Quantitative Observation of Magnetic Flux Distribution in New Magnetic Films for Future High Density Recording Media
Aurélien Masseboeuf, A. Marty, Pascale Bayle-Guillemaud, Christophe Gatel, E. Snoeck
aa r X i v : . [ c ond - m a t . m t r l - s c i ] A ug Quantitative observation of magnetic fluxdistribution in new magnetic films for future highdensity recording media.
Aurélien Masseboeuf, ∗ , † , § Alain Marty, † Pascale Bayle-Guillemaud, † ChristopheGatel, ‡ and Etienne Snoeck ‡ CEA,INAC, Grenoble (FR), and CEMES-CNRS, Toulouse (FR)
E-mail: [email protected]
Abstract
Off-axis electron holography was used to observe and quantify the magnetic microstructureof a perpendicular magnetic anisotropic (PMA) recording media. Thin foils of PMA materialsexhibit an interesting Up and Down domain configuration. These domains are found to be verystable and were observed at the same time with their stray field, closing magnetic flux in thevacuum. The magnetic moment can thus be determined locally in a volume as small as fewtens of nm . † CEA,Institut Nanosciences et Cryogénie - SP2M, 17 rue des Martyrs, 38054 Grenoble Cedex 09, France ‡ CEMES-CNRS, BP 94347, 29 rue Jeanne Marvig, 31055 Toulouse Cedex, France § New ad. : CEMES-CNRS, Toulouse, France Now,all improvements in new memory capacity are expected to be reached thanks to perpendicularrecording. While important studies are published on new materials design, studies on magneticinteraction between the recording head and the data bits are mostely concerned about the tip fieldleakage characterization.
Here we show that it is also possible to obtain nanometric and quantita-tive magnetic informations of stray field and magnetic induction at the same time on perpendicularmagnetic anisotropic (PMA) materials.Due to the strong stray fields perpendicular to their surface, PMA materials have been exten-sively studied by Magnetic Force Microscopy (MFM), micro-magnetic simulations, and othermagnetic characterization techniques. However these techniques were not suitable to study theinner magnetic configuration of materials and inner magnetization at the same time. Thus, forma-tion of stray field distribution with respect inner magnetic parameters of the material have not beenyet confirmed by experimental results. This could be of great importance to evaluate the abilityof a reading head to flip a bit.Electron Holography (EH) in a Transmission Electron Microscope (TEM) is a powerful tech-nique which enables observation of electrostatic and magnetic fields at the nanoscale by a electronwave phase retrieval process. Through the so-called Aharonov-Bohm effect, it is known thatan electron wave is sensitive to the electric and magnetic potential. As a consequence, it can beused to investigate magnetic properties of materials. Electron holography has thus been used tostudy many magnetic materials, for example, the analysis of stray fields around a MFM tip, orthe study of the magnetic configuration of magnetic nanoparticles, magnetic films, magnetic2urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscaletunnel junctions, or even magnetite core of magnetotactic bacteria. The phase shift of the exitelectron wave,
D f , travelling along the z direction across the sample, which interacted with theelectromagnetic field (electrostatic and magnetic potentials from the sample) can be expresses as : D f ( x , y ) = C E Z V int ( x , y , z ) dz − e ¯ h Z A z ( x , y , z ) dz (1)where V ( x ) int represents the electrostatic contribution to the phase shift (in the case of a materialit is mainly its Mean Inner Potential or MIP), and A z is the z -component of the magnetic vectorpotential describing the magnetic induction distribution in a plane (for a given z) pendicular tothe optic axis. A z is related to the magnetic induction ~ B by means of the Maxwell’s equation : ~ B = ( B x , B y ) = ( (cid:209) y A z , − (cid:209) x A z ) . C E is an electron energy related constant.The keypoint is the separation of the magnetic and electrostatic contributions in the reconstructedphase shift and is extensively discussed elsewhere. We have used for our purpose a method consisting of recording two holograms before (
D f + )and after ( D f − ) removing and inverting the sample. The electrostatic contribution to the phaseshift remains similar in the two holograms while the magnetic contribution changes in sign. Themagnetic contribution ( D f magn . ) can thus be obtained by evaluating half of the difference of the twophase images calculated from the two holograms. The MIP contribution ( D f
MIP ) is then half ofthe sum. To account for the sample reversal, it is necessary to reverse one of the two phase imagesand align them.
D f magn . = ∗ [ D f + − D f − ] = e ¯ h [ A z D t ( x , y )] D f M . I . P . = ∗ [ D f + + D f − ] = C E V ( x , y ) int D t ( x , y ) (2)EH has been used to study the magnetic configuration and measure the remanent magnetizationof a FePd L /FePd disord . stack, grown on MgO(001), which exhibits a strong PMA. The tetragonalaxis of the chemically ordered FePd L crystalline structure lies along the growth direction cor-responding to an alternate stacking of pure Iron and Palladium planes. This chemical anisotropyalong the z -axis induces an easy magnetic axis in the same direction which gives rise to the up and3urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscaledown magnetic domain configuration. The main purpose of the second "soft" layer, having a van-ishing anisotropy, is usually described as the main factor for increasing the recording efficiency. The expected magnetic configuration of the FePd L /FePd disord . bilayer is presented in Figure1-A.Domains in the FePd L layer are separated by Bloch walls where the magnetization lies in theplane of the foil, surrounded by a Néel Cap in which the magnetization runs around the Bloch wallaxis. The bottom FePd disord . layer gives rise to in-plane components allowing a flux closure withinthe bilayer, and enables the domains to be aligned in a parallel stripes configuration. This magneticconfiguration is confirmed by studying the external stray field by MFM experiment as shown inFigure 1-B.Figure 1: A. Magnetic configuration expected for the foil. B. MFM view of the sample in its stripeconfiguration. Black contrast is down domains, bright contrast is up ones. The dashed area showsthe geometry used for TEM sample preparation, across magnetic orientation. C. Fresnel view ofthe sample showing black and white contrast where the beam are overlaping.Our purpose is to analyse in more detail the inner magnetic configuration with higher resolu-tion. Figure 1-C shows a Fresnel TEM micrograph of the sample. The thin foil used for TEMexperiments has been prepared in cross-section in order to observe the magnetic structure by theside instead of the top in the MFM geometry. In this micrograph, domains are clearly defined,separated from one another by a bright or dark line localised at the position of domain walls. Thedomain periodicity is 100 nm which fully agree with MFM measurement. EH experiments werecarried out on the same area of the PMA magnetic film.4urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscaleFigure 2: Mean Inner Potential ( A ) and magnetic ( B ) contribution to the phase of the electron wave.The color scale used here is a temperature scale described near each picture. The contour linesare for equi-phase lines and represent 1 radian for MIP contribution and 1/4 radian for magneticcontribution.Figure 2-A shows the deduced MIP contribution to the phase shift, and the magnetic contribu-tion is shown in Figure 2-B. The iso-phase contours displayed on both phase images directly relateto thickness variations (in the MIP phase image) and to magnetic flux (in the magnetic one). Thevariation of the MIP contribution, exhibits that the TEM sample increases uniformly in thicknesswhile magnetic contribution highlights vortices corresponding to the Bloch walls. Between thesevortices are areas where the magnetic flux is parallel or anti-parallel to the growth direction. Thesecorrespond to the "up" and "down" magnetic domains. Stray fields close the flux in the vacuumand inside the stack. However, the vortices appear to be flatter at the bottom (close to the FePddisordered layer) than at the top (close to the vaccum). This asymmetrical shape of the vortices isdue to the disordered FePd layer which forces the magnetic induction to lie within the foil plane.From equation 2, quantitative values of magnetic induction can be extracted provided that the MIP5urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscaleor the thickness of the different layers are known: D t ( x ) = D f M . I . P . C E V ( x ) int B ⊥ = ¯ he · D t ( x ) (cid:209) [ D f magn . ] (3)Figure 3: Mean Inner Potential contribution to the phase analysis. This is a plot profile along thedashed line in Figure 2-A. Right (yellow) label and solid line is the phase profile used to deduce thedifferent layers in the foil. Dashed lines present linear interpolation variations for each differentlayers. Thickness profile is presented in the colored (blue) area and is labeled on the left.Figure 3 shows in yellow the experimental profile of the electrostatic contribution to the phaseextracted along the dashed line in Figure 2. To deduce the thickness profile (in blue), we have firstcalculated the mean inner potential values for the Pd, FePd L and FePd disord . layers.According to Equation 3, this thickness profile is used to calculate the B x and B y inside the layer(Figure 4, A and B). Neglecting the demagnetizing field within the material, the measured mag-netic induction is directly related to the magnetization in the material. The value of the magneticinduction modulus in the FePd L region ( i.e. inside the domains) gives rise to a magnetic inductionof 1 . ± . disord . area (under domain walls)gives an aeeeveraged values of 1 . ± . m M s of the FePd L layerare the same as those expected for FePd. The variations observed in the different domains comefrom a variation in the evaluation of the local thickness, due to small deformation of the crystal,or amorphization during ion milling. The difference found between the two layers is negligible. Itshould comes from the smaller area of magnetization purely perpendicular to the electron path in6urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscalethe soft layer under the domain wall. This implies a variation of the magnetization direction due tothe presence of the domain wall and the in-plane magnetization under the domains. The measuredvalue is then no longer a pure magnetic moment but a projection of it.More accurate measurements of the FePd magnetic properties can be done performing micromag-Figure 4: A and B are the X and Y components (left and right respectively) of the magneticinduction deduced from phase gradients. Profile for quantitative interpration is displayed on eachfigure. C is a zoomed view of Fig. 2-B compared with a micro-magnetic simulation in D . In bothimages, iso-phase lines represent 0 . rad and the color scale used is described.netic simulations. We have used a code based on LLG temporal integration, GL_FFT, to simulatethe magnetic flux (i.e. the iso-phase contours) observed by EH. Calculated magnetic phase shift(using bulk values ) and experimental data are compared in Figure 4, C and D.It is seen that the Bloch walls are much wider in the experimental data which could be explain bya slight decrease of the thickness due to the thinning process (see also Supporting Informations).The magnetic flux can be quantitatively measured, both in and outside the sample. The stray fieldcan then be related to the magnetization moment in the domains. This is potentially of great inter-est for the design of a reader of this kind of material.Electron holography was used to highlight the magnetic structure of FePd L -FePd disord . mag-7urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscalenetic bilayer exhibiting PMA, with a resolution closed to the nanometer and an accurate measure-ment of the local magnetic induction. The magnetic configuration was then successfully comparedto micromagnetic calculations. Moreover, the quantitative informations given by the technique canbe directly related to the stray field of the materials, which are the bit information for reading headsin hard drives. Acknowledgement
Thanks are due to Dr. P. Cherns for critics and usefull discussions about this manuscript.
Supporting Information Available
The sample was grown on MgO (001) substrate by Molecular Beam Epitaxy (MBE) according tothe following sequence : Cr (2.5 nm) in order to initiate the epitaxial growth, Pd (48 nm), FePd (15nm) co-deposited at room temperature, FePd (37 nm) co-deposited at 450 ◦ C and a 1.5 nm capingof Pt was added to avoid oxydation.The sample has been prepared for electron microscopy using mechanical polishing and ion milling.The layer is thus exhibiting a double wedge geometry along the observation plane. The microscopeused for the holography experiments is a FEI Tecnai F20 field-emission-gun TEM fitted with a Cscorrector (CEOS). A FEI Titan FEG TEM fitted with a dedicated Lorentz lens was used for Fresnelimaging. A Gatan Imaging Filter was also used for zero loss filtering for the Fresnel images.Holograms are recorded using off-axis electron holography with a rotatable biprism located in theSA aperture. The biprism is aligned along the foil direction x . The fringe spacing is 1.8 nm,the fringe contrast is 12 %. For calculating the phase image we perform a Fourier transform ofthe hologram and apply a mask of 0.25 nm − on the side-band spot before calculating an inverseFourier transform.To separate the electrostatic and magnetic contributions to the phase shfit, two holograms wererecorded before and after inverting the sample. Image calculations were then performed to alignedthe two images. The phase images have been digitally flipped for accordance with the physical8urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscaleinversion of the sample. After data acquisition, an accurate correction of the drift, rotation andscaling between the two images has been performed using recently developped scripts.Mean Inner Potentials have been calculated using the Doyle and Turner scattering amplitude cor-rected with the Ross and Stobbs equation (see chapter 12 of. ) We calculate : V FePd L o = .
73 V, V FePd dis . = .
67 V, V Pd = .
37 V.Micro-magnetical simulation has been carried out using the bulk FePd following parameters :Exchange constant A = . − J.m − , Uniaxial Anisotropy K=1 .
03 10 J.m − , Saturated Mag-netization m M S = 1 .
294 Tesla . The cells are 0 . nm × . nm and infinite along the z direction(considered as invariant). References (1) Iwasaki, S.; Y. Nakamura, K.
IEEE Transactions on Magnetics , , 1272–1277.(2) Kryder, M. H.; Gustafson, R. W. Journal of Magnetism and Magnetic Materials , ,449–458.(3) Richter, H. J. Journal of Physics D: Applied Physics , , R149.(4) Signoretti, S.; Beeli, C.; Liou, S.-H. Journal of Magnetism and Magnetic Materials , , 2167 – 2168.(5) Kim, J. J.; Hirata, K.; Ishida, Y.; Shindo, D.; Takahashi, M.; Tonomura, A. Applied PhysicsLetters , , 162501.(6) Gehanno, V.; Samson, Y.; Marty, A.; Gilles, B.; Chamberod, A. Journal of Magnetism andMagnetic Materials , , 26–40.(7) Attane, J. P.; Samson, Y.; Marty, A.; Halley, D.; Beigne, C. Applied Physics Letters , ,794–796. 9urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscale(8) Toussaint, J. C.; Marty, A.; Vukadinovic, N.; Youssef, J. B.; Labrune, M. ComputationalMaterials Science , , 175–180.(9) Dürr, H. A.; Dudzik, E.; Dhesi, S. S.; Goedkoop, J. B.; van der Laan, G.; Belakhovsky, M.;Mocuta, C.; Marty, A.; Samson, Y. Science , , 2166–2168.(10) Beutier, G.; Marty, A.; Chesnel, K.; Belakhovsky, M.; Toussaint, J. C.; Gilles, B.; derLaan, G. V.; Collins, S.; Dudzik, E. Physica B , , 143–147.(11) Hubert, A.; Shäfer, R. Magnetic Domains ; Springer, 1998.(12) Gabor, D.
Nature , , 777–778.(13) Aharonov, Y.; Bohm, D. Physical Review , , 485–491.(14) Streblechenko, D. G.; Scheinfein, M. R.; Mankos, M.; Babcock, K. IEEE Transactions onMagnetics , , 4124–4129.(15) Hytch, M. J.; Dunin-Borkowski, R. E.; Scheinfein, M. R.; Moulin, J.; Duhamel, C.; Maza-leyrat, F.; Champion, Y. Phys Rev Lett , , 257207.(16) McCartney, M. R.; Smith, D. J.; Farrow, R. F. C.; Marks, R. F. Journal of Applied Physics , , 2461–2465.(17) Snoeck, E.; Baules, P.; BenAssayag, G.; Tiusan, C.; Greullet, F.; Hehn, M.; Schuhl, A. Jour-nal of Physics: Condensed Matter , , 055219–.(18) Dunin-Borkowski, R. E.; McCartney, M. R.; Frankel, R. B.; Bazylinski, D. A.; Pósfai, M.;Buseck, P. R. Science , , 1868–1870.(19) Dunin-Borkowski, R. E.; Kasama, T.; Wei, A.; Tripp, S. L.; Hytch, M. J.; Snoeck, E.; Harri-son, R. J.; Putnis, A. MICROSCOPY RESEARCH AND TECHNIQUE , , 390–402.(20) Khizroev, S.; Litvinov, D. Perpendicular Magnetic Recording ; Kluwer Academic Publishers,2004; p 174. 10urélien Masseboeuf et al. Quantitative magnetic observation of PMA alloys at the nanoscale(21) Chapman, J. N.