Upload
epics-qt-collaboration
View
503
Download
5
Embed Size (px)
Citation preview
Ignace Jarrige Beamline SIX Group Leader
National Synchrotron Light Source II Brookhaven National Laboratory
BNL Summer Sundays July 13, 2014
SIX: A Looong Beamline at NSLS-II to Probe Electrons
A B C D
1
2
SIX BEAMLINE
PD-SIX-1000 SHEET SIZE DRAWING/PART NUMBER REVISION
SHEET OF1 4
DRAWN BY
CHECKED BY
VACUUMAPPROVALENGINEERAPPROVAL
SUPERVISORAPPROVAL
A. King
EWBS#
UNLESS OTHERWISE SPECIFIEDALL DIMENSIONS ARE IN INCHES
DIMENSIONS IN BRACKETS [xx.xx] (WHERE PRESENT)ARE MILLIMETERS AND ARE FOR REFERENCE ONLY
INTERPRET DRAWING AS PER ASME Y14.5-1994 OR Y32.2-1975
DIMENSIONAL TOLERANCES X. 0.060.X 0.030
.XX 0.015.XXX 0.005
ANGULAR TOLERANCE .5FINISH
THIRD ANGLEPROJECTION
BREAK EDGES & SHARP CORNERS 0.005 MIN. TO 0.030 MAX
3/11/2013
BROOKHAVEN NATIONAL LABORATORYBROOKHAVEN SCIENCE ASSOCIATES
UPTON, NEW YORK 11973
C:\M
ech-
Eng\
LT\P
D\S
IX\P
D-S
IX-1
000.
idw
Exploring Life's Mysteries, Protecting its Future
SCALE:SEE DWG VIEW
125
PROJECT: ESH&QRISK LEVEL
ES&H APPROVAL
NEXT ASSY: QA APPROVAL
PHOTON SCIENCESSIX BEAMLINE
Where is SIX? Tell me Google Map…
1/6/14 11973 - Google Maps
https://maps.google.com/maps?q=11973&ie=UTF-8&hq=&hnear=0x89e85b57fc4df783:0x63be313f85cebba9,Upton,+NY+11973&gl=us&ei=QSvLUo-KBMHZ2… 1/2
To see all the details that are visible on thescreen, use the "Print" link next to the map.
SIX
1/6/14 Ridge, NY - Google Maps
https://maps.google.com/maps?q=11973&ie=UTF-8&hq=&hnear=0x89e85b57fc4df783:0x63be313f85cebba9,Upton,+NY+11973&gl=us&ei=QSvLUo-KBMHZ2… 1/2
To see all the details that are visible on thescreen, use the "Print" link next to the map.
CFN
CMPMSD
NSLS-II
NSLS
SIX
1/6/14 Ridge, NY - Google Maps
https://maps.google.com/maps?q=11973&ie=UTF-8&hq=&hnear=0x89e85b57fc4df783:0x63be313f85cebba9,Upton,+NY+11973&gl=us&ei=QSvLUo-KBMHZ2… 1/2
To see all the details that are visible on thescreen, use the "Print" link next to the map.
RHIC
SIX
NSLS-II
Who is SIX? Tell me Tracom…
Driving
Amiable
Analytical
Tells Asks
Con
trol
s Em
otes
Yi Yi Zhu Mechanical
Engineer (Beamline)
Bill Joseph
Bill Leonhardt Mechanical
Engineer (Endstation)
Joe Dvorak Optics Scientist
Amanda
Ignace
Ignace Jarrige Group Leader
Amanda King Designer
Valentina
Valentina Bisogni Beamline Scientist
Expressive
Synchrotrons produce light: Why?
An extremely powerful source of light:
Providing many different ‘colors’: Synchrotron
range
� A synchrotron produces extremely bright light which is used in research. � The light comes in different wavelengths, x-rays, ultraviolet, visible, infrared.
Particles called electrons are accelerated to 99.99….% of the speed of light, injected in the synchrotron ring
As the electrons pass through magnets around the ring, they loose energy in the form of light, emitted as a narrow pencil directed forward They are then re-accelerated using RF cavities
We make the electrons glow!
Synchrotrons produce very bright pinpoint beams of light
This light is channeled out of the ring into beamlines, where it is tailored to accommodate specific needs of the research conducted
�
All beamlines operate simultaneously �
Each beamline is designed for use for a specific type of research
�
Experiments run throughout the day and night �
Synchrotrons: what kind of research?
Light sources are used to explore pretty much any type of matter, so they have a wide range of applications:
the spin crossover in the ferropericlase observedin our current XES and diffraction experimentslikely involves a mixed population of high-spinand low-spin states in the same crystal structure,an isosymmetric transition involving the substi-tution of low-spin ions for high-spin ions withoutchange in the structure of the host (6, 11).
Comparison with the model geotherm of theEarth’s lower mantle (20) indicates that the high-spin to low-spin crossover of iron likely occursfrom themiddle part to the lower part of the lowermantle from ~1000 km in depth and 1900 K to~2200 km and 2300 K, and that the low-spinferropericlase with the B1 structure exists in thelower mantle, that is, at depths below ~2200 km.The observed width of the spin crossover in fer-ropericlase is much narrower than that predicted bythe existing theoretical models (6, 11). Althoughthe temperature effect on the spin transitions inthe silicate perovskite and post-perovskite is yetto be studied (2–4, 8, 12), we propose that thisspin-crossover phenomenon should also occur insilicate perovskite because it is subject to similarthermal energy in the lower mantle as that re-quired to overcome the spin-pairing energy.
The spin crossover of iron in the lower-mantle phases substantially affects its implica-
tions for the geophysics and geodynamics ofEarth’s lower mantle. The continuous nature ofthe spin crossover observed here explains why nosignificant change in iron partitioning betweenferropericlase and perovskite has been observedin recent high pressure-temperature experimentswith a pyrolitic and olivine composition (16, 17),as opposed to a proposed dramatic change inpartitioning and chemical layering in the lowermantle (1, 2). Because the low-spin ferropericlaseexhibits relatively high density (5, 11), fast soundvelocities (5, 9), and lower radiative thermalconductivity (7) than the high-spin ferropericlase,the spin crossover in ferropericlase would resultin continuously enhanced density and reducedradiative thermal conductivity of ferropericlasefrom the middle part to the lower part of thelower mantle. However, slowing in soundvelocities and lowering in pressure derivativesof the sound velocities are expected within thetransition region (9). The observed increase ofthe low-spin (Mg,Fe)O at the mid-lower man-tle conditions would manifest seismically as alower-mantle spin transition zone (STZ), char-acterized by a steeper-than-normal density gra-dient between ~1000 km and 2200 km in depth(13–15) (Fig. 4). Spin transition may therefore
account for some of the seismic wave hetero-geneity in that region, and the existence of thelow-spin (Mg,Fe)O at the lowermost mantleconditions may affect the thermal stability ofthe mantle upwellings (18, 19).
Because the spin crossover of iron occurs inthe lower-mantle minerals such as ferropericlaseat high pressures and temperatures, the thermalcompression curves and sound velocities of thelower-mantle minerals will be continuouslyinfluenced by the ratio of the high-spin andlow-spin states along the lower-mantle geotherm(Figs. 3 and 4). This renders the use of theclassical equations of state with lattice finitestrain theory unreliable for modeling the densityand sound velocity behavior across the spincrossover (24, 27–29). For example, a densityincrease of ~3 to 4% is observed across the high-spin to low-spin transition in ferropericlase at~50GPa (5, 11, 22). Using a pyrolite lower-mantlecomposition model (29, 30) with ~33% of ferro-periclase and assuming a similar thermal compres-sion behavior in high-spin/low-spin (Mg,Fe)Owith various iron content (22, 31, 32), we predictthe spin crossover of iron in (Mg,Fe)O wouldresult in a density difference of ~1% between theextrapolated high-spin and low-spin densityprofiles, affecting our understanding of thelower-mantle chemistry. Such density increase isequivalent to the addition of ~5.0% FeO intoMgO in ferropericlase (32) and may be furtherenhanced if the stiffer low-spin ferropericlase (5)exhibits less thermal expansion than that extrapo-lated using the thermal equation of state of thehigh-spin ferropericlase (31). Therefore, knowl-edge of the ratio of the high-spin to low-spin statesin ferropericlase aswell as in the silicate perovskiteand post-perovskite is essential to evaluate reliablythe composition, geophysics, and dynamics ofEarth’s lower mantle.
References and Notes1. J. Badro et al., Science 300, 789 (2003).2. J. Badro et al., Science 305, 383 (2004).3. J. Li et al., Proc. Natl. Acad. Sci. U.S.A. 101, 14027
(2004).4. J. M. Jackson et al., Am. Miner. 90, 199 (2005).5. J. F. Lin et al., Nature 436, 377 (2005).6. W. Sturhahn, J. M. Jackson, J. F. Lin, Geophys. Res. Lett.
32, L12307 (2005).7. A. F. Goncharov, V. V. Struzhkin, S. D. Jacobsen, Science
312, 1205 (2006).8. A. M. Hofmeister, Earth Planet. Sci. Lett. 243, 44
(2006).9. J. F. Lin et al., Geophys. Res. Lett. 33, L22304 (2006).
10. K. Persson, A. Bengtson, G. Ceder, D. Morgan, Geophys.Res. Lett. 33, L16306 (2006).
11. T. Tsuchiya, R. M. Wentzcovitch, C. R. S. da Silva,S. de Gironcoli, Phys. Rev. Lett. 96, 198501 (2006).
12. F. Zhang, A. R. Oganov, Earth Planet. Sci. Lett. 249, 436(2006).
13. L. H. Kellogg, B. H. Hager, R. D. van der Hilst, Science283, 1881 (1999).
14. R. D. van der Hilst, H. Kárason, Science 283, 1885(1999).
15. J. Trampert, F. Deschamps, J. Resovsky, D. Yuen, Science306, 853 (2004).
16. Y. Kobayashi et al., Geophys. Res. Lett. 32, L19302 (2005).17. M. Murakami, K. Hirose, N. Sata, Y. Ohishi, Geophys. Res.
Lett. 32, L03304 (2005).
Fig. 3. Isosymmetric spincrossover of Fe2+ in (Mg0.75,Fe0.25)O. The phase diagramis constructed from the inter-polation and extrapolation ofthe derived fractions of thehigh-spin state in the sample(fig. S2). Colors in the verticalcolumn on the right repre-sent fractions of the high-spiniron, gHS, in (Mg0.75,Fe0.25)O.
2000
1800
1600
1400
1200
1000
800
600
400
20 30 40 50 60P (GPa)
Tem
pera
ture
(K)
70 80 900
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
!HS
Fig. 4. Derived fractions of thelow-spin ferropericlase (A) anddensity variation (B) along amodel lower-mantle geotherm(20). Fraction of the low-spinferropericlase is derived from anextrapolation of the experimentaldata in Fig. 3. Density variationsin ferropericlase across the spin-crossover region assume that thedensity varies linearly with thefraction of the low-spin iron (22).Dashed line and dash-dotted linerepresent derived density varia-tions using maximum variations of 2.8% and 4.2% across the spin-crossover region from evaluationof recent experimental (5) and theoretical (11) data, respectively. Vertical bars represent the densityvariations caused by 2% and 5% perturbation of total iron content in ferropericlase, respectively, atambient conditions (32).
21 SEPTEMBER 2007 VOL 317 SCIENCE www.sciencemag.org1742
REPORTS
on
Sept
embe
r 21,
200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
All these lines are blurred! Lots of interdisciplinary research (biophysics, physical chemistry,
geophysics…)
From NSLS…
X-Ray 2.8 GeV 300 mA
VUV-IR 0.8 GeV
1.0 A
Booster Ring
Electron Gun
• One of 4 DOE-supported synchrotron facilities
• 2 electron storage rings that produce synchrotron light
• 59 beamlines operate simultaneously • Operates 24/7, 10 months per year • Running today • 2,400 users every year • Users typically stay 2-4 days in on-site
housing
• Will bring best scientists to do experiments not possible today
• World’s finest capabilities for x-ray imaging and high-resolution energy analysis
• X-rays 10,000 times brighter than current NSLS
• $912-million project, early completion in September 2014, funded by U.S. Department of Energy
• Tax dollars at work: thank you everybody for your investment in science!
…to NSLS-II
SIX: Designed to look at what?
• Electrons!
• What are they: All matter is made of atoms, which are themselves like tiny solar systems (a billionth in size), with a nucleus and electrons rotating about the nucleus in orbits:
From the solar system…. …to a lithium atom Electron
Nucleus
Sun Planet Size/1021
• What do they do: Electrons have an electric charge and a magnetic moment, they confer the electric and magnetic properties of matter. Atoms, to bond with each other, share electrons:
H Cl
• The researcher’s tool box: Telescope for the solar system… Microscope for atoms… Beamline for electrons
Soft Inelastic X-ray scattering: What technique? The goal of the SIX beamline is to let us play billiards with light and electrons! � The setup:
� The process:
The beamline
The incoming light
Electrons in the sample
The detector
Our sample is a set of balls. Imagine that we want to find out their number and color, but have no mean to look at them directly.
We hit the cue ball (incoming light from beamline) which hits the set of balls and sends them around the table. We have disturbed, or
excited our sample.
Upon impact the cue ball is deflected and looses speed. By measuring the deflection angle and speed of the cue ball after
impact, we can find out about the number and color of balls
Soft Inelastic X-ray scattering: What technique? SIX will simultaneously probe electronic, magnetic, and structural properties!
Momentum transfer (deflection angle of
cue ball)
Energy transfer (speed of cue ball)
SIX
SC gap
Lower-energy spin waves
Phonons
NOW
Higher-energy spin waves
dd interband transitions
Charge transfer
~0.1 eV
Why SIX needs to be bigger (and costlier) than a billiard table • The more accurate the measurement, the finer our understanding about matter.
Without a world-leading instrument, most often impossible to do world-leading research!
• Ongoing world-wide race to improve the ‘color’ resolution of instruments. How? • First ingredient: To split the colors, focus, deflect the beam of light, need perfectly flat mirrors. Also
needs to hold these mirrors with an extreme stability.
1- Our goal: LIE (0.8 miles) Shore (8 miles) LIRR (1 mile) 2- Our environment:
SIX needs to be stable
3- Our design:
Concrete slab isolated ‘from the rest of the world’
Sample and optics on granite blocks, decoupled from chamber
Why SIX needs to be bigger (and costlier) than a billiard table • The more accurate the measurement, the finer our understanding about matter.
Without a world-leading instrument, most often impossible to do world-leading research!
• Ongoing world-wide race to improve the ‘color’ resolution of instruments. How? • First ingredient: To split the colors, focus, deflect the beam of light, need perfectly flat mirrors. Also
needs to hold these mirrors with an extreme stability.
• Second ingredient: Looking through a finite aperture, the more the colors split up, the easier it gets to pick up one particular color. Needs distance from the rainbow splitter!
SIX needs to be looong
Why SIX needs to be bigger (and costlier) than a billiard table • The more accurate the measurement, the finer our understanding about matter.
Without a world-leading instrument, most often impossible to do world-leading research!
• Ongoing world-wide race to improve the ‘color’ resolution of instruments. How? • First ingredient: To split the colors, focus, deflect the beam of light, need perfectly flat mirrors. Also
needs to hold these mirrors with an extreme stability.
• Second ingredient: Looking through a finite aperture, the more the colors split up, the easier it gets to pick up one particular color. Needs distance from the rainbow splitter! SOFT X-RAY EMISSION SPECTROMETER
SCIENTA XES 350
Scienta XES 350 is a state-of-the-art soft X-ray emissionspectrometer. It is a grazing incidence spectrometer cover-ing a wide energy range, 50 -1000 eV, at high resolution andsensitivity.* The Scienta XES 350 is easily adapted to diffe-rent excitation sources, since the instrument is flange-mounted and has an optical axis that is easily adjusted tothe excitation source.
The Scienta XES 350 optical arrangement consists of avariable entrance slit, two moveable shutters for gratingselection, three spherical gratings, and a 2-D detector thatcan be moved in a three-axis coordinate system. The ScientaXES 350 can be described as three spectrometers mergedinto one by having a common entrance slit and a detectorthat can be aligned to the focal curve (Rowland circle) ofthe selected grating.
X-ray emission spectroscopy, XES, measures the intensitydistribution of soft X-rays emitted due to radiative decay ofa core hole. With an attenuation length of photons in thisenergy range of typically hundreds of nanometers, themethod is inherently bulk sensitive.
The Scienta XES 350 is ultrahigh vacuum compatible, but itcan also be used with relatively high pressure gas systemssuch as vapor deposition equipment. Scienta XES 350 cantherefore be used for in-situ characterization in thin filmdeposition. It also makes it possible to study liquids andsolid/liquid interfaces. Since the X-ray emission processfollows the dipole selection rule, XES offers detailed infor-mation about the valence band electronic structure. Forsolids, essentially a partial density-of-state (PDOS) mappingis obtained.
In XES problems encountered with electron spectroscopy,such as charging or disturbance from electromagnetic fields,are not present. XES makes measurements of insulators,large bandgap semiconductors and ferromagnets feasible.The bulk sensitivity of XES makes it possible to study buriedlayers such as corrosion-sensitive films covered by an inertcapping layer.
*) J. Nordgren et al., Rev. Sci. Instr. 60, 1690 (1989)
© V
G S
cien
ta, 2
006-
06-2
8, v
1.1
1 m
15 m
5 m
Commercial spectrometer
ADRESS spectrometer, Swiss Light Source
SIX spectrometer
Why SIX needs to be bigger (and costlier) than a billiard table
Ideal (theoretical) RIXS spectrum
At XFEL: Get RIXS to the Heisenberg limit in time and energy
elastbimag
energy loss (eV)
-0.3 -0.2 -0.1 0.0
20000
40000
f
E/'E
mag ph2 ph1
7000
10000
30000
't (fs)@ 1 keV
165
't = h/'E = E/'E * O/c
123
82
41
28
't (fs)@ 300 eV
550
410
273
137
93
Adress (SLS)
eRIXS (ESRF) hRIXS (XFEL.EU) I21 (Diamond)
ARHEA (SOLEIL) Hornet (SPring-8) qRIXS (ALS)
Energy transfer (eV) (speed of cue ball)
SIX (NSLS-II) 100000
Courtesy of A. Föhlisch
But really… how big?
• Total length 120 m does not fit in NSLS-II’s experimental hall
• 15-m long spectrometer needs to rotate from (120° range) to measure deflection angle, bringing the building footprint to 9000 square feet
• Spectrometer splits light in the vertical, brings roof height to 22 feet
Beamline 105 m
Spectrometer
Detector
Sample chamber
Cue ball before ‘impact’
Cue ball after
‘impact’ 15 m
The SIX beamline • NSLS-II design (V. Bisogni, J. Dvorak, I. Jarrige, B. Leonhardt, Y. Zhu)
• Will use 15 state-of-the-art mirrors to focus, bounce, color-split the beam
• Will require the most stable mechanical systems ever built in the field
• Will outperform the best current resembling beamline in the world by a factor 10 in terms of ‘color’ resolution, for a similar flux of light
• 10/13 approval to start early procurement, 10/16 early completion, 10/17 start of operations
The SIX external building • BNL design (T. Joos, O. Dyling et al.)
• Engineered for low vibrations (28” thick slab isolated from ‘the rest of the world’) and high thermal stability (±0.3°)
• Contractor: Construction Consultants of Long Island (Riverhead, NY)
• 05/13 start of contract, 08/13 footings, 09/13 floor slab, 11/13 steel, 03/14 weather tight, 06/14 construction complete
The chronology of the construction, in photos…
The SIX external building in photos
Early September Late September Late October
Early November Late October Early November
Mid November Early January Late January
The SIX external building in photos: Outside
The SIX external building in photos: Inside
What scientific challenges for SIX? Energetic! • High-temperature superconductivity: Make it hot
� Superconducting copper wires = zero resistance, can move power over long distances without any loss � But need LOTS of liquid nitrogen to be chilled to superconduct � Crucial to understand superconductivity and design room-temperature superconductors
The LIPA-DOE Holbrook Superconductor Project: 600-m long cable, powers 300,000 homes, needs 13,000 gallons of LN2
Three Brookhaven Physicists Receive DOE Early Career Research Program Funding Recognition for explorations that peer into the heart of nuclear matter, the inner workings of superconductors, and the elusive mixing of different types of neutrinos Monday, May 12, 2014
Mark Dean, selected by the Office of Basic Energy Science for: "Probing the Magnetic Excitations in Complex Oxide Interfaces and Heterostructures"
LETTERS
PUBLISHED ONLINE: 2 SEPTEMBER 2012 | DOI: 10.1038/NMAT3409
Spin excitations in a single La2CuO4 layerM. P. M. Dean1*, R. S. Springell2,3, C. Monney4, K. J. Zhou4†, J. Pereiro1†, I. Bo�ovic1, B. Dalla Piazza5,H. M. Rønnow5, E. Morenzoni6, J. van den Brink7, T. Schmitt4 and J. P. Hill1*
Cuprates and other high-temperature superconductors consistof two-dimensional layers that are crucial to their properties.The dynamics of the quantum spins in these layers lie at theheart of the mystery of the cuprates1–7. In bulk cuprates suchas La2CuO4, the presence of a weak coupling between the two-dimensional layers stabilizes a three-dimensional magneticorder up to high temperatures. In a truly two-dimensionalsystem however, thermal spin fluctuations melt long-rangeorder at any finite temperature8. Here, we measure the spinresponse of isolated layers of La2CuO4 that are only one-unit-cell-thick. We show that coherent magnetic excitations,magnons, known from the bulk order, persist even in asingle layer of La2CuO4, with no evidence for more complexcorrelations such as resonating valence bond correlations9–11.These magnons are, therefore, well described by spin-wavetheory (SWT). On the other hand, we also observe a high-energy magnetic continuum in the isotropic magnetic responsethat is not well described by two-magnon SWT, or indeedany existing theories.
The simplest model for describing the magnetic excitations ofundoped cuprates is SWT (ref. 12). Coherent transverse magneticexcitations correspond to spin waves—magnons—with a well-defined energy; whereas longitudinal magnetic excitations result ina high-energy continuum of multi-magnons. Although measure-ments of the long-wavelength magnetic excitations of La2CuO4(ref. 13) can be understood in terms of a renormalized classicalmodel14, the short-range correlations remain controversial6,9,10,15–17as quantum fluctuations can transfer spectral weight out ofthe magnon peak into a high-energy continuum. Furthermore,the magnetic excitation spectrum of a one-unit-cell-thick (1 uc)La2CuO4 layer containing twoCuO2 planes, where spin fluctuationsare expected to be enhanced, has not beenmeasured. This is becausemost of what we know about the spin excitation spectrum of thecuprates has come from inelastic neutron scattering. Unfortunately,such experiments require large samples and are often challenging athigh-energy transfers. In recent years, however, resonant inelasticX-ray scattering (RIXS) has achieved sufficient resolution to accessmagnetic excitations7,18–21 and RIXS is well suited to measuringhigh-energy magnetic excitations in the range 100–1,000meV.Furthermore, the high sensitivity of the technique allows us to lookat nanostructured samples and this in turn opens up the excitingpossibility of measuring the spin response of a 1 uc La2CuO4layer for the first time.
1Department of Condensed Matter Physics and Materials Science, Brookhaven National Laboratory, Upton, New York 11973, USA, 2London Centre forNanotechnology and Department of Physics and Astronomy, University College London, London WC1E 6BT, UK, 3Royal Commission for the Exhibition of1851 Research Fellow, Interface Analysis Centre, University of Bristol, Bristol BS2 8BS, UK, 4Swiss Light Source, Paul Scherrer Institut, CH-5232 Villigen PSI,Switzerland, 5Laboratory for Quantum Magnetism, École Polytechnique Fédérale de Lausanne (EPFL), CH-1015, Switzerland, 6Laboratory for Muon SpinSpectroscopy, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland, 7Institute for Theoretical Solid State Physics, IFW Dresden, D01171 Dresden, Germany.†Present addresses: Diamond Light Source, Harwell Science and Innovation Campus, Didcot OX11 0DE, UK (K.J.Z.); Department of Physics, University ofCalifornia, San Diego, La Jolla, California 92093, USA (J.P.). *e-mail: [email protected]; [email protected].
Inte
nsity
(s¬1
)
La2CuO4LaAlO3
i!
!2
C
"
Q
Elastic + phonon line
Electronicexcitations
Magneticscattering
Energy loss (eV)x25 x15 x40
1 uc 2 uc Bulk
10
20
00 1 2 3
a
c
b
Figure 1 | The scattering geometry, a schematic of the samples and atypical RIXS spectrum. a, The experimental scattering geometry. The931 eV � -polarized X-rays are incident at an angle ✓i and are scatteredthrough a fixed angle 2✓ = 130�. Large Q corresponds to near-grazingincidence (✓i ! 0). b, The multilayer films studied, composed of13.2 Å La2CuO4 layers (red blocks) containing two CuO2 planes and3.8 Å LaAlO3 (blue blocks). We label the films, on the basis of the thicknessof La2CuO4, as 1 uc, 2 uc and bulk. The arrows denote the repeat unit of thefilms (⇥25, ⇥15, ⇥40). c, A representative RIXS spectrum of the 1 ucLa2CuO4 film (25⇥ [LaAlO3 +La2CuO4]) at Q= (0.77⇡,0) identifying themain spectral features: the elastic and phonon scattering aroundzero-energy transfer, the magnetic scattering around 300 meV and theelectronic (dd) excitations from 1 to 3 eV.
We performed RIXS measurements on bulk and single-layerLa2CuO4 films at 15 K using the scattering geometry shown inFig. 1a. The sample was rotated about the vertical axis to vary Q,the projection of the total scattering vector in the ab plane. Toprovide a sufficient scattering volume of isolated La2CuO4 layers,we prepared heterostructures based on 1 uc layers of La2CuO4and LaAlO3. Note that 1 uc of La2CuO4 contains two CuO2 layers.The samples are depicted in Fig. 1b as 1 uc = [1 uc LaAlO3 +1 uc La2CuO4] ⇥ 25, 2 uc = [2 uc LaAlO3 + 2 uc La2CuO4] ⇥ 15,and bulk = [1 uc La2CuO4] ⇥ 40. The films were characterizedusing muon spin rotation (see Supplementary Information for adiscussion). These results, and the RIXS results (discussed later),show that the correlated patches of spins are randomly orientatedasmight be expected for an isolated La2CuO4 layer.
The RIXS spectra of the three samples were measured from(0.14⇡,0) to (0.8⇡,0) and (0.1⇡,0.1⇡) to (0.6⇡,0.6⇡). Figure 1c
850 NATUREMATERIALS | VOL 11 | OCTOBER 2012 | www.nature.com/naturematerials
LETTERS
PUBLISHED ONLINE: 4 AUGUST 2013 | DOI: 10.1038/NMAT3723
Persistence of magnetic excitations inLa2�x
Srx
CuO4 from the undoped insulator to theheavily overdoped non-superconducting metalM. P. M. Dean1*, G. Dellea2, R. S. Springell3, F. Yakhou-Harris4, K. Kummer4, N. B. Brookes4, X. Liu1,5,Y-J. Sun1,5, J. Strle1,6, T. Schmitt7, L. Braicovich2,8, G. Ghiringhelli2,8, I. Bo�ovic1 and J. P. Hill1*
One of the most intensely studied scenarios of high-temperature superconductivity (HTS) postulates pairing byexchange of magnetic excitations1. Indeed, such excitationshave been observed up to optimal doping in the cuprates2–7.In the heavily overdoped regime, neutron scattering mea-surements indicate that magnetic excitations have effectivelydisappeared8–10, and this has been argued to cause thedemise of HTS with overdoping1,8,10. Here we use resonantinelastic X-ray scattering, which is sensitive to complementaryparts of reciprocal space, to measure the evolution of themagnetic excitations in La2�x
Srx
CuO4 across the entire phasediagram, from a strongly correlated insulator (x = 0) to anon-superconducting metal (x= 0.40). For x= 0, well-definedmagnon excitations are observed11. These magnons broadenwith doping, but they persist with a similar dispersion andcomparable intensity all the way to the non-superconducting,heavily overdoped metallic phase. The destruction of HTSwith overdoping is therefore caused neither by the generaldisappearance nor by the overall softening of magneticexcitations. Other factors, such as the redistribution of spectralweight, must be considered.
The undoped high-Tc cuprates such as La2CuO4 are antiferro-magnetic (Néel-ordered) insulators, withmagnetic Bragg peaks andwell-defined high-energy magnetic excitations termed magnons11.As shown in Fig. 1a, doping rapidly destroys the Néel ordering,leading to the emergence of the pseudogap state and superconduc-tivity. In the underdoped and optimally doped cuprates, supercon-ductivity is accompanied by an ‘hour-glass’-shaped dispersion ofmagnetic excitations around the scattering vectorQAFM = (0.5,0.5)in Fig. 1b (refs 2–6,12). In the lightly overdoped, but still super-conducting regime, high-energy magnetic excitations have beenobserved in La1.78Sr0.22CuO4 (ref. 13) and YBa2Cu3O7 (ref. 7). Farless work has been done on the magnetic excitations in the heavilyoverdoped region of the phase diagram. Neutron scattering studiesof La1.70Sr0.30CuO4 report that the Q-integrated magnetic dynamicstructure factor S(!) ismuch reduced by x=0.25 and thatmagneticexcitations have effectively disappeared by x = 0.30 (ref. 9), wherex is the doping level. This observation has been used in supportof proposals that spin fluctuations mediate the electron pairing
1Department of Condensed Matter Physics and Materials Science, Brookhaven National Laboratory, Upton, New York 11973, USA, 2Dipartimento di Fisica,Politecnico di Milano, Piazza Leonardo da Vinci 32, I-20133 Milano, Italy, 3Royal Commission for the Exhibition of 1851 Research Fellow, Interface AnalysisCentre, University of Bristol, Bristol BS2 8BS, UK, 4European Synchrotron Radiation Facility (ESRF), BP 220, F-38043 Grenoble Cedex, France, 5BeijingNational Laboratory for Condensed Matter Physics, and Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China, 6Department forComplex Matter, Jo�ef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia, 7Swiss Light Source, Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland,8CNR-SPIN, Consorzio Nazionale Interuniversitario per le Scienze Fisiche della Materia, Italy. *e-mail: [email protected]; [email protected]
in high-Tc superconductors1. A necessary, although not sufficient,condition for such scenarios is that spin fluctuations persist acrossthe superconducting portion of the phase diagram while retainingappreciable spectral weight. For this reason it was suggested thatthe destruction of HTS in the overdoped cuprates is due to thedisappearance of magnetic excitations8.
Resonant inelastic X-ray scattering (RIXS) at the Cu L3 edgehas recently emerged as a new experimental method for measuringmagnetic excitations in the cuprates7,14–19. RIXS is particularly wellsuited to measuring high-energy magnetic excitations and requiresonly very small sample volumes14,15. As explained in ref. 7, thissets it apart from current neutron scattering experiments, whichrequire large single crystals of several cm3 in volume that are usuallyvery difficult to synthesize, especially in the heavily overdopedregion. We also note that Cu L3 edge RIXS experiments focus on acomplementary region of the Brillouin zone in Fig. 1b compared tomost Q-resolved neutron scattering experiments20. RIXS typicallymeasures from (0,0) towards (0.5,0); whereas neutron scatteringexperiments focus around (0.5,0.5), where themagnetic excitationsare strongest. In the absence of a universally accepted, quantitativetheory of high-Tc superconductivity, it is essential to consider theexcitation spectrum over the whole Brillouin zone.
We synthesized La2�x
Srx
CuO4 films with x = 0, 0.11, 0.16,0.26 and 0.40 using molecular beam epitaxy. These films, unlikebulk samples, have atomically smooth surfaces (root mean squareroughness, as measured by atomic force microscopy, down to a fewÅ), which reduce the diffuse elastic scattering contribution to thespectra20. We chose the doping levels to span the La2�x
Srx
CuO4phase diagram, as indicated by the solid black squares in Fig. 1a.
RIXS spectra for these samples are shown in Fig. 1c. Themost intense feature corresponds to optically forbidden dd orbitalexcitations in which the valence band hole, primarily of Cu d
x
2�y
2
character, is promoted into higher energy orbitals21. The intensity ofthese excitations can provide a reference to compare different RIXSspectra22. In the mid-infrared energy scale (50–500meV) singlespin-flip excitations can be excited owing to the spin–orbit couplingof the Cu 2p3/2 core hole23,24. A broad, flat background of intensityarises from charge-transfer excitations of the Cu d
x
2�y
2 hole into theO 2p states. As x increases the dd excitations are seen to broaden,
NATUREMATERIALS | VOL 12 | NOVEMBER 2013 | www.nature.com/naturematerials 1019
© 2013 Macmillan Publishers Limited. All rights reserved
…more to come at SIX!
What scientific challenges for SIX? Energetic! • High-temperature superconductivity: Make it hot
� Superconducting copper wires = zero resistance, can move power over long distances without any loss � But need LOTS of liquid nitrogen to be chilled to superconduct � Crucial to understand superconductivity and design room-temperature superconductors
The LIPA-DOE Holbrook Superconductor Project: 600-m long cable, powers 300,000 homes, needs 13,000 gallons of LN2
• Lithium and fuel cell batteries: Long live
� Ageing is the result of several physicochemical processes � Study mechanism of lithium, hydrogen and oxygen ion transport in batteries � Understand how to improve efficiency and lifetime
• Spintronics for Computing: Making memories
� Magnetic processing of information, rather than charge, yields smaller, faster data storage � Spintronics transistors to create ultra-fast, low-consumption computer chips � Understand the nature of magnetic properties to help the design of future devices
Conclusions
• X-rays are not just for looking at broken bones
We watch, poke and control
atoms and electrons,
…without breaking them
• SIX is a big tool to look at small things that have a big impact
That control the bigger phenomena
We look at the tiny things
• SIX to do clean science for a cleaner future
Synchrotron light is clean…
…But why stop there?
At SIX, clean energy is our ultimate goal!
Thank you!