Origin of the hump anomalies in the Hall resistance loops of ultrathin SrRuO 3 /SrIrO 3 multilayers
Lin Yang, Lena Wysocki, Jörg Schöpf, Lei Jin, András Kovács, Felix Gunkel, Regina Dittmann, Paul H. M. van Loosdrecht, Ionela Lindfors-Vrejoiu
OOrigin of the hump anomalies in the Hall resistance loops of ultrathin SrRuO / SrIrO multilayers Lin Yang, Lena Wysocki, J¨org Sch¨opf, Lei Jin, Andr´as Kov´acs, Felix Gunkel, Regina Dittmann, Paul H. M. van Loosdrecht, and Ionela Lindfors-Vrejoiu University of Cologne, Institute of Physics II, 50937 Cologne, Germany Ernst Ruska-Centre for Microscopy and Spectroscopy with Electrons,Forschungszentrum Jülich GmbH, 52425 Jülich, Germany PGI-7, Forschungszentrum Jülich GmbH, 52428 Jülich, Germany (Dated:
September 22 , The proposal that very small N´eel skyrmions can form in SrRuO / SrIrO epitaxial bilayers andthat the electric field-effect can be used to manipulate these skyrmions in gated devices stronglystimulated the recent research of SrRuO heterostructures. A strong interfacial Dzyaloshinskii-Moriya interaction, combined with the breaking of inversion symmetry, was considered as thedriving force for the formation of skyrmions in SrRuO / SrIrO bilayers. Here, we investigatednominally symmetric heterostructures in which an ultrathin ferromagnetic SrRuO layer is sand-wiched between large spin-orbit coupling SrIrO layers, for which the conditions are not favorablefor the emergence of a net interfacial Dzyaloshinskii-Moriya interaction. Previously the formation ofskyrmions in the asymmetric SrRuO / SrIrO bilayers was inferred from anomalous Hall resistanceloops showing humplike features that resembled topological Hall effect contributions. SymmetricSrIrO / SrRuO / SrIrO trilayers do not show hump anomalies in the Hall loops. However, theanomalous Hall resistance loops of symmetric multilayers, in which the trilayer is stacked severaltimes, do exhibit the humplike structures, similar to the asymmetric SrRuO / SrIrO bilayers. Theorigin of the Hall effect loop anomalies likely resides in unavoidable differences in the electronic andmagnetic properties of the individual SrRuO layers rather than in the formation of skyrmions. I. INTRODUCTION
Topologically protected magnetic whirls, dubbed asmagnetic skyrmions, are considered to be ideal candi-dates for the potential application in future data storage[1]. This primarily derives from their small size and roomtemperature stability [2–4], low energy consumption [3–7], and topological protection [8–10]. Epitaxial per-ovskite oxide heterostructures, such as SrRuO / SrIrO are considered to have strong interfacial Dzyaloshin-skii–Moriya interaction (DMI) due to the broken spatialinversion symmetry and the strong spin-orbit couplingin SrIrO and it was reported that N´eel skyrmions formin these heterostructures [11, 12]. The insulating natureof perovskite oxide heterostructures, such as ultrathinSrRuO / SrIrO , makes them promising systems in termsof electric field manipulation as well as the ability to en-gineer their magnetic properties. Magnetic skyrmionscan in principle be observed in real space by magneticforce microscopy (MFM) [13], Lorentz transmission elec-tron microscopy (LTEM) [14], scanning transmission x-ray microscopy (STXM) [3], spin-polarized scanning tun-neling microscopy (SP-STM) [15], x-ray magnetic circu-lar dichroism based photoemission electron microscopy(XMCD-PEEM) [16], spin-polarized low energy electronmicroscopy (SPLEEM) [17], and in reciprocal-space bysmall-angle neutron scattering (SANS) [18] and reso-nant small-angle X-ray scattering (SAXS) [19]. How-ever, for epitaxial oxide heterostructures, by using thesetechniques, the direct observation of sub-100 nm sizeskyrmions and their characterization become extremelychallenging. Therefore, the possibility of examining the formation of skyrmions by magnetotransport measure-ments is very attractive as Hall resistivity investigationsare rather easy to perform in any solid state research lab-oratory. Recently, there were many reports in which theformation of skyrmions was inferred from the observa-tion of humplike anomalies of Hall resistance loops thatwere attributed to the manifestation of the topologicalHall effect (THE). This is the case of quite a few reportsrelated to epitaxial SrRuO heterostructures and to bareSrRuO thin films [20–22]. Matsuno et al . [11] attributedthe observation of such features of Hall loops measuredfor ultrathin ferromagnetic SrRuO (thinner than 6 pseu-docubic unit cells (uc)) capped by 2 uc SrIrO thick layerto the formation of skyrmions. Many similar publica-tions followed shortly. There were reports of the hump-like features observed in Hall resistance loops of a varietyof SrRuO based samples: SrRuO /SrIrO multilayerswith relatively thick layers (10 uc thick SrRuO ) [23],BaTiO /SrRuO bilayers [24], SrRuO (5 uc)/ SrIrO (2uc) in which the iridate layer was the bottom layer onthe SrTiO substrate [25], SrRuO (8 uc)/BaTiO (2 uc)bilayers on SrTiO [26], SrIrO (2 uc)/SrRuO (10 uc)bilayers for which MFM experiments were also performed[12], or relatively thick SrRuO (3-6 nm) films grown inlow oxygen pressure [27]. Different mechanisms for theoccurrence of the interfacial DMI were proposed in thesepapers, adapted to the particular interfaces and samplepeculiarities.However, the interpretation of the observed humps inanomalous Hall effect (AHE) resistance loops as a finger-print of the THE contribution due to skyrmions is cur-rently under debate. Other reports addressed the possi-ble role played by SrRuO layer inhomogeneity, such as,Ru/O vacancies [28], thickness variations [29–31], crys-tal structure distortions [26, 32], and intermixing [33]in the occurrence of the THE-like features of the AHEloops. This division of opinions concerning the origin ofthe THE-like structures of the Hall resistance loops callsfor a careful analysis and understanding of the electronicand magnetotransport properties of SrRuO -based het-erostructures. We stress that there are no direct mea-surements of the magnitude of the interfacial DMI insuch epitaxial SrRuO -based heterostructures, but onlythe theoretical proposal from Ref. [11], which does nothowever provide any quantitative microscopic descrip-tion of how the DMI is generated at the SrIrO /SrRuO interfaces. There exists very little insight in the inter-facial DMI at epitaxial oxide interfaces [34], althoughhints for the existence of an interfacial DMI in SrIrO (2uc)/SrRuO (10 uc) bilayers were inferred from the anal-yses of the magnetic domain wall chirality [12]. We pre-viously studied asymmetric SrZrO /SrRuO /SrIrO andSrHfO /SrRuO /SrZrO multilayers in which we aimedto observe the possible effects of the net interfacial DMIon the magnetotransport properties and magnetic do-main formation [35]. However, these SrRuO multilay-ers, with insulating spacers, were magnetically only veryweakly coupled [36] and did not permit a conclusive in-vestigation of the magnetic domains by magnetic forcemicroscopy.Here we deliberately consideredSrIrO /SrRuO /SrIrO epitaxial trilayers and mul-tilayers with several repeats of the trilayer, in orderto have interfaces as symmetric as possible in thismaterial system. We aimed to eliminate, or at leastminimize, the role of interfacial DMI. In a perfectlysymmetric ultrathin film heterostructure the interfacialDMI should cancel out. However, for epitaxial interfacesof perovskite oxides (ABO ), the interfaces are likely tobe asymmetric due to the AO/BO stacking imposedby epitaxial growth, due to asymmetric intermixing ordifferent oxygen octahedral rotations (OOR) angles atthe upper and lower interface. For example, the stronginfluence of the type of interface stacking on the physicalproperties of perovskite oxide heterostructures was re-cently demonstrated for SrIrO -La . Sr . MnO bilayers[37]. For our trilayers and multilayers, because A = Sr forboth SrRuO and SrIrO , the interfaces are either of thetype SrO/IrO //SrO/RuO or IrO /SrO//RuO /SrO,and from this viewpoint the interfaces are equivalent.Prior investigations demonstrated that SrRuO lay-ers separated by 2 uc thick SrIrO non-magnetic layersare magnetically decoupled [38]. Therefore, the over-all conditions in these SrIrO /SrRuO /SrIrO multilay-ers strongly disfavor the formation of N´eel skyrmions.The trilayer SrIrO /SrRuO /SrIrO samples did not ex-hibit any humplike anomalies in the Hall effect loops. Incontrast, humplike anomalies were observed over a largetemperature range in Hall effect loops of nominally sym-metric multilayer samples, in which a SrRuO /SrIrO bilayer was stacked 3 or 6 times. The Hall effect loopswith hump anomalies can be the result of inhomogeneousmagnetic and electronic properties of the SrRuO layersin the multilayers. The inhomogeneous properties possi-bly arise from layer thickness variation, different degreeof intermixing of Ir on the Ru-site, and oxygen octahe-dron deformations that can be different for the SrRuO layers next to the substrate and for the layers at the topof the multilayer [39]. II. METHODA. Sample growth
The heterostructures studied here,SrIrO /[SrRuO /SrIrO ] m ( m = 1, 6) were grown onSrTiO (001) by pulsed-laser deposition (PLD) using aKrF excimer laser ( λ = 248 nm). SrTiO (001) single-crystal substrate was used for the deposition afterNH F-buffered HF etching for 2 - 2.5 min and annealingat 1000 ◦ C for 2 hours in air. The oxygen partialpressure and deposition temperature were optimizedat 0.133 mbar and 650 ◦ C for the growth of all thelayers. The pulse repetition rate of the laser was 5Hz and 2 Hz for the SrRuO layers and SrIrO layers,respectively. The growth of each layer was monitoredby high oxygen pressure reflective high-energy electrondiffraction (RHEED). The thickness of each SrRuO layer is nominally 6 uc and the thickness of each SrIrO layer is nominally 2 uc (1 uc layer is ∼ ◦ C down to room temperaturewith a rate of 10 ◦ C/minute. A multilayer with m =3, SrIrO /[SrRuO /SrIrO ] , was grown in a secondRHEED-PLD system (made by SURFACE systems+technology GmbH und Co. KG), under similar growthconditions, except for a higher laser fluence and target-to-substrate distance. The properties of the sample with m = 3, along with the properties of a second trilayer( m = 1) reference sample made in this PLD system,are discussed in the supplementary online material (seesection 2 and section 3). B. Sample characterization
The surface morphology of our samples was charac-terized by atomic force microscopy (AFM), as shownin the supplementary material. The microstructure ofthe multilayers, in terms of sharpness of the interfaces,layer thickness and element distribution, was investi-gated by high-angle annular dark field scanning trans-mission electron microscopy (HAADF-STEM) of crosssection specimens. The distribution of the atomic el-ements was observed with high resolution energy dis-persive X-ray spectroscopy (EDX). Both STEM and
FIG. 1. Microstructure investigations by scanning transmission electron microscopy. (a) Schematics of sampleSrIrO /[SrRuO /SrIrO ] m ( m = 1, 3, 6) grown on SrTiO (100) substrates. (b) An overview HAADF-STEM micrograph ofsample SrIrO /[SrRuO /SrIrO ] indicates the layers are uniform (except for the top layer that was damaged during the FIBprocessing of the specimen). (c) Schematics of the structure of the trilayer SrIrO /[SrRuO /SrIrO ] , for which a 6 ucSrRuO layer is inserted between two SrIrO layers (both 2 uc thick). Green, orange, blue, and red dots represent Sr, Ir, Ru,and O atomic column positions, respectively, in the crystal structure drawn using VESTA [40]. In (d) and (e) high magnifica-tion micrographs show the quality of the interfaces. EDX elemental mapping across the entire stacks shown in (e) probed theuniformity of chemical element distribution. (f) FFT pattern obtained from the image shown in (d), which shows the spotsdue to the reflections originating from the orthorhombic distortion (marked by red circles), and confirms the in-plane c -axisorientation of the layers [white arrow in (d)]. EDX were performed using an electron probe aberra-tion corrected FEI Titan 80-200 ChemiSTEM microscopeequipped with in-column EDX detectors. Hall effect mea-surements were carried out in the four-point geometry(van der Pauw), with permutating contacts for antisym-metrization. Hall resistance loops were recorded bothwith a Physical Property Measurement System (PPMS,Quantum Design Inc.) and with a home-made setup.The home setup enables the simultaneous measurementof transverse Hall resistance and magneto-optical Kerreffect (MOKE). The polar MOKE studies were per-formed with the magnetic field applied perpendicular tothe thin film surface with incoherent light from a Xelamp. The probe wavelength was chosen individually foreach sample to reduce the contributions of optical ar-tifacts, such as interference effects that can be presentin heterostructures with ultrathin films of dissimilar ox-ides. Light of 491-520 nm wavelength was used for theSrIrO /[SrRuO /SrIrO ] trilayers, 630 nm wavelengthwas used SrIrO /[SrRuO /SrIrO ] , and 610 nm wave-length was used SrIrO /[SrRuO /SrIrO ] multilayer.The magnetic moment of the samples was measuredas a function of temperature and magnetic field using asuperconducting quantum interference device (SQUID)magnetometer (MPMS XL-7 from Quantum Design).The magnetic background due to the diamagnetic SrTiO substrates was subtracted from the total magnetic re- sponse and often also corrections for a ferromagnetic im-purity contribution had to be applied [35]. III. RESULTS AND DISCUSSIONA. Microstructure investigations
The results of microstructure investigations byHAADF-STEM and high-resolution EDX are summa-rized in
Fig. /[SrRuO /SrIrO ] m ( m =1, 3, 6) is shown in Fig.
Fig.
Fig.
Fig. /[SrRuO /SrIrO ] multilayer at lowand high magnification, respectively. The stacking startswith a SrIrO and individual SrRuO and SrIrO lay-ers have thicknesses that match fairly well the expectedthickness values from the in situ RHEED monitoring ofthe layer deposition (see supplementary material, Fig.S1(c)). The high resolution image [
Fig.
Fig. layer asthe first layer on the substrate. We could not analyzequantitatively the exact stoichiometry of the individuallayers. Line profiles confirmed that the SrRuO layers areabout 6-7 uc thick (as the number of individual Ru-O planes varies between 6 and 7), while the SrIrO layersare 2-3 uc thick (as the number of individual Ir-O planesvaries between 2 and 3) (see supplementary material, Fig.S2). Concerning the intermixing, because the individualSrIrO layers are much too thin, no quantitative analy-ses of the possible intermixing at the interfaces with theSrRuO layers are feasible. Achieving atomic resolutionin EDX investigations, due to the electron beam chan-neling, volume and spectrum background effects, is veryproblematic.The structure was analyzed by fast Fourier trans-form (FFT) images [ Fig. c -axis orientation of the layers (see white arrow)and demonstrate the expected orthorhombic distortions(due to A-site atom displacements of the pseudocubicperovskite ABO ) by the presence of extra reflections,marked by the red circles in Fig.
B. Magnetic properties
We measured the dependence of the out-of-plane to-tal magnetic moment as a function of temperature, un-der zero-field cooling (ZFC, measured while heating upin 0.1 Tesla (T) after cooling the sample with no ap-plied field) and field cooled (FC) with a 0.1 T fieldapplied perpendicular to the sample surface. The re-sults for the SrIrO /[SrRuO /SrIrO ] trilayer and theSrIrO /[SrRuO /SrIrO ] multilayer are summarized in Fig.
2. For the trilayer sample, the Curie temperature( T c ) is 126 K, which was determined from the derivativeof the FC magnetic moment curve as a function of tem-perature (see inset in Fig. T c of the 6 uc thickSrRuO layer of this sample is lower than for the bulkSrRuO single crystals ( T c = 160 K), which is typicalfor ultrathin films, due to epitaxial strain and disorderand stoichiometry effects, which are more pronouncedthe thinner the SrRuO layers are[41]. The magnitudeof the magnetic moment for SrIrO /[SrRuO /SrIrO ] isalmost 6 times as large as SrIrO /[SrRuO /SrIrO ] (seethe red dotted curve in Fig. T c1 (120 K) and T c2 (140 K) occur for the SrIrO /[SrRuO /SrIrO ] epitaxialmultilayers. We assume that the occurrence of two tran-sition temperatures originates from the inhomogeneousmagnetic properties of the SrRuO layers. Most likelythe six SrRuO layers of the SrIrO /[SrRuO /SrIrO ] have all slightly different Curie temperatures distributedin the interval between T c1 and T c2 . Comparing withthe transition temperature of the trilayer sample, which FIG. 2. Temperature dependence of the magnetic mo-ment of the samples (a) SrIrO /[SrRuO /SrIrO ] and (b)SrIrO /[SrRuO /SrIrO ] under zero field cooling (ZFC, blueplot) and field cooling (FC, red plot, 0.1 T applied perpendic-ular to the sample surface) conditions. The dotted red curveshown in (b) is the FC curve of the trilayer sample from (a)multiplied by 6 and plotted for the sake of comparison. Theinsets in (a) and (b) show the first derivative of the magne-tization with respect to temperature, used to determine theCurie transition temperatures of the SrRuO layers. is 126 K, we are led to consider that the bottom mostSrRuO layer has the lowest Curie temperature, whilethe top SrRuO layers have the largest Curie tempera-ture. It is likely that the bottommost SrRuO layers aremost affected by the epitaxial strain and oxygen octahe-dral accommodation to the conditions of the SrTiO sub-strate, resulting in suppressed Curie temperature. Thetopmost SrRuO layers of the multilayer may have struc-tures that are more relaxed towards the bulk SrRuO structure, approaching the OOR values of the bulk, andthus have larger Curie temperature.Two ferromagnetic transition temperatures were re-ported recently for (SrRuO ) n /(SrIrO ) n superlatticeswith ultrathin individual layers ( n ≤ 3) [42]. The hightemperature transition, occurring also at 140 K as forour samples, was attributed to the interesting possibilitythat the ultrathin SrIrO layers undergo a canting anti-ferromagnetic transition. This transition vanished for thesuperlattices with thicker layers, n ≥ 4. As stressed inthis reference, no X-ray circular magnetic dichroism spec-troscopy (XCMD) measurements at the Ru and Ir edgeshave been performed yet to test this proposal. Thereare however XMCD studies of LaMnO /SrIrO super-lattices, which demonstrate the formation of interfacial FIG. 3. Magnetic moment hysteresis loops (measured bySQUID magnetometry, red loops) and MOKE rotation an-gle loops for samples (a) SrIrO /[SrRuO /SrIrO ] and (b)SrIrO /[SrRuO /SrIrO ] , measured in perpendicular mag-netic field. (c) Comparison of the coercive fields of thesetwo samples at different temperatures, as obtained from theSQUID and MOKE hysteresis loops. The lines are guide forthe eye. Ir-Mn molecular orbitals and ferromagnetic order of theIr magnetic moments [43].The comparison of the out-of-plane total magnetic mo-ment hysteresis loops, measured with the SQUID mag-netometer, and of the MOKE rotation angle loops of thesamples with m = 1 and m = 6 is shown in Fig.
Fig.
Fig.
C. Anomalous Hall resistance and MOKEhysteresis loops
For the particular samples under study, the measuredtotal Hall voltage has a contribution from the ordinaryHall effect and a contribution from the anomalous Halleffect, at temperatures below the Curie temperature ofthe SrRuO layers. The total Hall voltage V yx was mea-sured in van der Pauw configuration (as shown in theschematic inset of Fig. R yx as the ratio of the Hall voltage andthe excitation current I: R yx = V yx / I . For the SrIrO /[SrRuO /SrIrO ] m multilayers, themetallic SrRuO layers are magnetically decoupled [36]and are electrically connected in parallel. The SrRuO layers have very similar resistances, because they havenominally the same thickness and similar interfaces [44].The contribution of the ordinary Hall effect to the mea-sured Hall voltage V yx was subtracted from all the Hallloops shown in the paper: we assumed that in the highmagnetic field range, when the magnetization of the sam-ple gets saturated, the only field dependence comes fromthe linear contribution of the ordinary Hall effect. There-fore, in the following R yx reflects the anomalous Hall ef-fect of the samples and we refer to it as the anomalousHall resistance in discussing the data presented in Fig. R yx at fixed tem-perature in the range 10 K-110 K/120K of the SrIrO /[SrRuO /SrIrO ] andSrIrO /[SrRuO /SrIrO ] samples are plotted in Fig.
Fig. epitaxial films as well as single crystals [45], con-sistent with previous experimental data and theoreticalpredictions [11, 46–51].This peculiar sign change, fromnegative to positive as the temperature increases, comesfrom the change of the sign of the intrinsic anomalousHall conductivity. The latter is the result of the presenceof Weyl like nodes, acting as magnetic monopoles, inthe electronic band structure of SrRuO , combinedwith changes in the band structure as a function of themagnetization (and thus of the temperature). Althoughthe existence of magnetic monopoles in SrRuO isnot experimentally unambiguously proved yet, thethree-dimensional bulk SrRuO has been consideredas a system for which a large intrinsic AHE driven by FIG. 4. Summary of the anomalous Hall effect (AHE) resistance R yx loops of the SrIrO /[SrRuO /SrIrO ] m ( m = 1, 6)samples, as a function of temperature: (a) for SrIrO /[SrRuO /SrIrO ] and (b) for SrIrO /[SrRuO /SrIrO ] . In (c) and(d) the anomalous Hall resistance loops (black) and the MOKE rotation angle loops (red) measured at 80 K for sampleSrIrO /[SrRuO /SrIrO ] and SrIrO /[SrRuO /SrIrO ] , respectively, are compared. topological band structure can be observed [45]. As theenergies of nodal points and lines are different, whenthe Berry curvatures from them have opposite signs, themagnitude and the sign of intrinsic AHE conductivitycan be tuned by changing the position of the Fermi level[49, 50, 52]. We note that measurements of a secondtrilayer sample, made in another PLD chamber, agreequalitatively with the AHE and MOKE loops data ofthe trilayer discussed here [see supplementary material,Fig. S3 and Fig. S5]. The most important observationis that the trilayer samples do not exhibit any humplikeanomalies in the as measured Hall effect loops, as this isexpected for symmetric interfaces.Interestingly, the AHE resistance loops of the multi-layer sample ( m = 6) do show humplike features withina broad temperature range from 70 K - 110 K (see Fig.
Fig. layers of themultilayer.We made a symmetric SrIrO /[SrRuO /SrIrO ] m ( m = 3) multilayer in the second PLD system, withthe same PLD parameters as for the second trilayer SrIrO /[SrRuO /SrIrO ] . The AHE and MOKE loopsof this multilayer with m = 3, at different temperatures,are summarized in the supplementary material [see Fig.S4]. The behavior of the AHE loops was very intricatefor this particular multilayer and humplike features occurin the temperature range of 10-80 K. We thus confirmedthat the multilayers did have in common the appearanceof the hump anomalies, in contrast to the bare trilayers.Our symmetric multilayers, SrIrO /[SrRuO /SrIrO ] and SrIrO /[SrRuO /SrIrO ] , had a geometry that min-imizes a net interfacial DMI. The lack of net DMI is astrong indication that other mechanisms than skyrmionsand their topological Hall effect have to be consideredfor the hump anomalies of the AHE hysteresis loops. Amore plausible explanation is that the individual SrRuO layers of the multilayer SrIrO /[SrRuO /SrIrO ] haveslightly different magnetic properties (i.e., saturationmagnetization, coercive field, T c ), as a result of chem-ical and structural differences among each other (origi-nating from slight layer thickness variation, different de-gree of intermixing of Ir on the Ru-site, and oxygen octa-hedron deformations). These differences, though proba-bly minute, are of great importance for the temperaturedependence and the magnitude of the intrinsic anoma-lous Hall resistivity of each layer. Hence, the individ-ual ferromagnetic SrRuO layers generate several inde-pendent magnetotransport channels leading to the ob-served hump-anomalies of the AHE loops. As proposedin several papers [28, 33, 52–54] and in our previous re-ports [32, 35, 38], the humplike anomalies of the AHEhysteresis loops in SrRuO -based heterostructure can bewell explained by a model of several independent mag-netic channels, with distinct coercive fields and differenttemperatures at which the intrinsic AHE conductivitychanges sign. IV. SUMMARY
In epitaxial asymmetric SrRuO /SrIrO bilayersa strong interfacial Dzyaloshinskii-Moriya interaction(DMI) was proposed to emerge and to be the drivingforce for the formation of skyrmions. These skyrmionswould result in a topological Hall effect, whose mani-festation was considered to be spotted as humplike fea-tures, developing while the magnetization of the SrRuO layer reversed between saturated states. We studiedhere heterostructures in which an ultrathin ferromag-netic SrRuO layer was sandwiched between SrIrO lay-ers. Principally, this geometry disfavors the occurrence ofa net interfacial DMI and thus the formation of skyrmionswould be exceptional. SrIrO /SrRuO /SrIrO trilay-ers did not have hump anomalies of the Hall resistanceloops. However, the Hall resistance loops of multilayers,in which the trilayer was stacked several times, did ex-hibit the humplike structures, similar to the asymmetricSrRuO /SrIrO bilayers. The magnetization as a func-tion of temperature indicated that the multilayers had aspread of the Curie temperatures, hinting to differencesin the magnetic properties of the individual SrRuO lay-ers.The origin of the Hall effect anomalies likely stemsfrom unavoidable structural differences between the in-dividual SrRuO layers stacked in epitaxial multilayers.The minute structural differences (oxygen octahedra ro-tation angles, bond lengths) of the individual ruthen-ate layers result in inhomogeneous magnetic and elec- trical properties across the multilayer. It is possible thatthe individual SrRuO layers generate several indepen-dent magnetotransport channels leading to the observedanomalous features of the Hall effect loops. The rela-tion of the hump anomalies to the skyrmion formationcannot be ruled out, however our data strongly supportthe interpretation in terms of multiple magnetotransportchannels present in multilayers. V. ACKNOWLEDGEMENT
We thank Michael Ziese for constant valuable advicewith the physical properties of SrRuO samples andwith the SQUID and Hall measurements (University ofLeipzig). We are grateful to Achim Rosch for fruit-ful discussions and insightful suggestions, Susanne Hei-jligen for kind assistance with SQUID measurements,and Andrea Bliesener for AFM and assistance withthe PPMS measurements (University of Cologne). Wethank Ren´e Borowski for etching the STO substrates(FZ Julich). This work was supported by the GermanResearch foundation (DFG) (projects number 335038432and 403504808) and through CRC1238 (Project No.277146847). PvL and ILV thank DFG for funding thepurchase of the PLD-RHEED system at University ofCologne, with which some of the investigated sampleswere grown (Project No.407456390). Also support fromthe German Excellence Initiative via the key profile area“quantum matter and materials” (QM2) of the Univer-sity of Cologne is gratefully acknowledged. L. Y. thanksfor financial support from the China Scholarship Council.(File No.201706750015). Correspondence to : L.Y. ([email protected])and I. L.-V. ([email protected]) [1] N. Kiselev, A. Bogdanov, R. Sch¨afer, and U. R¨oßler,Journal of Physics D: Applied Physics , 392001 (2011).[2] A. Soumyanarayanan, M. Raju, A. L. G. Oyarce, A. K. C.Tan, M.-Y. Im, A. P. Petrovi´c, P. Ho, K. H. Khoo,M. Tran, C. K. Gan, F. Ernult, and C. Panagopoulos,Nature materials , 898 (2017).[3] S. Woo, K. Litzius, B. Kru¨ger, M.-Y. Im, L. Caretta,K. Richter, M. Mann, A. Krone, R. M. Reeve,M. Weigand, P. Agrawal, I. Lemesh, M.-A. Mawass,P. Fischer, M. Kl¨aui, and G. S. D. Beach, Nature ma-terials , 501 (2016).[4] S. Das, Y. Tang, Z. Hong, M. Gonc¸alves, M. Mc-Carter, C. Klewe, K. Nguyen, F. G´omez-Ortiz, P. Shafer,E. Arenholz, V. A. Stoica, S.-L. Hsu, B. Wang, C. Ophus,J. F. Liu, C. T. Nelson, S. Saremi, B. Prasad, A. B. Mei,D. G. Schlom, D. A. M. J. ´In˜iguez, P. Garc´ıa-Fern´andez,L. Q. Chen, J. Junquera, L. W. Martin, and R. Ramesh,Nature , 368 (2019).[5] X. Yu, N. Kanazawa, W. Zhang, T. Nagai, T. Hara,K. Kimoto, Y. Matsui, Y. Onose, and Y. Tokura, Na-ture communications , 988 (2012). [6] A. Fert, V. Cros, and J. Sampaio, Nature nanotechnology , 152 (2013).[7] J. Iwasaki, M. Mochizuki, and N. Nagaosa, Nature com-munications , 1463 (2013).[8] C. Reichhardt, D. Ray, and C. O. Reichhardt, Physicalreview letters , 217202 (2015).[9] W. Jiang, P. Upadhyaya, W. Zhang, G. Yu, M. B.Jungfleisch, F. Y. Fradin, J. E. Pearson, Y. Tserkovnyak,K. L. Wang, O. Heinonen, S. G. E. te Velthuis, andA. Hoffmann, Science , 283 (2015).[10] W. Jiang, G. Chen, K. Liu, J. Zang, S. G. te Velthuis,and A. Hoffmann, Physics Reports , 1 (2017).[11] J. Matsuno, N. Ogawa, K. Yasuda, F. Kagawa,W. Koshibae, N. Nagaosa, Y. Tokura, and M. Kawasaki,Science Advances , e1600304 (2016).[12] K.-Y. Meng, A. S. Ahmed, M. Ba´cani, A.-O. Mandru,X. Zhao, N. Bagu´es, B. D. Esser, J. Flores, D. W. Mc-Comb, H. J. Hug, and F. Y. Yang, Nano letters , 3169(2019).[13] A. Fert, N. Reyren, and V. Cros, Nature Reviews Mate-rials , 1 (2017). [14] X. Yu, J. P. DeGrave, Y. Hara, T. Hara, S. Jin, andY. Tokura, Nano letters , 3755 (2013).[15] S. Heinze, K. Von Bergmann, M. Menzel, J. Brede,A. Kubetzka, R. Wiesendanger, G. Bihlmayer, andS. Blu¨gel, Nature Physics , 713 (2011).[16] O. Boulle, J. Vogel, H. Yang, S. Pizzini,D. de Souza Chaves, A. Locatelli, T. O. Mente¸s,A. Sala, L. D. Buda-Prejbeanu, O. Klein, M. Belmegue-nai, Y. Roussign´ea, A. Stashkevich, S. M. Ch´erif,L. Aballe, M. Foerster, M. Chshiev, S. Auffret, I. M.Miron, and G. Gaudin, Nature nanotechnology , 449(2016).[17] G. Chen, A. Mascaraque, A. T. N’Diaye, and A. K.Schmid, Applied Physics Letters , 242404 (2015).[18] S. Mu¨hlbauer, B. Binz, F. Jonietz, C. Pfleiderer,A. Rosch, A. Neubauer, R. Georgii, and P. B¨oni, Science , 915 (2009).[19] Y. Yamasaki, D. Morikawa, T. Honda, H. Nakao, Y. Mu-rakami, N. Kanazawa, M. Kawasaki, T. Arima, andY. Tokura, Physical Review B , 220421 (2015).[20] H. Huang, S.-J. Lee, B. Kim, B. Sohn, C. Kim, C.-C. Kao, and J.-S. Lee, arXiv preprint arXiv:2004.08092(2020).[21] P. Zhang, A. Das, E. Barts, M. Azhar, L. Si, K. Held,M. Mostovoy, and T. Banerjee, Phys. Rev. Research ,032026 (2020).[22] C. Wang, C.-H. Chang, A. Herklotz, C. Chen, F. Ganss,U. Kentsch, D. Chen, X. Gao, Y.-J. Zeng, O. Hellwig,M. Helm, S. Gemming, Y.-H. Chu, and S. Q. Zhou, Ad-vanced Electronic Materials , 2000184 (2020).[23] B. Pang, L. Zhang, Y. Chen, J. Zhou, S. Yao, S. Zhang,and Y. Chen, ACS Applied Materials & Interfaces , 3201(2017).[24] L. F. Wang, Q. Y. Feng, Y. Kim, R. Kim, K. H. Lee,S. D. Pollard, Y. J. Shin, H. B. Zhou, W. Peng, D. Lee,W. J. Meng, H. Yang, J. H. Han, M. Kim, Q. Y. Lu, andT. W. Noh, Nature materials , 1087 (2018).[25] Y. Ohuchi, J. Matsuno, N. Ogawa, Y. Kozuka,M. Uchida, Y. Tokura, and M. Kawasaki, Nature com-munications , 213 (2018).[26] Y. Gu, Y.-W. Wei, K. Xu, H. Zhang, F. Wang, F. Li,M. S. Saleem, C.-Z. Chang, J. Sun, C. Song, et al. , Jour-nal of Physics D: Applied Physics , 404001 (2019).[27] Q. Qin, L. Liu, W. Lin, X. Shu, Q. Xie, Z. Lim, C. Li,S. He, G. M. Chow, and J. Chen, Advanced Materials , 1807008 (2019).[28] D. Kan and Y. Shimakawa, physica status solidi (b) ,1800175 (2018).[29] L. Wang, Q. Feng, H. G. Lee, E. K. Ko, Q. Lu, and T. W.Noh, Nano Letters , 2468 (2020).[30] P.-C. Wu, H. Song, Y. Yuan, B. Feng, Y. Ikuhara,R. Huang, P. Yu, C.-G. Duan, and Y.-H. Chu, PhysicalReview Materials , 014401 (2020).[31] G. Kimbell, P. M. Sass, B. Woltjes, E. K. Ko, T. W. Noh,W. Wu, and J. W. Robinson, Physical Review Materials , 054414 (2020).[32] M. Ziese, L. Jin, and I. Lindfors-Vrejoiu, Journal ofPhysics: Materials , 034008 (2019).[33] D. J. Groenendijk, C. Autieri, T. C. van Thiel,W. Brzezicki, J. Hortensius, D. Afanasiev, N. Gauquelin,P. Barone, K. van den Bos, S. van Aert, J. Verbeeck,A. Filippetti, S. Picozzi, M. Cuoco, and A. D. Caviglia, Physical Review Research , 023404 (2020).[34] N. Mohanta, E. Dagotto, and S. Okamoto, Physical Re-view B , 064429 (2019).[35] L. Wysocki, J. Sch¨opf, M. Ziese, L. Yang, A. Kov´acs,L. Jin, R. B. Versteeg, A. Bliesener, F. Gunkel, L. Ko-rnblum, R. Dittmann, P. H. M. van Loosdrecht, andI. Lindfors-Vrejoiu, ACS omega , 5824 (2020).[36] L. Wysocki, R. Mirzaaghayev, M. Ziese, L. Yang,J. Sch¨opf, R. B. Versteeg, A. Bliesener, J. Engelmayer,A. Kov´acs, L. Jin, F. Gunkel, R. Dittmann, P. H. M.van Loosdrecht, and I. Lindfors-Vrejoiu, Applied PhysicsLetters , 192402 (2018).[37] L. Bergmann, P. Du¨ring, S. Agrestini, A. Efimenko, S.-C. Liao, Z. Hu, P. Gargiani, C.-J. Choi, H. Baik, D.-S.Park, K. D¨orr, and A. D. Rata, AIP Advances , 035132(2020).[38] L. Wysocki, L. Yang, F. Gunkel, R. Dittmann, P. H.van Loosdrecht, and I. Lindfors-Vrejoiu, Physical ReviewMaterials , 054402 (2020).[39] K. Samanta, M. Leˇzai´c, M. Merte, F. Freimuth, S. Blu¨gel,and Y. Mokrousov, Journal of Applied Physics ,213904 (2020).[40] K. Momma and F. Izumi, Journal of applied crystallog-raphy , 1272 (2011).[41] Y. J. Chang, C. H. Kim, S.-H. Phark, Y. Kim, J. Yu, andT. Noh, Physical review letters , 057201 (2009).[42] Z. T. Zeng, J. T. Feng, X. Zheng, C. H. Wang, J. W. Liu,Z. X. Lu, F.-X. Jiang, X.-H. Xu, Z. M. Wang, and R.-W.Li, Applied Physics Letters , 142401 (2020).[43] Y. J. Zhang, Y. Z. Luo, L. Wu, M. Suzuki, Y. Hi-rata, K. Yamagami, K. Takubo, K. Ikeda, K. Yamamoto,A. Yasui, N. Kawamura, C. Lin, K. Koshiishi, X. Liu,J. X. Zhang, Y. Hotta, R. Wang, A. Fujimori, Y. H.Lin, C. W. Nan, L. Shen, and H. Wadat, arXiv preprintarXiv:2004.06864 (2020).[44] S. L. Zhang, Y. Liu, L. J. Collins-McIntyre, T. Hesjedal,J. Y. Zhang, S. G. Wang, and G. H. Yu, Scientific Reports , 2087 (2013).[45] Z. Fang, N. Nagaosa, K. S. Takahashi, A. Asamitsu,R. Mathieu, T. Ogasawara, H. Yamada, M. Kawasaki,Y. Tokura, and K. Terakura, Science , 92 (2003).[46] N. Haham, Y. Shperber, M. Schultz, N. Naftalis,E. Shimshoni, J. W. Reiner, and L. Klein, Physical Re-view B , 174439 (2011).[47] I. Lindfors-Vrejoiu and M. Ziese, physica status solidi (b) , 1600556 (2017).[48] G. Malsch, D. Ivaneyko, P. Milde, L. Wysocki, L. Yang,P. H. van Loosdrecht, I. Lindfors-Vrejoiu, and L. M. Eng,ACS Applied Nano Materials , 1182 (2020).[49] X. J. Wang, D. Vanderbilt, J. R. Yates, and I. Souza,Physical Review B , 195109 (2007).[50] Y. Chen, D. L. Bergman, and A. A. Burkov, PhysicalReview B , 125110 (2013).[51] D. Vanderbilt, I. Souza, and F. D. M. Haldane, PhysicalReview B , 117101 (2014).[52] B. Sohn, E. Lee, W. Kyung, M. Kim, H. Ryu, J. S.Oh, D. Kim, J. K. Jung, B. Kim, M. Han, et al. , arXivpreprint arXiv:1912.04757 (2019).[53] A. Gerber, Physical Review B , 214440 (2018).[54] D. Kan, T. Moriyama, K. Kobayashi, and Y. Shimakawa,Physical Review B , 180408 (2018). upplementary online mat e rialsOrigin of the hump anomalies in the Hall resistance loops of ultrathin SrRuO / SrIrO multilayers Lin
Yang, Lena
Wysocki, J ö rg Sch ö pf, Lei
Jin, Andr á s Kov á cs, Felix
Gunkel, Regina Dittmann, Paul H. M. van Loosdrecht, and Ionela Lindfors-Vrejoiu University of Cologne,
Institute of Physics
II,
Cologne,
Germany Ernst
Ruska-Centre for
Microscopy and
Spectroscopy with
Electrons,
Forschungszentrum
Jülich
GmbH,
Jülich,
Germany PGI-7,
Forschungszentrum
Jülich
GmbH,
Jülich,
Germany
I. STRUCTURAL CHARACTERIZATION: IN SITU RHEED AND AFM INVESTIGATIONS
We monitored the growth mode and the thickness of the individual layers of the multilayers by employing in situ highoxygen pressure reflective high-energy diffraction (RHEED). The average intensity of the RHEED specular spot as afunction of time is shown in
Fig.
S1(a) and
Fig.
S1(c) for SrIrO /[SrRuO /SrIrO ] and SrIrO /[SrRuO /SrIrO ] ,respectively, both grown in the same PLD chamber at FZ Ju¨lich under the same PLD parameters. The oscillations inthe RHEED intensity-time curves during the SrIrO deposition show that the iridate layers grew in a layer-by-layergrowth mode: see a first clear oscillation followed by a more damped second oscillation, marked by the small blackarrows, corresponding to the growth of two monolayers of SrIrO . SrRuO layers grew in step-flow growth regime [1].The similar RHEED intensity behavior indicates the homogeneous thickness of sample layers. The surface morphologyof both SrIrO /[SrRuO /SrIrO ] and SrIrO /[SrRuO /SrIrO ] samples was investigated by atomic force microscopy(AFM) and is shown in Fig.
S1(b) and
Fig.
S1(d), respectively. Non-continuous terrace-like structure is shownby the trilayer (see
Fig.
S1(b)) and the multilayer has a rather large density of tiny holes, coming from probablyincomplete coverage of the top most SrIrO layer (see Fig.
S1(d)).As shown in the main paper a SrIrO /[SrRuO /SrIrO ] multilayer was investigated by scanning transmissionelectron microscopy (STEM) in the high-angle annular dark field mode (HAADF) (see Fig. 1 of the paper). We madeenergy dispersive x-ray spectroscopy (EDX) maps of the chemical elements Sr, Ti, Ru and Ir across the multilayer,as shown in Fig.
S2(a). Line profiles for each element were acquired and are shown in
Fig.
S2(b): they confirm thatthe SrRuO layers are about 6-7 uc thick (as the number of individual Ru-O planes varies between 6 and 7), whilethe SrIrO layers are 2-3 uc thick (as the number of individual Ir-O planes varies between 2 and 3), in agreementwith our RHEED observations. In Fig.
S2(b), we drew four lines, two in orange and the other two in green. Thefirst orange (green) line marks the position of Ir (Ru), the second orange (green) line marks the neighboring B-sitepositions. For both cases, the net count drops from 80 to 30. That means no matter what the element is, thedecrease is the same. Only the thickness of the individual layers in the growth direction matters. The SrIrO layers,being only 2-3 uc thick, are much too thin to allow quantitative analyses of the possible intermixing at the interfaceswith the SrRuO layers. To achieve atomic resolution in EDX investigations, due to the electron beam channeling,volume and spectrum background effects, is very problematic. II. MOKE AND HALL LOOPS OF A SECOND SrIrO /[SrRuO /SrIrO ] TRILAYER AND ASrIrO /[SrRuO /SrIrO ] MULTILAYER MADE IN A SECOND PLD SYSTEM
To investigate the reproducibility of the behavior of the AHE loops, we grew the trilayer sample SrIrO /[SrRuO /SrIrO ] and a multilayer sample with m = 3 SrIrO /[SrRuO /SrIrO ] , in our PLD system at University of Cologne. TheMOKE and Hall effect loops of the samples are summarized in Fig.
S3 and in
Fig.
S4, respectively. The Kerrrotation angle and AHE resistance measurements were performed simultaneously and both type of loops show similarcoercive fields at all temperatures. In the ferromagnetic phase of SrRuO layer, as the temperature increases, theAHE resistance changes sign, from negative to positive, somewhat below 80 K. There are no humplike features in theAHE resistance loops also for this second SrIrO /[SrRuO /SrIrO ] . These results are consistent with the trilayersample reported in the main paper (see Fig. 4 of the paper).For the multilayer SrIrO /[SrRuO /SrIrO ] (see Fig.
S4), the MOKE and Hall loops show striking differencesin the temperature range 10 K - 80 K. For the MOKE measurement, the open MOKE loops are obtained up toabout 110 K, indicating that the Curie temperature is at least 110 K. For the anomalous Hall effect resistancemeasurement, the clear humps exist from the lowest temperature we can measure at, 10 K, to about 80 K. The
Dated: September 22, 2020 evolution of the sign of anomalous Hall constant is quite different from the previous sample SrIrO /[SrRuO /SrIrO ] and bilayer sample in Matsuno et al. paper [2]. The sign of the total anomalous Hall resistance (voltage) of thesample SrIrO /[SrRuO /SrIrO ] is positive down to about 10 K. The multiple peaks of the AHE loops at 10 K and30 K may be explained, if the global loop is decomposed in three independent loops generated by the three SrRuO layers. The three separated magnetic layers possess slightly different magnetic and AHE properties [3]. III. AHE RESISTANCE AND MOKE ROTATION ANGLE FIELD LOOPS FOR TWO SYMMETRICSrIrO /[SrRuO /SrIrO ] TRILAYERS
The data of AHE resistance and MOKE loop measurement at different temperatures for the trilayer made in anotherPLD system and its reference trilayer sample (studied in the main paper) are summarized in
Fig.
S5. It should benoted that we obtained these data by simultaneous measurement of MOKE and Hall effect resistance loops in ourcombined MOKE-Hall setup, with the sample in the same cryostat. In general, for each sample, the AHE loopsscale with the MOKE loops fairly well. The most striking differences between the two samples are the magnitude ofthe coercive fields and the temperature dependence of the anomalous Hall effect resistance. The second trilayer hasmuch smaller coercive field at low temperatures, see for instance the loops measured at 10 K: the coercive field ofthe reference layer is almost twice as large. The AHE changes sign well below 80 K for the second trilayer, while thereference trilayer changes the sign of the AHE from negative to positive at 86 K. We grew the second trilayer with theintention to obtain a fairly comparable sample, i.e. with 2 uc thick SrIrO and 6 uc thick SrRuO layers. However,the second PLD chamber has major differences (such as target-to-substrate distance, perpendicular geometry, laserfluence measurement), which made it not possible to have the same PLD parameters for the growth. Thus, we stresshow important the growth conditions are for the magnetic and electronic properties of SrRuO /SrIrO oxide thinfilms. However, the most important similarity between the two trilayer samples is that both do not show any humplikeanomalies of the AHE resistance loops, demonstrating a consistent behavior for the expected symmetric trilayers. Correspondence to : L.Y. ([email protected]) and I. L.-V. ([email protected]) [1] J. Choi, C.-B. Eom, G. Rijnders, H. Rogalla, and D. H. Blank, Applied physics letters , 1447 (2001).[2] J. Matsuno, N. Ogawa, K. Yasuda, F. Kagawa, W. Koshibae, N. Nagaosa, Y. Tokura, and M. Kawasaki, Science Advances , e1600304 (2016).[3] L. Wysocki, L. Yang, F. Gunkel, R. Dittmann, P. H. van Loosdrecht, and I. Lindfors-Vrejoiu, Physical Review Materials ,054402 (2020). FIG. S1. RHEED and AFM investigations of SrIrO /[SrRuO /SrIrO ] and SrIrO /[SrRuO /SrIrO ] , which were grownunder the same PLD conditions: (a) and (c) deposition time dependence of RHEED intensity; (b) and (d) AFM topographyimages (5 µ m × µ m scans) of the top surface of the as grown samples. The small black arrows in (a) and (c) mark the topof the oscillations of the RHEED intensity signal during the growth of the SrIrO layers. The inset in (c) shows the RHEEDsignal recorded during the growth of the first three layers of the multilayer SrIrO /[SrRuO /SrIrO ] .FIG. S2. (a) HAADF-STEM and EDX analyses of a SrIrO /[SrRuO /SrIrO ] multilayer. The line profiles across themultilayer, starting from the substrate upwards in the growth direction, for the atomic column with the Sr, Ti, Ru, Ir ions areshown in (b). FIG. S3. Anomalous Hall effect resistance loops and Kerr rotation angle loops for the second SrIrO /[SrRuO /SrIrO ] trilayer at different temperatures from 10 K to 120 K: AHE resistance loops (red line) and the MOKE loops (black line withsolid square dots). The AHE changes sign to positive above 60-70 K.FIG. S4. Anomalous Hall effect resistance loops and Kerr rotation angle loops for the multilayer SrIrO /[SrRuO /SrIrO ] ( m = 3), at different temperatures from 10 K to 120 K: AHE resistance loops (red line) and the MOKE loops (black line). FIG. S5. Comparison of AHE and MOKE loops for the two SrIrO /[SrRuO /SrIrO ] trilayers (the reference is the trilayerpresented in the main paper), made in different PLD chambers, with differing PLD conditions. We compare the loops atdifferent temperatures, capturing the change of sign of AHE for both samples: (a) 10 K, (b) 80 K, (c) 100 K. Anomalous Halleffect (AHE) resistance loops are plotted in red and green and the Kerr rotation angle black and blue. The magnitude of Hallloop for SrIrO /[SrRuO /SrIrO ]1