Epitaxial growth of ultrathin MgO layers on Fe3O4(0 0 1) films

Accepted Manuscript Title: Epitaxial Growth of Ultrathin MgO Layers on Fe3 O4 (001) Films Author: T. Nordmann O. Kuschel J. Wollschl¨ager PII: DOI: Reference:

S0169-4332(16)30311-7 http://dx.doi.org/doi:10.1016/j.apsusc.2016.02.133 APSUSC 32651

To appear in:

APSUSC

Received date: Revised date: Accepted date:

11-10-2015 11-2-2016 13-2-2016

Please cite this article as: T. Nordmann, O. Kuschel, J. Wollschl¨ager, Epitaxial Growth of Ultrathin MgO Layers on Fe3 O4 (001) Films, (2016), http://dx.doi.org/10.1016/j.apsusc.2016.02.133 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

T. Nordmann et al.

AIP/123-QED

Epitaxial Growth of Ultrathin MgO Layers on Fe3 O4 (001) Films T. Nordmann,1, 2 O. Kuschel,1, 2 and J. Wollschl¨ager1, 2, a) 1)

Fachbereich Physik, Universit¨ at Osnabr¨ uck, Barbarastr. 7, 49069 Osnabr¨ uck,

Germany 2)

Center of Physics and Chemistry of New Materials, Universit¨at Osnabr¨ uck,

ip t

Barbarastr. 7, 49069 Osnabr¨ uck, Germany (Dated: 11 February 2016)

cr

The initial growth stages of MgO on Fe3 O4 films are studied by means of x-ray photoelectron spectroscopy and low energy electron diffraction to clarify stoiThis bilayer

us

chiometric and structural properties of these layered structures.

structure is important to fabricate high quality magnetic tunnel junctions based

an

on Fe3 O4 electrodes and MgO tunnling barriers. For this purpose, the deposition temperature of MgO has been varied between 100◦ C and 250◦ C. Initially, MgO grows layer-by-layer on Fe3 O4 /MgO(001) forming a wetting layer. Depending on

M

the growth temperature, after growth of a 2-3 nm thick laminar wetting layer, the MgO films finally start to roughen during growth. Thus the growth of MgO on

d

Fe3 O4 /MgO(001) is described by a Stranski-Krastanov growth mode. Diffraction √ √ experiments show that the magnetite ( 2 × 2)R45◦ superstructure is removed

Ac ce pt e

already during the initial stages of MgO deposition. Furthermore, these experiments show that the MgO film are rougher for growth at low deposition temperatures.

PACS numbers: Valid PACS appear here Keywords: ultra thin films, x-ray photo electron spectroscopy, XPS, low energy electron diffraction, LEED, MgO, magnetite, Fe3 O4 , Stranski Krastanov growth mode, tunneling barrier

a)

Electronic mail: [email protected].

1

Page 1 of 12

T. Nordmann et al. I.

INTRODUCTION In the field of spintronics, magnetic tunnel junctions (MTJs), which consist of two fer-

romagnetic electrodes separated by an insulating tunneling barrier, play an important role. These junctions are based on the effect of tunnel magnetoresistance (TMR) which was first observed by Julliere1 . MTJs with ferromagnetic electrodes of 3d metals and tunneling bar-

ip t

riers made by amorphous aluminia were reported by Moodera et al.2,3 and Miyazaki et al.4 . For crystalline MgO tunnel barriers, however, a giant TMR effect was predicted from first

cr

principles calculations due to coherent tunneling5,6 . Giant TMR effects obtained with crystalline MgO tunnel barriers are reported using electrodes of elemental Fe7 or Co8 as well

us

as CoFe alloys9 . The effect has also been observed using CoFeB electrodes and textured polycrystalline MgO tunnel barriers10–12 .

an

For Fe/MgO/Fe MTJs, a drastically increased effect has been calculated if crystalline Fe electrodes are used instead of amorphous ones13 . Here, the interface structure plays an important role. Recently, it has been reported that an enhanced TMR effect has been obtained

M

for increased interface roughness14 . However, disadvantageous effects are also reported since Fe can be oxidized at the interface and antiferromagnetic FeO is occasionally formed, which

d

reduces or destroys the TMR effect15 .

Ac ce pt e

Therefore, there is a demand on all-oxide devices16 . Here, Fe3 O4 electrodes combined with MgO tunneling barriers seem to be an adequate choice of materials for high-quality MTJs since Fe3 O4 is a ferrimagnetic halfmetal with proposed high spin polarization at the Fermi level17,18 . Furthermore, Fe3 O4 has a high Curie temperature (860 K) and a saturation moment beyond 4µB 19 . It crystallizes in an inverse spinel structure with a cubic lattice constant of 839.63 pm at 300 K. MgO, however, is an insulator (band gap of 7.8 eV) with the rocksalt structure and cubic lattice constant of 421.17 pm. It is expected that MgO/Fe3 O4 structures grow well due to the very small lattice mismatch of 0.3% between both materials taking into account the cubic lattice constant of Fe3 O4 and the doubled lattice constant of MgO. In addition, it has to be emphasized that the oxygen sublattices of both materials have fcc symmetry. Thus, the oxygen lattice can be continued from one material to the other with only very small distortion making the MgO/Fe3 O4 structures unique. Indeed, the high quality of Fe3 O4 films deposited on MgO(001) substrates has been reported very often if the deposition temperature does not exceed 250◦ C20–22 . A 2

Page 2 of 12

T. Nordmann et al. strong magnetic coupling between Fe3 O4 electrodes separated by a MgO spacer is reported if the MgO thickness is below 1.3 nm23 . Beyond 250◦ C interdiffusion of the Mg and Fe cations is significantly increased and the MgO and Fe3 O4 films start to intermix close to the interface24–26 . Asymmetric MTJs structures with Fe and Fe3 O4 electrodes separated by MgO tunneling barriers have been reported27,28 . Here, the tunneling structure was prepared by pulsed

ip t

laser deposition (PLD) and the structural and magnetic properties were mainly studied. Recently, we reported on another asymmetric Fe3 O4 (bottom electrode) based MTJ with

cr

MgO tunneling barrier and CoFeB top electrode29 . Here, the effect of post deposition annealing (PDA) of the entire CoFeB/MgO/Fe3 O4 structure has been studied. Furthermore,

us

an inversion of the TMR effect is demonstrated if the entire structure is annealed above 250◦ C. This effect may be attributed to interface effects due to intermixing.

an

The study presented here sheds light on the initial stages of growth of MgO on Fe3 O4 films via reactive molecular beam epitaxy (RMBE) where Mg is evaporated in diluted oxygen atmosphere. The goal of this study is to clarify the growth mode of MgO/Fe3 O4 using RMBE

M

since the formation of high quality MgO films of homogeneous thickness are important for the performance of intended Fe3 O4 /MgO/Fe3 O4 MTJs. Therefore, both the stoichiometry and

d

structure of MgO films deposited on Fe3 O4 /MgO(001) are investigated in-situ by means

Ac ce pt e

of x-ray photoelectron spectroscopy (XPS) and low energy electron diffraction (LEED), respectively. Especially, we clarify here the role of deposition temperature on the growth of ultrathin MgO films

II.

METHODS

Samples were prepared and studied in an ultra high vacuum (UHV) chamber consisting of a preparation and an analysis chamber. MgO(001) substrates were annealed at 400◦ C in an atmosphere of 10−4 mbar O2 for 60 min to obtain well-ordered and clean surfaces. Cleaness and structure of the samples were proved by XPS and LEED. The XPS system operates with an Al Kα x-ray anode (1486.6 eV) and a hemispherical analyzer. Since the kinetic energy of the photoelectrons may vary from sample to sample due to charging effects, all photoelectron spectra were calibrated that the O 2p photoelectron peak is at 530 eV. First, magnetite (Fe3 O4 ) films were deposited using RMBE by evaporating Fe in 5×10−6 mbar 3

Page 3 of 12

cr

ip t

T. Nordmann et al.

us

FIG. 1. Photoelectron spectrum of Fe 2p obtained after growth of a 47 nm iron oxide film (red line). For comparison the Fe 2p spectra of FeO, Fe3 O4 and Fe2 O3 are also presented (black lines)30,31 .

an

Obviously, best matching of the experimental data is obtained for the Fe3 O4 spectrum considering both the binding energies of the Fe 2p1/2 and Fe 2p3/2 photoelectron peaks as well as the fact that

M

no apparent charge transfer satellites can be observed.

O2 at 250◦ C. The deposition process was controlled by use of a quartz balance. After deposi-

d

tion, both LEED and XPS were used to determine the surface structure and the surface near

Ac ce pt e

stoichiometry of the iron oxide films. Furthermore, the Fe3 O4 film thickness of 47(±5) nm was controlled a posteriori by means of x-ray reflectometry (XRR) using Cu Kα radiation. √ √ The LEED pattern of all iron oxide films showed the typical ( 2 × 2)R45◦ superstructure of Fe3 O4 (001) (cf. Fig.4(a)).

The stoichiometry of the iron oxide films has been controlled by means of the Fe 2p photoelectron spectra. Fig.1 shows a typical spectrum (red) obtained from one of the iron oxide films. Comparing this spectrum with reference spectra (black), it can clearly be concluded that the surface near region of the iron oxide film has Fe3 O4 stoichiometry since the Fe 2p1/2 and Fe 2p3/2 photoelectron peaks are located at 724.1 eV and 710.6 eV binding energies, respectively32 . In addition, the charge transfer satellites observed for pure Fe2+ and Fe3+ spectra (cf. Fig.1 for reference spectra of pure FeO and Fe2 O3 , respectively) overlap in a way that no apparent satellite is visible and the region between the Fe 2p1/2 and Fe 2p3/2 photoelectron peaks is structureless. This effect has been reported before in literature32,33 . After preparation of the Fe3 O4 films, MgO films were also deposited by RMBE (deposition 4

Page 4 of 12

T. Nordmann et al. a)

b)

Mg 2p

data fit background Mg 2p Fe 3p 2+ Fe 3p 3+

Fe 2p 26nm 12nm 4nm

45

45 [eV]

cr

[eV]

ip t

1.6nm 0.5nm 0.3nm Fe3O4

us

FIG. 2. Photoelectron spectra including Fe 3p and Mg 2p photoelectrons. (a) Sequence of spectra showing the increasing Mg 2p and the decreasing Fe 3p intensity due to successively increased MgO

an

film thickness for deposition at 250◦ C. (b) Comparison of experimental data (black line) and fitted spectrum (red line) containing photoelectron peaks due to Fe3+ , Fe2+ and Mg2+ with binding

M

energies of 55.6 eV, 53.7 eV, and 50.0 eV.

of Mg in 10−6 mbar O2 at different substrate temperatures). The MgO films were deposited

d

step by step. The structure of the MgO films and their stoichiometry were determined via

Ac ce pt e

LEED and XPS, respectively, after each deposition step. The final MgO film thickness was determined by means of x-ray reflectometry (XRR), too.

III.

RESULTS AND DISCUSSION

Quantitative analysis of the Mg 2p and Fe 3p photoelectron signal was used to obtain deeper insight in the growth process of MgO on Fe3 O4 . Fig.2(a) shows a sequence of combined Fe 3p and Mg 2p spectra for deposition of MgO at 250◦ C. Clearly, the intensity of the Fe 3p photo emission line (convolution of Fe2+ at 53.7 eV binding energy and Fe3+ at 55.6 eV binding energy) decreases while the intensity of the Mg 2p photoelectron emission line (50.0 eV binding energy) increases. A Shirley background was substracted from the photoelectron emission spectra to determine the photoemission intensity of both lines34 . Furthermore, all photoelectron emission lines were fitted by pseudo Voigt peaks keeping the splitting of 1.9 eV binding between both 5

Page 5 of 12

photoelectron yield YMg

T. Nordmann et al.

10 6

250°C 200°C 100°C

4 2 0 0

5 10 15 20 25 film thickness D [nm]

ip t

σ (D)

8

cr

film thickness D [nm]

FIG. 3. Thickness dependence of the Mg yield. While the yield follows the expected behavior

us

during initial stages for layer-by-layer growth (dashed line, cf. eq. (2)), it deviates in later stages. This result is attributed to the growth of the MgO film in the Stranski-Krastanov growth mode.

an

Inset: film roughness determined from the difference between experimental data and expected values for perfect layer-by-layer growth. At the beginning, the film roughness is less than 1 nm but

M

increases drastically for films with thickness beyond the critical thickness of 2-3 nm.

Fe cation components fixed. Fig.2(b) presents a fit to a typical photoelectron emission spec-

Ac ce pt e

Fe3+ peak is at 55.6 eV.

d

trum where the Fe2+ photoelectron emission peak is at 53.7 eV binding energy while the

Following Newberg et al.35 , the integrated intensities IFe 3p and IMg 2p of both photoelectron peaks Fe 3p and Mg 2p, respectively, have been used to determine the effective stoichiometry of the MgO/Fe3 O4 (001) system via the Mg yield

YMg =

IMg 2p . IMg 2p + CMg/Fe IFe 3p

(1)

Here, the factor CMg/Fe includes the ratio of the number of atoms per unit volume NMg /NFe , the ratio of the photoelectron cross sections σFe 3p /σMg 2p and the ratio λMg /λFe of the inelastic mean free path of the constituents MgO and Fe3 O4 . We used the photoelectron cross sections calculated by Scofield36 and, according to Tanuma et al.37 , assumed that the inelastic mean free path of the Fe 3p and Mg 2p photoelectrons is λ=3.2 nm for both species. Fig. 3 presents the dependence of the Mg yield on the thickness D. For comparison, the dashed line presents the yield 6

Page 6 of 12

T. Nordmann et al.

YMg = 1 − exp (−D/λ)

(2)

expected for perfect laminar growth (layer-by-layer growth) of the MgO film on the Fe3 O4 film. Obviously, the MgO films grow layer-by-layer up to thickness of approximately 23 nm for all studied deposition temperatures. After this wetting layer is formed, the Mg yield increases less than expected for the layer-by-layer growth mode (dashed line). This

ip t

effect points to the formation of islands on top of the wetting layer38 . Thus, we conclude that MgO grows on Fe3 O4 in the Stranski-Krastanov growth mode. AFM micrographs (not

cr

shown here) recorded ex-situ after the final growth step of the MgO film confirmed that MgO islands are formed on top of the wetting layer.

us

Therefore, we analysed the Mg yield in more detail considering a Gaussian height distribution of the film thickness. For this prupose, eq.(2) has to be modified by substituting the

an

(average) film thickness D by an effective film thickness Deff = D − σ 2 /2λ where σ denotes the root-mean-square (rms) roughness of the MgO film. The result is presented in the inset

M

of Fig.3. This analysis also shows the initial formation of smooth MgO film and the drastic increase of the roughness if the film thickness is beyond the critical film thickness of 2-3 nm. The initial wetting behavior of the MgO film is explained by its lower surface energy

d

compared to the surface energy of Fe3 O4 (001)39 . Surface energies in the range of 1.00 J/m2

Ac ce pt e

and 1.16 J/m2 have been determined for MgO(001)40,41 while the surface energy 1.45 J/m2 has been calculated for Fe3 O4 (001)42 .

Obviously, the roughening process is slightly stronger for deposition at 100◦ C than for higher temperatures. This effect is surprising since, in contrary, the de-wetting should be stronger at higher temperatures since this process is driven by the thermodynamic equilibrium morphology. Thus, the de-wetting should be diminished at lower temperatures where the kinetics of the process is slowed down. The difference between growth at 200◦ C and 250◦ C, however, is not significant. Both curves run parallel up to more than 10 nm film thickness.

As pointed above, besides the homogeneity of the tunneling barrier, the interface structure is also important for the tunneling process. Therefore, we studied the growth of the MgO also by LEED. Selected LEED pattern are presented in Fig.4. √ √ The clean Fe3 O4 (001) surface shows a ( 2 × 2)R45◦ which is unique for Fe3 O4 (001) and is not observed for FeO(001) nor for Fe2 O3 (001). The superstructure is attributed 7

Page 7 of 12

us

cr

ip t

T. Nordmann et al.

an

FIG. 4. Sequence of LEED pattern recorded at 120 eV depositing MgO on an Fe3 O4 (001) film at √ √ 100◦ C. (a) Clean Fe3 O4 (001) surface with (1 × 1) (large square) and ( 2 × 2)R45◦ (small square) reciprocal unit cells. (b) After depositon of 0.3 nm MgO: Superposition of (1 × 1) of MgO(001) √ √ (large square) and (1 × 1) of Fe3 O4 (001) (small square). The ( 2 × 2)R45◦ superstructure has

M

vanished. (c) After depositon of 1.2 nm MgO: (1 × 1) of MgO(001) with sharp diffraction spots.

d

(d) After depositon of 4.1 nm MgO: (1 × 1) of MgO(001) with broadened diffraction spots.

Ac ce pt e

to periodically arranged subsurface Fe vacanacies43 . Therefore, combined with the XPS study presented above, this diffraction pattern emphasizes the Fe3 O4 (001) stoichiometry and structure of the initial iron oxide film. √ √ After deposition of 0.3 nm MgO on the Fe2 O3 (001) at 100◦ C the ( 2 × 2)R45◦ has vanished. The deposited amount of MgO is equivalent to 1.5 ML. The electron beam used for this experiment can penetrated the entire MgO film and interacts with the MgO/Fe3 O4 interface. Therefore, one can conclude that the superstructure is destroyed during the initial stages of MgO growth.

Further MgO growth cause a gradual decrease of the Fe3 O4 diffraction peaks until only the MgO(001)-(1×1) is observed [cf. Fig.4(c) for 1.2 nm MgO]. Initially, the MgO diffraction peaks are sharp. However, they broaden with increasing MgO deposition showing that the MgO film becomes rougher [cf. Fig.4(d) for 4.1 nm MgO, cf. Fig.3 for equivalent XPS results]. This effect is reduced for deposition at higher temperatures (not shown here). Varying the temperature from 100◦ C to 250◦ C for deposition of 6 nm MgO reduced the 8

Page 8 of 12

T. Nordmann et al. halfwidth of the diffraction peaks by 50%.

IV.

CONCLUSION

In summary, we have shown that MgO films grow on Fe3 O4 /MgO(001) in the Stranski Krastanov growth mode. The initial wetting layer has a thickness of 2-3 nm. For the use of

ip t

tunneling barriers, one has to fabricate very homogeneous insulating ultra thin films since tunneling electrons are affected by local thickness variations, e.g., due to scattering processes.

cr

Therefore, the use of MgO barriers on Fe3 O4 films is restricted to the initial wetting layer. We conclude that the wetting layer of MgO on Fe3 O4 can be used as tunneling barrier since,

us

depending on the band gap and band alignments, the typical thickness of tunneling barriers is less than 2 nm.

an

Furthermore, the LEED pattern obtained from the MgO wetting layer show sharp diffraction peaks pointing to a high crystalline quality. In addition, the LEED investigations show that the initial superstructure of the Fe3 O4 (001) bottom electrode is removed during the ini-

M

tial growth stages of MgO. Therfore, this more complicated surface structure has not to be considered for the interface of Fe3 O4 /MgO/Fe3 O4 MTJs to describe the tunneling process.

d

In conclusion, MgO tunneling barriers can also alternatively be formed by means of

V.

Ac ce pt e

reactive molecular-beam epitaxy to be part of Fe3 O4 /MgO/Fe3 O4 MTJs.

ACKNOWLEDGEMENTS

We gratefully acknowledge supporting AFM and XRR experiments by A. Boehnke, T. Kuschel and G. Reiss from Bielefeld University.

REFERENCES 1

M. Julliere, Phys. Lett. 54 (1975) 225.

2

J.S. Moodera, L.R. Kinder, T.M. Wong, R. Meservey, Phys. Rev. Lett. 74 (1995) 3273.

3

J.S. Moodera, L.R. Kinder, J. Appl. Phys. 79 (1996) 4724.

4

T. Miyazaki, N. Tezuka, J. Magn. Magn. Mater. 139 (1995) L231. 9

Page 9 of 12

T. Nordmann et al. 5

W. Butler, X.-G. Zhang, T.C. Schulthess, J.M. MacLaren, Phys. Rev. B63 (2001) 054416.

6

J. Mathon, A. Umersky, Phys. Rev. B63 (2001) 220403R.

7

S. Yuasa, T. Nagahama, A. Fukushima, Y. Suzuki, K. Ando, Nature Materials 3 (2004) 868.

8

S. Yuasa, A. Fukushima, H. Kubota, Y. Suzuki, K. Ando, Appl. Phys. Lett. 89 (2006) 042505. S.S.P. Parkin, C. Kaiser, A. Panchula, P.M. Rice, B. Hughes, M. Samant, S.-H. Yang,

ip t

9

Nature Materials 3 (2004) 862.

D.D. Djayaprawira, K. Tsunekawa, M. Nagai, H. Maehara, S. Yamagata, N. Watanabe,

cr

10

S. Yuasa, Y. Suzuki, K. Ando, Appl. Phys. Lett. 86 (2005) 092502.

S. Ikeda, J. Hayakawa, Y. Ashizawa, Y.M. Lee, K. Miura, H. Hasegawa, M. Tsunoda, F.

us

11

Matsukura, H. Ohno, Appl. Phys. Lett. 93 (2008) 082508.

S. Ikeda, K. Miura, H. Yamamoto, K. Mizunuma, H.D. Gan, M. Endo, S. Kanai, J.

an

12

Hayakawa, F. Matsukura, H. Ohno, Nature Materials 9 (2010) 721. 082508. C. Heiliger, M. Gradhand, P. Zahn, I. Mertig, Phys. Rev. Lett. 99 (2007) 066804.

14

A. Duluard, C. Bellourad, Y. Lu, M. Hehn, D. Lacour, F. Montaigne, G. Lengaigne, S.

M

13

Andrieu, F. Bonell, C. Tiusan, Phys. Rev. B91 (2015) 174403. C. Heiliger, P. Zahn, I. Mertig, J. Magn. Magn. Mater. 316 (2007) 478.

16

M. Bibes, J.E. Villegas, A. Barthelemy, Adv. Phys. 60 (2011) 5.

17

Z. Zhang, S. Satpathy, Phys. Rev. B 44 (1991) 13319.

18

Z. Szotek, W.M. Temmerman, D. Kodderitzsch, A. Svane, L. Petit, H. Winter, Phys.

Ac ce pt e

d

15

Rev.B74 (2006) 174431. 19

P. I. Slick, in: Ferromagnetic Materials: A Handbook on the Properties of Magnetically Ordered Substances, E.P. Wohlfarth (ed.), North-Holland Publishing Company, Amsterdam, 1980.

20

F. Bertram, C. Deiter, O. Hoefert, T. Schemme, F. Timmer, M. Suendorf, B. Zimmermann, J. Wollschl¨ager, J. Phys. D: Appl. Phys. 45 (2012) 395302.

21

F. Bertram, C. Deiter, T. Schemme, S. Jentsch, J. Wollschl¨ager, J. Appl. Phys. 113 (2013) 184103.

22

T. Schemme, N. Path, K. Kuepper, G. Niu, F. Bertram, T. Kuschel, J. Wollschl¨ager, Mater. Res. Express 2 (2015) 016101.

10

Page 10 of 12

T. Nordmann et al. 23

P.A.A. van der Heijden, P.J.H. Bloemen, J.M. Metselaar, R.M. Wolf, J.M. Gaines, J.T.W.M. van Eermeren, P.J. van der Zaag, W.J.M. de Jonge, Phys. Rev. B55 (1997) 11569.

24

Y. Gao, Y. J. Kim, and S. A. Chambers, J. Mater. Res. 13 (1998) 2003.

25

K. A. Shaw, E. Lochner, and D. M. Lind, J. Appl. Phys. 87 (2000) 1727.

26

N.-T. H. Kim-Ngan, A. G. Balogh, J. D. Meyer, J. Brotz, S. Hummelt, M. Zajac, T.

27

ip t

Slezak, J. Korecki, Nucl. Instrum. Methods Phys. Res. B267 (2009) 1484.

W. Kim, K. Kawaguchi, N. Koshizaki, M. Sohma, T. Matsumoto, J. Appl. Phys. 93 (2003)

28

cr

8032.

J. Orna, L. Moerllon, P.A. Algarabel, J.A. Pardo, S. Sangiao, C. Magen, E. Snoeck, J.M.

29

us

De Teresa, M.R. Ibarra, IEEE Transact. Magnetics 44 (2008) 2862.

L. Marnitz, K. Rott, S. Nieh¨orster, C. Klewe, D. Meier, S. Fabretti, M. Witziok, A.

an

Krampf, O. Kuschel, T. Schemme, K. Kuepper, J. Wollschl¨ager, A. Thomas, G. Reiss, T. Kuschel, AIP Advances 5 (2015) 047103.

S. Uhlenbrock, PhD thesis, Universit¨at Osnabr¨ uck (1994).

31

K. C. Prince, M. Matteucci, K. Kuepper, S. G. Chiuzbaian, S. Bartkowski, M. Neumann, Phys. Rev. B71 (2005) 085102.

M

30

T. Yamashita, P. Hayes, Appl. Surf. Sci. 254 (2008) 2441.

33

T. Fujii, F.M.F. de Groot, G.A. Sawatzky, F.C. Voogt, T.Hibma, K. Okada, Phys. Rev. B

Ac ce pt e

d

32

59 (1999) 3195. 34

D. A. Shirley, Phy. Rev. B5 (1972) 4709.

35

J.T. Newberg, D.E. Starr, S. Yamamoto, S. Kaya, T. Kendelewicz, E.R. Mysak, S. Porsgaard, M.B. Salmeron, G.E. Brown Jr., A. Nilsson, H. Bluhm, Surf. Sci. 605 (2011) 89.

36

J.H. Scofield, J. Electron Spectroscopy Relat. Phenomena 8 (1976) 129.

37

S. Tanuma, C. J. Powell, and D. R. Penn, Surf. Interface Analysis 35 (2003) 268.

38

S. Meroz, Progr. Surf. Sci. 46 (1994) 377.

39

E. Bauer, Z. Kristallog. 110 (1958) 372.

40

G. Jura, C.W. Garland, J. Am. Chem. Soc. 74 (1952) 6033.

41

P.W. Tasker, D.M. Duffy, Surf. Sci. 137 (1984) 92.

42

R.K. Mishra, G. Thomas, J. Appl. Phys. 48 (1977) 4576.

43

R. Bliem, E. McDermott, P. Ferstl, M. Setvin, O. Gamba, J. Pavelec, M. A. Schneider, M. Schmid, U. Diebold, P. Blaha, L. Hammer, and G. S. Parkinson, Science 346 (2014) 1215. 11

Page 11 of 12

Highlights

ce pt

ed

M

an

us

cr

ip t

LEED/XPS characterization of MgO/Fe3O4 Stranski-Krastanov growth mode of MgO/Fe3O4 determined by quantitative XPS Gradual structural degradation observed by LEED Temperature dependent thickness of wetting layer possible use for magnetic tunnel junctions certified

Ac

    

Page 12 of 12