# Fileset

[138-Appl. Phys. Lett. 126, 022407 (2025)-Sihombing.pdf](https://mdr.nims.go.jp/filesets/84b7b45a-2a3b-40c3-ae90-122bdb15c6b5/download)

## Creator

[Rombang Rizky Sihombing](https://orcid.org/0000-0002-8461-7352), [Thomas Scheike](https://orcid.org/0000-0002-9163-5524), [Jun Uzuhashi](https://orcid.org/0000-0003-2023-8158), [Tadakatsu Ohkubo](https://orcid.org/0000-0003-3548-1951), [Zhenchao Wen](https://orcid.org/0000-0001-7496-1339), [Seiji Mitani](https://orcid.org/0000-0002-1348-0774), [Hiroaki Sukegawa](https://orcid.org/0000-0002-4034-7848)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Enhanced tunnel magnetoresistance of Fe/MgGa2O4/Fe(001) magnetic tunnel junctions by interface-tuning with atomic-scale MgO insertion layers](https://mdr.nims.go.jp/datasets/e8a4316d-5a04-4fa1-b2fc-a6953f7bc17b)

## Fulltext

Enhanced tunnel magnetoresistance of Fe/MgGa2O4/Fe(001) magnetic tunnel junctionsby interface-tuning with atomic-scale MgO insertion layersViewOnlineExportCitationRESEARCH ARTICLE |  JANUARY 15 2025Enhanced tunnel magnetoresistance of Fe/MgGa2O4/Fe(001)magnetic tunnel junctions by interface-tuning with atomic-scale MgO insertion layers Rombang Rizky Sihombing  ; Thomas Scheike  ; Jun Uzuhashi  ; Tadakatsu Ohkubo  ;Zhenchao Wen  ; Seiji Mitani  ; Hiroaki Sukegawa  Appl. Phys. Lett. 126, 022407 (2025)https://doi.org/10.1063/5.0247660Articles You May Be Interested InKoopmanLab: Machine learning for solving complex physics equationsAPL Mach. Learn. (September 2023)Experimental realization of a quantum classification: Bell state measurement via machine learningAPL Mach. Learn. (September 2023) 15 January 2025 17:03:23https://pubs.aip.org/aip/apl/article/126/2/022407/3331518/Enhanced-tunnel-magnetoresistance-of-Fe-MgGa2O4-Fehttps://pubs.aip.org/aip/apl/article/126/2/022407/3331518/Enhanced-tunnel-magnetoresistance-of-Fe-MgGa2O4-Fe?pdfCoverIconEvent=citejavascript:;https://orcid.org/0000-0002-8461-7352javascript:;https://orcid.org/0000-0002-9163-5524javascript:;https://orcid.org/0000-0003-2023-8158javascript:;https://orcid.org/0000-0003-3548-1951javascript:;https://orcid.org/0000-0001-7496-1339javascript:;https://orcid.org/0000-0002-1348-0774javascript:;https://orcid.org/0000-0002-4034-7848https://crossmark.crossref.org/dialog/?doi=10.1063/5.0247660&domain=pdf&date_stamp=2025-01-15https://doi.org/10.1063/5.0247660https://pubs.aip.org/aip/aml/article/1/3/036110/2910717/KoopmanLab-Machine-learning-for-solving-complexhttps://pubs.aip.org/aip/aml/article/1/3/036111/2910912/Experimental-realization-of-a-quantumhttps://e-11492.adzerk.net/r?e=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&s=YqGfELRSYvq37y2k-7pZQLR3U7gEnhanced tunnel magnetoresistanceof Fe/MgGa2O4/Fe(001) magnetic tunneljunctions by interface-tuning withatomic-scale MgO insertion layersCite as: Appl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660Submitted: 7 November 2024 . Accepted: 1 January 2025 .Published Online: 15 January 2025Rombang Rizky Sihombing,1,2 Thomas Scheike,1 Jun Uzuhashi,1 Tadakatsu Ohkubo,1 Zhenchao Wen,1Seiji Mitani,1,2 and Hiroaki Sukegawa1,a)AFFILIATIONS1National Institute for Materials Science (NIMS), 1-2-1 Sengen, Tsukuba 305-0047, Japan2Graduate School of Science and Technology, University of Tsukuba, Tsukuba 305-8577, Japana)Author to whom correspondence should be addressed: sukegawa.hiroaki@nims.go.jpABSTRACTWe demonstrate a significant effect of atomic-scale MgO insertion layers on the tunnel magnetoresistance (TMR) in epitaxial magnetic tun-nel junctions (MTJs) using a small bandgap oxide MgGa2O4. An enhanced TMR ratio of 151% at room temperature (resistance area product,RA: 23 kX � lm2) and 291% at 5K (RA: 26 kX � lm2) were observed using 0.3 nm MgO insertion layers at the bottom and top barrier interfa-ces in Fe/MgGa2O4/Fe(001) MTJs with a total barrier thickness of 2.3 nm. The TMR showed a strong MgO thickness dependence.Microstructure analyses revealed that after MgO insertion, a homogeneous rock-salt structured Mg0.55Ga0.45O(001) barrier is formed, whichdiffers from the nominal spinel crystal MgGa2O4. Elemental mapping of the MTJ showed that Ga diffusion into the adjacent Fe can be effec-tively suppressed while maintaining perfect lattice-matching at the Fe/barrier interfaces, thereby improving effective tunneling spin polariza-tion through the barrier. The RA of the Mg0.55Ga0.45O (2.3 nm) MTJ is smaller than that of a comparable MgAl2O4 barrier (2.3 nm), thanksto the lower barrier height of the Mg0.55Ga0.45O as confirmed by the current–voltage characteristics.VC 2025 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/5.0247660Magnetic tunnel junctions (MTJs), consisting of a ferromagneticlayer (FM)/ultra-thin insulator (barrier)/FM structure, are widely usedin spintronic applications, including read heads of hard disk drives(HDDs) and magnetoresistive random access memory (MRAM) cells.1Recently, new MTJ applications, such as ultra-highly sensitive mag-netic sensors and neuromorphic devices, have also attracted muchattention.2–6 Most practical MTJs use insulating MgO as a barrierlayer.7–10 However, state-of-the-art spin-transfer-torque (STT)-MRAMs and HDD heads require very low resistance area product(RA) less than a few X � lm2 by reducing the MgO thickness to about1 nm (�5 monolayers) or less.11 Ultra-thin barriers are sensitive toimperfections with small error margins with respect to their crystallin-ity and flatness, leading to reduced reliability and increased risk of fail-ure of the MTJ when, i.e., a moderate bias voltage is applied. By usingnew barrier materials with a lower barrier height, the thickness of thebarrier can be increased while keeping RA low.A magnesium gallium spinel oxide, MgGa2O4 (MGO), is a prom-ising MTJ barrier due to its bandgap of�4.7 eV, which is much smallerthan that of typical barriers, such as MgO and MgAl2O4 (�8 eV).12,13Sukegawa et al.14 reported a relatively large tunnel magnetoresistance(TMR) ratio of 121% in an epitaxial Fe/MGO/Fe MTJ at room temper-ature (RT). It was also shown that the RA of the Fe/MGO/Fe MTJ wassignificantly lower than that of an Fe/MgAl2O4/Fe MTJ with the samebarrier thickness. First-principles calculations predicted that MGO-basedMTJs exhibit a large TMR ratio and low RA,15 making the MGObarrier a good option to be developed for futureMTJ devices.One of the issues with MGO-based MTJs is that achieving alarger TMR ratio at RT is necessary for practical applications. Fine-tuning of the barrier interface, such as nano-layer insertions and addi-tional oxidation processes, would be promising for improving theTMR ratio in MGO-based MTJs, as demonstrated in MgO-basedMTJs.16,17 Mertens et al.18 reported that the insertion of ultra-thinAppl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660 126, 022407-1VC Author(s) 2025Applied Physics Letters ARTICLE pubs.aip.org/aip/apl 15 January 2025 17:03:23https://doi.org/10.1063/5.0247660https://doi.org/10.1063/5.0247660https://doi.org/10.1063/5.0247660https://www.pubs.aip.org/action/showCitFormats?type=show&doi=10.1063/5.0247660http://crossmark.crossref.org/dialog/?doi=10.1063/5.0247660&domain=pdf&date_stamp=2025-01-15https://orcid.org/0000-0002-8461-7352https://orcid.org/0000-0002-9163-5524https://orcid.org/0000-0003-2023-8158https://orcid.org/0000-0003-3548-1951https://orcid.org/0000-0001-7496-1339https://orcid.org/0000-0002-1348-0774https://orcid.org/0000-0002-4034-7848mailto:sukegawa.hiroaki@nims.go.jphttps://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://doi.org/10.1063/5.0247660pubs.aip.org/aip/aplMgO layers at the bottom and top-MGO interfaces of CoFeB/MGO/CoFeB polycrystalline MTJs, i.e., CoFeB/MgO/MGO/MgO/CoFeB,significantly improves the perpendicular magnetic anisotropy of theCoFeB layers due to the improved interface composition. It is sug-gested that Ga diffusion is a major problem for MGO-based MTJs,and how to control it is key to obtain larger TMR ratios.In this study, we fabricated epitaxial Fe/MGO/Fe(001) MTJ stacksand investigated theMgO insertion effect onmagnetotransport propertiesat the MGO barrier interface. A strong dependence of TMR on the MgOinsertion layer thickness was found for the MTJs. The optimum MgOthickness of 0.3nm for both the bottom and top interfaces resulted in asignificant enhancement in the TMR ratio, up to 151% at RT and 297% at5K. Second, microstructural analysis revealed that instead of a spinelstructure, the MGO barrier formed a rock-salt crystal. Additionally,atomic-scale MgO insertion effectively suppresses Ga atomic diffusion atthe barrier interfaces, maintaining the MGO barrier’s high tunneling spinpolarization as well as the low barrier height properties.MTJ stacks were deposited on MgO(001) single crystal substratesusing an ultra-high-vacuum multi-chamber magnetron sputtering sys-tem (EIKO Corp.) with a base pressure of �5� 10�7Pa. The typicalstack structure is MgO(001) substrate//Cr (60)/Fe (50)/bottom-MgO(tbot-MgO¼ 0–1.0)/MgGa2O4 (MGO) (1.7)/top-MgO (ttop-MgO¼ 0 or0.3)/Fe (5)/Ir20Mn80 (IrMn) (8)/Ru (10) [numbers in parentheses innm, see Fig. 1(a)]. The MgO substrate was annealed in situ at 700 �Cto remove surface contamination prior to deposition. All metalliclayers were deposited using DC sputtering at RT followed by in situpost-annealing to improve the flatness and crystallinity of each layer.MGO and MgO were deposited from a 76.2mm diameter MgGa2O4and MgO sintered targets by RF sputtering using an input power of100W and an Ar pressure of 2.75 and 1.00 Pa, respectively. The MGOlayer was post-annealed at 400 �C, followed by 300 s of natural oxida-tion using pure O2 gas (99.999%, �1Pa) after cooling down to RT totune the oxidation state at the top barrier interface. The bottom-MgOlayer thickness was varied during deposition using a linear motionshutter. The MgO insertion layers were in situ post-annealed at250 �C. High energy electron diffraction (RHEED) was used to evalu-ate the surface crystal structure of each layer. The MTJ stacks wereannealed ex situ in a magnetic field of 2 kOe at 200 �C along the Feeasy axis direction, i.e., MgO[110] jj Fe[100], and patterned into4� 4–8� 8 lm2 square MTJs using laser and photolithography, andAr ion milling. Magnetotransport properties of the patterned MTJswere characterized using a conventional DC 4-probe method at RT(Keithley 2400 sourcemeter and Keithley 2182A nanovoltmeter). TheTMR ratios were measured with a bias voltage of �10mV. The TMRratio (%) is defined as 100�(RAP� RP)/RP, where RAP [RP] is the resis-tance in the antiparallel (AP) [parallel (P)] magnetization configura-tion. The temperature dependence of the TMR ratio and RA from RTto 5K was characterized using physical property measurement system(PPMS, Quantum Design, Dynacool). High-resolution annular dark-field scanning transmission electron microscopy (ADF-STEM), nano-beam electron diffraction (NBED), and energy dispersive x-rayspectroscopy (EDS) (Titan G2 80-200 TEM) were used to investigatethe microstructure of the MTJ cross section.The RHEED patterns at each growth stage for the epitaxial Fe/MGO/Fe with MgO insertion layers are shown in Fig. 1(b) for theMgO[100] azimuth and Fig. 1(c) for the MgO[110] azimuth, respec-tively. Figure 1(d) schematically shows the crystal structures and orien-tations reconstructed from the RHEED patterns. The pattern of theFIG. 1. (a) Schematic illustration of MTJ stacks. (b) and (c) RHEED patterns at each deposition stage for (b) MgO[100] azimuth and (c) MgO[110] azimuth. (d) Correspondingatomic structure models.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660 126, 022407-2VC Author(s) 2025 15 January 2025 17:03:23pubs.aip.org/aip/aplbottom Fe shows sharp streaks, indicating highly (001)-oriented bccepitaxial growth and a flat surface. Weaker streaks marked by yellowarrows were also observed, suggesting a c(2� 2) reconstructed surfacedue to oxygen adsorption during the in situ annealing.19 Nevertheless,the bottom (tbot-MgO � 0.5 nm) and top-MgO insertion layer exhibitsepitaxial growth with a (001)-oriented rock-salt (RS) structure with45� in-plane lattice rotation on the bottom Fe, as shown in Fig. 1(d).The MGO layer on the lower MgO also shows epitaxial growth,although it has a lower RHEED intensity than that of the lower MgOlayer. The MGO also has an RS-like structure (cation-disordered spi-nel)20,21 as indicated by the absence of superlattice streaks resulting inhalf the unit cell size of an ordered spinel MGO. The top Fe layershows bcc (001) epitaxial growth. Thus, all layers are epitaxially grownwith (001) orientation.Figure 2(a) shows the tbot-MgO dependence of the TMR ratio andRA at RT of the MTJs with [without] the top-MgO (ttop-MgO¼ 0.3 nm)[(ttop-MgO¼ 0 nm)]. In both cases, the TMR ratio increases rapidlywith tbot-MgO, reaching a maximum between 0.3 and 0.4 nm followedby a decrease for larger tbot-MgO. The maximum TMR ratio of the MTJwith (without) top-MgO insertion reaches 151% (102%) at tbot-MgO¼ 0.3 nm (0.4 nm). The MgO insertion layers can effectively increasethe maximum TMR ratio. In the previous report of Fe/MGO/FeMTJs,14 a 0.6 nm Mg insertion at the bottom side of the MGO barrierincreased the TMR ratio: 121% (85%) with (without) the insertion.Similarly, the MgO insertion may improve the MGO interface qualityin this study. The value of 0.3 nm corresponds to 1.5 monolayers of anMgO(001) lattice; thus, this thickness may be sufficient to completelycover the Fe interfaces with MgO to block the Ga diffusion, as shownlater in the nanostructural analysis. The decrease in TMR ratioobserved at larger tbot-MgO may be due to the increase in effective in-plane lattice mismatch between the barrier and the Fe electrodes; themismatch between MgO(001) and Fe(001) (�3.8%) is larger than thatbetween MGO(001) and Fe(001) (�2.2%) for bulk.13 As seen inFig. 2(b), the log(RA) increases linearly with tbot-MgO around the largeTMR regions, i.e., tbot-MgO � 0.25–0.55 nm. In this thickness range, a0.3 nm increase in tbot-MgO results in an almost tenfold increase in RA.This increase is almost the same as the increase due to the 0.3 nm top-MgO insertion. This means that the barrier thickness increases linearlywith tbot-MgO and ttop-MgO for both the structures. Therefore, both theMgO/MGO and MgO/MGO/MgO barriers work as a single layer tun-nel barrier for MTJs.Figures 3(a) and 3(b) show the TMR ratio and RA as a functionof the in-plane magnetic field of the Fe/MgO (tbot-MgO¼ 0.3 nm)/MGO/MgO (ttop-MgO¼ 0.3nm)/Fe at RT and 5K, respectively. A max-imum TMR ratio of 151% and an RA of 23kX � lm2 at RT wereobserved with clear magnetic switching. At 5K, the TMR ratio almostdoubles to 291% (RA: 26 kX � lm2), demonstrating a significant spin-dependent coherent tunneling through the MgO/MGO/MgO barrier.These TMR ratios are much larger than those reported values in a Fe/Mg (0.6 nm)/MGO/Fe MTJ (121% at RT and 165% at low tempera-tures).14 The effective tunneling spin polarization Peff at 5K is calcu-lated to be 0.77 based on the Julliere formula, TMR ratio ¼ 100� 2P2eff=ð1� P2eff Þ, assuming Peff is the same for both interfaces.22Figures 3(c) and 3(d) show the temperature dependences of the RA forthe P (RAP) and AP (RAAP) states, and the TMR ratio of the MTJ,respectively. The RAAP decreases significantly with increasing temper-ature, while the RAP is almost constant. Therefore, the RAAP primarilydetermines the temperature dependence of the TMR ratio. This behav-ior can be attributed to the occurrence of spin-dependent coherenttunneling though the D1 state for the P state, as observed in MgO-based and MgAl2O4-based MTJs with large TMR ratios.16,23–25Figures 4(a) and 4(b) show the cross-sectional ADF-STEMimages of Fe/MgO (0.3 nm)/MGO (1.6 nm) MgO (0.3nm)/Fe with RAof 16.5 kX � lm2. The image shows the formation of high-quality epi-taxial Fe/MgO/MGO/MgO/Fe with atomically flat interfaces for boththe bottom-Fe/MgO/MGO and MGO/MgO/top-Fe sides. Only a fewmisfit dislocations were observed at the interfaces, indicating nearlyperfect lattice-matching between Fe and the barrier. The NBED pat-terns taken from the bottom Fe and top Fe electrodes show their goodcrystallinity with bcc(001) orientation, as expected from the RHEEDpatterns shown in Figs. 1(b) and 1(c). The pattern of the MgO/MGO/MgO barrier [Fig. 4(d)] shows an RS(001) structure, consistent with theRHEED patterns of the MGO layer. The EDS elemental maps of Mg,Ga, O, and Fe shown in Figs. 4(f)–4(j) indicate that the element distri-bution in the barrier is homogeneous. The barrier composition wasdetermined to be (Mg0.55Ga0.45)0.47O0.53 (hereafter, Mg0.55Ga0.45O)using Gaussian fits of Mg, Ga, and O as shown in the supplementarymaterial, Fig. S1. Therefore, the formation of the RS structure is mainlyFIG. 2. tbot-MgO dependences of (a) TMR ratio and (b) RA at RT for MTJs withbottom-MgO (tbot-MgO)/MGO/top-MgO (0.3 nm) barrier (blue symbols) and bottom-MgO (tbot-MgO)/MGO barrier (red symbols).Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660 126, 022407-3VC Author(s) 2025 15 January 2025 17:03:23https://doi.org/10.60893/figshare.apl.c.7605968https://doi.org/10.60893/figshare.apl.c.7605968pubs.aip.org/aip/apldue to the off-stoichiometric MGO composition by mixing of Mg andGa atoms in the MgO/MGO/MgO trilayer. The EDS line profilesshown in Fig. 4(k) indicate that there is no significant interdiffusion ofGa atoms into the bottom- and top-Fe side. This means that MgOinsertion layers with optimum thickness of 0.3 nm act as good Ga diffu-sion barriers to make both the Fe interfaces chemically sharp, resultingin TMR enhancements. In MgAl2O4(001)-based MTJs, the cation sitedisorder that halves the unit cell size, i.e., the change from a spinel struc-ture (aspinel) to an RS-like structure (aRS � aspinel/2), can effectivelyimprove the TMR ratios due to suppression of the band-foldingeffect.20,26 Therefore, the formation of RS-like MGO barrier may alsobe one of the possible origins of the improved TMR ratios. The esti-mated barrier thickness was�2.3nm, which is close to the design struc-ture of the MgO/MGO/MgO (¼2.2 nm). The RA of 16.5 kX � lm2 ofthe Fe/Mg0.55Ga0.45O (2.3 nm)/Fe is 8 times higher than that of Fe/MgGa2O4 (2.3 nm)/Fe, but it is still 3 times lower than that of Fe/MgAl2O4 (2.3nm)/Fe.14 Therefore, RA reduction is observed even afterthe 0.6nmMgO insertion.We evaluated the current–voltage (I–V) curves and bias voltagedependence of the TMR ratio of the Fe/MgO (0.3 nm)/MGO (1.7 nm)/MgO (0.3 nm)/Fe MTJ at RT, and 5K. Figures 5(a) and 5(d) show I–Vcurves and Figs. 5(b) and 5(e) show the dI/dV curves at RT (5K),respectively. Positive bias is defined as the direction in which electronstunnel from the top electrode to the bottom electrode, as shown in theinset of (a). The dI/dV curves are obtained by numerical differentiationof the I–V curves. In both the P and AP states, jIj and dI/dV curvesincrease significantly when the bias voltage exceeds j0.5Vj for bothbias polarities. In MgO- and MgAl2O4-based MTJs, such a significantcurrent acceleration is typically not observed below j1Vj. In MgO-and MgAl2O4-based MTJs with large TMR ratios, significantly largelocal minima can be observed in the dI/dV curves in the P state.16,17,23However, our MGO-MTJ does not clearly show such structures in thedI/dV curves even at 5K.We fitted the I–V curves using the Simmons equation to estimatethe effective barrier height of the Mg0.55Ga0.45O.27 The fitting resultsare shown in Table I, in addition to the values in the previousreports.14 Typical log(I)–log(V) plots with fitting curves of theSimmons equation are shown in the supplementary material, Fig. S2.The experimental data were well fitted by the equation for both posi-tive and negative bias. For the positive bias, the effective barrier height(/eff) was 0.86 eV in the P state and 0.70 eV in the AP state at 300K.For the negative bias, /eff was 1.1 eV in the P state and 0.82 eV in theAP state at 300K. These values are comparable to the barrier height ofthe previous Fe/MgGa2O4/Fe MTJ and much smaller than that of theFe/MgAl2O4/Fe MTJ.14 We also found a large temperature dependencein the barrier height of Fe/Mg0.55Ga0.45O/Fe [see Figs. S2(b) andS2(d)], indicating the low barrier feature.27 The low barrier height ismaintained even after the insertion of MgO.Figures 5(c) and 5(f) show the bias voltage dependences of thenormalized TMR ratio by its zero bias value of the Fe/MgO (0.3 nm)/FIG. 3. (a) and (b) TMR ratio and RA as a function of magnetic field of Fe/MgO (0.3)/MGO (1.7)/MgO (0.3 nm)/Fe MTJ measured at (a) RT and (b) 5 K. (c) and (d)Temperature dependences of (c) RA for P and AP states and (d) TMR ratio.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660 126, 022407-4VC Author(s) 2025 15 January 2025 17:03:23https://doi.org/10.60893/figshare.apl.c.7605968pubs.aip.org/aip/aplFIG. 4. (a) Low magnification and (b) high magnification ADF-STEM images of Fe/MgO (0.3)/MGO (1.6)/MgO (0.3 nm)/Fe MTJ. (c)–(e) NBED patterns taken from (c) top Fe,(d) barrier, and (e) bottom Fe. (f)–(j) EDS maps for (f) ADF image, (g) Mg, (h) Ga, (i) O, and (j) Fe. (k) Corresponding EDS profiles.FIG. 5. (a) and (d) I–V curves, (b) and (e) dI/dV curves, (c) and (f) bias voltage dependence of normalized TMR of Fe/MgO (0.3)/MGO (1.7)/MgO (0.3 nm)/Fe MTJ (devicearea: 8� 8 lm2) for RT and 5 K, respectively. The insets of (b) and (e) are close-ups of the dI/dV curves for the P state.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660 126, 022407-5VC Author(s) 2025 15 January 2025 17:03:23pubs.aip.org/aip/aplMGO (1.7 nm)/MgO (0.3 nm)/Fe MTJ at RT and 5K, respectively.The asymmetry of the TMR ratio with the bias polarity is larger thanthat of the I–V characteristics. Due to the low barrier height, a largerbias voltage dependence of the TMR ratio is expected: Vhalf, the biasvoltage where the TMR becomes half the value of zero bias, is 0.43V(�0.26V) for the positive (negative) bias at RT, which is less than halfthe values in Fe/MgAl2O4/Fe MTJs.24,25 At 5K, the asymmetric featureis more pronounced. Note that the asymmetry in the I–V characteris-tics and the differences in barrier height and Vhalf between the positiveand negative bias are attributed to the slight difference in interfacestates between the top and bottom Mg0.55Ga0.45O interfaces, whichcannot be detected by the present STEM analysis. Therefore, furtherimprovement of the bottom and top interfaces can significantlyimprove the TMR ratio of MGO-based MTJs.In summary, we investigated the effect of MgO insertion layerson magnetotransport properties using epitaxial Fe/MGO/Fe(001).Nanostructural analysis revealed the formation of a uniform barrierlayer with an RS structure with a Mg0.55Ga0.45O composition. TheMgO layers effectively suppress Ga interdiffusion into the top and bot-tom Fe layers, resulting in the formation of a lattice-matched Fe/Mg0.55Ga0.45O/Fe(001) MTJ. The interfacial modification by MgOinsertions significantly improves the TMR ratio up to 151% at RT(291% at 5K) while maintaining the low barrier height. Our resultsindicate that an MGO-based barrier is promising as a low RA barrierwith significant coherent tunneling effect, which can be beneficial fornext generation spintronic applications such as ultra-high densityMRAMs.See the supplementary material for the EDS Gaussian fit results forthe barrier and the Simmons fits for the log(I)–log(V) characteristics.The authors thank Hiromi Ikeda and Chika Shigaki for theirtechnical support on device microfabrication. This work waspartially supported by KIOXIA Corporation, MEXT Program: DataCreation and Utilization-Type Material Research and DevelopmentProject (Grant No. JPMXP1122715503) and JSPS KAKENHI(Grant Nos. 21H01750, 22H04966, and 24H00408). R.R.S. thanksthe National Institute for Materials Science for the provision of aNIMS Junior Research Assistantship.AUTHOR DECLARATIONSConflict of InterestThe authors have no conflicts to disclose.Author ContributionsRombang Rizky Sihombing: Data curation (lead); Formal analysis(equal); Investigation (lead); Methodology (equal); Visualization(lead); Writing – original draft (equal); Writing – review & editing(equal). Thomas Scheike: Methodology (supporting); Validation(equal); Writing – review & editing (equal). Jun Uzuhashi: Data cura-tion (equal); Investigation (equal); Validation (equal); Visualization(equal); Writing – review & editing (supporting). Tadakatsu Ohkubo:Investigation (equal); Validation (equal); Writing – review & editing(supporting). ZhenchaoWen: Investigation (supporting); Methodology(supporting); Writing – review & editing (supporting). Seiji Mitani:Funding acquisition (equal); Methodology (supporting); Supervision(equal); Validation (lead);Writing – original draft (supporting);Writing –review & editing (lead). Hiroaki Sukegawa: Conceptualization (lead);Formal analysis (equal); Funding acquisition (lead); Investigation (equal);Methodology (equal); Supervision (lead); Visualization (equal); Writing –original draft (lead);Writing – review& editing (equal).DATA AVAILABILITYThe data that support the findings of this study are availablewithin the article and from the corresponding author upon reasonablerequest.REFERENCES1S. Yuasa and D. D. Djayaprawira, J. Phys. Appl. Phys. 40, R337 (2007).2P. P. Freitas, R. Ferreira, and S. Cardoso, Proc. IEEE 104, 1894 (2016).3J. Torrejon, M. Riou, F. A. Araujo, S. Tsunegi, G. Khalsa, D. Querlioz, P.Bortolotti, V. Cros, K. Yakushiji, A. Fukushima, H. Kubota, S. Yuasa, M. D.Stiles, and J. Grollier, Nature 547, 428 (2017).4N. Maciel, E. Marques, L. Naviner, Y. Zhou, and H. Cai, Sensors 20, 121 (2019).5A. Fukushima, T. Yamamoto, T. Nozaki, K. Yakushiji, H. Kubota, and S.Yuasa, APL Mater. 9, 030905 (2021).6S. 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).7W. Butler, X.-G. Zhang, T. Schulthess, and J. MacLaren, Phys. Rev. B 63,054416 (2001).8J. Mathon and A. Umerski, Phys. Rev. B 63, 220403 (2001).9S. S. P. Parkin, C. Kaiser, A. Panchula, P. M. Rice, B. Hughes, M. Samant, andS.-H. Yang, Nat. Mater. 3, 862 (2004).10S. Yuasa, T. Nagahama, A. Fukushima, Y. Suzuki, and K. Ando, Nat. Mater. 3,868 (2004).11G. Albuquerque, S. Hernandez, M. T. Kief, D. Mauri, and L. Wang, IEEE Trans.Magn. 58, 1 (2022).12Z. Galazka, D. Klimm, K. Irmscher, R. Uecker, M. Pietsch, R. Bertram, M.Naumann, M. Albrecht, A. Kwasniewski, R. Schewski, and M. Bickermann,Phys. Status Solidi A 212, 1455 (2015).13M. Coll, J. Fontcuberta, M. Althammer, M. Bibes, H. Boschker, A. Calleja, G.Cheng, M. Cuoco, R. Dittmann, B. Dkhil, I. E. Baggari, M. Fanciulli, I. Fina, E.Fortunato, C. Frontera, S. Fujita, V. Garcia, S. T. B. Goennenwein, C. G. Granqvist,J. Grollier, R. Gross, A. Hagfeldt, G. Herranz, K. Hono, E. Houwman, M. Huijben,A. Kalaboukhov, D. J. Keeble, G. Koster, L. F. Kourkoutis, J. Levy, M. Lira-Cantu, J.L. MacManus-Driscoll, J. Mannhart, R. Martins, S. Menzel, T. Mikolajick, M.Napari, M. D. Nguyen, G. Niklasson, C. Paillard, S. Panigrahi, G. Rijnders, F.S�anchez, P. Sanchis, S. Sanna, D. G. Schlom, U. Schroeder, K. M. Shen, A. Siemon,M. Spreitzer, H. Sukegawa, R. Tamayo, J. van den Brink, N. Pryds, and F. M.Granozio, Appl. Surf. Sci. 482, 1 (2019).14H. Sukegawa, Y. Kato, M. Belmoubarik, P. H. Cheng, T. Daibou, N.Shimomura, Y. Kamiguchi, J. Ito, H. Yoda, T. Ohkubo, S. Mitani, and K. Hono,Appl. Phys. Lett. 110, 122404 (2017).15K. Nawa, K. Masuda, and Y. Miura, Phys. Rev. B 102, 144423 (2020).TABLE I. Summary of Simmons fit results for Fe/barrier/Fe(001) MTJs for the posi-tive bias direction.Barrier structure (Unit: nm) State teff (nm) /eff (eV)MgO (0.3)/MgGa2O4 (1.7)/MgO (0.3) (present study)P 1.42 0.86AP 1.65 0.70MgGa2O4 (2.4) (Ref. 14)P 0.96 0.96AP 1.1 1.0MgAl2O4 (2.4) (Ref. 14)P 0.89 3.0AP 1.1 2.1Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660 126, 022407-6VC Author(s) 2025 15 January 2025 17:03:23https://doi.org/10.60893/figshare.apl.c.7605968https://doi.org/10.1088/0022-3727/40/21/R01https://doi.org/10.1109/JPROC.2016.2578303https://doi.org/10.1038/nature23011https://doi.org/10.3390/s20010121https://doi.org/10.1063/5.0038974https://doi.org/10.1038/s41586-021-04196-6https://doi.org/10.1103/PhysRevB.63.054416https://doi.org/10.1103/PhysRevB.63.220403https://doi.org/10.1038/nmat1256https://doi.org/10.1038/nmat1257https://doi.org/10.1109/TMAG.2021.3081042https://doi.org/10.1109/TMAG.2021.3081042https://doi.org/10.1002/pssa.201431835https://doi.org/10.1016/j.apsusc.2019.03.312https://doi.org/10.1063/1.4977946https://doi.org/10.1103/PhysRevB.102.144423pubs.aip.org/aip/apl16T. Scheike, Q. Xiang, Z. Wen, H. Sukegawa, T. Ohkubo, K. Hono, and S.Mitani, Appl. Phys. Lett. 118, 042411 (2021).17T. Scheike, Z. Wen, H. Sukegawa, and S. Mitani, Appl. Phys. Lett. 122, 112404(2023).18S. Mertens, S. Couet, R. Carpenter, J. Swerts, D. Crotti, and G. S. Kar, Appl.Phys. Lett. 118, 172402 (2021).19J. W. Koo, S. Mitani, T. T. Sasaki, H. Sukegawa, Z. C. Wen, T. Ohkubo, T.Niizeki, K. Inomata, and K. Hono, Appl. Phys. Lett. 103, 192401 (2013).20H. Sukegawa, Y. Miura, S. Muramoto, S. Mitani, T. Niizeki, T. Ohkubo, K. Abe,M. Shirai, K. Inomata, and K. Hono, Phys. Rev. B 86, 184401 (2012).21J. P. Hadorn, H. Sukegawa, T. Ohkubo, S. Mitani, and K. Hono, Acta Mater.145, 306–315 (2018).22M. Julliere, Phys. Lett. A 54, 225 (1975).23V.Drewello, J. Schmalhorst, A. Thomas, andG. Reiss, Phys. Rev. B 77, 014440 (2008).24H. Sukegawa, H. Xiu, T. Ohkubo, T. Furubayashi, T. Niizeki, W. Wang, S.Kasai, S. Mitani, K. Inomata, and K. Hono, Appl. Phys. Lett. 96, 212505 (2010).25M. Belmoubarik, H. Sukegawa, T. Ohkubo, S. Mitani, and K. Hono, Appl. Phys.Lett. 108, 132404 (2016).26Y. Miura, S. Muramoto, K. Abe, and M. Shirai, Phys. Rev. B 86, 024426 (2012).27J. G. Simmons, J. Appl. Phys. 34, 1793 (1963).Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 126, 022407 (2025); doi: 10.1063/5.0247660 126, 022407-7VC Author(s) 2025 15 January 2025 17:03:23https://doi.org/10.1063/5.0037972https://doi.org/10.1063/5.0145873https://doi.org/10.1063/5.0046554https://doi.org/10.1063/5.0046554https://doi.org/10.1063/1.4828658https://doi.org/10.1103/PhysRevB.86.184401https://doi.org/10.1016/j.actamat.2017.12.018https://doi.org/10.1016/0375-9601(75)90174-7https://doi.org/10.1103/PhysRevB.77.014440https://doi.org/10.1063/1.3441409https://doi.org/10.1063/1.4945049https://doi.org/10.1063/1.4945049https://doi.org/10.1103/PhysRevB.86.024426https://doi.org/10.1063/1.1702682pubs.aip.org/aip/apl