Synchrotron X-ray imaging of nanomagnetism in meteoritic …...generalized to electric-multipole...

15
Synchrotron X-ray imaging of nanomagnetism in meteoritic metal AGU Fall meeting 2013, Rock Magnetism: Beyond the State-of-the-Art James Bryson , Claire Nichols, Julia Herrero Albillos, Florian Kronast, Tolek Tyliszczak, Simon Redfern, Gerrit van der Laan and Richard Harrison

Transcript of Synchrotron X-ray imaging of nanomagnetism in meteoritic …...generalized to electric-multipole...

  • Synchrotron X-ray imaging of nanomagnetism in meteoritic metal !

    AGU Fall meeting 2013, Rock Magnetism: Beyond the State-of-the-Art!

    James Bryson, Claire Nichols, Julia Herrero Albillos, Florian Kronast, Tolek Tyliszczak, Simon Redfern, Gerrit van der Laan and Richard Harrison !

  • Introduction!

    To date, nanopaleomagnetism has been studied mostly using high-resolution magnetic microscopies, for example electron holography:!
Advantages!-  Very-high spatial resolution (1 - 2 nm)!-  Provides images of magnetic dipolar interactions !!Disadvantages!-  Excessive off-line processing; inhibits measurements of

    dynamic phenomena!-  Sample experiences large magnetic fields during

    preparation; restricts study of natural remanences!-  Images magnetic induction, not magnetisation!!Poster, Today: Nanopaleomagnetism of Meteoritic Fe-Ni: the Potential

    for Time-Resolved Remanence Records within the Cloudy Zone.!GP41D: Understanding Planetary and Stellar Magnetic Fields !

    Bryson et al., EPSL (2014)!

  • X-ray Magnetic Circular Dichroism (XMCD)!Principles of X-Ray Magnetic Dichroism Spectromicroscopy 1301

    Fig. 4. Principles of X-ray magnetic circular dichroism spectroscopy, illustrated for the case of L edge absorption in a dband transition metal. In a magnetic metal the d valence band is split into spin-up and spin-down states with differentoccupation. Absorption of right (left) circularly polarized light mainly excites spin-up (spin-down) photoelectrons.Since spin flips are forbidden in X-ray absorption, the measured resonance intensity directly reflects the number ofempty d band states of a given spin. In XMCD spectroscopy it is equivalent whether the photon polarization is changedand the magnetization direction is kept fixed or whether the magnetization direction is changed and the photon helicityis fixed. The corresponding XMCD spectra for Fe metal44 are shown on the right for three different orientations of themagnetization directions relative to the fixed photon spin (right circular polarization).

    photon spin and magnetization directions are quan-titatively related by sum rules to the size of thespin and orbital magnetic moments15,16 and to theanisotropies of the spin density and orbital mo-ment.17 XMCD spectroscopy can therefore deter-mine the sizes, the directions, and anisotropies(sizes in different directions) of the atomic magneticmoments.

    2.2. XMLD spectroscopy

    It is well known that linearly polarized X-rays canprobe the orientation of molecular orbitals. Thisis one of the strengths of near edge X-ray absorp-tion fine structure, (NEXAFS), spectroscopy.18 Moregenerally, polarized X-ray absorption can sense the

    charge anisotropy of the valence states involved inthe core excitation process.19,20 The electric fieldvector E of the linearly polarized X-rays acts asa searchlight for the number of valence holes indifferent directions of the atomic volume or theWigner–Seitz cell. In most cases the anisotropy ofthe charge in the atomic volume is caused by ananisotropy in the bonding, i.e. by the electrostaticpotential.

    A lesser-known effect, which is the basis ofXMLD spectroscopy, arises from the presence ofa magnetic anisotropy in the sample.7–10 Let usconsider the case of NiO. Because of the cubicsymmetry of the lattice the charge distributionaround the atoms is nearly spherical and no lineardichroism effect exists in the absence of magnetic

    Nanomagnetism can be imaged directly by utilising circularly polaised X-rays:!⁃  XMCD intensity is a

    measure of the projection the of magnetisation onto the X-ray beam direction!

    ⁃  Can study dynamic processes!

    ⁃  Element specific!

    Stohr et al., Surface review and Letters (1998)!

    (c) L3

    L2 Fe XMCD

    Figure 1. (a) Magneto-optical Faraday and Kerr e↵ect. (b) XAS/XMCD e↵ect in core-levelspectroscopy. (c) Relative contributions to the energy dependence of the Fe 2p XMCD spectrumarising from the di↵erent ground state moments for iron metal. n

    h

    = number of holes, ` · s =scalar part of spin-orbit interaction, L

    z

    = orbital moment, Sz

    = spin moment and Tz

    = magneticdipole term. The red and blue spectra relate to the sum rules for the orbital and spin magneticmoments, respectively. The upper spectrum shows the sum of all contributions, which resemblesthe experimental Fe L2,3 XMCD spectrum. For further details see Ref. [27].

    interaction, ` · s [28, 29, 30, 31]. This sum rule for the isotropic spectrum has recently beengeneralized to electric-multipole transitions, which are important in nonresonant inelastic x-rayscattering (NIXS) [32].

    Sum rules for XMCD enable us to separate the orbital moment from the spin moment[3, 33, 34]. Although in 3d transition metals the orbital moment is usually small (⇠10% ofthe total magnetic moment), it represents that part of the magnetization which is coupledto the lattice and, therefore, determines magnetic properties such as the easy-direction ofmagnetization, the magneto-crystalline anisotropy and the magnetostriction.

    It is useful to put the sum rules in a more general perspective. X-ray absorption [35, 34] andphotoemission [36] provide a whole set of sum rules, relating integrated spectral intensities toground state properties . The reason that excited states (typically at 1 keV) still give informationabout ground state properties is that by integrating over all possible states of the core hole itsinfluence is averaged out. A necessary condition for this is that the core level is characterized bya good quantum number, such as the total angular momentum j. In other words, the jj mixingbetween the L3 and L2 edges must be small, which is fulfilled when the core spin-orbit splittingis much larger than the core-valence electrostatic interaction [37].

    The moment transferred by the light determines which combinations of ground state operatorsare accessed. The integral over the isotropic intensity reveals the number of valence holes. Theintegral over the circular dichroism is related to the orbital magnetic moment, and the linear

    4

    Relative photon energy (eV)

    Figure 1. (a) Magneto-optical Faraday and Kerr e↵ect. (b) XAS/XMCD e↵ect in core-levelspectroscopy. (c) Relative contributions to the energy dependence of the Fe 2p XMCD spectrumarising from the di↵erent ground state moments for iron metal. n

    h

    = number of holes, ` · s =scalar part of spin-orbit interaction, L

    z

    = orbital moment, Sz

    = spin moment and Tz

    = magneticdipole term. The red and blue spectra relate to the sum rules for the orbital and spin magneticmoments, respectively. The upper spectrum shows the sum of all contributions, which resemblesthe experimental Fe L2,3 XMCD spectrum. For further details see Ref. [27].

    interaction, ` · s [28, 29, 30, 31]. This sum rule for the isotropic spectrum has recently beengeneralized to electric-multipole transitions, which are important in nonresonant inelastic x-rayscattering (NIXS) [32].

    Sum rules for XMCD enable us to separate the orbital moment from the spin moment[3, 33, 34]. Although in 3d transition metals the orbital moment is usually small (⇠10% ofthe total magnetic moment), it represents that part of the magnetization which is coupledto the lattice and, therefore, determines magnetic properties such as the easy-direction ofmagnetization, the magneto-crystalline anisotropy and the magnetostriction.

    It is useful to put the sum rules in a more general perspective. X-ray absorption [35, 34] andphotoemission [36] provide a whole set of sum rules, relating integrated spectral intensities toground state properties . The reason that excited states (typically at 1 keV) still give informationabout ground state properties is that by integrating over all possible states of the core hole itsinfluence is averaged out. A necessary condition for this is that the core level is characterized bya good quantum number, such as the total angular momentum j. In other words, the jj mixingbetween the L3 and L2 edges must be small, which is fulfilled when the core spin-orbit splittingis much larger than the core-valence electrostatic interaction [37].

    The moment transferred by the light determines which combinations of ground state operatorsare accessed. The integral over the isotropic intensity reveals the number of valence holes. Theintegral over the circular dichroism is related to the orbital magnetic moment, and the linear

    4

    van der Laan, Journal of Physics: Conference Series (2013)!

  • X-ray Microscopy

    1304 J. Stöhr et al.

    Fig. 6. Principles of scanning X-ray microscopy, shownin (a), and two imaging X-ray microscopy techniques,shown in (b) and (c). In the scanning mode (a) a smallX-ray spot is formed by a suitable X-ray optic, for ex-ample a zone plate as shown, and the sample is scannedrelative to the X-ray focal spot. The spatial resolution isdetermined by the spot size. The intensity of the trans-mitted X-rays or the fluorescence or electron yield fromthe sample are detected as a function of the sample po-sition and thus determine the contrast in the image. Inimaging transmission X-ray microscopy, shown in (b), acondensor zone plate in conjunction with a pinhole be-fore the sample produces a monochromatic photon spoton the sample. A micro-zone-plate generates a magnifiedimage of the sample which can be viewed in real timeby an X-ray-sensitive CCD camera. The spatial resolu-tion is determined by the width of the outermost zonesin the micro-zone-plate. In imaging X-ray photoelectronmicroscopy, shown in (c), the X-rays are only moder-ately focused in order to match the field of view of anelectron microscope. Electrons emitted from the sampleare projected with magnification onto a phosphor screenand the image can be viewed in real time at video rates.The spatial resolution is determined by the electron op-tics within the microscope, the size of the aperture andthe operation voltage.

    Scanning Transmission X-ray Microscopy (STXM)!

    Advantages:!-  High spatial

    resolution (best 25 nm)!

    -  High energy resolution!

    Disadvantages:!-  Cannot study natural

    remanences!-  Long acquisitions!-  Studies relatively

    small areas!

    Stohr et al., Surface review and Letters (1998)!

    Eisebitt et al., Nature (2004)!

    Out-of-plane worm-like magnetic domains – Co/Pt

    bilayer at Co absorption edge!

    250 nm

  • X-ray Microscopy

    1304 J. Stöhr et al.

    Fig. 6. Principles of scanning X-ray microscopy, shownin (a), and two imaging X-ray microscopy techniques,shown in (b) and (c). In the scanning mode (a) a smallX-ray spot is formed by a suitable X-ray optic, for ex-ample a zone plate as shown, and the sample is scannedrelative to the X-ray focal spot. The spatial resolution isdetermined by the spot size. The intensity of the trans-mitted X-rays or the fluorescence or electron yield fromthe sample are detected as a function of the sample po-sition and thus determine the contrast in the image. Inimaging transmission X-ray microscopy, shown in (b), acondensor zone plate in conjunction with a pinhole be-fore the sample produces a monochromatic photon spoton the sample. A micro-zone-plate generates a magnifiedimage of the sample which can be viewed in real timeby an X-ray-sensitive CCD camera. The spatial resolu-tion is determined by the width of the outermost zonesin the micro-zone-plate. In imaging X-ray photoelectronmicroscopy, shown in (c), the X-rays are only moder-ately focused in order to match the field of view of anelectron microscope. Electrons emitted from the sampleare projected with magnification onto a phosphor screenand the image can be viewed in real time at video rates.The spatial resolution is determined by the electron op-tics within the microscope, the size of the aperture andthe operation voltage.

    X-ray Photoemission Electron Microscopy (XPEEM)!

    Advantages:!-  Can study natural

    remanences!-  Capable of producing

    maps of all three spatial components of magnetisation!

    Disadvantages:!-  Lower spatial

    resolution (best 40 nm)!

    -  Surface sensitive (top 5 - 30 nm)!

    -  Sample surface must be conductive (coating)!

    Wu et al., Journal of Physics: Condensed Matter (2010)!

    Kittel structure – Co/Cu/Py/Cu(001) trilayer at Co absorption edge!

    500 nm!

    X-ray !direction!

    Stohr et al., Surface review and Letters (1998)!

  • 50 µm

    Kamacite!

    Tetrataenite (TT) rim!

    Cloudy Zone (CZ)!

    Plessite!

    Meteoritic Metal!

    Tazewell IIICD Iron meteorite!

    Widmanstatten microstructure !

  • The Cloudy Zone!

    500 nm!TT rim! CZ!

    Bryson et al., EPSL (2014)

    3D spinodal nanostructure, intergrown islands of tetrataenite (FeNi) and matrix (ordered Fe3Ni)!Over time islands and matrix evolve and coarsen !

  • The Cloudy Zone!

    500 nm!TT rim! CZ!

    Bryson et al., EPSL (2014)

    Tetrataenite (L10 superstructure) is an extremely hard magnetic phase and adopts one of 6 possible magnetisation directions. The proportion of the magnetisation directions depends on external fields.!

    Fe-Ni metals in meteorites - terminology

    Kamacitebcc

    Soft ferromagnet

    Taenitefcc

    Soft ferromagnet(> 25 wt% Ni)

    TetrataeniteOrdered FeNi (L10)Hard ferromagnetEasy axis // c

    AwaruiteOrdered FeNi3

    (L12)Soft

    ferromagnet

    ???Ordered Fe3Ni

    (L12)Soft

    ferromagnet?

    Fe-Ni metals in meteorites - terminology

    Kamacitebcc

    Soft ferromagnet

    Taenitefcc

    Soft ferromagnet(> 25 wt% Ni)

    TetrataeniteOrdered FeNi (L10)Hard ferromagnetEasy axis // c

    AwaruiteOrdered FeNi3

    (L12)Soft

    ferromagnet

    ???Ordered Fe3Ni

    (L12)Soft

    ferromagnet?

    Fe-Ni metals in meteorites - terminology

    Kamacitebcc

    Soft ferromagnet

    Taenitefcc

    Soft ferromagnet(> 25 wt% Ni)

    TetrataeniteOrdered FeNi (L10)Hard ferromagnetEasy axis // c

    AwaruiteOrdered FeNi3

    (L12)Soft

    ferromagnet

    ???Ordered Fe3Ni

    (L12)Soft

    ferromagnet?

  • Scanning transmission X-ray microscopy

    KamaciteTetrataenite

    rim Cloudy zone

    Easy axis 1Easy X-ray directionAxis 3

    EasyAxis 2

    Tazewell IIICD Iron meteorite!Magnetisation projected onto X-ray beam!

    White – zero signal, easy axis 1 !Blue – positive signal, easy axis 2!Red – negative signal, easy axes 3!

    1 µm!

    Advanced Light Source, Lawrence Berkeley National Laboratory!

  • a b c

    5 µm 1 µm 1 µm1 µm

    Photoemission electron microscopy!

    Easy axis 1

    Easy axis 2Easy axis 3

    X-ray direction

    Kamacite

    Tetrataenite rim

    Intermediate cloudy zone

    16º 35º

    35º

    35º

    Plessite

    35º

    35º 35º

    -35º

    -35º

    -35º

    Easy axis 1

    Easy axis 3 Easy axis 2

    Coarse cloudy zone

    Fine cloudy zone

    Plessite

    Kamacite

    Magnetisation projected onto X-ray beam !Blue – positive signal, easy axis 1!

    Red – negative signal, easy axes 2 and 3!

    Kamacite!

    Tazewell IIICD Iron meteorite!

    Bryson et al., (in prep)

    Tetrataenite rim!

    CZ!

    Plessite!

    b

    c

    BESSY II, HZB, Berlin!

  • Quantitative Analysis!

    Simulated nanostructure!

    Bryson et al., (in prep)

    In-plane magnetisation!

    Out-of-plane magnetisation!

    Magnetisation projection onto X-ray direction!

    Projected magnetisation at experimental resolution!

    Realistic PEEM image!

  • 140

    120

    100

    80

    60

    40

    20

    0Fr

    eque

    ncy

    -1.0 -0.5 0.0 0.5 1.0Pixel intensity

    Experimental data Simulated data

    a b

    c500 nm

    700

    600

    500

    400

    300

    200

    100

    0

    Freq

    uenc

    y

    -1.0 -0.5 0.0 0.5 1.0Pixel intensity

    Experimental data Simulated data

    a

    b

    c500 nm

    25

    20

    15

    10

    5

    0

    Freq

    uenc

    y

    8006004002000Cluster size (nm)

    Experimental data Simulated data

    500

    400

    300

    200

    100

    0

    Freq

    uenc

    y

    -1.0 -0.5 0.0 0.5 1.0Pixel intensity

    Experimental data Simulated data

    a

    b

    c

    d

    500 nm

    14

    12

    10

    8

    6

    4

    2

    0

    Freq

    uenc

    y

    8006004002000Cluster size (nm)

    Experimental data Simulated data

    250

    200

    150

    100

    50

    0

    Freq

    uenc

    y

    -1.0 -0.5 0.0 0.5 1.0Pixel intensity

    Experimental data Simulated data

    a

    b

    c

    d

    500 nm

    Quantitative nanopaleomagnetism

    8% easy axis 1!

    Simulation!

    Experiment!Intermediate CZ! Fine CZ!

    8% easy axis 1!

    Simulation!

    Experiment! Experiment!

    Simulation!

    Experiment!

    Coarse and intermediate CZ – equal easy axis proportions!

    Fine CZ – unequal easy axis proportions!Entire CZ – short range order present!

    0.33:0.33:0.33 easy axis ratio! 0.38:0.31:0.31 easy axis ratio !

    Coarse CZ!Simulation!

    Magnetic wall!

    Bryson et al., (in prep)

  • Quantitative natural nanopaleomagnetism

    Imilac pallasite!

    Experimental data!

    Simulated data!

    Esquel pallasite!

    Experimental data!

    Simulated data!

    8% +ve easy axis 1!17% -ve easy axis 1!29% +ve easy axis 2!

    8% -ve easy axis 2!23% +ve easy axis 3!

    15% -ve easy axis 3!

    14% +ve easy axis 1!0% -ve easy axis 1!32% +ve easy axis 2!

    24% -ve easy axis 2!17% +ve easy axis 3!

    14% -ve easy axis 3!

  • Quantitative natural nanopaleomagnetism

    Imilac pallasite!

    Experimental data!

    Simulated data!

    Esquel pallasite!

    Experimental data!

    Simulated data!

    8% +ve easy axis 1!17% -ve easy axis 1!29% +ve easy axis 2!

    8% -ve easy axis 2!23% +ve easy axis 3!

    15% -ve easy axis 3!

    14% +ve easy axis 1!0% -ve easy axis 1!

    32% +ve easy axis 2!

    24% -ve easy axis 2!17% +ve easy axis 3!

    14% -ve easy axis 3!

    Field intensity! > 3.5 µT!

    Field intensity! > 1.5 µT!

    8x10-3

    6

    4

    2

    0

    Norm

    alise

    d fre

    quen

    cy

    -0.1 0.0 0.1XMCD signal

    Experimental dataSimulated data

    8x10-3

    6

    4

    2

    0

    Nor

    mal

    ised

    freq

    uenc

    y

    -0.1 0.0 0.1XMCD signal

    Experimental data Simulated data

  • Conclusions 
 !

    Demonstrated the capabilities of synchrotron X-ray imaging in geomagnetism by studying novel nanomagnetic structures in meteoritic metal:!

    -  Images magnetisation directly!-  High spatial resolution (best 30nm) and large field of view (15 µm)!-  Sub-micron-scale paleomagnetism!

    Coarse CZ in the Tazewell iron meteorite displays equal proportions of the possible easy axes, suggesting an absence of external fields during CZ formation, while in the Imilac and Esquel pallasites there are unequal proportions of the possible magnetisation directions, suggesting there were magnetised in fields of > 3.5 µT and > 1.5 µT respectively.!

    !