Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Cation- and lattice-site-selective magnetic depth profiles of ultrathin $\mathrm{Fe_3O_4}$(001) films

Cation- and lattice-site-selective magnetic depth profiles of ultrathin $\mathrm{Fe_3O_4}$(001)... Cation- and lattice-site-selective magnetic depth pro les of ultrathin Fe O (001) lms 3 4 1, 2, 3 1 1 1 Tobias Pohlmann, Timo Kuschel, Jari Rodewald, Jannis Thien, Kevin Ruwisch, Florian 2 4 5 1,y 1,z Bertram, Eugen Weschke, Padraic Shafer, Joachim Wollschl ager, and Karsten Kupp  er Department of Physics, Osnabruck  University, Barbarastr. 7, 49076 Osnabruck,  Germany DESY Photon Science, Notkestr. 85, 22607 Hamburg, Germany Center for Spinelectronic Materials and Devices, Department of Physics, Bielefeld University, Universitatsstr. 25, 33615 Bielefeld, Germany Helmholtz-Zentrum Berlin fur  Materialien und Energie, Wilhelm-Conrad-Rontgen-Campus BESSY II, Albert-Einstein-Strasse 15, 12489 Berlin, Germany Advanced Light Source, Lawrence Berkeley National Laboratory, 6 Cyclotron Rd, Berkeley, CA, 94720, USA (Dated: May 5, 2020) A detailed understanding of ultrathin lm surface properties is crucial for the proper interpretation of spectroscopic, catalytic and spin-transport data. We present x-ray magnetic circular dichroism (XMCD) and x-ray resonant magnetic re ectivity (XRMR) measurements on ultrathin Fe O lms 3 4 to obtain magnetic depth pro les for the three resonant energies corresponding to the di erent 2+ 3+ 3+ cation species Fe , Fe and Fe located on octahedral and tetrahedral sites of the inverse spinel oct tet oct structure of Fe O . By analyzing the XMCD spectrum of Fe O using multiplet calculations, 3 4 3 4 the resonance energy of each cation species can be isolated. Performing XRMR on these three resonant energies yields magnetic depth pro les that correspond each to one speci c cation species. 3+ The depth pro les of both kinds of Fe cations reveal a 3:9  1 A-thick surface layer of enhanced 2+ magnetization, which is likely due to an excess of these ions at the expense of the Fe species in oct 3+ the surface region. The magnetically enhanced Fe layer is additionally shifted about 3  1:5 A tet 3+ farther from the surface than the Fe layer. oct Introduction.Magnetite, Fe O , is one of the most (XPS), to explain the behavior of the bulk material [14]. 3 4 frequently investigated transition-metal oxides, since it is In particular, drawing conclusions about the cation dis- a key material in spintronics [1], spin caloritronics [2] and tribution of magnetite requires caution, because the bulk material chemistry [3]. Fe O thin lms were considered material of the inverse spinel Fe O should contain di- 3 4 3 4 2+ highly suitable as electrode material for magnetic tunnel valent Fe , as well as trivalent ions in both octahedral oct 3+ 3+ junctions [4, 5] due to their predicted half-metallic behav- and tetrahedral coordination, Fe and Fe . However, tet oct ior with 100% spin polarization [6]. However, the promise the DFT+U calculations of the SCV structure predict 3+ was never quite met, with modest tunnel magnetoresis- the rst four atomic layers to only contain Fe ions and tance ratios ranging from -26% to 18% [4, 5, 7]. In or- to have a formal stoichometry of Fe O , in agreement 11 16 3+ der to test its half-metallicity, spin-resolved x-ray photo- with earlier reports on an excess of Fe at the (001) electron spectroscopy on Fe O (111) lms found a spin- surface [15]. A subsequent study of how the magnetic 3 4 polarization of about 80% [8], while on Fe O (001) the properties of Fe O are a ected at the surface used x-ray 3 4 3 4 same technique yielded polarizations of only  40%70% magnetic circular dichroism (XMCD) with varying prob- [9{11]. ing depth and found a reduced magnetic moment at the surface of a natural magnetite crystal [16]. Both the reduced tunnel magnetoresistance and the In this letter we report an investigation of the mag- deviations from 100% spin-polarization in spin-resolved netic surface properties of ultrathin Fe O (001) lms, in XPS were argued to emerge from interface and surface 3 4 contrast to the bulk, by recording magnetooptical depth e ects, respectively [4, 5, 8, 10, 11]. For Fe O (111) 3 4 pro les of the three cation species in Fe O . Using pho- lms deposited on semiconducting ZnO(0001), lattice- 3 4 tons with resonant energies, we employed XRMR to de- site-selective depth pro les obtained by x-ray resonant termine the magnetooptical depth pro les at the energies magnetic re ectivity (XRMR) and electron energy loss 2+ 3+ 3+ characteristic for Fe , Fe and Fe in magnetite's spectroscopy did not nd a notable surface modi ca- oct tet oct tion apart from a Fe termination [12]. Reduction of L XMCD spectrum individually for three ultrathin 2;3 oct the spin-polarization measured at the Fe O (001) sur- Fe O =MgO(001) lms of varying thicknesses. We nd a 3 4 3 4 face was typically considered to originate from a sur-  3:9 A layer of enhanced magnetooptical absorption at 3+ face reconstruction, the existence of which has long been the surface at the resonant energies of both Fe species 2+ 3+ known but only recently has been resolved as a subsur- but not for Fe , suggesting an Fe -rich surface. face cation vacancy (SCV) structure [13]. This revela- Experimental and theoretical details.We prepared tion highlights the issues that might arise from general- the Fe O =MgO(001) samples in a multichamber ultra- 3 4 izing results from surface-sensitive techniques, such as high-vacuum system using reactive molecular beam x-ray absorption spectroscopy (XAS) in total electron epitaxy (RMBE). Their chemical composition and p p yield (TEY) mode and x-ray photoelectron spectroscopy ( 2 2)R45 superstructure was con rmed by in-situ arXiv:2005.01657v1 [cond-mat.mes-hall] 4 May 2020 2 4 0 XPS and low-energy electron di raction (LEED), respec- d a t a 2 5 n m F e O ( c ) ( a ) 1 . 0 3 4 s i m u l a t i o n tively. For details on the deposition and characterization i 3 5 % 2 + n 3 5 F e o c t B = 4 T u 3 3 % methods please see Refs. [17] or [18]. . 3 + F e 3 1 % t e t r 0 . 5 3 + 3 0 For the XAS and XMCD study, the samples were a F e o c t % transfered from our lab under ambient conditions to the y 2 5 t 0 . 0 Superconducting Vector Magnet Endstation at beamline 4.0.2 of the Advanced Light Source (ALS). All samples 2 0 ( b ) were measured in a magnetic eld of 4 T along the x-ray t 0 . 0 1 5 beam at room temperature. The incidence angle of the x-rays was 30 from the [100] direction of Fe O , and r 3 4 1 0 ( - 0 . 2 the degree of circular polarization was 90%. The XAS and XMCD spectra were measured across the Fe L 2;3 - 0 . 4 absorption edges (690 750 eV). All XAS spectra were X 2 + 3 + 3 + 7 0 0 7 1 0 7 2 0 7 3 0 F e F e F e measured in the TEY mode, which has a probing depth o c t t e t o c t e n e r g y ( e V ) in magnetite of about 3 nm. The XAS and XMCD data were analyzed by apply- FIG. 1. (a) XAS and (b) XMCD spectrum at the Fe L edge 2;3 ing the sum rules [19{21] and charge-transfer multiplet for the 25 nm Fe O lm, taken at 4 T external magnetic eld, 3 4 calculations using the Thole code [22] with assistance of at room temperature and in TEY mode. A step function was CTM4XAS [23, 24]. For the sum rules, we took into ac- subtracted from the XAS spectrum. Black dots represent count a correction factor of 1.142 derived by Teramura data; green, red and blue spectra are multiplet calculations 2+ holes for the three cation species of Fe O , the grey line is their 3 4 et al. for Fe [21] and assumed 14 . For the mul- f:u: sum. The cation spectra are o set for better visibility. (c) tiplet calculations, we assumed the three-cation model, Cation stoichometry used to obtain the t in (a) and (b). using crystal eld and charge-transfer parameters as de- scribed in Ref. [17]. The parameters and more details regarding the multiplet calculations can also be found in Figure 1 shows the XAS and XMCD spectra of the 25 nm the Supplemental Materials [25] (Chapter A, including Fe O lm, recorded at ALS. Corresponding data mea- 3 4 Refs. [17, 26, 27]). The multiplet states resulting from sured at BESSY II, under the same conditions in which these calculations were compared to the data by assum- the XRMR was performed, can be found in the Supple- ing a Gaussian instrumental broadening of 0:2 eV, and a mental Material [25] (Chapter A). The spin and orbital Lorentzian lifetime broadening of 0:3 eV at L and 0:6 eV moments  and  obtained from the sum rules are spin orb at L . given in Tab. I. Their sum is slightly reduced compared The samples were transfered to BESSY II under ambi- to the bulk value of magnetite of  =  +  = spin orb ent conditions and x-ray re ectivity (XRR) and XRMR 4:07  [32]. They also show a tendency toward a slightly were performed in the XUV Di ractometer at beamline higher moment for the thicker lms. This behavior of UE46 PGM-1 [28]. The samples were placed between magnetite lms has also been observed previously [17], two permanent magnets in a magnetic eld of 200 mT and can be explained by a higher density of anti-phase at room temperature, with a degree of circular polariza- boundaries (APBs) for thinner lms due to the antifer- tion of 90%. First, we characterized the structural prop- romagnetic coupling of APBs reducing the average mag- erties of the sample (thickness d, roughness ) by XRR netic moment of the lm [33{35]. at o -resonant energies (680 eV, 1000 eV). Second, XAS Additionally, multiplet simulations of the three-cation and XMCD were measured in order to select suitable model are tted to the XMCD data (cf Fig. 1). By energies for XRMR. Finally,  2 scans in the range 2 = 0 140 at resonant energies E with extrema in the XMCD signal (maximum at 708:4 eV, minimum surf at 709:5 eV, maximum at 710:2 eV) were performed with TABLE I. Thicknesses d, surface roughnesses  and inter- int face roughnesses  of the investigated samples as obtained both right and left circularly polarized x-rays, to obtain by tting o -resonant XRR curves, as well as their spin mo- the XRMR asymmetry ratios I (E ; q ). These curves i z ments  and orbital moments  from the sum-rule anal- spin orb were then tted with the Zak matrix formalism using the ysis. software ReMagX [29] to determine the depth pro les sample 13 nm 25 nm 50 nm of the magnetooptical absorption  (z) and dispersion d (nm) 13:5  0:5 25:2  0:3 52:8  0:3 (z) along the lm height z. A detailed review of the surf (nm) 0:22  0:05 0:33  0:05 0:34  0:05 XRMR method and the software is given in Ref. [29], int (nm) 0:35  0:05 0:35  0:05 0:37  0:05 and a conclusive recipe for tting XRMR data can be ( =f:u:) 3:2  0:3 3:5  0:3 3:7  0:3 spin B found in Refs. [30, 31]. ( =f:u:) 0:07  0:02 0:09  0:02 0:11  0:02 orb B Results.Structural parameters obtained from the o - ( =f:u:) 3:27  0:3 3:59  0:3 3:81  0:3 resonant XRR measurements are displayed in Tab. I. 8 0 weighting the individual spectra with respect to the 2 5 n m F e O d a t a f i t 3 4 cation stoichometry given in Fig. 1(c), the XAS and 6 0 XMCD data can be described well by our model (cf. grey 4 0 line). Thus, the cation distribution on di erent sites al- most follows the ideal stoichiometry of 1:1:1. 2 0 One feature of this kind of modelling is the fact that i 0 7 1 0 . 2 e V each of the three extrema observed in the XMCD spec- 0 7 0 9 . 5 e V trum can mainly be attributed to one cation spectrum. cation Table II shows the contributions r (E ) of each cation - 2 0 spectrum at the resonant energies E in the XMCD spec- s 2 0 trum, according to 7 0 8 . 4 e V cation I (E ) cation - 2 0 r (E ) = ; 2+ 3+ 3+ 0 . 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 0 . 6 Fe Fe Fe oct tet oct jI (E )j +jI (E )j +jI (E )j i i i s c a t t e r i n g v e c t o r q ( 1 / Å ) (1) cation with I (E ) being the XMCD signal of the corre- FIG. 2. XRMR data (open circles) and corresponding ts sponding cation spectrum in Fig. 1(b) at energies E = (solid lines) from the 25 nm Fe O lm, recorded at the 3 4 708:4 eV; 709:5 eV; 710:2 eV. While there still is a con- three resonant energies of the XMCD L edge, using the siderable mixing, at least 64% of each extremum can be modeled magnetooptical depth pro les of Fig. 3(a). Data attributed to its dominant cation. were recorded with a magnetic eld of 200 mT along the 2+ 3+ Fe O (001) direction at room temperature. The distinction between the Fe and Fe ions is a 3 4 oct oct noteworthy issue for several reasons. Above the Ver- wey transition temperature, these ions should be ran- two samples can be found in the Supplemental Mate- domly distributed on octahedral sites as shown by neu- rial [25] (Chapter B). The magnetooptical depth pro les tron di raction [36]. Thus, the model could potentially 2:5+ (z) that produce the ts are shown in Fig. 3(a). The be simpli ed by describing them as Fe , e ectively re- oct most striking feature of all three samples is the behavior ducing the number of cation species. However, only mul- 2+ at the surface: at the Fe resonance energy (708:4 eV, tiplet calculations based on a three-cation model describe oct room-temperature spectra from XPS, XAS and XMCD green), the magnetooptical depth pro les in fact appear at the Fe L absorption edge with sucient agreement to be just homogeneous for all samples. However, at both 3+ 3+ [17, 37{40]. Note that this extremum-cation assignment the Fe and the Fe resonance energies, there are no- tet oct ticeable changes to the  depth pro les. In order to t qualitatively also holds true in LSDA+U calculations, their asymmetry ratios, we must include a thin surface but they predict a stronger overlap of the three cation layer of enhanced magnetooptical absorption. The two spectra [41]. A possible explanation for this behavior thinner Fe O lms - 13 nm and 25 nm - are quantita- may well be the di erent electronic structure at the sur- 3 4 3+ 3+ tively very similar for the Fe and Fe magnetoopti- face of magnetite. In that case, we would expect di erent oct tet cal depth pro les, with only minor di erences in the en- spectra from the surface and the bulk. In fact, a re- hanced amplitude at the surface, and the  in the bulk cent study using hard x-ray photoelectron spectroscopy matches between the samples. In contrast, the 50 nm (HAXPES) reported on a bulk-exclusive state not ob- 3+ Fe O lm shows slightly higher  at the Fe res- servable in surface-sensitive soft XPS [42]. 3 4 oct 3+ onance, and smaller  at the Fe resonance. Also, Accordingly, the strategy to obtain cationic depth pro- tet 2+ the magnetooptical absorption at the Fe resonance be- les is to pick the three corresponding XMCD resonant oct comes larger with increasing lm thickness, in agreement energies and perform XRMR measurements at these res- with our results from the sum-rule analysis (see Tab. onances. Figure 2 shows the asymmetry ratios and I). The obvious choice of magnetooptical depth pro les their ts at the three XMCD resonant energies for the which are simply homogeneous through the entire lm 25 nm Fe O lm. Corresponding gures for the other 3 4 did not provide satisfactory ts to the data. This is dis- cussed in more detail in the Supplemental Material [25] (Chapter C). TABLE II. Contributions of the three cation species to the In order to highlight this phenomenon, Fig. 3(b) shows extrema in the XMCD spectrum in Fig. 1(b), as obtained by the surface region of the 25 nm Fe O lm, together with the multiplet analysis using Eq. (1). 3 4 the density depth pro le obtained from o -resonant XRR 2+ 3+ 3+ Energy Fe Fe Fe oct tet oct (grey line). The edge of the magnetooptical depth pro- 708:4 eV 73  5% 8  3% 19  5% 2+ le of the Fe resonance roughly matches the location of oct 709:5 eV 18  3% 64  3% 18  3% 3+ the magnetically enhanced Fe layer. The thickness of 710:2 eV 4  3% 16  8% 80  10% tet the magnetically enhanced layers is about 3:5 A for both Δ 4 (a) enh TABLE III. Thicknesses d and vertical shifts of the en- 3+ 3+ hanced magnetic layers at the Fe and Fe resonances. tet oct enh enh d d vertical shift 3+ 3+ Fe Fe oct tet 13 nm 4:7 A 3:2 A 2 A 25 nm 3:5 A 3:4 A 4:5 A 50 nm 3:8 A 6:4 A 2:5 A enh ples, both the thicknesses d and the depth o sets are similar in magnitude to the surface roughness, making it dicult to resolve the exact distances with any greater precision. (b) Discussion.Within the three-cation picture, we can 4.5 Å discuss the magnetooptical depth pro les obtained for the resonant energies. The magnetooptical depth pro- les are not identical with the depth distribution of the cations: As quanti ed in Tab. II, however, the signal 3.5 Å on each resonance is a mixture of contributions from all three cations. For the magnetooptical depth pro le at 710:2 eV approximately 80% of the signal originates from 3+ the Fe and can be regarded as an almost pure e ect oct from that species. And since the position of the layer of enhanced magnetization at 709:5 eV does not match the position of the 710:2 eV layer, we can conclude it to 3.5 Å 3+ be a distinct physical feature, stemming from the Fe tet species. One ansatz is to take into account rearranged cation (c) 2+ Fe oct distributions due to the Fe O (001) surface as proposed 3 4 3+ Fe by the SCV model [13, 14]. The SCV model predicts oct 3+ that, in order to achieve polarity compensation, the rst Fe tet 3+ 3+ unit cell contains only Fe species, with the rst Fe 2- tet 3+ (001) layer lying about 1 A deeper than the Fe . This model oct matches surprisingly well some aspects of our ndings. FIG. 3. (a)  (z) depth pro les for all three samples at the The rst Fe -O layer remains stoichometric, but the oct 2+ 3+ three resonant energies, extracted from the XRMR ts. (b) Fe changes valency to Fe , e ectively doubling the oct oct Close-up of the surface magnetooptical depth pro le of the 3+ 3+ Fe density. In the second layer, an additional Fe oct tet 25 nm Fe O lm, together with the optical density obtained 3 4 3+ ion is added, increasing the Fe density by 50%. How- tet from o -resonant XRR ts (grey line). (c) (Bulk-terminated) 2+ ever, we do not observe the depletion of Fe cations oct model of the magnetite unit cell, in scale with Fig. 3(b). in the rst 8:4 A. This agreement is surprising because Comparison with the model in (c) illustrates the sizes of the enhanced regions being roughly half a unit cell of magnetite Fe O surfaces tend to hydroxilate on ambient conditions 3 4 p p (four cation layers). and do not show the ( 2 2)R45 LEED pattern, but instead a (1  1) pattern [14]. This, however, may be attributed to disorder at the surface with loss of long- 3+ range order while the local order of vacancies and inter- Fe species. This corresponds to slightly less than half stitials is kept. Our results now suggest that at least the a bulk unit cell of magnetite (a=2 = 4:2 A), as illustrated 3+ Fe -enrichment of the surface remains intact under am- by Fig. 3(c). Furthermore, the magnetically enhanced bient conditions. A more detailed comparison of the SCV layers are not colocated at the same depth: the magnet- 3+ model to our ndings can be found in the Supplemental ically enhanced Fe layer is shifted about 4:5 A deeper tet 3+ Material [25] (Chapter D). into the lm than the magnetically enhanced Fe layer. oct Table III summarizes the individual thicknesses of the Taking the model of occupation of octahedral sites 2:5+ enh magnetically enhanced layers d and their o set from by Fe cations as an alternative, the agreement of oct one another for the di erent samples. All three samples the XAS/XMCD spectra with the multiplet calculations show comparable, but not quite identical results. While might be merely valid at the surface, while both octahe- the model qualitatively holds up well among the sam- drally coordinated Fe species are identical in the bulk. In (010) 5 that case, the discrepancy between surface and bulk of the magnetooptical depth pro les would represent the transition from the surface electronic structure to the [email protected] bulk structure. Using bulk-sensitive HAXPES, Taguchi [email protected] et al. report on a bulk feature at 708:5 eV, which is in- [email protected] visible for surface-sensitive soft XPS [42]. In this picture, [1] J.-B. Moussy, Journal of Physics D: Applied Physics 46, we could interpret the magnetooptical depth pro les and 143001 (2013). [2] R. Ramos, T. Kikkawa, K. Uchida, H. Adachi, I. Lucas, the XMCD spectra as follows: the top 4 6 A, consisting 3+ 3+ M. H. Aguirre, P. Algarabel, L. Morelln, S. Maekawa, of Fe and Fe ions, give rise to the XMCD extrema oct tet E. Saitoh, and M. R. Ibarra, Applied Physics Letters at 709:5 eV and 710:2 eV, as re ected in the enhanced 102, 072413 (2013). layers at those energies. Deeper into the Fe O lm, 3 4 [3] M. Lin and H. Leu, Electroanalysis 17, 2068 (2005). 3+ 2:5+ the Fe species vanishes in favor of the Fe species, oct oct [4] T. Kado, Applied Physics Letters 92, 092502 (2008). and consequently, the magnetic dichroism at 710:2 eV [5] L. Marnitz, K. Rott, S. Nieh orster, C. Klewe, D. Meier, S. Fabretti, M. Witziok, A. Krampf, O. Kuschel, is reduced. Instead, the bulk-feature at 708:4 eV be- 2:5+ T. Schemme, K. Kuepper, J. Wollschl ager, A. Thomas, comes prominent, related to the Fe species. In the oct 2+ G. Reiss, and T. Kuschel, AIP Advances 5, 047103 XMCD signal, it is visible as the supposed Fe peak, but oct (2015). strongly suppressed due to the surface sensitivity of the [6] Z. Zhang and S. Satpathy, Phys. Rev. B 44, 13319 (1991). TEY detection. In this interpretation, our results agree [7] G. Hu and Y. Suzuki, Phys. Rev. Lett. 89, 276601 (2002). with the more pronounced signal at 708:5 eV far from the [8] Y. S. Dedkov, U. Rudiger,  and G. Gun  therodt, Phys. surface reported in Ref. [42], and can explain the obser- Rev. B 65, 064417 (2002). [9] J. G. Tobin, S. A. Morton, S. W. Yu, G. D. Waddill, vation of the two distinct Fe species at the surface, oct I. K. Schuller, and S. A. Chambers, Journal of Physics: which should not be distinguishable at room tempera- Condensed Matter 19, 315218 (2007). ture. [10] M. Fonin, Y. S. Dedkov, R. Pentcheva, U. Rudiger,  and G. Gun  therodt, Journal of Physics: Condensed Matter 20, 142201 (2008). Summary.In conclusion, we fabricated three [11] W. Wang, J.-M. Mariot, M. C. Richter, O. Heck- Fe O =MgO(001) ultrathin lms of di erent thicknesses mann, W. Ndiaye, P. De Padova, A. Taleb-Ibrahimi, 3 4 P. Le F evre, F. Bertran, F. Bondino, E. Magnano, by RMBE. We recorded XAS/XMCD at ALS beamline J. Krempasky,  P. Blaha, C. Cacho, F. Parmigiani, and 4.0.2 as well as XAS/XMCD and XRMR measurements K. Hricovini, Phys. Rev. B 87, 085118 (2013). at BESSY II beamline UE46 PGM-1, in order to obtain [12] S. Bruc  k, M. Paul, H. Tian, A. Muller,  D. Kufer, C. Prae- magnetooptical depth pro les. By tting multiplet torius, K. Fauth, P. Audehm, E. Goering, J. Verbeeck, calculations to the XMCD data, we disentangle the G. Van Tendeloo, M. Sing, and R. Claessen, Applied cation contributions at the three resonant energies of Physics Letters 100, 081603 (2012). the XMCD spectrum, and use XRMR at those energies [13] R. Bliem, E. McDermott, and G. S. Parkinson, Science 346 (2014). in order to resolve the magnetooptical depth pro les of 3+ [14] G. S. Parkinson, Surface Science Reports 71, 272 (2016). the three iron species in Fe O . We nd that both Fe 3 4 [15] S. Chambers and S. Joyce, Surface Science 420, 111 species show an enhanced signal in the surface-near (1999). 3+ region in a  3:9 1 A thick layer, with the Fe layer tet [16] L. Mart n-Garc a, R. Gargallo-Caballero, M. Monti, 3+ located about 3  1:5 A underneath the Fe layer. oct M. Foerster, J. F. Marco, L. Aballe, and J. de la Figuera, We attribute this to the rst unit cell from the surface Phys. Rev. B 91, 020408 (2015). 3+ containing an excess of Fe cations. [17] K. Kuepper, O. Kuschel, N. Path e, T. Schemme, J. Schmalhorst, A. Thomas, E. Arenholz, M. Gor- goi, R. Ovsyannikov, S. Bartkowski, G. Reiss, and J. Wollschl ager, Phys. Rev. B 94, 024401 (2016). Acknowledgements.Financial support from the Bun- [18] O. Kuschel, R. Bu, W. Spiess, T. Schemme, desministerium fur  Bildung und Forschung (FKZ J. W ollermann, K. Balinski, A. T. N'Diaye, T. Kuschel, 05K16MP1) is gratefully acknowledged. We are also J. Wollschl ager, and K. Kuepper, Phys. Rev. B 94, grateful for the kind support from the Deutsche 094423 (2016). [19] C. Piamonteze, P. Miedema, and F. M. F. de Groot, Forschungsgemeinschaft (DFG under No. KU2321/6-1, Phys. Rev. B 80, 184410 (2009). and No. WO533/20-1). This research used resources [20] C. T. Chen, Y. U. Idzerda, H.-J. Lin, N. V. Smith, of the Advanced Light Source, a DOE Oce of Science G. Meigs, E. Chaban, G. H. Ho, E. Pellegrin, and User Facility under contract no. DE-AC02-05CH11231, F. Sette, Phys. Rev. Lett. 75, 152 (1995). by recording XMCD at at beamline 4.0.2. (ALS-10261). [21] Y. Teramura, A. Tanaka, and T. Jo, Journal of the Phys- We thank HZB for the allocation of synchrotron beam- ical Society of Japan 65, 1053 (1996). time at beamline UE46 PGM-1 (181/06266ST/R) where [22] G. van der Laan, Journal of Electron Spectroscopy and Related Phenomena 86, 41 (1997). we recorded the XRMR and additional XMCD measure- [23] F. de Groot, Coordination Chemistry Reviews 249, 31 ments. 6 (2005). Bataille, M. Gautier-Soyer, M. Viret, C. Gatel, P. Bayle- [24] E. Stavitski and F. M. de Groot, Micron 41, 687 (2010). Guillemaud, and E. Snoeck, Journal of Applied Physics [25] See Supplemental Material for details on the multiplet 100, 103902 (2006). calculations, as well as those XMCD, XRR, and XRMR [35] J.-B. Moussy, S. Gota, A. Bataille, M.-J. Guittet, data which were not included in the main text. It also fur- M. Gautier-Soyer, F. Delille, B. Dieny, F. Ott, T. D. ther illustrates the necessity to include the layers of en- Doan, P. Warin, P. Bayle-Guillemaud, C. Gatel, and hanced magnetization in order to properly t the XRMR E. Snoeck, Phys. Rev. B 70, 174448 (2004). data, and contains a thorough comparison of our model [36] J. Garc a, G. Sub as, M. G. Proietti, H. Renevier, Y. Joly, to the SCV model. J. L. Hodeau, J. Blasco, M. C. S anchez, and J. F. B erar, [26] T. J. Regan, H. Ohldag, C. Stamm, F. Nolting, J. Luning,  Phys. Rev. Lett. 85, 578 (2000). J. St ohr, and R. L. White, Phys. Rev. B 64, 214422 [37] F. M. F. de Groot, J. C. Fuggle, B. T. Thole, and G. A. (2001). Sawatzky, Phys. Rev. B 42, 5459 (1990). [27] C. T. Chantler, J. Phys. Chem Ref. Data 24, 71 (1995). [38] P. Kuiper, B. G. Searle, L.-C. Duda, R. M. Wolf, and [28] E. Weschke and E. Schierle, Journal of large-scale re- P. J. van der Zaag, Journal of Electron Spectroscopy and search facilities 4, A127 (2018). Related Phenomena 86, 107 (1997). [29] S. Macke and E. Goering, Journal of Physics: Condensed [39] P. S. Miedema and F. M. de Groot, Journal of Electron Matter 26, 363201 (2014). Spectroscopy and Related Phenomena 187, 32 (2013). [30] T. Kuschel, C. Klewe, J.-M. Schmalhorst, F. Bertram, [40] D. Phase, G. Panchal, R. Rawat, S. Tiwari, R. Prakash, O. Kuschel, T. Schemme, J. Wollschl ager, S. Francoual, D. Jain, and R. Choudhary, Journal of Magnetism and J. Strempfer, A. Gupta, M. Meinert, G. G otz, D. Meier, Magnetic Materials 482, 296 (2019). and G. Reiss, Phys. Rev. Lett. 115, 097401 (2015). [41] V. N. Antonov, B. N. Harmon, and A. N. Yaresko, Phys. [31] C. Klewe, T. Kuschel, J.-M. Schmalhorst, F. Bertram, Rev. B 67, 024417 (2003). O. Kuschel, J. Wollschl ager, J. Strempfer, M. Meinert, [42] M. Taguchi, A. Chainani, S. Ueda, M. Matsunami, and G. Reiss, Phys. Rev. B 93, 214440 (2016). Y. Ishida, R. Eguchi, S. Tsuda, Y. Takata, M. Yabashi, [32] Weiss, Pierre and Forrer, R., Ann. Phys. 10, 279 (1929). K. Tamasaku, Y. Nishino, T. Ishikawa, H. Daimon, [33] S. Celotto, W. Eerenstein, and T. Hibma, European S. Todo, H. Tanaka, M. Oura, Y. Senba, H. Ohashi, and Physical Journal B 36, 271 (2003). S. Shin, Phys. Rev. Lett. 115, 256405 (2015). [34] A. V. Ramos, J.-B. Moussy, M.-J. Guittet, A. M. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Cation- and lattice-site-selective magnetic depth profiles of ultrathin $\mathrm{Fe_3O_4}$(001) films

Download PDF
6 pages

Loading next page...
 
/lp/arxiv-cornell-university/cation-and-lattice-site-selective-magnetic-depth-profiles-of-ultrathin-Ug0FvK2Mrx

References (14)

ISSN
2469-9950
eISSN
ARCH-3331
DOI
10.1103/PhysRevB.102.220411
Publisher site
See Article on Publisher Site

Abstract

Cation- and lattice-site-selective magnetic depth pro les of ultrathin Fe O (001) lms 3 4 1, 2, 3 1 1 1 Tobias Pohlmann, Timo Kuschel, Jari Rodewald, Jannis Thien, Kevin Ruwisch, Florian 2 4 5 1,y 1,z Bertram, Eugen Weschke, Padraic Shafer, Joachim Wollschl ager, and Karsten Kupp  er Department of Physics, Osnabruck  University, Barbarastr. 7, 49076 Osnabruck,  Germany DESY Photon Science, Notkestr. 85, 22607 Hamburg, Germany Center for Spinelectronic Materials and Devices, Department of Physics, Bielefeld University, Universitatsstr. 25, 33615 Bielefeld, Germany Helmholtz-Zentrum Berlin fur  Materialien und Energie, Wilhelm-Conrad-Rontgen-Campus BESSY II, Albert-Einstein-Strasse 15, 12489 Berlin, Germany Advanced Light Source, Lawrence Berkeley National Laboratory, 6 Cyclotron Rd, Berkeley, CA, 94720, USA (Dated: May 5, 2020) A detailed understanding of ultrathin lm surface properties is crucial for the proper interpretation of spectroscopic, catalytic and spin-transport data. We present x-ray magnetic circular dichroism (XMCD) and x-ray resonant magnetic re ectivity (XRMR) measurements on ultrathin Fe O lms 3 4 to obtain magnetic depth pro les for the three resonant energies corresponding to the di erent 2+ 3+ 3+ cation species Fe , Fe and Fe located on octahedral and tetrahedral sites of the inverse spinel oct tet oct structure of Fe O . By analyzing the XMCD spectrum of Fe O using multiplet calculations, 3 4 3 4 the resonance energy of each cation species can be isolated. Performing XRMR on these three resonant energies yields magnetic depth pro les that correspond each to one speci c cation species. 3+ The depth pro les of both kinds of Fe cations reveal a 3:9  1 A-thick surface layer of enhanced 2+ magnetization, which is likely due to an excess of these ions at the expense of the Fe species in oct 3+ the surface region. The magnetically enhanced Fe layer is additionally shifted about 3  1:5 A tet 3+ farther from the surface than the Fe layer. oct Introduction.Magnetite, Fe O , is one of the most (XPS), to explain the behavior of the bulk material [14]. 3 4 frequently investigated transition-metal oxides, since it is In particular, drawing conclusions about the cation dis- a key material in spintronics [1], spin caloritronics [2] and tribution of magnetite requires caution, because the bulk material chemistry [3]. Fe O thin lms were considered material of the inverse spinel Fe O should contain di- 3 4 3 4 2+ highly suitable as electrode material for magnetic tunnel valent Fe , as well as trivalent ions in both octahedral oct 3+ 3+ junctions [4, 5] due to their predicted half-metallic behav- and tetrahedral coordination, Fe and Fe . However, tet oct ior with 100% spin polarization [6]. However, the promise the DFT+U calculations of the SCV structure predict 3+ was never quite met, with modest tunnel magnetoresis- the rst four atomic layers to only contain Fe ions and tance ratios ranging from -26% to 18% [4, 5, 7]. In or- to have a formal stoichometry of Fe O , in agreement 11 16 3+ der to test its half-metallicity, spin-resolved x-ray photo- with earlier reports on an excess of Fe at the (001) electron spectroscopy on Fe O (111) lms found a spin- surface [15]. A subsequent study of how the magnetic 3 4 polarization of about 80% [8], while on Fe O (001) the properties of Fe O are a ected at the surface used x-ray 3 4 3 4 same technique yielded polarizations of only  40%70% magnetic circular dichroism (XMCD) with varying prob- [9{11]. ing depth and found a reduced magnetic moment at the surface of a natural magnetite crystal [16]. Both the reduced tunnel magnetoresistance and the In this letter we report an investigation of the mag- deviations from 100% spin-polarization in spin-resolved netic surface properties of ultrathin Fe O (001) lms, in XPS were argued to emerge from interface and surface 3 4 contrast to the bulk, by recording magnetooptical depth e ects, respectively [4, 5, 8, 10, 11]. For Fe O (111) 3 4 pro les of the three cation species in Fe O . Using pho- lms deposited on semiconducting ZnO(0001), lattice- 3 4 tons with resonant energies, we employed XRMR to de- site-selective depth pro les obtained by x-ray resonant termine the magnetooptical depth pro les at the energies magnetic re ectivity (XRMR) and electron energy loss 2+ 3+ 3+ characteristic for Fe , Fe and Fe in magnetite's spectroscopy did not nd a notable surface modi ca- oct tet oct tion apart from a Fe termination [12]. Reduction of L XMCD spectrum individually for three ultrathin 2;3 oct the spin-polarization measured at the Fe O (001) sur- Fe O =MgO(001) lms of varying thicknesses. We nd a 3 4 3 4 face was typically considered to originate from a sur-  3:9 A layer of enhanced magnetooptical absorption at 3+ face reconstruction, the existence of which has long been the surface at the resonant energies of both Fe species 2+ 3+ known but only recently has been resolved as a subsur- but not for Fe , suggesting an Fe -rich surface. face cation vacancy (SCV) structure [13]. This revela- Experimental and theoretical details.We prepared tion highlights the issues that might arise from general- the Fe O =MgO(001) samples in a multichamber ultra- 3 4 izing results from surface-sensitive techniques, such as high-vacuum system using reactive molecular beam x-ray absorption spectroscopy (XAS) in total electron epitaxy (RMBE). Their chemical composition and p p yield (TEY) mode and x-ray photoelectron spectroscopy ( 2 2)R45 superstructure was con rmed by in-situ arXiv:2005.01657v1 [cond-mat.mes-hall] 4 May 2020 2 4 0 XPS and low-energy electron di raction (LEED), respec- d a t a 2 5 n m F e O ( c ) ( a ) 1 . 0 3 4 s i m u l a t i o n tively. For details on the deposition and characterization i 3 5 % 2 + n 3 5 F e o c t B = 4 T u 3 3 % methods please see Refs. [17] or [18]. . 3 + F e 3 1 % t e t r 0 . 5 3 + 3 0 For the XAS and XMCD study, the samples were a F e o c t % transfered from our lab under ambient conditions to the y 2 5 t 0 . 0 Superconducting Vector Magnet Endstation at beamline 4.0.2 of the Advanced Light Source (ALS). All samples 2 0 ( b ) were measured in a magnetic eld of 4 T along the x-ray t 0 . 0 1 5 beam at room temperature. The incidence angle of the x-rays was 30 from the [100] direction of Fe O , and r 3 4 1 0 ( - 0 . 2 the degree of circular polarization was 90%. The XAS and XMCD spectra were measured across the Fe L 2;3 - 0 . 4 absorption edges (690 750 eV). All XAS spectra were X 2 + 3 + 3 + 7 0 0 7 1 0 7 2 0 7 3 0 F e F e F e measured in the TEY mode, which has a probing depth o c t t e t o c t e n e r g y ( e V ) in magnetite of about 3 nm. The XAS and XMCD data were analyzed by apply- FIG. 1. (a) XAS and (b) XMCD spectrum at the Fe L edge 2;3 ing the sum rules [19{21] and charge-transfer multiplet for the 25 nm Fe O lm, taken at 4 T external magnetic eld, 3 4 calculations using the Thole code [22] with assistance of at room temperature and in TEY mode. A step function was CTM4XAS [23, 24]. For the sum rules, we took into ac- subtracted from the XAS spectrum. Black dots represent count a correction factor of 1.142 derived by Teramura data; green, red and blue spectra are multiplet calculations 2+ holes for the three cation species of Fe O , the grey line is their 3 4 et al. for Fe [21] and assumed 14 . For the mul- f:u: sum. The cation spectra are o set for better visibility. (c) tiplet calculations, we assumed the three-cation model, Cation stoichometry used to obtain the t in (a) and (b). using crystal eld and charge-transfer parameters as de- scribed in Ref. [17]. The parameters and more details regarding the multiplet calculations can also be found in Figure 1 shows the XAS and XMCD spectra of the 25 nm the Supplemental Materials [25] (Chapter A, including Fe O lm, recorded at ALS. Corresponding data mea- 3 4 Refs. [17, 26, 27]). The multiplet states resulting from sured at BESSY II, under the same conditions in which these calculations were compared to the data by assum- the XRMR was performed, can be found in the Supple- ing a Gaussian instrumental broadening of 0:2 eV, and a mental Material [25] (Chapter A). The spin and orbital Lorentzian lifetime broadening of 0:3 eV at L and 0:6 eV moments  and  obtained from the sum rules are spin orb at L . given in Tab. I. Their sum is slightly reduced compared The samples were transfered to BESSY II under ambi- to the bulk value of magnetite of  =  +  = spin orb ent conditions and x-ray re ectivity (XRR) and XRMR 4:07  [32]. They also show a tendency toward a slightly were performed in the XUV Di ractometer at beamline higher moment for the thicker lms. This behavior of UE46 PGM-1 [28]. The samples were placed between magnetite lms has also been observed previously [17], two permanent magnets in a magnetic eld of 200 mT and can be explained by a higher density of anti-phase at room temperature, with a degree of circular polariza- boundaries (APBs) for thinner lms due to the antifer- tion of 90%. First, we characterized the structural prop- romagnetic coupling of APBs reducing the average mag- erties of the sample (thickness d, roughness ) by XRR netic moment of the lm [33{35]. at o -resonant energies (680 eV, 1000 eV). Second, XAS Additionally, multiplet simulations of the three-cation and XMCD were measured in order to select suitable model are tted to the XMCD data (cf Fig. 1). By energies for XRMR. Finally,  2 scans in the range 2 = 0 140 at resonant energies E with extrema in the XMCD signal (maximum at 708:4 eV, minimum surf at 709:5 eV, maximum at 710:2 eV) were performed with TABLE I. Thicknesses d, surface roughnesses  and inter- int face roughnesses  of the investigated samples as obtained both right and left circularly polarized x-rays, to obtain by tting o -resonant XRR curves, as well as their spin mo- the XRMR asymmetry ratios I (E ; q ). These curves i z ments  and orbital moments  from the sum-rule anal- spin orb were then tted with the Zak matrix formalism using the ysis. software ReMagX [29] to determine the depth pro les sample 13 nm 25 nm 50 nm of the magnetooptical absorption  (z) and dispersion d (nm) 13:5  0:5 25:2  0:3 52:8  0:3 (z) along the lm height z. A detailed review of the surf (nm) 0:22  0:05 0:33  0:05 0:34  0:05 XRMR method and the software is given in Ref. [29], int (nm) 0:35  0:05 0:35  0:05 0:37  0:05 and a conclusive recipe for tting XRMR data can be ( =f:u:) 3:2  0:3 3:5  0:3 3:7  0:3 spin B found in Refs. [30, 31]. ( =f:u:) 0:07  0:02 0:09  0:02 0:11  0:02 orb B Results.Structural parameters obtained from the o - ( =f:u:) 3:27  0:3 3:59  0:3 3:81  0:3 resonant XRR measurements are displayed in Tab. I. 8 0 weighting the individual spectra with respect to the 2 5 n m F e O d a t a f i t 3 4 cation stoichometry given in Fig. 1(c), the XAS and 6 0 XMCD data can be described well by our model (cf. grey 4 0 line). Thus, the cation distribution on di erent sites al- most follows the ideal stoichiometry of 1:1:1. 2 0 One feature of this kind of modelling is the fact that i 0 7 1 0 . 2 e V each of the three extrema observed in the XMCD spec- 0 7 0 9 . 5 e V trum can mainly be attributed to one cation spectrum. cation Table II shows the contributions r (E ) of each cation - 2 0 spectrum at the resonant energies E in the XMCD spec- s 2 0 trum, according to 7 0 8 . 4 e V cation I (E ) cation - 2 0 r (E ) = ; 2+ 3+ 3+ 0 . 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 0 . 6 Fe Fe Fe oct tet oct jI (E )j +jI (E )j +jI (E )j i i i s c a t t e r i n g v e c t o r q ( 1 / Å ) (1) cation with I (E ) being the XMCD signal of the corre- FIG. 2. XRMR data (open circles) and corresponding ts sponding cation spectrum in Fig. 1(b) at energies E = (solid lines) from the 25 nm Fe O lm, recorded at the 3 4 708:4 eV; 709:5 eV; 710:2 eV. While there still is a con- three resonant energies of the XMCD L edge, using the siderable mixing, at least 64% of each extremum can be modeled magnetooptical depth pro les of Fig. 3(a). Data attributed to its dominant cation. were recorded with a magnetic eld of 200 mT along the 2+ 3+ Fe O (001) direction at room temperature. The distinction between the Fe and Fe ions is a 3 4 oct oct noteworthy issue for several reasons. Above the Ver- wey transition temperature, these ions should be ran- two samples can be found in the Supplemental Mate- domly distributed on octahedral sites as shown by neu- rial [25] (Chapter B). The magnetooptical depth pro les tron di raction [36]. Thus, the model could potentially 2:5+ (z) that produce the ts are shown in Fig. 3(a). The be simpli ed by describing them as Fe , e ectively re- oct most striking feature of all three samples is the behavior ducing the number of cation species. However, only mul- 2+ at the surface: at the Fe resonance energy (708:4 eV, tiplet calculations based on a three-cation model describe oct room-temperature spectra from XPS, XAS and XMCD green), the magnetooptical depth pro les in fact appear at the Fe L absorption edge with sucient agreement to be just homogeneous for all samples. However, at both 3+ 3+ [17, 37{40]. Note that this extremum-cation assignment the Fe and the Fe resonance energies, there are no- tet oct ticeable changes to the  depth pro les. In order to t qualitatively also holds true in LSDA+U calculations, their asymmetry ratios, we must include a thin surface but they predict a stronger overlap of the three cation layer of enhanced magnetooptical absorption. The two spectra [41]. A possible explanation for this behavior thinner Fe O lms - 13 nm and 25 nm - are quantita- may well be the di erent electronic structure at the sur- 3 4 3+ 3+ tively very similar for the Fe and Fe magnetoopti- face of magnetite. In that case, we would expect di erent oct tet cal depth pro les, with only minor di erences in the en- spectra from the surface and the bulk. In fact, a re- hanced amplitude at the surface, and the  in the bulk cent study using hard x-ray photoelectron spectroscopy matches between the samples. In contrast, the 50 nm (HAXPES) reported on a bulk-exclusive state not ob- 3+ Fe O lm shows slightly higher  at the Fe res- servable in surface-sensitive soft XPS [42]. 3 4 oct 3+ onance, and smaller  at the Fe resonance. Also, Accordingly, the strategy to obtain cationic depth pro- tet 2+ the magnetooptical absorption at the Fe resonance be- les is to pick the three corresponding XMCD resonant oct comes larger with increasing lm thickness, in agreement energies and perform XRMR measurements at these res- with our results from the sum-rule analysis (see Tab. onances. Figure 2 shows the asymmetry ratios and I). The obvious choice of magnetooptical depth pro les their ts at the three XMCD resonant energies for the which are simply homogeneous through the entire lm 25 nm Fe O lm. Corresponding gures for the other 3 4 did not provide satisfactory ts to the data. This is dis- cussed in more detail in the Supplemental Material [25] (Chapter C). TABLE II. Contributions of the three cation species to the In order to highlight this phenomenon, Fig. 3(b) shows extrema in the XMCD spectrum in Fig. 1(b), as obtained by the surface region of the 25 nm Fe O lm, together with the multiplet analysis using Eq. (1). 3 4 the density depth pro le obtained from o -resonant XRR 2+ 3+ 3+ Energy Fe Fe Fe oct tet oct (grey line). The edge of the magnetooptical depth pro- 708:4 eV 73  5% 8  3% 19  5% 2+ le of the Fe resonance roughly matches the location of oct 709:5 eV 18  3% 64  3% 18  3% 3+ the magnetically enhanced Fe layer. The thickness of 710:2 eV 4  3% 16  8% 80  10% tet the magnetically enhanced layers is about 3:5 A for both Δ 4 (a) enh TABLE III. Thicknesses d and vertical shifts of the en- 3+ 3+ hanced magnetic layers at the Fe and Fe resonances. tet oct enh enh d d vertical shift 3+ 3+ Fe Fe oct tet 13 nm 4:7 A 3:2 A 2 A 25 nm 3:5 A 3:4 A 4:5 A 50 nm 3:8 A 6:4 A 2:5 A enh ples, both the thicknesses d and the depth o sets are similar in magnitude to the surface roughness, making it dicult to resolve the exact distances with any greater precision. (b) Discussion.Within the three-cation picture, we can 4.5 Å discuss the magnetooptical depth pro les obtained for the resonant energies. The magnetooptical depth pro- les are not identical with the depth distribution of the cations: As quanti ed in Tab. II, however, the signal 3.5 Å on each resonance is a mixture of contributions from all three cations. For the magnetooptical depth pro le at 710:2 eV approximately 80% of the signal originates from 3+ the Fe and can be regarded as an almost pure e ect oct from that species. And since the position of the layer of enhanced magnetization at 709:5 eV does not match the position of the 710:2 eV layer, we can conclude it to 3.5 Å 3+ be a distinct physical feature, stemming from the Fe tet species. One ansatz is to take into account rearranged cation (c) 2+ Fe oct distributions due to the Fe O (001) surface as proposed 3 4 3+ Fe by the SCV model [13, 14]. The SCV model predicts oct 3+ that, in order to achieve polarity compensation, the rst Fe tet 3+ 3+ unit cell contains only Fe species, with the rst Fe 2- tet 3+ (001) layer lying about 1 A deeper than the Fe . This model oct matches surprisingly well some aspects of our ndings. FIG. 3. (a)  (z) depth pro les for all three samples at the The rst Fe -O layer remains stoichometric, but the oct 2+ 3+ three resonant energies, extracted from the XRMR ts. (b) Fe changes valency to Fe , e ectively doubling the oct oct Close-up of the surface magnetooptical depth pro le of the 3+ 3+ Fe density. In the second layer, an additional Fe oct tet 25 nm Fe O lm, together with the optical density obtained 3 4 3+ ion is added, increasing the Fe density by 50%. How- tet from o -resonant XRR ts (grey line). (c) (Bulk-terminated) 2+ ever, we do not observe the depletion of Fe cations oct model of the magnetite unit cell, in scale with Fig. 3(b). in the rst 8:4 A. This agreement is surprising because Comparison with the model in (c) illustrates the sizes of the enhanced regions being roughly half a unit cell of magnetite Fe O surfaces tend to hydroxilate on ambient conditions 3 4 p p (four cation layers). and do not show the ( 2 2)R45 LEED pattern, but instead a (1  1) pattern [14]. This, however, may be attributed to disorder at the surface with loss of long- 3+ range order while the local order of vacancies and inter- Fe species. This corresponds to slightly less than half stitials is kept. Our results now suggest that at least the a bulk unit cell of magnetite (a=2 = 4:2 A), as illustrated 3+ Fe -enrichment of the surface remains intact under am- by Fig. 3(c). Furthermore, the magnetically enhanced bient conditions. A more detailed comparison of the SCV layers are not colocated at the same depth: the magnet- 3+ model to our ndings can be found in the Supplemental ically enhanced Fe layer is shifted about 4:5 A deeper tet 3+ Material [25] (Chapter D). into the lm than the magnetically enhanced Fe layer. oct Table III summarizes the individual thicknesses of the Taking the model of occupation of octahedral sites 2:5+ enh magnetically enhanced layers d and their o set from by Fe cations as an alternative, the agreement of oct one another for the di erent samples. All three samples the XAS/XMCD spectra with the multiplet calculations show comparable, but not quite identical results. While might be merely valid at the surface, while both octahe- the model qualitatively holds up well among the sam- drally coordinated Fe species are identical in the bulk. In (010) 5 that case, the discrepancy between surface and bulk of the magnetooptical depth pro les would represent the transition from the surface electronic structure to the [email protected] bulk structure. Using bulk-sensitive HAXPES, Taguchi [email protected] et al. report on a bulk feature at 708:5 eV, which is in- [email protected] visible for surface-sensitive soft XPS [42]. In this picture, [1] J.-B. Moussy, Journal of Physics D: Applied Physics 46, we could interpret the magnetooptical depth pro les and 143001 (2013). [2] R. Ramos, T. Kikkawa, K. Uchida, H. Adachi, I. Lucas, the XMCD spectra as follows: the top 4 6 A, consisting 3+ 3+ M. H. Aguirre, P. Algarabel, L. Morelln, S. Maekawa, of Fe and Fe ions, give rise to the XMCD extrema oct tet E. Saitoh, and M. R. Ibarra, Applied Physics Letters at 709:5 eV and 710:2 eV, as re ected in the enhanced 102, 072413 (2013). layers at those energies. Deeper into the Fe O lm, 3 4 [3] M. Lin and H. Leu, Electroanalysis 17, 2068 (2005). 3+ 2:5+ the Fe species vanishes in favor of the Fe species, oct oct [4] T. Kado, Applied Physics Letters 92, 092502 (2008). and consequently, the magnetic dichroism at 710:2 eV [5] L. Marnitz, K. Rott, S. Nieh orster, C. Klewe, D. Meier, S. Fabretti, M. Witziok, A. Krampf, O. Kuschel, is reduced. Instead, the bulk-feature at 708:4 eV be- 2:5+ T. Schemme, K. Kuepper, J. Wollschl ager, A. Thomas, comes prominent, related to the Fe species. In the oct 2+ G. Reiss, and T. Kuschel, AIP Advances 5, 047103 XMCD signal, it is visible as the supposed Fe peak, but oct (2015). strongly suppressed due to the surface sensitivity of the [6] Z. Zhang and S. Satpathy, Phys. Rev. B 44, 13319 (1991). TEY detection. In this interpretation, our results agree [7] G. Hu and Y. Suzuki, Phys. Rev. Lett. 89, 276601 (2002). with the more pronounced signal at 708:5 eV far from the [8] Y. S. Dedkov, U. Rudiger,  and G. Gun  therodt, Phys. surface reported in Ref. [42], and can explain the obser- Rev. B 65, 064417 (2002). [9] J. G. Tobin, S. A. Morton, S. W. Yu, G. D. Waddill, vation of the two distinct Fe species at the surface, oct I. K. Schuller, and S. A. Chambers, Journal of Physics: which should not be distinguishable at room tempera- Condensed Matter 19, 315218 (2007). ture. [10] M. Fonin, Y. S. Dedkov, R. Pentcheva, U. Rudiger,  and G. Gun  therodt, Journal of Physics: Condensed Matter 20, 142201 (2008). Summary.In conclusion, we fabricated three [11] W. Wang, J.-M. Mariot, M. C. Richter, O. Heck- Fe O =MgO(001) ultrathin lms of di erent thicknesses mann, W. Ndiaye, P. De Padova, A. Taleb-Ibrahimi, 3 4 P. Le F evre, F. Bertran, F. Bondino, E. Magnano, by RMBE. We recorded XAS/XMCD at ALS beamline J. Krempasky,  P. Blaha, C. Cacho, F. Parmigiani, and 4.0.2 as well as XAS/XMCD and XRMR measurements K. Hricovini, Phys. Rev. B 87, 085118 (2013). at BESSY II beamline UE46 PGM-1, in order to obtain [12] S. Bruc  k, M. Paul, H. Tian, A. Muller,  D. Kufer, C. Prae- magnetooptical depth pro les. By tting multiplet torius, K. Fauth, P. Audehm, E. Goering, J. Verbeeck, calculations to the XMCD data, we disentangle the G. Van Tendeloo, M. Sing, and R. Claessen, Applied cation contributions at the three resonant energies of Physics Letters 100, 081603 (2012). the XMCD spectrum, and use XRMR at those energies [13] R. Bliem, E. McDermott, and G. S. Parkinson, Science 346 (2014). in order to resolve the magnetooptical depth pro les of 3+ [14] G. S. Parkinson, Surface Science Reports 71, 272 (2016). the three iron species in Fe O . We nd that both Fe 3 4 [15] S. Chambers and S. Joyce, Surface Science 420, 111 species show an enhanced signal in the surface-near (1999). 3+ region in a  3:9 1 A thick layer, with the Fe layer tet [16] L. Mart n-Garc a, R. Gargallo-Caballero, M. Monti, 3+ located about 3  1:5 A underneath the Fe layer. oct M. Foerster, J. F. Marco, L. Aballe, and J. de la Figuera, We attribute this to the rst unit cell from the surface Phys. Rev. B 91, 020408 (2015). 3+ containing an excess of Fe cations. [17] K. Kuepper, O. Kuschel, N. Path e, T. Schemme, J. Schmalhorst, A. Thomas, E. Arenholz, M. Gor- goi, R. Ovsyannikov, S. Bartkowski, G. Reiss, and J. Wollschl ager, Phys. Rev. B 94, 024401 (2016). Acknowledgements.Financial support from the Bun- [18] O. Kuschel, R. Bu, W. Spiess, T. Schemme, desministerium fur  Bildung und Forschung (FKZ J. W ollermann, K. Balinski, A. T. N'Diaye, T. Kuschel, 05K16MP1) is gratefully acknowledged. We are also J. Wollschl ager, and K. Kuepper, Phys. Rev. B 94, grateful for the kind support from the Deutsche 094423 (2016). [19] C. Piamonteze, P. Miedema, and F. M. F. de Groot, Forschungsgemeinschaft (DFG under No. KU2321/6-1, Phys. Rev. B 80, 184410 (2009). and No. WO533/20-1). This research used resources [20] C. T. Chen, Y. U. Idzerda, H.-J. Lin, N. V. Smith, of the Advanced Light Source, a DOE Oce of Science G. Meigs, E. Chaban, G. H. Ho, E. Pellegrin, and User Facility under contract no. DE-AC02-05CH11231, F. Sette, Phys. Rev. Lett. 75, 152 (1995). by recording XMCD at at beamline 4.0.2. (ALS-10261). [21] Y. Teramura, A. Tanaka, and T. Jo, Journal of the Phys- We thank HZB for the allocation of synchrotron beam- ical Society of Japan 65, 1053 (1996). time at beamline UE46 PGM-1 (181/06266ST/R) where [22] G. van der Laan, Journal of Electron Spectroscopy and Related Phenomena 86, 41 (1997). we recorded the XRMR and additional XMCD measure- [23] F. de Groot, Coordination Chemistry Reviews 249, 31 ments. 6 (2005). Bataille, M. Gautier-Soyer, M. Viret, C. Gatel, P. Bayle- [24] E. Stavitski and F. M. de Groot, Micron 41, 687 (2010). Guillemaud, and E. Snoeck, Journal of Applied Physics [25] See Supplemental Material for details on the multiplet 100, 103902 (2006). calculations, as well as those XMCD, XRR, and XRMR [35] J.-B. Moussy, S. Gota, A. Bataille, M.-J. Guittet, data which were not included in the main text. It also fur- M. Gautier-Soyer, F. Delille, B. Dieny, F. Ott, T. D. ther illustrates the necessity to include the layers of en- Doan, P. Warin, P. Bayle-Guillemaud, C. Gatel, and hanced magnetization in order to properly t the XRMR E. Snoeck, Phys. Rev. B 70, 174448 (2004). data, and contains a thorough comparison of our model [36] J. Garc a, G. Sub as, M. G. Proietti, H. Renevier, Y. Joly, to the SCV model. J. L. Hodeau, J. Blasco, M. C. S anchez, and J. F. B erar, [26] T. J. Regan, H. Ohldag, C. Stamm, F. Nolting, J. Luning,  Phys. Rev. Lett. 85, 578 (2000). J. St ohr, and R. L. White, Phys. Rev. B 64, 214422 [37] F. M. F. de Groot, J. C. Fuggle, B. T. Thole, and G. A. (2001). Sawatzky, Phys. Rev. B 42, 5459 (1990). [27] C. T. Chantler, J. Phys. Chem Ref. Data 24, 71 (1995). [38] P. Kuiper, B. G. Searle, L.-C. Duda, R. M. Wolf, and [28] E. Weschke and E. Schierle, Journal of large-scale re- P. J. van der Zaag, Journal of Electron Spectroscopy and search facilities 4, A127 (2018). Related Phenomena 86, 107 (1997). [29] S. Macke and E. Goering, Journal of Physics: Condensed [39] P. S. Miedema and F. M. de Groot, Journal of Electron Matter 26, 363201 (2014). Spectroscopy and Related Phenomena 187, 32 (2013). [30] T. Kuschel, C. Klewe, J.-M. Schmalhorst, F. Bertram, [40] D. Phase, G. Panchal, R. Rawat, S. Tiwari, R. Prakash, O. Kuschel, T. Schemme, J. Wollschl ager, S. Francoual, D. Jain, and R. Choudhary, Journal of Magnetism and J. Strempfer, A. Gupta, M. Meinert, G. G otz, D. Meier, Magnetic Materials 482, 296 (2019). and G. Reiss, Phys. Rev. Lett. 115, 097401 (2015). [41] V. N. Antonov, B. N. Harmon, and A. N. Yaresko, Phys. [31] C. Klewe, T. Kuschel, J.-M. Schmalhorst, F. Bertram, Rev. B 67, 024417 (2003). O. Kuschel, J. Wollschl ager, J. Strempfer, M. Meinert, [42] M. Taguchi, A. Chainani, S. Ueda, M. Matsunami, and G. Reiss, Phys. Rev. B 93, 214440 (2016). Y. Ishida, R. Eguchi, S. Tsuda, Y. Takata, M. Yabashi, [32] Weiss, Pierre and Forrer, R., Ann. Phys. 10, 279 (1929). K. Tamasaku, Y. Nishino, T. Ishikawa, H. Daimon, [33] S. Celotto, W. Eerenstein, and T. Hibma, European S. Todo, H. Tanaka, M. Oura, Y. Senba, H. Ohashi, and Physical Journal B 36, 271 (2003). S. Shin, Phys. Rev. Lett. 115, 256405 (2015). [34] A. V. Ramos, J.-B. Moussy, M.-J. Guittet, A. M.

Journal

Condensed Matter – arXiv (Cornell University)

Published: May 4, 2020

There are no references for this article.