Legal forms

Fundamental studies of magnetism down to the atomic scale: present status and future perspectives of spin-polarized scanning tunneling microscopy

Description
Fundamental studies of magnetism down to the atomic scale: present status and future perspectives of spin-polarized scanning tunneling microscopy
Categories
Published
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  Journal of Magnetism and Magnetic Materials 272–276 (2004) 2115–2120 Fundamental studies of magnetism down to the atomic scale:present status and future perspectives of spin-polarizedscanning tunneling microscopy R. Wiesendanger*, M. Bode, A. Kubetzka, O. Pietzsch, M. Morgenstern,A. Wachowiak, J. Wiebe Institute of Applied Physics and Microstructure Research Center,University of Hamburg, Jungiusstr. 11, Hamburg D-20355, Germany Abstract Recent developments in spin-polarized scanning tunneling microscopy and spectroscopy have led to anunprecedented insight into magnetism at the nanometer length scale and, in some cases, even down to the atomiclevel. The correlation between structural, local electronic and local magnetic structure can now be studied beyond theexchange length. Most striking recent results include the discovery of atomically sharp magnetic domain walls in atomiclayers of iron and the determination of the intrinsic width of magnetic vortex cores in three-dimensional iron islands. r 2003 Elsevier B.V. All rights reserved. PACS:  75.75.+a; 75.25.+z; 75.60.Ch Keywords:  Spin-polarized scanning tunneling microscopy; Nanomagnetism; Magnetic domain walls; Magnetic vortices; Spin structure 1. Introduction Spin-polarized scanning tunneling microscopy (SP-STM) and spectroscopy (SP-STS) based on the use of magnetic probe tips has become one of the mostpowerful tools to probe magnetism beyond the exchangelength scale. While SP-STM operated in the constantcurrent mode [1] has proven its atomic resolutioncapability for ferrimagnetic [2,3] as well as for anti-ferromagnetic systems [4,5], the corresponding spectro-scopic technique of SP-STS allows the correlation of structural, local electronic and local magnetic propertiesdown to the atomic level. This has been demonstratedfor ferromagnetic rare earth [6] and transition metal [7] systems as well as for antiferromagnets [8,9]. Asignificant breakthrough has been the tailoring of themagnetic properties of SP-STM tips by thin film coatingtechniques including an appropriate choice of materials,coating thickness and measurement temperature. Forinstance, Fe-coated W tips proved to be magnetizedperpendicular to the tip axis at the tip apex and aretherefore sensitive to the in-plane component of thesample’s magnetization [6,8,10]. In contrast, FeGd-coated probe tips exhibit a perpendicular magneticanisotropy and are therefore sensitive to the out-of-plane component of the sample’s magnetization [11].However, ferromagnetic coatings might not be theoptimum choice if samples with a low coercivity shouldbe investigated because the resulting stray field from thetip might disturb the magnetization state of the sample.To solve this problem, antiferromagnetic coatings, suchas Cr, can be used as has successfully been demonstratedon magnetic nanowires [11] and nano-scale islands [12]. In this case, the magnetic anisotropy of the Cr-coated tipcan simply be tailored by an appropriate choice of thethickness of the Cr coating [12]. In contrast to manyother magnetic imaging techniques, SP-STM and SP-STS can even be performed in the presence of highexternal magnetic fields [13]. If antiferromagnetic probe ARTICLE IN PRESS *Corresponding author. Tel.: +49-40-42838-5244; fax: +49-40-42838-6188. E-mail address:  wiesendanger@physnet.uni-hamburg.de(R. Wiesendanger).0304-8853/$-see front matter r 2003 Elsevier B.V. All rights reserved.doi:10.1016/j.jmmm.2003.12.817  tips are used, magnetic switching of the tip can beprevented while sweeping the external magnetic field. Onthe other hand, SP-STM and SP-STS can be applied tostudy antiferromagnetic samples which has proven to bedifficult by other magnetic imaging techniques. Moststriking results in this field have been the experimentalproof of the topological antiferromagnetic order of theCr(001) surface [1,8,14,15] and of the existence of two-dimensional antiferromagnetic order for a single atomiclayer of Mn on a W(110) substrate [4,16].In the following, the focus will be on recent SP-STSstudies of atomically sharp domain walls in singleatomic layers of Fe on W(110) as well as of magneticvortex core states in three-dimensional Fe islands. 2. Experimental The experiments were performed with a low-tempera-ture SP-STM in a multi-chamber ultrahigh vacuum(UHV) system [17] at a typical tip and sampletemperature of 14 7 1K and a background pressure inthe 10  11 mbar range. The electrochemically etched Wtips were flashed in situ to high temperatures in order toremove oxide layers before the deposition of magneticmaterial (either Fe or Cr). The differential tunnelingconductance d I  = d U   was measured by modulating theapplied bias voltage  U  :  This differential conductanceobtained with a magnetic probe tip at the location  ~ rr  onthe sample surface for a bias voltage  U  0  can be writtento first approximation asd I  d u ð  r ! ; U  0 Þ SP  ¼  C  ð 1 þ P  T P  S  cos  y Þ ; where  C   is the spin-averaged differential conductance, P  T  ¼  P  T ð E  F Þ  is the spin polarization of the tip at theFermi energy  E  F ;  and  P  S  ¼  P  S ð E  F  þ e U  0 Þ  is the spinpolarization of the sample at the energy ( E  F  þ e U  0 ). Theangle  y  ¼  y ð  ~ M M  T ; ~ M M  S ð ~ rr ÞÞ  is enclosed by the tip magneti-zation  ~ M M  T  and the local sample magnetization  ~ M M  S ð ~ rr Þ below the tip apex. For an electronically homogeneoussurface,  C   and  P  S  are independent of the location  ~ rr : Therefore, any lateral variation of the d I  = d U   signal willbe caused by the (cos  y ) term in this case, which—for afixed tip magnetization  ~ M M  T  —will be determined by thelocal orientation of the sample magnetization  ~ M M  S ð ~ rr Þ : For our study we have chosen the model system Fe/W(110) at different coverages. Here, we will focusexclusively on the pseudomorphically grown monolayerof Fe on W(110) and on three-dimensional Fe islandsgrown in the Stranski–Krastanov mode on W(110).The pseudomorphically grown monolayer (ML) of Feon W(110) was already known to exhibit a large in-plane uniaxial magnetic anisotropy which should favorthe existence of very narrow domain walls. It was aprimary goal to prove the existence of such narrow wallsby a high-resolution magnetic imaging technique. On theother hand, three-dimensional islands of appropriatesize exhibit a magnetic vortex state with a curling in-plane magnetization around and an out-of-plane mag-netization inside the core region. Due to the small size of this core region, it could not be resolved so far by othermagnetic imaging techniques. 3. Results and discussion First, we studied 1.25ML Fe on a stepped W(110)substrate. Fig. 1 shows the topography (a) and the spin-resolved d I  = d U  -map (b,c) for a 100nm  100nm surfacearea. The domain structure of the Fe monolayer stripes is ARTICLE IN PRESS Fig. 1. (a) Topography and (b) spin-resolved d I  = d U  -map of 1.25ML Fe on stepped W(110) as measured with an Fe-coatedprobe tip being sensitive to the in-plane component of thesample magnetization. Adjacent monolayer stripes exhibitopposite in-plane magnetization directions. Obviously, thedomain walls in the monolayer are extremely sharp. (c)Rendered perspective representation of the topographic datacolorized by the magnetic d I  = d U  -signal. (d) Spin-resolvedd I  = d U  -signal as measured along the line shown in (b). Thetransition width between two adjacent domains is found to beas small as 0.6 7 0.2nm. R. Wiesendanger et al. / Journal of Magnetism and Magnetic Materials 272–276 (2004) 2115–2120 2116  clearly visible as dark and bright regions in Figs. 1(b) and(c). Since the Fe coverage exceeds one monolayer, there arealso narrow double-layer stripes present on this surface.In contrast to the Fe monolayer, these double-layerstripes exhibit a perpendicular anisotropy. Therefore,with a tip being sensitive to the in-plane component of magnetization only the domain walls of the Fe doublelayer would produce magnetic contrast. This can nicelybe seen in Fig. 1(b) where the location of the domainwalls of the Fe double layer is strongly correlated with thelocation of the domain walls of the Fe monolayer. Inaddition, dark spots are visible on the Fe monolayerwhich are caused by residual surface contamination.From STM data the actual amount (1–2%) of surfaceimpurities is usually overestimated because a singleimpurity distorts the local electronic structure of thesample surface with a decay length being a factor of 2–3larger than the actual size of the impurity. To analyze thedomain wall width for the Fe monolayer, a line sectionhas been drawn along the line indicated in Fig. 1(b). Thecorresponding data are plotted in Fig. 1(d). The regionaround the position of the wall is plotted with anexpanded lateral scale in the inset. A fit on the basis of micromagnetic theory would result in a domain wallwidth  w  ¼  0 : 6 7 0 : 2nm corresponding to approximatelytwo lattice constants. However, since the STM has beenoperated at relatively low tunneling currrent of  I   ¼  0 : 8nA and a relatively high bias voltage of  U   ¼  130mV, the Tersoff–Hamann theory [18] wouldpredict a lateral resolution of STM on the order of 0.4– 0.5nm. This value is close to the measured transitionwidth and therefore the instrumental resolution functioncannot be neglected in this case. Several arguments can begiven supporting the existence of atomically sharpdomain walls in the Fe monolayer: First, if we assumethat the 180  domain wall in the Fe monolayer onW(110) extends over several lattice sites then themagnetization in the center of the wall must beperpendicular to the  ½ 1  % 1 0  -direction which is the easyaxis of the Fe monolayer. Therefore, the magnetizationwould either point along the surface normal [110]- oralong the [001]-direction for the two possible wall types,i.e. Bloch or N ! eel walls, respectively. In both cases, thewalls in the monolayer would appear as bright or darklines when a tip is used being sensitive to the respectivemagnetization direction in the center of the wall. Anobservation like this has never been made which—due tothe large number of experiments that have beenperformed with many different tips—excludes the aboveconsidered hypothesis. Second, atomically sharp domainwalls in the Fe monolayer can also be expected on atheoretical basis. The Fe monolayer exhibits an extremelylarge value of the anisotropy energy [10] being aboutthree orders of magnitude larger than in bulk Fe. Thespin configuration of domain walls in materials with largemagnetic anisotropy has been theoretically investigatedby Hilzinger and Kronm . uller [19]. It was found that themicromagnetic continuum theory is a good approxima-tion of the spin configuration within the domain wall aslong as the width of the domain wall exceeds three latticeconstants. A completely different behavior was proposed,however, if the domain wall width becomes smaller thantwo atomic distances which is the case for the pseudo-morphic Fe monolayer on W(110). In this case, thecalculations suggest that the spin rotation does not takeplace continuously, but by an abrupt 180  rotationbetween two lattice sites, probably half-way between theatoms where the total spin-density is minimal. We wouldlike to point out that atomically sharp domain walls areonly observed on the Fe monolayer.If the Fe coverage is increased to two monolayers theobserved wall width is on the order of a few nanometersas shown in Fig. 2. On the other hand, the existence of  ARTICLE IN PRESS Fig. 2. (a) Topography and (b) spin-resolved d I  = d U  -map of 2ML Fe on stepped W(110) as measured with a Gd-coated tipbeing sensitive to the perpendicular component of the samplemagnetization. The sample exhibits a stripe domain phase alongthe  ½ 1  % 1 0  -direction. (c) Rendered perspective image of thetopographic data shown in (a) colorized with the magneticd I  = d U  -signal. (d) Spin-resolved d I  = d U  -signal measured alongthe line shown in (b). The magnetic periodicity is E 50nm andthe transition width is typically on the order of severalnanometers. R. Wiesendanger et al. / Journal of Magnetism and Magnetic Materials 272–276 (2004) 2115–2120  2117  atomically sharp domain walls in the Fe monolayer doesnot depend on the details of the surface morphology asexpected.In Fig. 3 we present the topography (a) and the spin-resolved d I  = d U  -map (b) of 5ML Fe on W(110) grownin the Stranski–Krastanov mode. Between the 3D Fe-islands a wetting layer of single atomic height is presentwhich clearly reveals a magnetic domain structure inFig. 3(b).Again, the transitions between the magnetic domainsin the Fe monolayer appear atomically sharp similar tothe case of the nanostripe system (Fig. 1(b)). Interest-ingly, the magnetization state of the single-domainislands in Fig. 3(b) is strongly correlated with thedomain structure of the Fe monolayer. In this case, thetypical island height is 2–5nm. If this island height isincreased to 8–9nm the magnetic ground state becomesa vortex (Fig. 4). The dimensions of the islands are nowtoo large to form a single-domain state because it wouldcost a relatively high stray field energy [20]. On the other hand, the islands are too small to form domains similarto those found in macroscopic pieces of iron because theadditional cost of domain wall energy cannot becompensated by the reduction of stray field energy.Instead, the magnetization continuously curls aroundthe center of the island, thereby reducing the stray fieldenergy and avoiding domain wall energy. However, inthe vortex core center, the magnetization has beenpredicted to turn into the surface normal because theresulting cost in stray field energy is overcompensated bythe gain in exchange energy. This has indeed beenconfirmed by magnetic force microscopy (MFM) onlithographically defined permalloy disks [21,22]. How-ever, the spatial resolution of MFM turned out to beinsufficient to resolve the internal spin structure of themagnetic vortex core. By making use of Cr-coatedSPSTM tips with a different thickness of the coatinglayer we could map out both the curling in-planemagnetization around the vortex core (Fig. 5(a)) as wellas the perpendicular magnetization within the vortexcore (Fig. 5(b)). The spin-resolved d I  = d U  -signal asmeasured along a circular path at a distance of 19nmfrom the vortex core center clearly reveals a cosine-likemodulation (Fig. 5(c)) indicating that the in-planecomponent of the sample magnetization indeed con-tinuously curls around the vortex core. Line sectionsgoing through the center of the vortex core for both in-plane and out-of-plane component images (Figs. 5(a)and (b)) are presented in Fig. 5(d). They result in a consistent value for the intrinsic width of the magnetic ARTICLE IN PRESS Fig. 3. (a) Topography and (b) spin-resolved d I  = d U   map of 5ML Fe on W(110) grown in the Stranski–Krastanov modeand measured with an Fe-coated tip being sensitive to the in-plane component of the sample magnetization. The Fe wettinglayer between the three-dimensional (3D) islands exhibitsatomically sharp magnetic domain boundaries similar to theFe nanostripe system shown in Fig. 1. The magnetization stateof the single-domain 3D islands closely follows the magnetiza-tion state of the Fe monolayer in the surrounding.Fig. 4. (a) Spin-resolved d I  = d U   map of 7ML Fe on W(110)grown in the Stranski–Krastanov mode and measured with aCr-coated tip being sensitive to the in-plane component of thesample magnetization. In contrast to Fig. 3 the 3D islands arehigher in this case and exhibit a magnetic vortex state, as clearlyvisible in (b). R. Wiesendanger et al. / Journal of Magnetism and Magnetic Materials 272–276 (2004) 2115–2120 2118  vortex core being  w  ¼  9 7 1nm. This experimentallydetermined value is in excellent agreement with theory[23] which predicted that the size of the vortex core isgoverned by only two material parameters, the exchangestiffness and the saturation magnetization that deter-mines the stray field energy. 4. Summary and conclusions Recent advances in SP-STM and SP-STS have led toan unprecedented insight into magnetic phenomenaoccuring on a scale comparable to or beyond theexchange length. Most striking examples are thediscovery of atomically sharp magnetic domain wallsin Fe monolayers and the determination of the internalspin structure of magnetic vortex cores. The firstexample is of relevance in view of future trends inmagnetic recording in which atomically sharp transi-tions between differently magnetized regions are neededto reach the ultimate limit of recording density. On theother hand, magnetic vortex states play a crucial role innon-volatile magnetic storage cells (MRAM) where thedetails of the spin structure of the vortex core have to beknown in order to control the stray-field couplingbetween neighboring cells. Acknowledgements We thank H.J. Elmers, U. Gradmann, and M. Pratzerfor valuable discussions. Financial support from theDeutsche Forschungsgemeinschaft (Grant Wi1277/19-1,SFB 508, and Graduiertenkolleg ‘‘Physik nanostruktur-ierter Festk . orper’’) is gratefully acknowledged. References [1] R. Wiesendanger, H.-J. G . untherodt, G. G . untherodt,R.J. Gambino, R. Ruf, Phys. Rev. Lett. 65 (1990) 247.[2] R. Wiesendanger, I.V. Shvets, D. B . urgler, G. Tarrach,H.-J. G . untherodt, J.M.D. Coey, S. Gr . aser, Science 255(1992) 583.[3] R. Koltun, M. Herrmann, G. G . untherodt, V.A.M. Brabers,Appl. Phys. A 73 (2001) 49.[4] S. Heinze, M. Bode, O. Pietzsch, A. Kubetzka, X. Nie,S. Bl . ugel, R. Wiesendanger, Science 288 (2000) 1805.[5] H. Yang, A.R. Smith, M. Prikhodko, W.R.L. Lambrecht,Phys. Rev. Lett. 89 (2002) 226101.[6] M. Bode, M. Getzlaff, R. Wiesendanger, Phys. Rev. Lett.81 (1998) 4256.[7] O. Pietzsch, A. Kubetzka, M. Bode, R. Wiesendanger,Phys. Rev. Lett. 84 (2000) 5212.[8] M. Kleiber, M. Bode, R. Ravli ! c, R. Wiesendanger, Phys.Rev. Lett. 85 (2000) 4606.[9] T.K. Yamada, M.M.J. Bischoff, G.M.M. Heijnen,T. Mizoguchi, H. van Kempen, Phys. Rev. Lett. 90(2003) 056803.[10] M. Pratzer, H.J. Elmers, M. Bode, O. Pietzsch,A. Kubetzka, R. Wiesendanger, Phys. Rev. Lett. 87(2001) 127201.[11] A. Kubetzka, M. Bode, O. Pietzsch, R. Wiesendanger,Phys. Rev. Lett. 88 (2002) 057201.[12] A. Wachowiak, J. Wiebe, M. Bode, O. Pietzsch, M.Morgenstern, R. Wiesendanger, Science 298 (2002) 577.[13] O. Pietzsch, A. Kubetzka, M. Bode, R. Wiesendanger,Science 292 (2001) 2053.[14] M. Kleiber, M. Bode, R. Ravli ! c, N. Tezuka,R. Wiesendanger, J. Magn. Magn. Mater. 240 (2002) 64. ARTICLE IN PRESS Fig. 5. Spin-resolved d I  = d U  -maps as measured with an (a) in-plane and (b) out-of-plane sensitive Cr-coated tip. The curlingin-plane magnetization around the vortex core region is visiblein (a), whereas the perpendicular magnetization componentwithin the central vortex core region is recognizable in (b) as abright area. (c) Spin-resolved d I  = d U  -signal around the vortexcore at a distance of 19nm (circle in (a)). (d) Spin-resolvedd I  = d U  -signal along the lines in (a) and (b) showing consistentlyan intrinsic width of the vortex core region of 9 7 1nm. R. Wiesendanger et al. / Journal of Magnetism and Magnetic Materials 272–276 (2004) 2115–2120  2119
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x