe(dtop-CoFe)/Fe (5)/Ir20Mn80 (10)/Ru (20) [values in parentheses in nm,see Fig. 1(a)]. All metallic layers were deposited by DC sputtering. TheMgO barrier was deposited using electron-beam (EB) evaporation of asintered MgO pellet and a linear shutter to create the wedge shape forthe x-direction of a 20� 10mm2 substrate [see Fig. 1(b)]. After depos-iting each layer, in situ postannealing was performed. More details onthe film preparation are described in Ref. 15. The single crystal cubicstructures of bottom- and top-Fe electrodes and the MgO barrier wereconfirmed using reflection high energy electron diffraction (RHEED)observation [Fig. 1(c)]. The multilayer wafers were ex situ annealed at200 �C in a 0.2-T magnetic field along the MgO[110] jj Fe[100] jjCoFe[100] direction. TMR ratios and resistance area products (RAs)of the wafers were evaluated using the current in-plane tunneling(CIPT) method (Capres A/S, CIPTech-SPM200 prober).17 From Fig.1(d), the multilayer was patterned into MTJs with a 10� 5lm2 ellip-soidal pillar using photolithography, Ar ion etching, and magnetronsputtering for the SiO2 insulator and Au electrodes. The magnetotran-sport properties were characterized by a DC 4-probe method using asourcemeter (Keithley, 2400) and nanovoltmeter (Keithley, 2182A) atRT and physical property measurement system (Quantum Design,Dynacool) for low temperature (LT) measurements. The TMR ratio isdefined as (RAP � RP)/RP � 100%, where RP [RAP] is the resistance inthe parallel (P) [antiparallel (AP)] magnetization state. At positive biasvoltage, electrons tunnel from the top to the bottom electrode.First, we investigated the bottom-CoFe insertion effect using thestack structure shown in the inset of Fig. 2(a), i.e., Fe/CoFe (dbot-CoFe)/Mg/MgO/Fe. In Fig. 2(a), the TMR ratio measured at RT with a biasvoltage less than 10mV as a function of dbot-CoFe in ML (bottom axis)and nm (top axis) of patterned MTJ pillars is shown. Each TMR ratiois the maximum value for the corresponding wafer. The TMR ratioincreases with dbot-CoFe, and it saturates at around 16 ML. The maxi-mum TMR ratio was 504% at dbot-CoFe¼ 24 ML.The MgO interfaces were further tuned by controlling the CoFelayer thickness at the top-side (dtop-CoFe) and the Mg layer thickness atthe bottom-side (dMg), as shown in the inset of Fig. 2(b), i.e., CoFe (16ML)/Mg (dMg)/MgO/CoFe (dtop-CoFe). Figure 2(b) shows the maximumRT-TMR ratio measured as a function of dtop-CoFe. When dtop-CoFe wasincreased with maintaining dMg¼ 0.5 nm (green line), the maximumTMR ratio of 551% was observed at dtop-CoFe¼ 4 ML. MTJs with largerdtop-CoFe showed smaller values, indicating that the CoFe insertion effecton the TMR for the top interface differs from that for the bottom one.This behavior is attributable to the difference in their growth mode,which is sensitive to the state of the interfaces, e.g., interfacial roughness,degree of (001)-orientation, and interfacial strain. Figure 2(b) alsoshows the relationship between the maximum TMR ratio and dMg. TheTMR ratio further increased by optimizing dMg; the maximum value of631% at RT was observed at dtop-CoFe¼ 4 ML and dMg¼ 0.6 nm [blackopen squares in Fig. 2(b)]. This TMR ratio is larger than the RT-TMRratio of 604% reported by Ikeda et al.4 using a CoFeB/MgO/CoFeBpseudo-spin-valve MTJ. Therefore, better MgO crystallinity than that ofthe CoFeB/MgO/CoFeB structure, which requires a high post-annealing temperature (525 �C), could be realized at a lower tempera-ture (up to 400 �C) in the epitaxial MTJ in this study. Because the Mginsertion mainly works as protection for the bottom electrode interfacefrom oxidation during the deposition of the MgO barrier and postan-nealing process in our MTJs, dMg¼ 0.6 nm is probably the optimumthickness that realizes favorable interface bonding states between CoFeand oxygen atoms at the MgO surface. A lower and higher dMg reducesthe TMR ratio due to slight over- and under-oxidation at the bottomMgO interface, respectively, showing the sensitivity of the bottom inter-face to the oxidation condition. Notably, for different electrode materi-als, the optimum Mg insertion thickness may change, e.g.,dMg¼ 0.5 nm for the Fe electrode.15 Note that the RA decreased slightlywhen the Mg thickness was increased from 0.5 to 0.6 nm and remainedunchanged when further increased to 0.7 nm.The resistance (left axis) and TMR ratio (right axis) as a functionof the magnetic field (H) at RT for the MTJ showing the maximumTMR ratio of 631% (dbot-CoFe¼ 16ML, dtop-CoFe¼ 4ML, dMg¼ 0.6 nm,and dMgO¼ 1.86nm) are shown in Fig. 2(c). The curve shows a typicalFIG. 1. (a) Schematic MTJ stacking structure and processes. (b) Images of a MTJ wafer. (c) RHEED patterns along the MgO[100] azimuth of top-Fe (upper), the MgO barrier(middle), and bottom-Fe (lower). (d) Schematic structure of a patterned MTJ pillar after microfabrication.Applied Physics Letters ARTICLE scitation.org/journal/aplAppl. Phys. Lett. 122, 112404 (2023); doi: 10.1063/5.0145873 122, 112404-2VC Author(s) 2023https://scitation.org/journal/aplexchange-biased hysteresis loop with stable P and AP states at RT. TheMTJ has a junction resistance of 70 X (RA¼ 2.74 kX�lm2) for the Pstate, which increases to 7.31 times (¼514 X) when switched to the APstate. To exclude measurement or microfabrication errors, we also eval-uated zero-bias TMR ratios and RA values by CIPT measurements ofthe unpatterned wafer with dbot-CoFe¼ 16 ML, dtop-CoFe¼ 4 ML, anddMg¼ 0.6 nm. Figures 2(d) and 2(e) show the CIPT results near thewafer position of dMgO¼ 1.9 nm: (d) the sheet resistance in the P state(Rlow( ) and (e) the current-in-plane TMR ratio (MRcip) vs the meanprobe pitch.17 The fits (circles) by the theoretical equations match wellwith the measured data (cross marks). We obtained reasonable valuesof 617% (RA¼ 3.4 kX�lm2) by the fit. Therefore, the RT-TMR ratioexceeding 600% was confirmed by both an unpatterned wafer (CIPT)and patterned MTJ pillars (DC 4-probe).Figure 3 shows the temperature dependences of the (a) TMRratio and (b) RP and RAP (bias voltage< 10mV). The correspondingconductance ratio [¼ (gP � gAP)/gP¼ (RAP � RP)/RAP, where gP(AP)� 1/RP(AP) is the DC conductance in the P (AP) state] is also plottedon the right axis. The TMR ratio and conductance ratio monotonicallyincrease with decreasing temperature. The TMR ratio reaches a maxi-mum of 1143% at 10K, which is much larger than the value of the pre-vious Fe/MgO/Fe (914%)15 and almost the same as the value of theCoFeB/MgO/CoFeB at LT (1144%) by Ikeda et al.4The inset of Figs. 3(a) and 3(b) shows the corresponding TMR-Hloop at 10K. Below 10K, the TMR ratio reduces slightly because of animperfect AP state, suggesting that the observed value is underesti-mated at the LT limit. Using the Julliere model1 and TMR ratio(%)¼ 100� 2 P2eff=ð1� P2eff Þ, where Peff is the effective spinFIG. 2. (a) Maximum RT-TMR ratio vs bottom-CoFe insertion thickness dbot-CoFe. (b) Maximum TMR ratio vs top-CoFe insertion thickness dtop-CoFe and Mg insertion thicknessdMg. Insets of (a) and (b) show schematic stacking structures. (c) Resistance (left axis) and TMR ratio (right axis) vs magnetic field l0H of a MTJ with maximum RT-TMR ratio(dbot-CoFe¼ 16 ML, dtop-CoFe¼ 4 ML, dMg¼ 0.6 nm, and dMgO¼ 1.86 nm). (d) and (e) CIPT results of the unpatterned wafer of (c); (d) Rlow( and (e) MRcip. Rt (Rb) indicates thesheet resistance of the top (bottom) electrode.Applied Physics Letters ARTICLE scitation.org/journal/aplAppl. Phys. Lett. 122, 112404 (2023); doi: 10.1063/5.0145873 122, 112404-3VC Author(s) 2023https://scitation.org/journal/aplpolarization, Peff at RT (LT) was calculated to be 0.871 (0.923) byassuming both interfaces had the same Peff. The ratio of Peff at RT toLT [Peff(RT)/Peff(LT)] was 0.94, which was higher than 0.91 of previ-ous Fe/MgO/Fe.15 The weaker temperature dependence is attributableto the increased interface Curie temperature by Co doping into Fe,18,19which may effectively improve the interlayer exchange stiffness con-stant.20 A recent theoretical work considering an intra-atomic s–dexchange interaction also predicted this improvement in CoFe/MgO/CoFe(001) with an increase in the Co composition.21 It was alsoreported for a bcc-Co/MgO/bcc-Co(001) MTJ [Peff(RT)/Peff(LT)¼ 0.97].22 Therefore, introducing CoFe at both MgO interfa-ces effectively suppressed the temperature dependence of Peff in addi-tion to the high Peff at LT, yielding a giant RT-TMR ratio in thepresent MTJ.The temperature dependence of the TMR ratio follows that ofRAP rather than RP, which is commonly observed in variousMTJs.23–25 Interestingly, the temperature dependence of RP shows acomplicated behavior with two slope changes, which is different fromthe dependence of RAP that shows a monotonic change. Similar behav-ior in RP has been reported in Co-based electrode MTJs with largeTMR ratios such as Co2(Mn,Fe)Si and Co2FeAl,24,25 which contradictspure Fe electrode MTJs that show a slight monotonic decrease in RPwith decreasing temperature,15,16,26–28 suggesting a difference in elec-tronic structures between Fe and Co at MgO interfaces.The bias voltage dependences of the differential conductance forAP (GAP) and P (GP) at 300 and 5K are shown in Figs. 3(c) and 3(d),respectively. Figures 3(e) and 3(f) show the bias voltage dependencesof the TMR normalized by its zero-bias value and the output voltageVout [� jVj � (RAP � RP)/RAP, where V is the bias voltage], respec-tively. The differential conductance G was obtained by the numericaldifferentiation of current–voltage curves. GP spectra are asymmetricand have clear minimum structures at �0.3, �0.7, and þ0.4V, whichare pronounced at 5K. The minimum structures in the GP spectra arelarger than those in typical CoFe-based MTJs.29–31 The minimumstructure of the positive bias is much deeper than that of the negativebias; the relative change from the zero-bias value reaches �34%(�29%) at 5K (300K). The minimum structures at negative bias areshallower than that at positive bias. However, the appearance of thetwo minima is similar to the case of the Fe/MgO/Fe(001) MTJs with alarge RT-TMR ratio.15 Tunneling electrons primarily sense the finalstate, i.e., the top- (bottom-) interface electronic structures at the nega-tive (positive) bias voltage. Thus, the GP spectra at the negative biasmay represent the electronic structure of the top-MgO/CoFe(001)interface, which resembles that of a pure Fe/MgO(001) interface.15 Incontrast, the deep minimum at the positive bias may be associatedwith specific electronic structures of the bottom-CoFe/MgO(001)interface. These asymmetric features are attributable to the thicknessdifference of the bottom- and top-CoFe insertion layers (i.e.,dbot-CoFe¼ 16 ML � dtop-CoFe¼ 4 ML), implying that furtherimprovement in the RT-TMR ratio can be expected if symmetric spec-tra are obtained by creating defect-free and well-balanced electronicstates between the top- and bottom-side interfaces.Vhalf values, the bias voltage where the TMR ratio reduces to halfof the zero-bias value, at 300 and 5K for the positive (negative) biasare 0.51V (�0.49V) and 0.28V (�0.30V), respectively. The curvesappear nearly symmetric because of the symmetric GAP feature. TheFIG. 3. Temperature dependences of TMR properties of the MTJ with dbot-CoFe¼ 16 ML, dtop-CoFe¼ 4 ML, dMg¼ 0.6 nm, and dMgO¼ 1.86 nm. (a) TMR ratio (left axis) andconductance ratio (right axis). (b) RP (right axis) and RAP (left axis). Inset: TMR ratio vs magnetic field l0H at 10 K. (c)–(f) Bias voltage dependence measured at 300 K (RT)and 5 K. (c) and (d) Differential conductances GAP and GP, respectively. (e) TMR normalized by zero-bias value. (f) Output voltage Vout � jVj � (RAP � RP)/RAP.Applied Physics Letters ARTICLE scitation.org/journal/aplAppl. Phys. Lett. 122, 112404 (2023); doi: 10.1063/5.0145873 122, 112404-4VC Author(s) 2023https://scitation.org/journal/apllower Vhalf compared with Fe/MgO/Fe32 is mainly due to the CoFeband structure, i.e., the effect of the lowered minority D1 band edge byCo doping into Fe, as seen in bcc-Co/MgO/bcc-Co.22 Vout vs bias volt-age shown in Fig. 3(f) is slightly asymmetric with respect to the biaspolarity reaching 0.68V at 300K (0.77V at 5K) at the positive biasregion. The value at 300K is larger than the values reported in CoFeB/MgO/CoFeB MTJs: 0.3833 and 0.56V.34 For our definition, Vout at lowbias regions nearly follows the line that assumes an infinite TMR(dashed–dotted line) due to the observed large TMR ratio.Figures 4(a) and 4(b), respectively, show the TMR ratio and RAin the P and AP states at RT for the MTJs with dbot-CoFe¼ 16 ML,dtop-CoFe¼ 4 ML, and dMg¼ 0.6nm as a function of dMgO (jVj< 10mV). To obtain consistent plots, we measured the MTJ seriesalong the x-direction of the wafer (MgO wedge direction) at the samey position [Fig. 1(b)]. From Fig. 4(a), the TMR ratio increased rapidlyfor dMgO > 1.2 nm and showed significant oscillation with dMgO. Theoscillation period was approximately 0.32 nm, which was almost iden-tical to the values in Fe/MgO/Fe.15,35 The maximum PV difference of141% is larger than previous reports.32,35,36 Suppression of the oscilla-tion toward a thicker barrier is attributed to the deviation from theoptimum interface condition due to the use of the constantdMg¼ 0.6 nm for the entire wafer. A 0.9-nm period oscillation, whichwas reported by Matsumoto et al.,35 was not observed in our MTJs.Notably, the significant TMR and RA reduction for dMgO < 1.3nmin Fig. 4 is mainly because the MTJ resistances in the low dMgO region(less than a few X) are too small to neglect the effect of the electroderesistance (several X). In fact, a CIPT result at dMgO   1.25 nm(RA¼ 69 X�lm2) showed a TMR ratio of 348%, which is larger thanthe value of a microfabricated pillar (�270%).From Fig. 4(b), the plots of ln(RA) show a linear increase forboth P and AP states in a dMgO range of 1.4–2.8nm. The oscillatorycomponents of the RAs were extracted by the slope-correction of theln(RA) plots.16,35,37 Here, RAr was fitted by exp(ardMgO þ br), wherear and br are the fit parameters and r¼AP or P. We obtain theslopes aP¼ 6.159 and aAP¼ 6.145 nm�1 by fits, indicating that theslopes are nearly identical. Figures 4(c) and 4(d) show the extractedoscillatory components, RAr/exp(ardMgO þ br), for the AP and Pstates, respectively. Note that scattering is likely caused by minor dam-age to the respective MTJ pillar during microfabrication. Both theplots show significant oscillatory behavior with dMgO, which deter-mines the TMR oscillations. Their periods are similar to the previousexperimental values for different MTJs with bcc-structured electrodes(0.306 0.02 nm).32,35 Notably, the oscillations no longer follow a sim-ple sine curve, as observed in the Fe/Mg4Al-Ox/Fe MTJs.16In summary, we observed a giant TMR ratio of up to 631%(1143%) at RT (10K) in a CoFe/MgO/CoFe(001) MTJ. The giantvalue is attributed to improved (001)-orientation and MgO barrierinterface crystallinity by introducing ultrathin Mg and CoFe insertionlayers and controlling the oxidation states. Furthermore, the observedTMR ratio showed a significant oscillatory behavior as a function ofdMgO; the oscillation PV difference reached 141% at RT. The giantTMR ratio and its significant oscillation could be correlated due tocommon underlying physics. Although we observed a large TMR ratioand its oscillation, asymmetric bias voltage dependences were stillobserved, indicating that further improvement in the RT-TMR ratio isexpected if more symmetric characteristics are achieved by tuning bar-rier interfaces such as improving the interface crystallinity by nanoin-sertions and using the lattice-matched barrier, e.g., MgAl2O4.The authors are grateful to Shinji Yuasa for his valuablecomments on the TMR oscillation effect of MgO-based MTJs. Theauthors thank Yoshio Miura and Keisuke Masuda for their fruitfuldiscussion from theoretical viewpoints, and Hiromi Ikeda for hertechnical support on device microfabrication. This paper is partlybased on the results obtained from a Project commissioned by the NewEnergy and Industrial Technology Development Organization(NEDO) via No. JPNP16007. This work was partly supported by JSPSKAKENHI via Grant Nos. 16H06332, 21H01750, 21H01397, and22H04966, and the ImPACT Program of the Council for Science,Technology and Innovation (Cabinet Office, Government of Japan).AUTHOR DECLARATIONSConflict of InterestThe authors have no conflicts to disclose.Author ContributionsThomas Scheike: Conceptualization (equal); Data curation (lead);Formal analysis (equal); Investigation (lead); Methodology (lead);Visualization (equal); Writing – original draft (lead); Writing –FIG. 4. dMgO dependences of (a) TMR ratio and (b) RA for P and AP states at RT.(c) and (d) Background corrected RA plots for P and AP states, respectively, withr¼ (P, AP).Applied Physics Letters ARTICLE scitation.org/journal/aplAppl. Phys. Lett. 122, 112404 (2023); doi: 10.1063/5.0145873 122, 112404-5VC Author(s) 2023https://scitation.org/journal/aplreview & editing (equal). Zhenchao Wen: Data curation (supporting);Formal analysis (supporting); Methodology (supporting); Validation(equal); Writing – original draft (supporting); Writing – review & edit-ing (supporting). Hiroaki Sukegawa: Conceptualization (lead);Funding acquisition (lead); Investigation (supporting); Methodology(equal); Supervision (lead); Validation (lead); Visualization (support-ing); Writing – original draft (equal); Writing – review & editing(lead). Seiji Mitani: Funding acquisition (lead); Methodology (sup-porting); Supervision (equal); Writing – review & editing (equal).DATA AVAILABILITYThe data that support the findings of this study are availablefrom the corresponding author upon reasonable request.REFERENCES1M. Julliere, Phys. Lett. A 54, 225 (1975).2H. J. M. Swagten, in Handbook of Magnetic Materials (Elsevier, 2007), pp.1–121.3S. Yuasa and D. D. Djayaprawira, J. Phys. D 40, R337 (2007).4S. Ikeda, J. Hayakawa, Y. Ashizawa, Y. M. Lee, K. Miura, H. Hasegawa, M.Tsunoda, F. Matsukura, and H. Ohno, Appl. Phys. Lett. 93, 082508 (2008).5A. V. Silva, D. C. Leitao, J. Valadeiro, J. Amaral, P. P. Freitas, and S. Cardoso,Eur. Phys. J. Appl. Phys. 72, 10601 (2015).6K. Yakushiji, A. Sugihara, A. Fukushima, H. Kubota, and S. Yuasa, Appl. Phys.Lett. 110, 092406 (2017).7S. Bhatti, R. Sbiaa, A. Hirohata, H. Ohno, S. Fukami, and S. N.Piramanayagam, Mater. Today 20, 530 (2017).8M. Romera, P. Talatchian, S. Tsunegi, F. Abreu Araujo, V. Cros, P. Bortolotti,J. Trastoy, K. Yakushiji, A. Fukushima, H. Kubota, S. Yuasa, M. Ernoult, D.Vodenicarevic, T. Hirtzlin, N. Locatelli, D. Querlioz, and J. Grollier, Nature563, 230 (2018).9A. Hirohata, K. Yamada, Y. Nakatani, I.-L. Prejbeanu, B. Di�eny, P. Pirro, andB. Hillebrands, J. Magn. Magn. Mater. 509, 166711 (2020).10A. Fukushima, T. Yamamoto, T. Nozaki, K. Yakushiji, H. Kubota, and S.Yuasa, APL Mater. 9, 030905 (2021).11S. Jung, H. Lee, S. Myung, H. Kim, S. K. Yoon, S.-W. Kwon, Y. Ju, M. Kim, W.Yi, S. Han, B. Kwon, B. Seo, K. Lee, G.-H. Koh, K. Lee, Y. Song, C. Choi, D.Ham, and S. J. Kim, Nature 601, 211 (2022).12G. Albuquerque, S. Hernandez, M. T. Kief, D. Mauri, and L. Wang, IEEETrans. Magn. 58, 1 (2022).13N. Maciel, E. Marques, L. Naviner, Y. Zhou, and H. Cai, Sensors 20, 121(2019).14D. Ielmini and S. Ambrogio, Nanotechnology 31, 092001 (2020).15T. Scheike, Q. Xiang, Z. Wen, H. Sukegawa, T. Ohkubo, K. Hono, and S.Mitani, Appl. Phys. Lett. 118, 042411 (2021).16T. Scheike, Z. Wen, H. Sukegawa, and S. Mitani, Appl. Phys. Lett. 120, 032404(2022).17D. C. Worledge and P. L. Trouilloud, Appl. Phys. Lett. 83, 84 (2003).18I. Ohnuma, H. Enoki, O. Ikeda, R. Kainuma, H. Ohtani, B. Sundman, and K.Ishida, Acta Mater. 50, 379 (2002).19M. A. Turchanin, L. A. Dreval, A. R. Abdulov, and P. G. Agraval, PowderMetall. Met. Ceram. 50, 98 (2011).20Y. Miura, K. Abe, and M. Shirai, Phys. Rev. B 83, 214411 (2011).21K. Masuda, T. Tadano, and Y. Miura, Phys. Rev. B 104, L180403 (2021).22S. Yuasa, A. Fukushima, H. Kubota, Y. Suzuki, and K. Ando, Appl. Phys. Lett.89, 042505 (2006).23S. G. Wang, R. C. C. Ward, G. X. Du, X. F. Han, C. Wang, and A. Kohn, Phys.Rev. B 78, 180411 (2008).24B. Hu, K. Moges, Y. Honda, H. Liu, T. Uemura, M. Yamamoto, J. Inoue, andM. Shirai, Phys. Rev. B 94, 094428 (2016).25T. Scheike, H. Sukegawa, T. Ohkubo, K. Hono, and S. Mitani, J. Phys. D 53,045001 (2020).26S. S. P. Parkin, C. Kaiser, A. Panchula, P. M. Rice, B. Hughes, M. Samant, andS.-H. Yang, Nat. Mater. 3, 862 (2004).27E. Y. Tsymbal and I. Zutic, Handbook of Spin Transport and Magnetism (CRCPress, 2011).28M. Belmoubarik, H. Sukegawa, T. Ohkubo, S. Mitani, and K. Hono, Appl.Phys. Lett. 108, 132404 (2016).29S. Ikeda, J. Hayakawa, Y. M. Lee, T. Tanikawa, F. Matsukura, and H. Ohno,J. Appl. Phys. 99, 08A907 (2006).30V. Drewello, M. Sch€afers, O. Schebaum, A. A. Khan, J. M€unchenberger, J.Schmalhorst, G. Reiss, and A. Thomas, Phys. Rev. B 79, 174417 (2009).31F. Bonell, T. Hauet, S. Andrieu, F. Bertran, P. Le Fèvre, L. Calmels, A. Tejeda, F.Montaigne, B. Warot-Fonrose, B. Belhadji, A. Nicolaou, and A. Taleb-Ibrahimi,Phys. Rev. Lett. 108, 176602 (2012).32S. Yuasa, T. Nagahama, A. Fukushima, Y. Suzuki, and K. Ando, Nat. Mater. 3,868 (2004).33D. D. Djayaprawira, K. Tsunekawa, M. Nagai, H. Maehara, S. Yamagata, N.Watanabe, S. Yuasa, Y. Suzuki, and K. Ando, Appl. Phys. Lett. 86, 092502(2005).34G. Feng, S. van Dijken, J. F. Feng, J. M. D. Coey, T. Leo, and D. J. Smith,J. Appl. Phys. 105, 033916 (2009).35R. Matsumoto, A. Fukushima, T. Nagahama, Y. Suzuki, K. Ando, and S. Yuasa,Appl. Phys. Lett. 90, 252506 (2007).36T. Ishikawa, S. Hakamata, K. Matsuda, T. Uemura, and M. Yamamoto, J. Appl.Phys. 103, 07A919 (2008).37T. Marukame, T. Ishikawa, T. Taira, K. Matsuda, T. Uemura, and M.Yamamoto, Phys. Rev. B 81, 134432 (2010).Applied Physics Letters ARTICLE scitation.org/journal/aplAppl. Phys. Lett. 122, 112404 (2023); doi: 10.1063/5.0145873 122, 112404-6VC Author(s) 2023https://doi.org/10.1016/0375-9601(75)90174-7https://doi.org/10.1088/0022-3727/40/21/R01https://doi.org/10.1063/1.2976435https://doi.org/10.1051/epjap/2015150214https://doi.org/10.1063/1.4977565https://doi.org/10.1063/1.4977565https://doi.org/10.1016/j.mattod.2017.07.007https://doi.org/10.1038/s41586-018-0632-yhttps://doi.org/10.1016/j.jmmm.2020.166711https://doi.org/10.1063/5.0038974https://doi.org/10.1038/s41586-021-04196-6https://doi.org/10.1109/TMAG.2021.3081042https://doi.org/10.1109/TMAG.2021.3081042https://doi.org/10.3390/s20010121https://doi.org/10.1088/1361-6528/ab554bhttps://doi.org/10.1063/5.0037972https://doi.org/10.1063/5.0082715https://doi.org/10.1063/1.1590740https://doi.org/10.1016/S1359-6454(01)00337-8https://doi.org/10.1007/s11106-011-9307-zhttps://doi.org/10.1007/s11106-011-9307-zhttps://doi.org/10.1103/PhysRevB.83.214411https://doi.org/10.1103/PhysRevB.104.L180403https://doi.org/10.1063/1.2236268https://doi.org/10.1103/PhysRevB.78.180411https://doi.org/10.1103/PhysRevB.78.180411https://doi.org/10.1103/PhysRevB.94.094428https://doi.org/10.1088/1361-6463/ab50d0https://doi.org/10.1038/nmat1256https://doi.org/10.1063/1.4945049https://doi.org/10.1063/1.4945049https://doi.org/10.1063/1.2176588https://doi.org/10.1103/PhysRevB.79.174417https://doi.org/10.1103/PhysRevLett.108.176602https://doi.org/10.1038/nmat1257https://doi.org/10.1063/1.1871344https://doi.org/10.1063/1.3068186https://doi.org/10.1063/1.2750398https://doi.org/10.1063/1.2843756https://doi.org/10.1063/1.2843756https://doi.org/10.1103/PhysRevB.81.134432https://scitation.org/journal/apl f1 f2 f3 l f4 l c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 c14 c15 c16 c17 c18 c19 c20 c21 c22 c23 c24 c25 c26 c27 c28 c29 c30 c31 c32 c33 c34 c35 c36 c37