1 st International Conference on Quantum Photonic Science

1 st International Conference on Quantum Photonic Science Nonlinear Magneto-Optical Studies in Magnetic Superlattices and Magnetic Nano Structures K. Sato, A. Kodama, M. Miyamoto, M. Tsuruga, T. Matsumoto, T. Ishibashi Y. Morishita, Department of Applied Physics, Tokyo University of Agriculture and Technology, Koganei, Tokyo, Japan K. Takanashi, S. Mitani, Institute of Materials Science, Tohoku University, Sendai, Miyagi, Japan TUAT COE Project “Future Nano Materials”

Nonlinear Magneto-optics • What is the nonlinear magneto-optical effect? Magnetization-induced nonlinear optics • What is the nonlinear Kerr rotation? When P-polarized primary light is incident both Pand S-polarized SH light emits: which leads to rotation of E vector from the plane of incidence. • In centrosymmetric materials such as Fe and Au no SHG occurs due to cancellation of P and –P.

Azimuthal angle dependence of SHG from Si and Ga. As wafer Si wafer (001) centrosymmetric Ga. As wafer (001) Non-centrosymmetric

Theoretical prediction and experimental verification • Nonlinear magneto-optical Kerr rotation larger than theoretically linear rotaion predicted 1), and was experimentally proved 2, 3). 1) W. Hübner and K. -H. Bennemann: Phys. Rev. B 40, 5973 (1989) 2) Th. Rasing et al. : J. Appl. Phys. 79, 6181 (1996) 3) Th. Rasing: J. Mag. Soc. Japan 20 (Suppl. S 1), 13 (1996)

Surface and interface sensitivity of MSHG • Application of MSHG Sensitive to Evaluation of the break of Multilayers symmetry at surface Imaging of domains ・This effect cannot be expected to be applied to some practical memory devices but is thought to be useful for characterization of surfaces and interfaces of materials.

Nonlinear magneto-optical effect ・ For weak incident laser field E(w) : linear e respons ・ For strong incident laser field E(w) : ar Nonline e respons Third rank tensor is not allowed in centrosymmetric materials. ・ Nonlinear polarization 　P(2) for incident field of E=E 0 sinwt Second harmonic generation (SHG)

Nonlinear polarization of 2 nd order parametric process Light rectification SHG process

Definition of nonlinear susceptibility Centrosymmetric materials: 　all the χijk(2) components vanish. 　　　　　　(from symmetry operations) Surfaces and interfaces: symmetry breaks, leading to appreciable amount of nonlinear magneto-optical effect even in the centrosymmetric materials

Wave equation of linear magnetooptical effect YK=f. K+ih. K　（複素カー回転角） c 1(1)=eyz, c 0(1)=exx-1=N 2 -1

Wave equation of nonlinear magneto-optics. Source term does not depend on optical constants of materials, leading to special solution associated with the second order susceptibility.

Nonlinear Kerr roation

Nonlinear Kerr rotaion Different from the linear case χ(2)odd/χ(2)even contributes。 This term is zero in centrosymmetric materials And takes a finite value at surfaces Surface sensitivity　　useful for surface magnetism studies！

Difference between linear and nonlinear Kerr rotation 　Linear： 　　　　　　factor reduces the magnitude Also χxy is order of magnitude smaller that χxx 　Nonlinear：no such factor exists Also χodd and χeven are of the same order 。

Microscopic origin of MSHG 3 photon proicess Here |k q//l >→|k+q//l‘>　|k+q//l’>→|k+2 q//l" > w w Ground state |kl> →|k+2 q//l”>　 2 w Intermediate state Excited state

Illustration of microscopic process of MSHG

Nonlinear Kerr rotation of Fe

Nonlinear Kerr rotation of Fe/Cu Nonlinaer Kerr effect nolinear

Cu cover layer-thickness dependence of Co/Cu SH信号 Cuの層厚

Superlattices ：〔Fe(x. ML)/Au(x. ML)〕N Integer : x=1, 2, 3, 4, 5, 6, 8, 10, 15 Non-integer : x=1. 25, 1. 75, 2. 25, 2. 75, 3. 25, 3. 75 N { 3 { 2 { 1 { Au(x. ML) Fe(x. ML) Au buffer Fe seed 2 Au 1 Fe Mg. O(100) 0 0. 25 0. 75 x=3. 75 ML 1

Fe(1 ML)/Au(1 ML) superlattice N Au(x. ML) 2 Fe(x. ML) Au buffer layer Fe seed layer 1 bcc-Fe (001) 4. 054Å Au Mg. O (100) Fe [Fe(1 ML)/Au(1 ML)] 　Schematic structure for the Fe/Au superlattice 2. 867Å L 10 fcc-Au (001) 4. 079Å Atomic arrangement in a unit cell of Fe-Au with a L 10 suructure.

MSHG Measureing System LD pump SHG laser Electromagnet =532 nm Filter Stage　 controller Ti: sapphire laser =810 nm Pulse=150 fs P=600 m. W rep 80 MHz Mirror Berek compensator Mirror Sample Analyzer Lens Filter PMT lens polarizer Chopper Photon counting Photon counter Computer

Laboratory • Experimental setup for MSHG measurement

Sample 試料回転 Sample stage Pole p ie w (810 nm) ce 45° Rotating analyzer Filter 2 w (405 nm) P-polarized or Spolarized light w (810 nm) Analyzer Optical Setups (Longitudinal Kerr)

lt u s Re Nonlinear Kerr Rotation and Kerr Ellipticity Electromagnet SHG intensity (counts/10 sec. ) 104 2 K(2)=34. 3 S-polarized light ω(810 nm) Rotating Analyzer 45° Analyzer Filter 2 w (405 nm) The curves show a shift for two opposite directions of magnetic field Analyzer angle (deg. ) Analyzer angle-dependence for [Fe(3. 5 ML)/Au(3. 5 ML)] (Sin) Nonlinear Kerr rotation & ellipticity K(2)= 17. 2 h. K(2)=３

Largest nonlinear Kerr rotation observed in the Fe/Au series SHG intensity (counts/10 sec. ) Df = 31. 1° Analyzer angle (deg. ) Fe(1. 75 ML)/Au(1. 75 ML)　Sin

Azimuthal Angle Dependence Electromagnet Rotation of sample ・　Linear optical response　( =810 nm) 　　　The isotropic response for the azimuthal angle ・　Nonlinear optical response ( =405 nm) 　　　The 4 -fold symmetry pattern 　　　Azimuthal pattern show 45 -rotation by reversing the magnetic field li Analyzer 45° Filter r 2 w (405 nm) a e nlin no (a)　Linear (810 nm) 45 SHG intensity (counts/10 sec. ) r a e n P-polarized light w (810 nm) (b)　SHG (405 nm) Azimthal angle-dependence of MSHG intensity for [Fe(3. 75 ML)/Au(3. 75 ML)] superlattice. 　 （Pin Pout）

ence hal Azimut pend e D e l g An Pin-Pout Sin-Sout

D n o i s s u c is The equation of the azimuthal angledependence by theoretical analysis term ris e tic e c v i a f gne igin g r u S or a : e l m A on ipo n ic d e to r lect al. e n g e Th c si i p 　・ ro isot term ic an an t e agn uses pic o r t iso m n o n ca n or. i s g i n r k upole o r rank te l u B: Bhe quadr n for fou T utio b i 　 ・ r t con tion a z i ith et w n m g d r e ma nte t e y b c m ti ple e ted f p i l n u is es ag b try d e m m e oul m h c y s s fa ersal r ons i u t a C: S e time rev etry oper n of M. Th mm rsio y e s 　・ v r l re irro a n M itio 　・ d d a an

Calculated azimuthal angle dependence of SHG and MSHG signals Kerr rotation calculated from parameters Axx, B, C Sin Pin

　Surface non-magnetic term ・SHG response causes an isotropic contribution only. 　Bulk non-magnetic term ・　For crystallographic contribution the electric quadrupole should be introduced to get four rank tensor. SHG response causes an anisotropic contribution (parameter B). Surface magnetization induced term ・　The surface magnetic response comes from the electric dipole term expanded by magnetization and contributes to the parameter C.

Calculated polar patterns of the azimuthal angle-dependence (Sin-Pout) ty i l a c o l t non u o h t i W (a) A=5, B=0, C=0. 85 With ty ali c o l n no The equation of the azimuthal angledependence by theoretical analysis (a) A=5, B=0, C=0. 85 　　・For B much smaller than C, the polar pattern shows 45 rotation for the magnetization reversal. (b) A=5, B=0. 85, C=0. 85 　　･For B comparable C, the polar pattern undergo a smaller rotation. (b) A=5, B=0. 85, C=0. 85 The azimuthal pattern was interpreted in terms of combination of B and C.

SHG intensity (counts/10 sec. ) Azimuthal angle-dependence of MSHG for a [Fe(3. 5 ML)/Au(3. 5 ML)] superlattice (Sin-Pout, Sin-Sout configuration) 103 The equation of the azimuthal angledependence by theoretical analysis Sin-Pout Sin-Sout SHG intensity (counts/10 sec. ) (a) Sin-Pout 103 (b) Sin-Sout ASP(surface nonmagnetic term) =　460 ASS(surface nonmagnetic term) =　100 B(bulk nonmagnetic term) =　26 C(surface magnetic term) =　-88

Calculated and experimental patterns : x=3. 5 SHG intensity (counts/10 sec. ) (a) Pin-Pout 103 (b) Pin-Sout 103 APP=1310, B=26, C=-88 APS=-300, B=26, C=-88 (c) Sin-Pout (d) Sin-Sout 103 ASP=460, B=26, C=-88 ASS=100, B=26, C=-88 Dots：exp. Solid curve：calc.

Nonlinear Kerr rotation (deg. ) (a) Experimental pattern (Sin) Nonlinear Kerr ellipticity Nonlinear Kerr rotation (deg. ) Calculated and experimental pattern of Nonlinear Kerr rotation and ellipticity (c) Exp. (d) Calc. Azimuthal angle (deg. ) (b) Calculated pattern (Sin) The azimuthal angle-dependences of nonlinear Kerr rotation angle and ellipticity in [Fe(3. 75 ML)Au(3. 75 ML)]

Experimental and calculated patterns of Fe(2. 75 ML)/Au(2. 75 ML) (b) Calculation Experiment Nonlinear Kerr rotation (deg. ) (a) Fe(3. 25 ML)/Au(3. 25 ML) Nonlinear Kerr rotation (deg. ) Fe(3. 25 ML)/Au(3. 25 ML) Fe(2. 5 ML)/Au(2. 5 ML) Nonlinear Kerr rotation (deg. ) Fe(2. 5 ML)/Au(2. 5 ML) (b) Calculation Nonlinear Kerr rotation (deg. ) (a) Experiment Nonlinear Kerr rotation (deg. ) Kerr rotation angle Fe(3. 5 ML)/Au(3. 5 ML) Sin configuration: (a) Experimental data, (b) Calculated using parameters determined by fitting to the azimuth patterns

Nonlinear Kerr rotation angle (deg. ) Nonlinear Kerr rotation angle of [Fe(x. ML)/Au(x. ML)] (1. 25 x 3. 75) superlattices (Sin) Calculation and experimental result K(2)=31. 1 (a) Exp. Calculated nonlinear Kerr rotation angle K(2) using the fitting parameter ASP, ASS, B, C of the azimuthal pattern (The maximum K(2) was selected for azimuth angle) (b) Calc. ･ The experimental maximum K(2) for x=1. 75 superlattice was 31. 1. ・ The calculated K(2) reproduced the muximum K(2) for x=1. 75 superlattice. x=1. 75 Modulated rate x (ML) Fig. Nonlinear Kerr rotation angle of [Fe(x. ML)/Au(x. ML)] (1. 25 x 3. 75) superlattices [(a)Calculation, (b)Experiment] The nonlinear Kerr rotation was explained by theoretical analysis.

Summary: MSHG of Fe/Au superlattice The four-fold pattern clearly reflects the symmetry of the Mg. O(100) substrate. This suggests that the Fe/Au superlattice is perfectly epitactic to the substrate. ・　The azimuthal angle dependence was analyzed in terms of nonlinear electrical susceptibility tensor taking into account the magnetic symmetry of the superlattice. ･The azimuthal pattern was explained by symmetry analysis, taking into account the surface nonmagnetic A, bulk non-magnetic B and surface magnetic C contributions.

Summary (cont’d) was shown to lead to a nonlinear Kerr rotation (2)K that can be orders of magnitude larger than its linear equivalent (0. 2 ), e. g. , (2)K for x=1. 75 was 31. 1 • We observed azimuthal angle-dependence of the nonlinear Kerr rotation for the first time. ・　The azimuthal angle-dependence of the nonlinear Kerr rotation were explained using parameters determined from azimuthal patterns of MSHG response • Modulation period dependence of parameters: • A (Surface nonmagnetic) is large for short period • B (Bulk nonmagnetic) is nearly constant • C (Surface magnetic) becomes larger with modulation Period. ・　MSHG

Fabrication of permalloy nanostructure by Damascene technique ①Preparation of substrate: Spin-coating of ZEP resist with high etching resistance ②EB-exposition: Write patterns by EB ③Development: Formation of mask-pattern by development ④Etching： By dry-etching process mask-pattern is transferred to the substrate ⑤Deposition of magnetic film: Deposition of magnetic films by sputter or evaporation ⑥Polishing: Obtain flat buried structure using chemicalmechanical polishing Process is simplified by abbreviation of lift-off and repeated spin-coating

EB-patterning process Spin coating of resist EB exposure Development Si substrate 〔１〕Dot size 　　　100 nm× 300 nm rectangular dot with 300 nmspacing 　　　100 nm square dot with 300 nm-spacing 〔２〕Patterned area: 3 mm× 3 mm 〔３〕EB-resist thickness: 300 nm 　　　・・・by spin-coating with 5000 rpm rotation 〔４〕Baking　160℃　20 min

Clean Room Laboratory • Electron beam lithography

Dry etching process Etching 〔１〕Etching gas: Resist removal CF 4 〔２〕Vacuum　3. 0× 10 -3 Pa 〔３〕Gas pressure 9. 2 Pa 〔４〕RF power: 400 W 〔５〕Etching rate: 0. 1μm/min 300 nm 100 nm Silicon surface after etching

Dry-etching

Laboratory EB deposition RF magnetron sputtering

Embedding of permalloy 〔１〕material: 　permalloy（Ni 80 Fe 20） 〔２〕Vacuum　3. 0× 10 -6 Torr 〔３〕Accelerating voltage 4 k. V Embedding of permalloy 〔４〕Deposition rate　 1. 0Å/sec film by electron beam deposition Chemical mechanical polishing 〔１〕Polishing chemicals: 　Si wafer flatting 　grain-size～ 20 nm 〔２〕p. H　11 〔３〕polishing rate: 　60 nm/min

Observation • ＡＦＭ/MFM FE-SEM

SEM observation 300 nm× 100 nmsquare dot, 300 nm space 3μm 0. 6μm

Cross sectional SEM observation Dot depth? 100 nm

Cross section SEM image of Line and space pattern (width =100 nm) ０．３μm

MFM observation of a permalloy film

AFM and MFM observation 1μm AFM Line scan ・・・Surface roughness~10 nm MFM image ・・・magnetization axis along the longer side direction

Comparison between two scans after magnetization in opposite direction 5 k. Oe

MFM-image for different scanning direction

Scan-direction dependence

Pattern variation with scan direction 0° 15° 30° 45° 60° 75° 90°

VSM measurement 0. 0004 M(emu) 0. 0002 0 -0. 0002 -0. 0004 -2 Longer axis Shorter axis -1 0 H(k. Oe) In-plane 1 2 Perpendicular

100 nm circular dots with 300 nm spacing 0. 5μm SEM AFM Surface roughness ~10 nm

VSM measurement of circular dot array Parallel to the plane Perpendicul ar to the plane

MFM measurement of circular dots Demagnetized Magnetic field applied Perpendicular to the plane

Influence of stray field from the MFM probe tip AFM sensing (23 nmlevitation) MFM probe A B Magnetization Recording by first scan 80 nm MFM measurenment A B Magnetization Reading by second scan

Models to explain MFM images MFM image A B A C B Magnetization MFM image Magnetization

1 m square dot array AFM MFM

MFM image of 300 nm x 100 nm dot with a low-moment probe tip AFM MFM

300 nm x 100 nm dot （wide scan） with a low-moment probe tip AFM MFM

Simulation by Nakatani

MSHG studies of Nano-structured magnetic patterns

Azimuthal angle dependence of SHG from unpatterned permalloy film Pin. Pout (counts/10 sec) Longitudinal Unstructured permalloy film: H=± 2. 5 k. Oe

Azimuthal angle dependence of SHG from unpatterned Si wafer (counts/10 sec) Pin. Pout H=± 2. 5 k. Oe

Azimuthal angle dependence of SHG from Ga. As wafer

Azimuthal angle dependence of MSHG from 1 m square dot array (counts/10 sec) Pin. Pout H=± 4 k. Oe Longitudinal Kerr configuration

Azimuthal angle dependence of MSHG from 300 nm x 100 nm rectangular dot array (Longitudinal) Sin. Pout (counts/10 sec) Pin. Pout H=± 4 k. Oe Longitudinal configuration

Azimuthal angle dependence of MSHG from 300 nm x 100 nm rectangular dot array (Polar) (counts/10 sec) Pin. Pout H=± 6 k. Oe polar Kerr configuration

(counts/10 sec) Sin. Sout (counts/10 sec) Sin. Pout (counts/10 sec) Pin. Sout (counts/10 sec) Pin. Pout 丸ドット(100 nm)の方位角依存性：磁場± 4 k. Oe 縦カー配置