# Fileset
[Takano_2026_Jpn._J._Appl._Phys._65_080907.pdf](https://mdr.nims.go.jp/filesets/04e8e6c3-e11b-4557-815e-83ce135c7e02/download)
## Creator
[Shun-ichi Takano](https://orcid.org/0009-0000-4654-2700), [Wenjin Zhang](https://orcid.org/0000-0002-3803-4770), [Yui Tamogami](https://orcid.org/0009-0000-8573-940X), [Takamasa Hirai](https://orcid.org/0000-0002-5577-8018), [Shunsuke Mori](https://orcid.org/0000-0002-1193-4322), [Yasumitsu Miyata](https://orcid.org/0000-0002-9733-5119), [Ken-ichi Uchida](https://orcid.org/0000-0001-7680-3051), [Yusuke Nakanishi](https://orcid.org/0000-0001-8782-9556)
## Rights
[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)
## Other metadata
[Direct thermal imaging of the influence of a transition metal dichalcogenide interlayer on a spin-current injection in a Pt/Y 3 Fe 5 O 12 heterostructure](https://mdr.nims.go.jp/datasets/fd469dcd-b130-4ded-be78-2b77c82cb3f1)
## Fulltext
Direct thermal imaging of the influence of a transition metal dichalcogenide interlayer on a spin-current injection in a Pt/Y3Fe5O12 heterostructureJapanese Journal ofApplied Physics LETTER • OPEN ACCESSDirect thermal imaging of the influence of atransition metal dichalcogenide interlayer on aspin-current injection in a Pt/Y3Fe5O12heterostructureTo cite this article: Shun-ichi Takano et al 2026 Jpn. J. Appl. Phys. 65 080907 View the article online for updates and enhancements.You may also likeSpin Peltier effect and its length scale inPt/YIG system at high temperaturesAtsushi Takahagi, Takamasa Hirai, RyoIguchi et al.-Charge-current angle and frequencydependences of the spin Peltier effectinduced by the spin Hall effectRyo Iguchi and Ken-ichi Uchida-Enhanced thermo-spin effects in iron-oxide/metal multilayersR Ramos, I Lucas, P A Algarabel et al.-This content was downloaded from IP address 202.220.232.52 on 05/05/2026 at 17:24https://doi.org/10.35848/1347-4065/ae620f/article/10.35848/1882-0786/ac6fae/article/10.35848/1882-0786/ac6fae/article/10.7567/JJAP.57.0902B6/article/10.7567/JJAP.57.0902B6/article/10.7567/JJAP.57.0902B6/article/10.1088/1361-6463/aabedb/article/10.1088/1361-6463/aabedbaaaDirect thermal imaging of the influence of a transition metal dichalcogenideinterlayer on a spin-current injection in a Pt/Y3Fe5O12 heterostructureShun-ichi Takano1,2 , Wenjin Zhang3 , Yui Tamogami3,4 , Takamasa Hirai2 , Shunsuke Mori1,2 ,Yasumitsu Miyata3,4 , Ken-ichi Uchida1,2* , and Yusuke Nakanishi1,2*1Department of Advanced Materials Science, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa 277-8561, Japan2Research Center for Magnetic and Spintronic Materials, National Institute for Materials Science, Tsukuba 305-0047, Japan3Research Center for Materials Nanoarchitectonics, National Institute for Materials Science, Tsukuba 305-0044, Japan4Department of Physics, Tokyo Metropolitan University, Hachioji 192-0397, Japan*E-mail: UCHIDA.Kenichi@nims.go.jp; naka24ysk@edu.k.u-tokyo.ac.jpReceived March 18, 2026; revised April 13, 2026; accepted April 20, 2026; published online May 5, 2026The spin Peltier effect (SPE) in Pt/WSe2/Y3Fe5O12 (YIG) and Pt/WS2/YIG heterostructures is investigated using lock-in thermography. Sinceinserting atomically thin transition metal dichalcogenides (TMDs), such as WSe2, between Pt and YIG is known to enhance the voltage induced bythe spin Seebeck effect, similar enhancement is expected for SPE. Contrary to this expectation, the SPE-induced temperature modulation isstrongly suppressed in the TMD-inserted regions compared with bare Pt/YIG regions. Raman spectroscopy reveals sputtering-induced structuraldegradation of WSe2, hindering spin-current injection into YIG. These results highlight the sensitivity of spin-current transport to interface qualityin spin-caloritronic devices incorporating atomically thin TMDs. © 2026 The Japan Society of Applied Physics. All rights, including for text anddata mining, AI training, and similar technologies, are reserved.T he integration of thermoelectric conversion intospintronics has led to the development of spincaloritronics,1–3) which provides novel energy con-version principles. The representative phenomena of spin-current transport in spin caloritronics include the spinSeebeck effect (SSE)4–6) and the spin Peltier effect(SPE).7,8) SSE and SPE enable interconversion betweenheat and spin currents in magnetic materials, functioning asa transverse thermoelectric conversion through the spin–orbitcoupling in metals attached to the magnetic materials.9,10)SSE and SPE have been extensively investigated in para-magnet/ferromagnet (or ferrimagnet) junctions such asPt/Y3Fe5O12 (YIG),7,8,11–13) where the strong spin–orbitcoupling in Pt enables interconversion between charge andspin currents via the (inverse) spin Hall effect.14–18) Theoutput voltage induced by SSE and temperature modulationinduced by SPE are much smaller than those of theconventional Seebeck and Peltier effects. However, SSEand SPE exhibit unique functionalities including the trans-verse thermoelectric conversion using insulators owing totheir magnon origin.SSE and SPE are strongly influenced by the efficiency ofspin-angular-momentum transfer across paramagnet/ferro-magnet interfaces,19–21) which is quantified by the spin-mixingconductance.22,23) Recently, the inverse spin Hall voltageinduced by SSE was found to be enhanced by insertion ofatomically thin transition metal dichalcogenides (TMDs)24,25)into the paramagnet/ferromagnet interfaces, e.g., inPt/WSe2/YIG heterostructures [Fig. 1(a)].26) However, theinfluence of the TMD-inserted interfaces on spin-currenttransport remains unclear since SSE signals reflect averageinformation of the entire sample and is determined by multipleparameters including the spin mixing conductance and spinHall angle. To investigate the effect of TMDs on SSE or SPEsignals, spatially resolved measurements are useful.In this letter, we investigate SPE in Pt/WSe2/YIG andPt/WS2/YIG heterostructures through active thermal imagingmeasurements based on lock-in thermography (LIT).8) Wefirst focus on Pt/WSe2/YIG, where enhanced SPE-inducedtemperature modulation is expected to occur owing to thereciprocity between SSE and SPE. The LIT measurementsshow that, contrary to the expectation, SPE-induced signalsare suppressed in the WSe2-inserted regions, suggestingsuppression of the spin-current injection from Pt into YIG.Raman spectroscopy shows that the inserted WSe2 layer maybe damaged during Pt sputtering. The same behaviors wereobserved in Pt/WS2/YIG. These observations indicate thatspin transport across the Pt/YIG interface is strongly affectedby the damage of atomically thin TMD interlayers, high-lighting the importance of careful interface engineering inthe integration of atomically thin materials into spin-calor-itronic devices.Our Pt/WSe2/YIG sample was prepared as follows. WSe2flakes were synthesized by salt-assisted chemical vapordeposition (CVD). An optical micrograph shows atomicallyflat triangular flakes [Fig. 2(a)], characteristic of CVD-grownmonolayer WSe2 with zigzag edges.27) The as-grown flakeswere characterized by Raman spectroscopy and atomic forcemicroscopy (AFM). The Raman spectrum exhibits a char-acteristic peak at ∼250 cm−1, corresponding to E′ and A1′modes of WSe2, respectively [Fig. 2(b)].28) An AFMtopographic image [Fig. 2(c)] and its line profile[Fig. 2(d)] reveal a thickness of the CVD-grown WSe2 tobe ∼0.7 nm, consistent with monolayer WSe2. The as-grownWSe2 flakes were subsequently transferred onto single-crystalline YIG, grown on a single-crystalline Gd3Ga5O12(111) substrate by a liquid phase epitaxy method, using apoly(methyl methacrylate)-assisted wet-transfer process. Thethickness of YIG was 17 μm, much larger than thecharacteristic length of SPE,29,30) and the substrate was cutinto a rectangular shape with a size of 2 mm in width and7 mm in length. Finally, a 5-nm-thick Pt layer was depositedon the YIG substrate with WSe2 flakes at room temperatureby RF magnetron sputtering with a power of 50W under anContent from this work may be used under the terms of the Creative Commons Attribution 4.0 license. Any further distribution ofthis work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.080907-1© 2026 The Japan Society of Applied Physics. All rights, including fortext and data mining, AI training, and similar technologies, are reserved.Japanese Journal of Applied Physics 65, 080907 (2026) LETTERhttps://doi.org/10.35848/1347-4065/ae620fhttps://crossmark.crossref.org/dialog/?doi=10.35848/1347-4065/ae620f&domain=pdf&date_stamp=2026-05-05https://orcid.org/0009-0000-4654-2700https://orcid.org/0000-0002-3803-4770https://orcid.org/0009-0000-8573-940Xhttps://orcid.org/0000-0002-5577-8018https://orcid.org/0000-0002-1193-4322https://orcid.org/0000-0002-9733-5119https://orcid.org/0000-0001-7680-3051https://orcid.org/0000-0001-8782-9556mailto:UCHIDA.Kenichi@nims.go.jpmailto:naka24ysk@edu.k.u-tokyo.ac.jphttps://creativecommons.org/licenses/by/4.0/https://doi.org/10.35848/1347-4065/ae620fAr pressure of 0.3 Pa. As shown in an optical micrograph inFig. 3(a), the transferred WSe2 flakes result in the formationof a WSe2 interlayer only on certain areas of YIG, where thepresence of WSe2 is confirmed by the Raman spectrum inFig. 3(b). This allows for simultaneous comparison of SPEsignals within a single sample between regions with andwithout the interlayer.Applying a charge current Jc to the Pt layer of thePt/WSe2/YIG sample generates a spin current perpendicularto the interface due to the spin Hall effect [Fig. 1(b)]. Thisspin current is injected into the YIG layer in areas whereWSe2 is absent. In areas where WSe2 is present, the spin-current injection is affected by the interlayer. The spincurrent injected into the YIG layer generates a heat currentnear the interface via SPE, resulting in a temperature changeon the sample surface that is linearly proportional to theapplied charge current. This process occurs when themagnetization of YIG is in-plane and perpendicular to Jc.8)The spatial distribution of the SPE-induced temperaturemodulation in the Pt/WSe2/YIG sample was measured usingLIT under an in-plane magnetic field H (with the magnitudeFig. 1. (a) Schematic of the crystal structure of WSe2 and thePt/WSe2/YIG heterostructure. (b) Schematic of SPE measurements usingLIT. Jc, Js, H, and ∇T denote the square-wave-modulated AC chargecurrent with amplitude Jc, spatial direction of the spin current, magneticfield, and temperature gradient, respectively.Fig. 2. (a) Optical microscope image of an as-grown WSe2 flake. (b)Raman spectrum of WSe2. (c) AFM image of WSe2. (d) Height profilealong the white line in (c).Fig. 3. (a) Optical microscope image of the Pt/WSe2/YIG samplecorresponding to the LIT measurement region. (b) Raman spectra ofWSe2 before and after Pt sputtering. (c), (d) H-odd-dependent componentsof the Aodd and fodd images for the Pt/WSe2/YIG sample at Jc = 60 mA,f = 5 Hz, and μ0|H| = 100 mT, where μ0 is the vacuum permeability. (e),(f) Jc dependence of the Aodd and fodd signals for the Pt/YIG(Pt/WSe2/YIG) region, marked with a pink (light blue) square in (c) and(d), at f = 5 Hz and μ0|H| = 100 mT. The dotted line in (e) represents alinear fit. (g), (h) A and f images at Jc = 60 mA, f = 25 Hz, andμ0|H| = 0 mT.080907-2© 2026 The Japan Society of Applied Physics. All rights, including fortext and data mining, AI training, and similar technologies, are reserved.Jpn. J. Appl. Phys. 65, 080907 (2026) S-i Takano et al.H) [Fig. 1(b)]. During the LIT measurements, a square-wave-modulated AC charge current with the amplitude Jcand frequency f was applied to the Pt layer in the directionperpendicular to H and the first harmonic response of thedetected signals was decomposed into lock-in amplitude Aand phase f images by Fourier analysis. When a chargecurrent with zero DC offset is applied, SPE-induced signals(∝Jc) can be detected separately from Joule-heating con-tributions (∝Jc2).8) In contrast, applying an ON-OFF-modu-lated charge current, i.e., the square-wave-modulated ACcurrent with a finite DC offset, produces Joule-heating-induced signals, which are utilized to evaluate the spatialdistribution of the charge current density (note that thecontribution of SPE in this measurement condition isnegligibly small compared to Joule-heating-induced signals).Since the SPE-induced signals exhibit an antisymmetricdependence on H, the odd components of the LIT amplitudeand phase were calculated from raw LIT images measuredunder a positive or negative field as: Aodd = |A(+H)e–if(+H)– A(–H)e–if(–H)|/2 and fodd = –arg[A(+H)e–if(+H) – A(–H)e–if(–H)]. By doing so, we can separate the H-independentPeltier signals at the boundaries between areas with andwithout WSe2. For the LIT measurements, insulating blackink was coated on the sample surface to obtain high (>0.94)and uniform infrared emissivity.Figures 3(c) and 3(d) display the Aodd and fodd images forthe Pt/WSe2/YIG sample, measured at Jc = 60 mA,f= 5.0 Hz, and μ0|H| = 100 mT, where μ0 denotes thevacuum permeability and the region of interest correspondsto the area shown the optical micrograph in Fig. 3(a). At thismagnetic field, the magnetization of the YIG layer isuniformly aligned perpendicular to the Jc direction, satis-fying the symmetry required for SPE signals to manifest.8) Inthe Pt/YIG region without the WSe2 interlayer, clear SPE-induced signals were observed, where the sign and magni-tude of the temperature modulation are consistent with theresults in Ref. 31. The magnitude of the Aodd signals wasproportional to Jc [Fig. 3(e)], while fodd ∼0° remainedunchanged as a function of Jc [Fig. 3(f)]. In contrast, asshown in Figs. 3(c)–3(f), tiny Aodd and undefined foddsignals appear in the Pt/WSe2/YIG region, indicating thedisappearance of SPE-induced temperature modulation (notethat, at a small magnetic field of 100 mT, the contributionfrom the ordinary Ettingshausen effect in the Pt layer isnegligibly small). The signals were reduced in all theWSe2-inserted areas, irrespective of the WSe2 thickness.The averaged magnitude of the Aodd/jc signal at μ0|H| = 100mT for the regions with and without the WSe2 interlayerswas estimated to be 0.04 × 10−13 Km2A−1 and 1.79 × 10−13Km2A−1, respectively.The non-uniform SPE signal is not due to non-uniformcharge-current distribution in the Pt layer. To verify thecurrent distribution within the sample, the Joule-heatingsignal was measured using LIT at a high f value.Increasing f suppresses the broadening of temperaturechanges caused by thermal diffusion, enabling visualizationof the charge current path. Figures 3(g) and 3(h) show the Aand f images for the same Pt/WSe2/YIG sample, measuredat Jc = 60 mA and f= 25.0 Hz in the absence of a magneticfield. The Joule-heating signals were observed to be nearlyuniform across the entire area, indicating the uniform charge-current distribution.To further confirm the effect of TMDs on spin-currenttransport, we performed SPE measurements on WS2-insertedPt/YIG heterostructures. First-principles calculations haveshown that WSe2 and WS2 possess spin–orbit interactions ofcomparable strength.32,33) In addition, inserting WS2 be-tween Pt and permalloy was shown to enhance the inversespin Hall voltage due to thermally induced spin currents.34)Therefore, WSe2 and WS2 are expected to exert similareffects on SSE and SPE as interlayers, but their insertion intoPt/YIG systems and SPE measurements have not beenreported. Figure 4 presents the measurement results for thePt/WS2/YIG sample. The triangular shape observed in theoptical micrograph [Fig. 4(a)] and the characteristic Ramanpeaks measured before forming the Pt layer [Fig. 4(b)]confirm the high crystallinity of CVD-grown WS2. As shownin Figs. 4(c) and 4(d), the SPE-induced signals weremarkedly suppressed in the dispersed WS2 regions (darkspots), similar to the behavior observed in Pt/WSe2/YIG. Wealso confirmed the uniform current distribution in thePt/WS2/YIG sample [Figs. 4(e) and 4(f)].In our experiments, the SPE-induced signals were sig-nificantly suppressed in the TMD-inserted regions. This isinconsistent with the previous studies reporting an increasein SSE-induced voltage due to TMD insertion.26) Onepossible origin of the suppression of the SPE-induced signalsin our samples is the formation of interfacial alloy layers atthe Pt/TMD interfaces. Although metals deposited on TMDsare generally separated by a van der Waals gap, densityfunctional theory calculations have suggested that certainmetals can form covalent bonds with TMDs,35) such as Pd onFig. 4. (a) Optical microscope image of as-grown WS2. (b) Ramanspectrum of as-grown WS2. (c), (d) H-odd-dependent components of theAodd and fodd images for the Pt/WS2/YIG sample at Jc = 50 mA,f = 5 Hz, and μ0|H| = 100 mT. (e), (f) A and f images at Jc = 50 mA,f = 25 Hz, and μ0|H| = 0 mT.080907-3© 2026 The Japan Society of Applied Physics. All rights, including fortext and data mining, AI training, and similar technologies, are reserved.Jpn. J. Appl. Phys. 65, 080907 (2026) S-i Takano et al.WSe2.36) During sputter deposition, the collision of high-energy atoms onto the TMD surface can induce structuraldamage37–39) and promote the formation of covalent metal–chalcogen bonds. For WSe2, Nakajima et al. reported theformation of a metallic alloy layer composed of WOx, SeOx,and PtSex at the interface between sputtered Pt and WSe2.40)In fact, the Raman peaks characteristic of WSe2 disappearedafter Pt deposition in our samples, suggesting significantdamage or structural modification of WSe2 during sputtering[Fig. 3(b)]. We repeated experiments while decreasing theDC power during sputter deposition to 20W, but the resultsremained unchanged. However, considering the uniformcharge-current distribution in our samples, the formation ofalloy layers may occur only near the interfaces and not affectthe electrical resistivity and spin Hall effect within the bulkof the Pt layer. Therefore, we conclude that the formation ofthe alloy layers significantly reduces the spin injectionefficiency at the interfaces,41,42) leading to a reduction inthe SPE-induced signals. Although Pt was sputtered also inRef. 26, the degree of damage to WSe2 and structure andcomposition of the alloys formed near the Pt/WSe2 interfacemay depend on deposition conditions, such as the sputteringpower and target–substrate distance (note that the sputteringpower is not provided in Ref. 26). Even with the samemethod, such variations may account for the discrepancybetween the present results and the results in Ref. 26. Sincematerials with strong spin–orbit coupling, such as Pt, aretypically deposited by high-energy sputtering due to theirhigh melting points, low-damage deposition processes arerequired to preserve the structural integrity of TMDs andensure efficient interfacial spin transport.In conclusion, we observed significant suppression of theSPE-induced temperature modulation signals in thePt/WSe2/YIG and Pt/WS2/YIG heterostructures by the LITtechnique. Raman spectroscopy indicates that WSe2 under-goes sputtering-induced damage, which hinders spin-currentinjection from Pt to YIG. Our results highlight the necessityof low-energy deposition processes for integrating TMDsinto spin-caloritronic devices.Acknowledgments The authors thank T. Endo and M. Isomura for thetechnical support and T. Kikkawa and S.-K. Lee for the valuable discussions.This work was supported by PRESTO “Nano Materials for New PrincipleDevices” (JPMJPR23H5), CREST “Nano-Material Semiconductors”(JPMJCR23A4), FOREST (JPMJFR213X), and ERATO “Magnetic ThermalManagement Materials Project” (JPMJER2201) from JST, Japan; Grant-in-Aidfor Scientific Research (S) (22H04965, 22H04957, 24H00044), Grant-in-Aid forTransformative Research Areas (A) (21H05232, 21H05234), Grant-in-Aid forScientific Research (B) (23K26500), and Grant-in-Aid for Early-Career Scientists(25K17940) from JSPS KAKENHI, Japan; Noguchi Shitagau Research(NJ202408); and NIMS Joint Research Hub Program. MANA is supported by theWorld Premier International Research Center Initiative (WPI), MEXT, Japan.ORCID iDs Shun-ichi Takano https://orcid.org/0009-0000-4654-2700Wenjin Zhang https://orcid.org/0000-0002-3803-4770Yui Tamogami https://orcid.org/0009-0000-8573-940XTakamasa Hirai https://orcid.org/0000-0002-5577-8018Shunsuke Mori https://orcid.org/0000-0002-1193-4322Yasumitsu Miyata https://orcid.org/0000-0002-9733-5119Ken-ichi Uchida https://orcid.org/0000-0001-7680-3051Yusuke Nakanishi https://orcid.org/0000-0001-8782-95561) G. E. W. Bauer, E. Saitoh, and B. J. van Wees, Nat. Mater. 11, 391 (2012).2) S. R. Boona, R. C. Myers, and J. P. Heremans, Energy Environ. Sci. 7, 885(2014).3) K. Uchida, Proc. Jpn. Acad. Ser. B 97, 69 (2021).4) K. Uchida, S. Takahashi, K. Harii, J. Ieda, W. Koshibae, K. Ando,S. Maekawa, and E. Saitoh, Nature 455, 778 (2008).5) C. M. Jaworski, J. Yang, S. Mack, D. D. Awschalom, J. P. Heremans, andR. C. Myers, Nat. Mater. 9, 898 (2010).6) K. Uchida et al., Nat. Mater. 9, 894 (2010).7) J. Flipse, F. K. Dejene, D. Wagenaar, G. E. W. Bauer, J. B. Youssef, and B.J. Van Wees, Phys. Rev. Lett. 113, 027601 (2014).8) S. Daimon, R. Iguchi, T. Hioki, E. Saitoh, and K. Uchida, Nat. Commun. 7,13754 (2016).9) K. Uchida and J. P. Heremans, Joule 6, 2240 (2022).10) H. Adachi, F. Ando, T. Hirai, R. Modak, M. A. Grayson, and K. Uchida,Appl. Phys. Express 18, 090101 (2025).11) K. Uchida, H. Adachi, T. Ota, H. Nakayama, S. Maekawa, and E. Saitoh,Appl. Phys. Lett. 97, 172505 (2010).12) H. Jin, S. R. Boona, Z. Yang, R. C. Myers, and J. P. Heremans, Phys. Rev.B 92, 054436 (2015).13) A. Kehlberger et al., Phys. Rev. Lett. 115, 096602 (2015).14) Y. K. Kato, R. C. Myers, A. C. Gossard, and D. D. Awschalom, Science306, 1910 (2004).15) E. Saitoh, M. Ueda, H. Miyajima, and G. Tatara, Appl. Phys. Lett. 88,182509 (2006).16) S. O. Valenzuela and M. Tinkham, Nature 442, 176 (2006).17) A. Hoffmann, IEEE Trans. Magn. 49, 5172 (2013).18) J. Sinova, S. O. Valenzuela, J. Wunderlich, C. H. Back, and T. Jungwirth,Rev. Mod. Phys. 87, 1213 (2015).19) P. Deorani and H. Yang, Appl. Phys. Lett. 103, 232408 (2013).20) D. Kikuchi, M. Ishida, K. Uchida, Z. Qiu, T. Murakami, and E. Saitoh,Appl. Phys. Lett. 106, 082401 (2015).21) H. Yuasa, K. Tamae, and N. Onizuka, AIP Adv. 7, 055928 (2017).22) J. Xiao, G. E. W. Bauer, K. Uchida, E. Saitoh, and S. Maekawa, Phys. Rev.B 81, 214418 (2010).23) H. Adachi, J. Ohe, S. Takahashi, and S. Maekawa, Phys. Rev. B 83, 094410(2011).24) E. C. Ahn, npj 2D Mater. Appl. 4, 17 (2020).25) J. F. Sierra, J. Fabian, R. K. Kawakami, S. Roche, and S. O. Valenzuela,Nat. Nanotechnol. 16, 856 (2021).26) S. Lee, W. Lee, T. Kikkawa, C. T. Le, M. Kang, G. Kim, A. D. Nguyen, Y.S. Kim, N. Park, and E. Saitoh, Adv. Funct. Mater. 30, 2003192 (2020).27) N. Wada et al., Adv. Funct. Mater. 32, 2203602 (2022).28) H. Terrones et al., Sci. Rep. 4, 4215 (2014).29) S. Daimon, K. Uchida, N. Ujiie, Y. Hattori, R. Tsuboi, and E. Saitoh, Appl.Phys. Express 13, 103001 (2020).30) A. Takahagi, T. Hirai, R. Iguchi, K. Nakagawara, H. Nagano, andK. Uchida, Appl. Phys. Express 15, 063002 (2022).31) T. Yamazaki, R. Iguchi, H. Nagano, and K. Uchida, Jpn. J. Appl. Phys. 59,050901 (2020).32) Z. Y. Zhu, Y. C. Cheng, and U. Schwingenschlögl, Phys. Rev. B 84, 153402(2011).33) K. Kośmider, J. W. González, and J. Fernández-Rossier, Phys. Rev. B 88,245436 (2013).34) G. Dastgeer, M. A. Shehzad, and J. Eom, ACS Appl. Mater. Interfaces 11,48533 (2019).35) A. Allain, J. Kang, K. Banerjee, and A. Kis, Nat. Mater. 14, 1195 (2015).36) J. Kang, D. Sarkar, W. Liu, D. Jena, and K. Banerjee, International ElectronDevices Meeting, 2012, p. 17.4.1.37) Z. H. Ni, H. M. Wang, Y. Ma, J. Kasim, Y. H. Wu, and Z. X. Shen, ACSNano 2, 1033 (2008).38) X. P. Qiu, Y. J. Shin, J. Niu, N. Kulothungasagaran, G. Kalon, C. Qiu,T. Yu, and H. Yang, AIP Adv. 2, 032121 (2012).39) J. Seo, I. Sohn, and H. Kim, ACS Mater. Lett. 7, 3190 (2025).40) R. Nakajima et al., ACS Omega 10, 42973 (2025).41) K. Gupta, R. J. H. Wesselink, R. Liu, Z. Yuan, and P. J. Kelly, Phys. Rev.Lett. 124, 087702 (2020).42) L. Liu, Y. Li, Y. Liu, T. Feng, J. Xu, X. R. Wang, D. Wu, P. Gao, and J. Li,Phys. Rev. B 102, 014411 (2020).080907-4© 2026 The Japan Society of Applied Physics. All rights, including fortext and data mining, AI training, and similar technologies, are reserved.Jpn. J. Appl. Phys. 65, 080907 (2026) S-i Takano et al.https://orcid.org/0009-0000-4654-2700https://orcid.org/0009-0000-4654-2700https://orcid.org/0000-0002-3803-4770https://orcid.org/0000-0002-3803-4770https://orcid.org/0009-0000-8573-940Xhttps://orcid.org/0009-0000-8573-940Xhttps://orcid.org/0000-0002-5577-8018https://orcid.org/0000-0002-5577-8018https://orcid.org/0000-0002-1193-4322https://orcid.org/0000-0002-1193-4322https://orcid.org/0000-0002-9733-5119https://orcid.org/0000-0002-9733-5119https://orcid.org/0000-0001-7680-3051https://orcid.org/0000-0001-7680-3051https://orcid.org/0000-0001-8782-9556https://orcid.org/0000-0001-8782-9556https://doi.org/10.1038/nmat3301https://doi.org/10.1039/c3ee43299hhttps://doi.org/10.1039/c3ee43299hhttps://doi.org/10.2183/pjab.97.004https://doi.org/10.1038/nature07321https://doi.org/10.1038/nmat2860https://doi.org/10.1038/nmat2856https://doi.org/10.1103/PhysRevLett.113.027601https://doi.org/10.1038/ncomms13754https://doi.org/10.1038/ncomms13754https://doi.org/10.1016/j.joule.2022.08.016https://doi.org/10.35848/1882-0786/adf700https://doi.org/10.1063/1.3507386https://doi.org/10.1103/PhysRevB.92.054436https://doi.org/10.1103/PhysRevB.92.054436https://doi.org/10.1103/PhysRevLett.115.096602https://doi.org/10.1126/science.1105514https://doi.org/10.1126/science.1105514https://doi.org/10.1063/1.2199473https://doi.org/10.1063/1.2199473https://doi.org/10.1038/nature04937https://doi.org/10.1109/TMAG.2013.2262947https://doi.org/10.1103/RevModPhys.87.1213https://doi.org/10.1063/1.4839475https://doi.org/10.1063/1.4913531https://doi.org/10.1063/1.4977496https://doi.org/10.1103/PhysRevB.81.214418https://doi.org/10.1103/PhysRevB.81.214418https://doi.org/10.1103/PhysRevB.83.094410https://doi.org/10.1103/PhysRevB.83.094410https://doi.org/10.1038/s41699-020-0152-0https://doi.org/10.1038/s41565-021-00936-xhttps://doi.org/10.1002/adfm.202003192https://doi.org/10.1002/adfm.202203602https://doi.org/10.1038/srep04215https://doi.org/10.35848/1882-0786/abb2b5https://doi.org/10.35848/1882-0786/abb2b5https://doi.org/10.35848/1882-0786/ac6faehttps://doi.org/10.35848/1347-4065/ab8025https://doi.org/10.35848/1347-4065/ab8025https://doi.org/10.1103/PhysRevB.84.153402https://doi.org/10.1103/PhysRevB.84.153402https://doi.org/10.1103/PhysRevB.88.245436https://doi.org/10.1103/PhysRevB.88.245436https://doi.org/10.1021/acsami.9b16476https://doi.org/10.1021/acsami.9b16476https://doi.org/10.1038/nmat4452https://doi.org/10.1021/nn800031mhttps://doi.org/10.1021/nn800031mhttps://doi.org/10.1063/1.4739783https://doi.org/10.1021/acsmaterialslett.5c01010https://doi.org/10.1021/acsomega.5c05764https://doi.org/10.1103/PhysRevLett.124.087702https://doi.org/10.1103/PhysRevLett.124.087702https://doi.org/10.1103/PhysRevB.102.014411 Acknowledgments A2