ESRF - HIGHLIGHTS 2014 The European Synchrotron The European Cedex 9, France - F38043 Grenoble CS 40220 www.esrf.eu HL2014-19-01-4-impression.pdf 1 26/01/15 16:35 1 HL2014-19-01-4-impression.pdf K M Y C CM MY CY CMY N J M C CJ MJ CM CMJ E u r o p e a n S y n c h r o t r o n R a d i a t i o n F a c i l i t y

Pages

Introduction 2

Status of the Upgrade Programme 3

Scientific Highlights 8 Electronic Structure and Magnetism 8 Soft Condensed Matter 28 X-ray Imaging 44 Dynamics and Extreme Conditions 64 Structural Biology 80 Structure of Materials 103

Enabling Technologies 125

Accelerator and X-ray Source 136

Facts and Figures 145 HIGHLIGHTS 2014 INTRODUCTION

Dear Reader

Twenty years after its inauguration, the ESRF, in 2014, welcomed the Russian Federation as a new Member State. It also reached important milestones in the delivery of Phase I of the Upgrade Programme (UP) and has proudly announced the launch of Phase II. In parallel, the excellent operation of the facility throughout the year enabled the Francesco ESRF users to carry out their scientific Sette programmes very successfully. Over 2000 proposals were submitted by implementation of UP Phase I. New President of the Rhône-Alpes Region ESRF users in 2014. This figure is the instruments were delivered, and in and the local authorities. The Science highest ever reached, and puts the particular, I wish to highlight the Building will foster new collaboration number of proposals at more than two restart of user operation at the new and closer working relations between and a half times the available beam beamlines ID01, ID02, ID16A/B, ID22, the two facilities, their scientific user time. ID30A and ID32. With respect to the communities as well as with industry. opening of ID30A, part of the MASSIF ESRF users can now access the Science The achievements of 2014 are the complex at ID30, the macromolecular Building from the ESRF thanks to a result of the hard work of thousands crystallography community will benefit covered footbridge. This footbridge of people among the ESRF Users and from the first public experimental provides access to the new library, new Staff, and would not have been possible station in the world that is completely chemistry labs and the new facilities without the support provided by our automated. In fact, beyond remote of the Partnership for Soft Condensed partner countries. Following on from access, users can now send in their Matter (PSCM). Thanks to the CPER these achievements, we are now excited samples, have them analysed Programme, the ESRF has also been about the prospects for 2015, which will automatically using instructions given made more accessible to users coming see the completion of UP Phase I, the in the ISPyB database, and then log from and going into town since the start of UP Phase II construction, and on to discover and review their results. completion, in September 2014, of an full-scale user operation for the ESRF This new facility has the potential extension of one of Grenoble’s tram accelerators and available beamlines. to save hours, or even days of users’ lines. valuable time. It has been designed The ESRF Council endorsed the UP Phase particularly for the initial screening of I wish to thank our colleagues in charge II deliverables, timescale and budget crystals and the collection of data sets of the operation of the storage ring for envelope in June 2014 following the from reasonably well-known samples. the outstanding performance in beam science case and construction planning Since it opened in late October, 500 delivery. Indeed, during 2014 they described in the UP Phase II Technical samples have already been processed. were able to establish a new world Design Study report, the ‘Orange Book’. ID30A has three end stations and, with record with 99.15% beam availability At the subsequent Council meeting BM29, replaces the highly successful with enhanced beam properties. The in December 2014, the first financial ID14 complex that was closed at the excellent beamline and beam delivery contribution to the UP Phase II budget end of 2012. performance have enabled excellent was approved. Over the next eight years scientific results. A selection of them is (2015-2022), the ESRF plans to bring UP Phase I continues to progress presented here in Highlights to record into operation a new synchrotron X-ray according to schedule, with the goal and applaud the wonderful scientific source, housed in the existing buildings of its completion by the end of 2015. advances of the ESRF Users. and replacing the present machine During 2015, therefore, the few that has been in operation since remaining Phase I beamlines will be Francesco Sette, 1992, as well as four new beamlines commissioned, and by early 2016 Director General, and an ambitious instrumentation the topping-up of the storage ring will ESRF programme. The objective is to deliver become part of normal user operation. a fully renewed facility by 2023 for the benefit of scientific users from around The new joint ESRF-ILL Science Building the world, and to enable the ESRF to and the new site entrance of the EPN remain a global leader in synchrotron Science Campus, entirely financed by X-ray science for many years to come. the French State (CPER Programme) to increase the visibility and attractiveness Alongside the preparation work for UP of the campus, was inaugurated in Phase II, in 2014, many developments February 2014 by the French Minister took place within the context of the for Higher Education and Research, the

2 ESRF status of the upgrade programme S tATUS OF THE Upgrade Programme

Following the occupation of the new buildings delivered within the EX2 project, 2014 has been the peak year for the implementation of the new beamline portfolio in the new premises. The year saw the inauguration of many new beamlines with unprecedented scientific capabilities, including ID02 (Figure 1), ID16A, ID16B, ID22, ID30A-1, and ID32. As we are entering the final year of Phase I of the Upgrade Programme, the construction of the new beamlines is approaching completion on time, within budget and in line with or exceeding specification.

The upgrade of the current accelerator complex is also heading towards its completion, with the latest step being preparation for top-up operation in the timing modes, scheduled for routine Fig. 1: Inauguration of beamline ID02 operation in 2016. Several tests on (UPBL9A) at the 69th meeting of the Science Advisory Committee, November 2014. top-up operation during machine Pictured from left to right: dedicated time have helped to identify F. Sette (Director General), the remaining issues and devise Prof. K. Hämäläinen (Chairman SAC), solutions. 2014 has also seen a large N. Theyencheri (Scientist in charge of ID02), effort from all Divisions of the ESRF in Dr A. Petukhov (SAC). the preparation of the Orange Book, with the proposal for the construction of a new low emittance storage ring into place in the straight section in light source at its core. Preliminary March 2015. Canted straight sections design work and cost estimates have have been successfully installed, and been concluded successfully and only ID15 remains to be modified led to the approval of the project by with this scheme, foreseen for Council in June 2014. Representing 2015. An improved electron beam our Users, we thank our staff for these position monitoring system and a achievements, delivered while the fast orbit feedback system have been facility was kept in full user operation. successfully developed and installed.

The top-up project is progressing Upgrade of the source well, and several tests were carried out with beamlines in 2014. Feedback 2014 has seen the successful from the beamlines has been positive completion of several upgrades to and, working with users, we have subsections of the accelerator system. implemented improvements following their requests. The new bunch cleaning The RF system has been renewed with system installed in the booster this the addition of solid state amplifiers year will greatly aid top-up operation, and HOM-damped cavities. The scheduled for the beginning of 2016, manufacture of 12 additional cavities once a formal agreement from the for Phase II is progressing well. All the authorities has been accorded. planned straight sections have been lengthened from 5 m to 6 m, with one The technical design study for the 7 m-long section in cell 23, enabling UP Phase II, based on a hybrid the installation of the RF cavities. 7-bend achromat lattice, has been completed and received very positive Following the construction of ID32, reports from the Accelerator Project the new helical undulator will be put Advisory Committee (APAC), the

The European Synchrotron 3 HIGHLIGHTS 2014 status of the upgrade programme

Scientific Advisory Committee (SAC) UPBL2 – High energies for interfaces and the Administration and Finance and materials processing. The Committee (AFC). In June 2014 it infrastructure for the beamline was submitted to the Council, which has been completed and first approved and launched the project. components installed in optics We entered the execution phase on hutch OH1. The design, assembly 1st January 2015 with leaders being and installation of the main optical appointed for the Accelerator Project components (transfocators, multilayer Office and Work Packages. The monochromator, double Laue lattice and engineering designs have monochromator) in OH2 is ongoing. advanced well, with some magnets The control infrastructure is under already at the prototype stage. preparation. The large granite detector gantry has been installed and aligned.

R enewal of beamlines UPBL4 – Nano-imaging and nano- analysis. The beamlines ID16A The Instrumentation Services and and ID16B have been installed and Development Division (ISDD) and commissioned. Both beamlines the Experiments Division (ExpD), resumed user operation in 2014 with with the assistance of the Technical some remaining work (cryo-cooling of Infrastructure Division (TID), deliver the sample stage and high energy KB new and refurbished beamlines. A system (34 keV)) at the endstation of significant number of beamlines are the ID16A-NI branch. still under construction or will receive NI- nanotomography endstation: The their finishing touches while they are installation of the endstation has been already in user operation. A total of five completed and the fixed-energy KB beamlines are still under construction system (17 keV) commissioned. Most with most of them returning to user of the detectors are operational while operation in 2015. Significant financial the remaining cSAXS detector is still in and human resources are also invested the assembly phase. After reaching a 2 in the continuous refurbishment and beam size of 20 x 40 nm the beamline further development of the remaining was successfully opened for user beamlines. operation. NA- nano-analysis endstation: The The main projects for beamline construction of the beamline was renewal are as follows: finished with all the detectors installed and the beamline has returned to UPBL1 – Diffraction for nano-analysis. user operation. The contaminated KB Most of the optical components, system, damaged by a defective gas including the white beam mirrors line, will be replaced in 2015. and multilayer monochromator, double crystal monochromator and UPBL6 – Hard X-ray spectroscopy transfocator have been assembled and (RIXS & XRS). The horizontally installed. With the completion of the deflecting (focusing) mirrors are installation of the nano-diffractometer, operational and five APD detectors the beamline received its first user have been installed. The beamline team in December 2014. The SAXS/ returned to full user operation. GISAXS end-station is in the assembly phase and will be commissioned early UPBL7 – Soft X-ray spectroscopy in 2015. (RIXS & XMCD). All components for the beamline optics have been received and installed. Existing ID08 experimental facilities have been moved to the new ID32 premises in the experimental hall extension.

Fig. 2: ID32 RIXS branch: refocusing optics, sample chamber, collimating optics and grating vessel (from right to left).

4 ESRF status of the upgrade programme

XMCD branch: The assembly of the Fig. 3: ID02 XMCD branch including the new detector wagon (front sample preparation facility has been view) inside finished and all components have been the USAXS commissioned. First user experiments tube. have been successfully performed with the XMCD setup as well as with a user- specific setup in the open access hutch attached to the XMCD branch. All specifications have been met. RIXS branch: The assembly of the 11m RIXS spectrometer arm is heading towards its final steps (Figure 2). All components have been received with only the gratings for the spectrometer still to be delivered.

UPBL9a – Time-resolved ultra- small-angle scattering. All optical components have been installed at beamline ID02 and the new 32 m long USAXS tube including the detector wagon has been assembled and aligned in the new premises (Figure 3). All SAXS (Figure 4) and WAXS detectors have been delivered and commissioned in the beam.

UPBL9b – Pump-probe and time- resolved experiments. The design of the new optical components for beamline ID09 has been finished and the orders placed. The high heatload chopper has been delivered and a new 3 kHz chopper is in the design stage. The beamline will remain fully operational throughout the upgrade. Fig. 4: Detectors inside the ID02 SAXS conveyor: FReLoN camera, beam viewer, UPBL10 – Automated crystallographic Rayonix MX170 CCD detector, Pilatus 300K evaluation. The new MASSIF sample hybrid pixel array detector (from left to handling facility including a high right). capacity dewar and RoboDiff setup (Figure 5) has been commissioned on ID30A-1 and has begun user operation. The microfocus beamline ID30A-3 has started user operation with a conventional experimental setup that will be replaced by a RoboDiff in 2015. The new Eiger 4M detector has been delivered and is being prepared for installation at the beamline. The installation of the new MAD beamline ID30B is in progress including the installation of a MD2S microdiffractometer, a new sample changer robot and a modified high

Fig. 5: Sample handling facility at ID30A-1 with high capacity dewar (left) and the new RoboDiff (centre) acting as sample changer and microdiffractometer.

The European Synchrotron 5 HIGHLIGHTS 2014 status of the upgrade programme

capacity dewar (in collaboration with a new data analysis strategy and the EMBL Grenoble). The new Pilatus3 technical solutions to provide better 6M detector has been received and service for ESRF users. commissioned. The year has also seen further UPBL11 – Time-resolved and extreme consolidation of the project conditions XAS. Work on several management organisation, with optical components continued while the implementation of an electronic keeping the upgraded ID24 beamline system for project submission and fully operational. New detectors associated approval workflow. This have been produced and are under system is now used by both the commissioning. Experiments Division and ISDD, and it is expected to be adopted by the ID22 – Powder diffraction. The ASD for the management of Phase II installation of the new beamline optics projects. and the transfer and refurbishment of the endstation have been completed. Several large-scale projects have After a short shutdown of only also aimed to strengthen the ESRF’s four months the powder diffraction capabilities in certain strategic areas. beamline successfully commenced One such highlight is a new laboratory user operation. which will be entirely dedicated to the development and production of ID19 – X-ray imaging and curved X-ray analyser crystals for palaeontology project. The upgrade spectroscopy applications. Similarly, and refurbishment of the beamline 2014 has seen the implementation of optics is nearly finished. The design, a new Fizeau interferometer in vertical procurement and installation of geometry with semi-automated detectors continues according to the stitching capability for optical TDR. metrology and the installation of a new system for cleaning carbonaceous ID15 – High pressure, materials contamination from optical surfaces. engineering, materials chemistry. The TDR for the two branches ID15A/B In 2014, ISDD and ExpD initiated an has been completed and approved. in-depth review of double-crystal The optical design and configuration monochromator (DCM) technologies has been validated and the order for for spectroscopic applications. A the construction of the new hutches dedicated working group is currently was placed. reflecting on the development of high- performance DCMs to cope with future In the context of the Upgrade scientific needs and also to anticipate Programme, the ISDD continues to the obsolescence of our aging DCMs. monitor its projects and associated resources, which are shared among As in previous years, special effort has several core activities, such as been devoted to developing generic completion of the UP Phase I, solutions and frameworks (see the daily operation of the facility chapter on Enabling Technologies). and development of the in-house Alongside the complete specifications R&D programme for scientific and initial design of a new DAnCE- instrumentation. In addition, in based encoder processing unit, several 2014, ISDD has focussed on defining software platforms have been further a roadmap for UP Phase II, which integrated and standardised, which includes strategic developments for include mxCuBE, LImA, PyFAI and future scientific instrumentation at PyMCA. the ESRF. This programme aims to provide engineering solutions for the Detector R&D activities have best scientific exploitation of the new continued to focus on the continuation storage ring properties. The two main or completion of several projects pillars of this programme are detector launched previously, including Medipix development and the implementation 3, HiZPAD2, XNAP and high efficiency of a new model for data analysis. A scintillators, whilst reflection began for special effort is foreseen to develop the Phase II development programme.

6 ESRF status of the upgrade programme

The instrumentation beamline, BM05, The Lab and Office Building is continued to be a strategic tool for now fully occupied and houses 30 instrumentation developments such laboratories for some of the beamlines as detector calibration, scintillator and the ISDD. Some of these new characterisation and testing of laboratories were built to compensate optical components. The beamline for space forfeited to the beamline DCM has been upgraded and a refurbishment programme, with new diffractometer will soon be beamlines now passing through the implemented to better serve those areas previously occupied by some instrumentation programmes. of the laboratories and into the EX2 extension. Others have enabled Finally, commercial activities and new activities to be re-grouped, such is the partnerships continued to support the case for the Electronics Group, which in European synchrotron community turn has freed up space in the Central with the procurement of specific Building. Laboratories are also being instruments such as a FReLoN-based modernised and relocated around detector and IcePAP controllers. the experimental hall for an optimal operation in proximity to the new and refurbished beamlines. Moving on The chemistry and surface science The extensive UP Phase I building laboratories have moved from the programme was completed in 2014 Central Building to the Science with the delivery of the EX2 extensions Building. The space freed up in the (ID16 satellite building, Belledonne Central Building will be converted into Hall, Chartreuse Hall, Lab and Office offices, allowing staff in dispersed Building) in parallel with the new site groups to be brought together. This entrance and joint ESRF-ILL Science concerns the Software, Mechanical Building. The availability of the new Engineering, Detector, and Alignment premises has inevitably triggered the groups. At the end of this operation, relocation of a large fraction of staff more than 50% of the ESRF staff will which will continue throughout 2015. have moved and will have gained From the outset, the ESRF was keen to improved working conditions and/ embrace a high quality environment or increased synergy to carry out the standard for the tertiary buildings, demanding work ahead to operate the thus contributing to sustainable facility and implement Phase II of the development and to provide a Upgrade Programme. comfortable working environment for the support laboratories of the R. Dimper, P. Raimondi, experimental programme. H. Reichert and J. Susini

The European Synchrotron 7 SCIENTIFIC HIGHLIGHTS Highlights 2014 Electronic structure and magnetism

One of the major events of the year 2014 was the start of operation of ID32, the upgrade beamline project UBPL7 dedicated to polarisation dependent studies using soft X-rays. This 120 m long beamline replacing the former ID08 features two branches, optimised for state-of-the-art X-ray magnetic circular dichroism (XMCD) experiments and resonant inelastic X-ray scattering (RIXS) with very high energy resolution. Among the upgraded features are a variable spot size down to 10x100 µm2 on the XMCD Inauguration ceremony at beamline ID32. Pictured from left to right: N. Brookes (Scientist in branch and a combined resolution charge of ID32), Prof. D. McMorrow (SAC member), Prof. K. Hämäläinen (Chairman SAC), F. Yakhou-Harris (ID32), Prof. M. Takata (SAC member), F. Sette (Director General). of 30000 on the RIXS branch. ID32 was inaugurated on November 6 (see photo below). The first user group At beamline ID12, the high pressure This year’s chapter reflects the large (Ghiringhelli et al., Politecnico di programme reached a milestone early variety of different spectroscopic and Milano, Italy) worked on the high field this year with the first ever measured scattering methods utilised at the magnet (9 T) end station on the XMCD chlorine K edge XAS at high pressure. ESRF to investigate the electronic and branch to investigate chiral effects in The B1(NaCl-type) - B2(CsCl-type) magnetic structure in materials that high temperature superconductors. structural phase transition (≈ 3 GPa) are becoming more and more diverse. The second group (Ogin et al., in a KCl crystal was recorded in total In addition to the five beamlines of the University of Exeter, UK) made use of fluorescence yield mode through a Electronic Structure and Magnetism the “free” experimental station, on the specially-designed diamond-anvil cell Group (ID12, ID24, ID26, ID32 and same branch, for time-resolved soft featuring a fully-perforated diamond BM23), our chapter also hosts science X-ray magnetic holography. coupled to a thin 30 µm diamond highlights from ID20, the new Inelastic window. The beam was focused in X-ray Scattering (IXS) beamline, and Another important result that marked both directions by Be refractive lenses. from BM28, the UK CRG beamline. the year 2014 was the success of This achievement opens exciting the first high-power laser-shock new opportunities for high pressure The science highlight from BM28 experiment at the energy-dispersive studies at low X-ray energies. shows how the interference between X-ray absorption spectroscopy (XAS) non-resonant magnetic X-ray beamline ID24, which was the upgrade At beamline ID26, the KB mirror scattering and the pure quadrupole beamline project UPBL11. For the assembly was replaced. This, resonance at the Fe K absorption first time at a synchrotron, iron was considerably improved the beam edge in the weak ferromagnet FeBO3 investigated while being dynamically stability and increased the brilliance carries the sign of the Dzyaloshinskii– compressed to extreme pressures and by a factor of 4. Following on from Moriya interaction. The selection temperatures (~4 Mbar, 9000 K). A pioneering experiments where X-ray from ID20 includes the investigation compact (30 J) focused laser beam magnetic circular dichroism was of high energy plasmonic excitations was used to produce a hot compressed combined with resonant inelastic in transition metal dichalcogenides state in the sample, of sufficient X-ray scattering, a chamber was by IXS and the study of the electronic duration to be probed with a single designed for permanent installation structure of the spin-orbit Mott

100 ps X-ray pulse. The feasibility at the beamline holding phase insulator CaIrO3, where RIXS at the Ir of a “single” shot scheme is very plates to generate circular polarised L3 edge has been applied to explain its important because the reproducibility light. Many samples measured at insulating behaviour. of such extreme conditions is never ID26 suffer radiation damage. This guaranteed. Usually, much larger, can be avoided by using a large RIXS in the soft X-ray regime has also very high energy lasers are required amount of sample that is moved been used for one of the last science to produce such extreme states through the beam. To permit this, highlights from beamline ID08 before making the experiments more costly ID26 has developed hardware and its closure, to reveal differences between and less accessible. The possibility software tools to map a large sample electronic excitations in electron- and of performing such experiments surface or liquid jets. Finally, ID26 is hole-doped cuprates, a prerequisite at a synchrotron beamline offers exploring the energy range from 2-5 for a complete understanding of the new opportunities to a large user keV in X-ray emission spectroscopy superconductivity in these systems. community. with a new instrument under design. The second selection from ID08 shows

8 ESRF ELECTRONIC STRUCTURE aND maGNETISm

an interesting application of soft XCMD Valence-to-core XES is highly sensitive behaviour in heterogeneous catalysts to investigate the magnetic properties to small variations of a transition by means of the combination of time- of surface-supported molecular systems metal’s valence electronic structure resolved EXAFS with synchronous employing transition metals dopants. induced by the ligand. In the fi rst diff use refl ectance infrared Fourier science highlight from beamline ID26, transform spectroscopy (DRIFTS) and XMCD with hard X-rays, the work this technique has been used as a mass spectroscopy. horse of the ID12 beamline, has also spectroscopic ruler for the quantitative been used to unveil a novel dilute determination of the NO oxidation To conclude, the two science highlights magnetic semiconductor, europium state in Fe-NO complexes. Two selected from the general purpose nitride, showing that it exhibits additional highlights focus on the use EXAFS beamline BM23 focus on the surprising properties that distinguish of HERFD-XAS to probe the chemically optimisation of Ru-doped spinels it from other materials in this class. A active 4f orbitals of Ce in colloidal Ce used as cathodes for high voltage second highlight from ID12 concerns oxide nanoparticles, and the oxidation lithium-ion batteries, and on revealing observation of strong magneto-chiral state of Mo in the active site of the icosahedral bonding confi gurations in dichroism at the K-edges of transition enzyme nitrogenase. highly undercooled metals. metals in paramagnetic molecular helices. This peculiar type of dichroism Remaining within the realm of catalysis, S. PASCARELLI is due to orbital toroidal currents which but moving from biology to inorganic are of relevance for many phenomena, chemistry, one of the science highlights ranging from multiferroicity to from the pre-upgrade ID24 beamline superconductivity. reports on the spontaneous oscillatory

MEASURING THE DZYALOSHINSKII– Principal publication and authors V.E. Dmitrienko (a), E.N. Ovchinnikova (b), S.P. Collins (c), MORIYA INTERACTION IN A WEAK G. Nisbet (c), G. Beutier (d), Y.O. Kvashnin (e), FERROMAGNET V.V. Mazurenko (f), A.I. Lichtenstein (g) and M.I. Katsnelson (f,h), Nature Physics 10, 202–206 (2014). The Dzyaloshinskii–Moriya interaction twisted magnetic systems, the sign (a) A.V. Shubnikov Institute of (DMI) plays a crucial role in several of the DM interaction has eluded Crystallography RAS, Moscow (Russia) (b) Faculty of Physics, M.V. Lomonosov of the most topical phenomena experimenters because it aff ects only Moscow State University (Russia) in magnetism, including the slow the phase of magnetic scattering, (c) Diamond Light Source Ltd, Didcot magnetic rotation found in skyrmion whereas experiments based on a single (UK) lattices and cycloidal magnetoelectrics. scattering mechanism determine (d) SIMaP, CNRS, Grenoble-INP & UJF, The original formulation of the DMI only intensity (a manifestation Saint-Martin d’Hères (France) (e) ESRF related to the occurrence of ‘weak of the famous ‘phase problem’ of (f) Department of Theoretical Physics ferromagnetism’ in nominally collinear crystallography). This meant that the and Applied Mathematics, Ural Federal antiferromagnetic materials such DMI in such weak ferromagnets could University, Ekaterinburg (Russia) (g) I. Institut für Theoretische Physik, as αFe2O3, MnCO3 and FeBO3 [1,2]. be determined in magnitude and axial Unfortunately, in these archetypal direction, but not in sign. Universität Hamburg (Germany) (h) Institute for Molecules and Materials, Radboud University Fig. 6: Spacegroup- Nijmegen (The forbidden diff raction Netherlands) intensities arising from pure non-resonant magnetic scattering (left ) and pure quadrupole resonant scattering (right – red line) as a function of photon fi eld angle and energy, respectively.

The European Synchrotron 9 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

References Following the suggestion that the sign the phase of the magnetic scattering [1] I. Dzyaloshinsky J. Chem. Phys. of the DMI might be determined by and fi nally the sign of the DMI. Solids 4, 241-255 (1958). measuring interference between non- [2] T. Moriya Phys. Rev. 120, 91–98 (1960). resonant magnetic X-ray scattering Armed with these results, we could [3] V.E. Dmitrienko et al., JETP Lett. 92, and the pure quadrupole resonance at provide a thorough test of state-of- 383–387 (2010). the K absorption edge [3], magnetic the-art fi rst-principles calculations. [4] M.I. Katsnelson, Y.O. Kvashnin, and resonant quadruple scattering A realistic simulation of the spin V.V. Mazurenko and A.I. Lichtenstein were fi rst investigated separately on canting in antiferromagnets leading Phys. Rev. B 82, 100403 (2010). a sample of FeBO3 (Figure 6). The to weak ferromagnetism requires an magnetic structure was observed in accurate account of the spin-orbit the conventional way, by rotating coupling, Coulomb correlations and the magnetism with respect to the hybridisation of the atomic states. To X-ray probe. However, with this class solve this problem we have developed of weak ferromagnets we were able to a method for calculating the DMI carry out such a rotation with a static based on the local force theorem, by crystal, using a small rotating magnet considering the change in energy with to turn the entire magnetic structure. respect to the twisting of the magnetic This approach allows very high-quality moments away from collinear

data to be obtained from a rapid series ordering, Dij = dE/dφij. We proposed [4] of measurements. The quadrupole a new nonlocal inter-site formulation resonance can be seen very clearly as of the DMI, a sharp peak in the pre-edge region i (Figure 6). D = – – Tr {N [^J, ^t ] } 2ij L,σ ji ij +

The fi nal step was to create the where Nji is the inter-site occupation conditions whereby these two scattering matrix, ^J is the total moment operator, ^ processes occur simultaneously and tji is the hopping matrix and [.,.]+ can interfere with each other. To this end, represents an anticommutator. Our measurements of the (009) resonant method leads to a compact and quadrupole refl ection were carried out transparent form that can take into at BM28, again rotating the magnet account hybridisation, correlation and to manipulate the antiferromagnetic temperature eff ects in the system. structure, causing the phase of the Moreover, it can be applied to a wide magnetic scattering to be reversed. range of the materials with diff erent This produced a remarkable jump in strengths of the spin-orbit coupling the energy of the resonance due to the and Coulomb correlations. reversal of the interference conditions on the low and high energy side of the The calculations of the isotropic peak (Figure 7). From the direction of exchange and DMI by this method the jump, we were able to determine have shown that weak ferromagnetism

in FeBO3 is fully described by the six exchange interactions between atoms belonging to the fi rst coordination sphere. Crystal symmetry dictates that the net DMI is along the c axis, which means that there is weak ferromagnetism in the ab plane. Both the sign and magnitude of the small twisting of the spins from collinear ordering are in excellent agreement Fig. 7: The jump in the with experimental data. quadrupole resonance as the magnetic structure was In conclusion, we have applied a rotated by 180 degrees. The relationship between novel experimental technique to a the twisting of the classic problem – the origin of weak magnetic moments and ferromagnetism in FeBO3 – and that of triangles of oxygen shown that the results are in excellent atoms between the iron agreement with a new theoretical planes is shown at the bottom. method that can be used to model the Dzyaloshinskii-Moriya interaction in a wide range of materials.

10 ESRF ELECTRONIC STRUCTURE aND maGNETISm

Principal publication and authors HIGH-ENERGY PLASMONIC P. Cudazzo (a,b), K.O. Ruotsalainen (c), C.J. Sahle (c), A. Al-Zein (d), H. Berger (e), E. Navarro-Moratalla (f), EXCITATIONS IN 2D TRANSITION-METAL S. Huotari (c), M. Gatti (b,g,h) and A. Rubio (a,b), Physical Review B 90, DICHALCOGENIDES 125125 (2014). (a) Nano-Bio Spectroscopy Group, Departamento Física de Materiales, Transition-metal dichalcogenides (IXS) spectroscopy. The results Universidad del País Vasco, Centro de Física de Materiales CSIC-UPV/EHU-MPC (TMDC) have attracted a tremendous represent an important step towards and DIPC, San Sebastián (Spain) interest in the last few years as the understanding of neutral excitations (b) European Theoretical Spectroscopy they exhibit many interesting in this class of materials. In fact, at Facility (ETSF) properties, such as the coexistence variance with the previous theoretical (c) Department of Physics, University of of two competing electronic orders: and experimental work that mainly Helsinki (Finland) (d) ESRF superconductivity and charge-density focused on the low energy plasmon, (e) École Polytechnique Fédérale de wave (CDW) as well as optical valley we investigated the dynamic density Lausanne (EPFL), Institut de Physique control and valleytronics in MoS2 response at higher energy up to 50 eV des Nanostructures (Switzerland) monolayers. for momentum transfer normal to the (f) Instituto de Ciencia Molecular (ICMol), plane of the layers. The dynamic range Universidad Valencia (Spain) (g) Laboratoire des Solides Irradiés, In particular, an intense debate in energy and momentum required by École Polytechnique, CNRS-CEA/DSM, has developed around the origin of this study is only accessible to IXS. Our Palaiseau (France) CDW order and its impact on the results are summarised in Figure 8 (h) Synchrotron SOLEIL, Saint Aubin spectroscopic properties of these where we compare the measured loss (France) materials. The dynamic density function obtained by IXS with state- References response is a key quantity that could of-the-art time-dependent density [1] J. van Wezel et al., Phys. Rev. Lett. help us to disentangle CDW and functional theory (TDDFT) calculations 107, 176404 (2011). superconductivity order. It is quantifi ed performed in adiabatic local density [2] P. Cudazzo, M. Gatti, A. Rubio, Phys. by the loss function, or equivalently approximation (ALDA). The agreement Rev. B 86, 075121(2012). the dynamic structure factor S(q,ω). between theory and experiment is good Its investigation has been the subject enough to show that TDDFT-ALDA of a very recent EELS study that is able to capture the physics has reported a negative plasmon of the collective excitations dispersion for in-plane momentum in this class of materials. At transfers of the loss function of the same time, we found that

2H-NbSe2 [1]. This result was diff erent the spectra of 2H-NbSe2 and from the expectations derived from 2H-Cu0.2NbS2 are similar not the homogeneous electron gas, where only at low energy and for in the plasmon dispersion is positive. plane momentum transfer, as The negative plasmon dispersion was shown in [2], but also at higher ascribed to correlation eff ects arising energy and large momentum from the incipient CDW [1]. transfer normal to the plane. This result further validates our In an earlier publication [2], we interpretation of the in-plane presented the results of an extensive negative dispersion of the low ab initio theoretical analysis of the energy plasmon reported in the loss functions of two prototypical previous work.

TMDC materials, namely 2H-NbSe2 and 2H-Cu0.2NbS2, which respectively do and do not exhibit a CDW ordered phase. We found a negative plasmon dispersion in both materials and proposed a very general and simple interpretation of the in-plane negative plasmon behaviour. Our interpretation Fig. 8: Comparison between links the peculiar anisotropic band experimental (bold line) and structure of those systems with narrow theoretical (thin line) spectra of d-symmetry bands crossing the Fermi Cu0.2NbS2 (a) and NbSe2 (b) at level. diff erent values of momentum transfer Q (in multiple of 2π/c) normal to the plane of the layers. Our study was conducted at ID20 using A broadening of 1.0 eV was used in nonresonant inelastic X-ray scattering the calculations.

The European Synchrotron 11 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

Principal publication and authors M. Moretti Sala (a), K. Ohgushi (b), CaIrO : A SPIN-ORBIT MOTT INSULATOR A. Al-Zein (a), Y. Hirata (b), 3 G. Monaco (a,c) and M. Krisch (a), Phys. Rev. Lett. 112, 176402 (2014). BEYOND THE je ff = 1/2 GROUND STATE (a) ESRF (b) Institute for Solid State Physics, University of Tokyo, Kashiwa (Japan) Spin-orbit coupling, when compared to Whether this scenario also holds true (c) Dipartimento di Fisica, Università di the energy scale of other interactions, for CaIrO3 is debatable because sizeable Trento, Povo (Italy) is usually neglected in the description octahedral distortions are expected to of correlated 3d transition metal drastically lower the symmetry of the

compounds. However, in compounds crystal fi eld compared to Sr2IrO4. The containing 5d elements its role grows robustness of the je ff = 1/2 ground state in importance and the combined against structural distortions has been infl uence of electron correlation and mostly tested by means of resonant spin-orbit coupling gives rise to novel X-ray magnetic scattering. However, quantum phases and transitions. the interpretation of RXMS has been strongly debated and, in the specifi c

The insulating behaviour of case of CaIrO3, a unifi ed picture of the antiferromagnetic Sr2IrO4, for example, ground state has not yet been reached: in is unexpected for a system with a fact, the interpretation of experimental half-fi lled outermost electronic shell results [2] has been contradicted by and spatially extended orbitals. The theoretical calculations [3]. conventional Mott mechanism oft en invoked to explain the insulating We used resonant inelastic X-ray

character of strongly correlated scattering (RIXS) at the Ir L3 edge to systems cannot be envisaged here as address the problem of the electronic

electron correlation is small. However, in structure of CaIrO3. rIXS measurements combination with a strong cubic crystal were carried out at ID20, the new fi eld, spin-orbit coupling enhances the inelastic X-ray scattering beamline.

eff ect of electronic correlation through Figure 9 shows a RIXS map of CaIrO3. the formation of the so-called je ff = 1/2 An elastic line, magnetic excitations ground state: Sr2IrO4 was therefore and possibly phonons are found close to termed a spin-orbit Mott insulator [1]. the zero energy loss line. At increasing energy losses, we assign features to Fig. 9: Ir L3 edge RIXS map of 5 5 intra-t2g (5d t2g → 5d t2g ), t2g-to-eg CaIrO3. 5 4 1 (5d t2g → 5d t2g eg ) and charge- transfer (CT) excitations, following previous RIXS studies. In this work,

we focus on intra-t2g excitations only, i.e. excitations in which electrons are

promoted within the 5d t2g manifold. These are composed by two peaks at 0.65 (B) and 1.22 eV (C), respectively.

To better understand the nature of these excitations and to estimate the eff ective tetragonal contribution to the crystal fi eld splitting and spin- Fig. 10: Schematic orbit coupling in CaIrO3, we adopted band structure of 5d5 a single-ion model [4]: the interacting iridates in the absence Hamiltonian for one hole in the 5d t2g of perturbations (a), 2 states is then written as H = ζL∙S - ΔLz , with spin-orbit in which the tetragonal crystal fi eld coupling (c), and spin-orbit coupling splitting Δ and spin-orbit coupling ζ are plus tetragonal crystal treated on equal footing. The transition fi eld splitting (e). energies are reproduced for ζ = 0.52 and Panels (b), (d), and Δ = –0.71 eV. The sign of Δ is consistent (f) correspond to with structural studies reporting a panels (a), (c) and (e), respectively, when compression of the IrO6 octahedrons the Hubbard term U is and implies a minor contribution of taken into account. the xy orbital to the ground state wave

12 ESRF Electronic structure and magnetism

function (only 10%). We conclude that with a bandwidth w < W (Figure 10c). References [1] B.J. Kim et al., Phys. Rev. Lett. 101, the wave function of CaIrO3 differs As now U > w, a lower (LHB) and an 076402 (2008). significantly from that of Sr IrO , for upper (UHB) Hubbard band are created, 2 4 [2] K. Ohgushi et al., Phys. Rev. Lett. which xy, yz and zx orbitals are equally and the system turns into an insulator 110, 217212 (2013). occupied. (Figure 10d): this is the mechanism [3] N.A. Bogdanov et al., Phys. Rev. B 85, introduced to explain the insulating 235147 (2012). [4] M. Moretti-Sala et al., Phys. Rev. Lett. How to explain the insulating behaviour behaviour of Sr2IrO4. In CaIrO3, 112, 026403 (2014). for CaIrO3 then? Consider the whole however, the tetragonal crystal field is 5d t2g manifold (Figure 10a): the comparable to the spin-orbit coupling bandwidth W would be too large for and three Kramers doublets are formed a reasonable Hubbard energy U (

Principal publication and authors Similarities and differences K. Ishii (a), M. Fujita (b), T. Sasaki (b), M. Minola (c), G. Dellea (c), between electron- and hole- C. Mazzoli (c), K. Kummer (d), G. Ghiringhelli (c), L. Braicovich (c), doped cuprate superconductors T. Tohyama (e), K. Tsutsumi (b), K. Sato (b), R. Kajimoto (f), K. Ikeuchi (g), K. Yamada (h), M. Yoshida (a,i), unveiled by inelastic X-ray M. Kurooka (i) and J. Mizuki (a,i), Nature Communications 5, 3714 (2014). scattering (a) Japan Atomic Energy Agency, Sayo (Japan) (b) Tohoku University, Sendai (Japan) (c) Politecnico di Milano, Milano (Italy) Superconductivity is a state of matter have been central in the study of the (d) ESRF where electrical resistance disappears cuprate superconductors over the last (e) Kyoto University, Kyoto (Japan) completely. This phenomenon is few decades. (f) J-PARC Center, Tokai (Japan) particularly fascinating and mysterious (g) Comprehensive Research Organization for Science and Society, in high-T superconducting cuprates, In this study, we have performed Cu L - c 3 Tokai (Japan) whose complicated electronic and edge resonant inelastic X-ray scattering (h) High Energy Accelerator Research magnetic structure has attracted great (RIXS) experiments of the electron- Organization, Tsukuba (Japan) interest since their discovery more than doped superconductor Nd2-xCexCuO4 (i) Kwansei Gakuin University, Sanda a quarter century ago. at beamline ID08 (now ID32) using the (Japan) AXES spectrometer. Very recently, high- The parent compound of resolution RIXS at the transition-metal cuprate superconductors is an L-edge has become a complementary antiferromagnetic insulator. The technique to the conventional inelastic material becomes a metal when either neutron scattering (INS) for measuring electrons or holes become mobile momentum-resolved spin excitations charge carriers through the effect of [1]. In addition, RIXS possesses doping. The superconductivity occurs sensitivity to charge excitations, which in the doped metallic state and the are hardly observed by neutron based antiferromagnetic interaction, which techniques. causes the spin order in the parent compound, is also intimately related to Figure 11 shows RIXS intensity maps the superconductivity. Therefore, both as a function of momentum in the spin and charge degrees of freedom CuO2 plane (q||) and energy. In the should be clarified comprehensively parent compound Nd2CuO4, spin-wave for a definitive understanding of the excitations with sinusoidal dispersion electron dynamics in the cuprates and are predominantly observed. Charge differences or similarities between excitations are missing in this energy electron- and hole-doped compounds region because the large charge transfer

The European Synchrotron 13 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

Fig. 11: Cu L3-edge RIXS intensity map of the parent antiferromagnetic insulator Nd2CuO4 and the electron-doped superconductor

Nd1.85Ce0.15CuO4. Blue and red circles respectively indicate the peak positions of the spin and charge excitations obtained by fi tting the spectra. The spectra were measured with a π polarised incident beam.

state, both in the electron- and hole- doped superconductors, they diff er substantially in their dispersion relation. In Figure 12, we summarise the spin and charge excitations in the parent antiferromagnetic insulator and the electron- and hole-doped superconductor. A high-energy shift of the spin excitations and their mixture with the charge excitations indicate that the electron dynamics of the electron-doped cuprates has a highly itinerant character in the sub-eV energy scale. In contrast, the spectral gap is about 2 eV. When electrons distribution of the spin excitations also are added by partial substitution of broadens but keeps its energy position trivalent Nd with tetravalent Ce, the almost unchanged upon hole doping spectral weight clearly moves to higher [2], which means that a more localised energy accompanied by the broadening picture is suitable. Our fi ndings impose of the width, as seen in the spectra constraints on theoretical models

of superconducting Nd1.85Ce0.15CuO4. and a comprehensive description We also confi rmed the high-energy of the electronic excitations in the shift of the spin excitations near the electron- and hole-doped cuprates is a magnetic zone centre by inelastic X-ray prerequisite for complete understanding scattering (INS). In addition, we found of the superconductivity. a distinct mode whose fast dispersion connects well to that of particle-hole Finally, we note that while polarisation of excitations observed in the Cu K-edge the scattered photons was not resolved RIXS spectra. We therefore attribute the in this work, its discrimination, which is dispersive mode to the same charge very useful for distinction between spin origin. and charge excitations, is becoming available [4]. Furthermore, this type From the Cu L References 3-edge RIXS, of experiments will soon become even [1] L.J.P. Ament et al., Phys. Rev. Lett. complemented by INS and Cu K-edge more enlightening thanks to the ERIXS 103, 117003 (2009). RIXS, we have clarifi ed the doping spectrometer installed at the new [2] M. Le Tacon et al., Nature Phys. 7, evolution of the spin and charge ID32 beamline, where a simultaneous 725-730 (2011). excitations in the electron-doped increase of energy resolution (x5) and [3] S. Wakimoto et al., Phys. Rev. B 87, cuprate superconductors. Although 104511 (2013). intensity (x3) is expected with respect [4] L. Braicovich et al., Rev. Sci. the spin excitations preserve their to the experimental conditions of this Instrum. 85, 115104 (2014). strength in the superconducting work.

Fig. 12: Schematic of spin and charge excitations in the copper oxides. In the present work, the parent antiferromagnetic insulator and an electron-doped superconductor were studied. Excitations in hole-doped superconductors have been reported in refs. [2,3].

14 ESRF ELECTRONIC STRUCTURE aND maGNETISm

Principal publication and authors TUNING THE MAGNETISM OF 3D-METAL S. Stepanow (a), A. Lodi Rizzini (b), C. Krull (b), J. Kavich (b), J.C. Cezar (c), F. Yakhou-Harris (c), PHTHALOCYANINE ADLAYERS BY P.M. Sheverdyaeva (d), P. Moras (d), C. Carbone (d), G. Ceballos (b), ELECTRON DOPING A. Mugarza (b) and P. Gambardella (a,b), J. Am. Chem. Soc. 136, 5451-5459 (2014). Thin phthalocyanine fi lms have dichroism (XMCD) at beamline ID08 (a) Department of Materials, ETH Zürich (Switzerland) been incorporated in various devices (now ID32) complemented by atomic (b) Catalan Institute of Nanoscience including fi eld eff ect transistors, light- multiplet calculations, which allowed us and Nanotechnology (ICN2), Barcelona emitting devices and photovoltaic cells. to accurately determine changes of the (Spain) Phthalocyanines can accommodate a spin, electronic valence, and crystal fi eld (c) ESRF variety of metal ions in their centre that in electron-doped phthalocyanines. (d) Istituto di Struttura della Materia, CNR, Trieste (Italy) have particular spin confi gurations Figure 13a shows a close-packed rendering this class of compounds array of CuPc. The addition of a small also interesting for magnetic materials amount of Li atoms leads to two major [1]. Due to the particular coordination types of Li-CuPc complexes where the Li geometry of the metal ions, the spin atoms can reside at the ligand position states of the complexes are sensitive (denoted L) or close to the metal centre to their local chemical environment. (denoted M). The relative proportion of Bringing the molecules into contact each type of complex depends on the Li with metal electrodes could therefore coverage. With increasing Li coverage, have signifi cant infl uence on the the preference changes from L-type magnetic properties of the metal to the M confi guration even though centres. Another eff ective route for there are more ligand sites available tailoring the electronic and magnetic than central positions (Figure 13b). characteristics of molecular magnets is The diff erent Li dopant types lead to off ered by direct chemical doping. site specifi c electron donation either to the metal d-states or to the organic The charging sequence of metal- backbone. phthalocyanines involves both ligand and metal orbitals and depends strongly on the nature of the central metal ion. Doping of phthalocyanine chemisorbed on a metal substrate is further complicated due to the interplay of additional substrate induced charging and screening eff ects that can alter the role of the charge donor towards the molecules. The step- wise attachment of electron donor Li atoms to chemisorbed CuPc and NiPc molecules was recently investigated by scanning probe microscopy and DFT calculations [2]. The eff ects of alkali metal doping on the magnetic properties of metal-phthalocyanines, however, have not been studied in a systematic way, neither in the bulk nor at interfaces. These eff ects are of particular importance for molecular electronic and spintronic applications.

This study focuses on the eff ect of Li doping on the electronic and magnetic states of the metal ions for a series of Fig. 13: STM topography of a CuPc adlayer 3d metal-phthalocyanine monolayers with two doped complexes of diff erent types (8.3 nm x 8.3 nm). Statistical analysis (MnPc, FePc, NiPc, CuPc) adsorbed on of the type of dopant as a function of Li Ag(001). Our systematic investigation coverage extracted from the STM data. Error employed X-ray magnetic circular bars account for undefi ned confi gurations.

The European Synchrotron 15 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

References Figure 14 shows the X-ray absorption function of doping, but a transition [1] J. Bartolomé, C. Monton and and XMCD spectra of NiPc as a from an intermediate S=3/2 to a high I.K. Schuller, in Molecular Magnets, function of Li doping. With increasing S = 5/2 spin state. This transition is due J. Bartolomé, F. Luis, J.F. Fernández (Eds.), Springer-Verlag, Berlin Li concentration, the pristine non- to a strong reduction of the ligand fi eld Heidelberg, 221-245 (2014). magnetic Ni ions, which have a spin induced by Li, which is also observed [2] C. Krull, R. Robles, A. Mugarza and singlet d8 confi guration, become for CuPc, NiPc, and FePc. P. Gambardella, Nature Mater. 12, 337 reduced and assume a d9 confi guration. (2013). A magnetic moment corresponding to When combined with STM S = 1/2 appears at the same time, as studies (Figure 13 and ref. [2]), evidenced by the XMCD. The Li doping these results provide a complete has the opposite eff ect For CuPc: the picture of chemically-doped metal- magnetic moment can be turned off by phthalocyanines on surfaces, in which Li doping, due to the complete fi lling of strong correlation eff ects induce site- the d-shell (from d9 to d10). In contrast, specifi c charging and magnetic eff ects. for MnPc, we observed no changes Moreover, these results demonstrate of the d5 metal ion valence state as a that diff erent spin states can be realised in metal-phthalocyanine layers interfaced with metals by doping with Li atoms, off ering a route to tune the magnetic properties of surface-supported molecular systems employing non-magnetic dopants.

Fig. 14: Evolution of NiPc XAS and XMCD at normal incidence with subsequent Li atom deposition.

Principal publication and authors EUROPIUM NITRIDE: A NOVEL DILUTED Do Le Binh (a), B.J. Ruck (a), F. Natali (a), H. Warring (a), MAGNETIC SEMICONDUCTOR H.J. Trodahl (a), E.-M. Anton (a), C. Meyer (b), L. Ranno (b), F. Wilhelm (c), and A. Rogalev (c), Phys. Rev. Lett. 111, 167206 (2013). Ferromagnetism, where a material off er exciting possibilities in the fi eld (a) The MacDiarmid Institute spontaneously forms a net of spintronics, where the functionality for Advanced Materials and magnetisation, is one of the oldest of electronic devices is enhanced Nanotechnology, Victoria University of and most striking examples of by exploiting the intrinsic magnetic Wellington, Wellington (New Zealand) correlated electron phenomena. Most moment of electrons in addition to (b) Institut Néel, Centre National de la Recherche Scientifi que and Université ferromagnets are metals, but under the their charge. The most well studied Joseph Fourier, Grenoble (France) right conditions some semiconductors diluted magnetic semiconductors (c) ESRF can also become ferromagnetic if they are formed by adding manganese to are doped with a small concentration of III-V semiconductors such as GaAs. magnetic ions. Such so-called diluted Mobile charges in the semiconductor magnetic semiconductors (DMSs) host mediate magnetic interactions

16 ESRF ELECTRONIC STRUCTURE aND maGNETISm

Fig. 15: Ferromagnetic response of a nitrogen defi cient europium nitride fi lm. The solid lines represent SQUID magnetometry measurements, while the red symbols represent element specifi c magnetisation measured by XMCD. The XMCD results prove that the ferromagnetism originates in the EuN rather than in an impurity phase.

Fig. 16: X-ray absorption and XMCD results from a ferromagnetic EuN fi lm. Magnetic polarisation of europium ions in both 2+ and 3+ charge between the manganese ions, allowing states is evident. them to form an ordered magnetic state. Understanding the details of the interactions remains a challenge even aft er more than two decades of study [1], and the fi eld remains an active research area. We have recently added a novel material, europium nitride, to the list of known DMSs, and we have shown that it exhibits some surprising dichroism (XMCD) results obtained properties that distinguish it from other at beamline ID12. This powerful materials in the class [2]. technique independently measured the magnetic contribution from the Eu2+ Europium nitride (EuN) is a member and Eu3+ (see Figure 16). The magnetic of the rare-earth nitride series, many signal from the Eu2+ component of the of which are ferromagnets at low EuN closely follows the magnetisation temperature. However, EuN is not of the whole sample (solid circles in expected to be magnetic because of Figure 15), confi rming the central the particular confi guration of the role of Eu2+ in the magnetic state. electrons in the Eu3+ ions. The team at More surprisingly, a strong magnetic Victoria University in Wellington had signal was also found on the Eu3+. This already discovered that in EuN thin proves that the Eu2+ remains coupled fi lms prepared with a slight defi cit of in the EuN matrix rather than forming nitrogen some of the europium ions a separate magnetic impurity phase. References convert to the 2+ charge state. Eu2+ Furthermore, it indicates that the Eu3+ [1] T. Dietl and H. Ohno, Rev. Mod. does carry a magnetic moment, but that forms the bulk of the material is Phys. 86, 187 (2014); T. Jungwirth, J. Wunderlich, V. Novák, K. Olejník, the moments of the individual 2+ ions playing far more than a passive role B.L. Gallagher, R.P. Campion, 3+ had not previously been observed to in the magnetism. Instead, the Eu is K.W. Edmonds, A.W. Rushforth, order into an overall magnetic state acting as a magnetically polarisable A.J. Ferguson and P. Němec, Rev. Mod. [3]. The new research showed that, background that enhances the Phys. 86, 855 (2014). once about 20% of the europium magnetic interactions between the Eu2+ [2] Do Le Binh, B.J. Ruck, F. Natali, H. Warring, H.J. Trodahl, E.-M. Anton, 2+ becomes Eu , magnetic ordering ions beyond that which can be provided C. Meyer, L. Ranno, F. Wilhelm and sets in below a critical temperature of by the small density of conduction A. Rogalev, Phys. Rev. Lett. 111, 167206 125 K (see Figure 15), which is one of electrons in the material. This is in (2013). the highest temperatures of confi rmed striking contrast to conventional DMS [3] B.J. Ruck, H.J. Trodahl, J. Richter, DMS systems. systems where the host semiconductor J. Criginski-Cezar, F. Wilhelm, A. Rogalev, V. Antonov, Do Le Binh, is magnetically inert, and thus F. Natali and C. Meyer, Phys. Rev. B 83, Proof that the magnetism is intrinsic to represents a new paradigm in the DMS 174404 (2011). EuN came from X-ray magnetic circular fi eld.

The European Synchrotron 17 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

Principal publication and authors R. Sessoli (a), M.-E. Boulon (a), STRONG MAGNETO-CHIRAL DICHROISM A. Caneschi (a), M. Mannini (a), L. Poggini (a), F. Wilhelm (b) and A. Rogalev (b), Nature Phys. 11, 69–74 IN A PARAMAGNETIC MOLECULAR HELIX (2015). (a) Department of Chemistry ‘University OBSERVED BY HARD X-RAYS of Florence (Italy) (b) ESRF Chirality is a phenomenon constantly This eff ect is in fact one of the few in front of us [1], as are our hands phenomena that is simultaneously from which its name originates parity odd and time-reversal odd. XMχD (χείρ=hand). It is well known that when has been considered as a possible origin inversion symmetry is absent and a of the homochirality of life on earth, structure is non superimposable on in alternative to the parity-violating its mirror image, a sample absorbs electroweak nuclear interactions [1]. right and left circularly polarised There are very few examples of MχD light diff erently giving rise to natural in literature, most of them reporting circular dichroism (NCD). When a a very weak eff ect. Here we report on magnetised sample interacts with detection of very strong magneto- circularly polarised light, another chiral dichroism in paramagnetic chiral variety of dichroism, magnetic circular complexes in the hard X-ray range. dichroism (MCD), can be observed. A chiral magnetised object interacts in Two molecular helices crystallising a diff erent way with unpolarised light in enantiopure trigonal crystals and depending on the sample chirality, comprising cobalt(II) or manganese(II) the magnetisation vector, and the ions bridged by stable organic radicals direction of propagation of light. This (see Figure 17) have been investigated non-reciprocal phenomenon called at the K-edge of the transition magneto-chiral dichroism (MχD) has metal ions by exploiting the unique been observed for the fi rst time only capabilities of the ID12 beamline. all in 1997 with visible light [2] and three dichroic spectra were measured even more recently with X-rays [3]. on pairs of opposite enantiomers, using single crystals with a volume of the order of mm3. The results for the cobalt derivatives are reported in Figure 17. By using both enantiomers, we were able to check the change of sign for XNCD and XMχD signals as well as the change in sign on reversal of the magnetic fi eld for XMCD and XMχD. It appears clearly from Figure 17 that the spectral shapes of the three dichroic contributions are very diff erent, the magneto-chiral dichroism being signifi cant only where electric quadrupolar transitions from 1s to 3d orbitals of the cobalt atoms are involved. It is also interesting to notice that the intensity of the magneto- chiral dichroism is indeed larger than the magnetic one, in contrast to the

Fig. 17: View of the helix formed by cobalt(II)-hexafl uroacetylacetonates and nitronyl-nitroxide radicals (Co in blue; methyl groups, fl uorine and hydrogen atoms omitted for clarity). The arrows indicate the non-collinear spin structure of the chain. The XANES and the three dichroic signals (XNCD, XMCD, and XMχD) normalised to the XANES intensity at the edge jump, are shown in the four panels respectively (in colour for the depicted helicity, in grey for the other enantiomer).

18 ESRF ELECTRONIC STRUCTURE aND maGNETISm

assumption that MχD is weak being the product of two other weak eff ects (NCD and MCD).

The novelty of the investigation also resides on the possibility of investigating two isostructural helices diff ering for their magnetic properties. In fact, while the cobalt(II) ions have strong magnetic anisotropy due to the orbital magnetic moments resulting in a non-collinear magnetic helix, the manganese(II) derivative, comprising spin-only d5 ions, is a pure Heisenberg spin chain. For the latter, the same hard X-ray investigation revealed comparable natural and magnetic Fig. 18: The XMχD signal detected at dichroisms, but a vanishingly small the K-edge of the cobalt atom provides magneto-chiral contribution. The information on the orbital toroidal currents References combination of magnetic anisotropy (drawn in red), or orbital anapole, while XMCD is indicative of the orbital dipolar [1] G.H. Wagnière, On Chiralty and the with helical structure resulting in a moment (in blue). Universal Asymmetry. (Verlag Helvetica non-collinear spin arrangement seems Chimica Acta, Zürich, 2 007). to be the key for a strong X-ray MχD. [2] G. Rikken and E. Raupach, Nature 390, 493-494 (1997). [3] J. Goulon, A. Rogalev, F. Wilhelm, The magneto-chiral dichroism has an C. Goulon-Ginet, P. Carra, D. Cabaret unusual symmetry, which is odd with and C. Brouder, Phys. Rev. Lett. 88, respect to both parity and time-reversal 237401 (2 002). symmetry but invariant with respect by XMCD are also shown. Orbital [4] V. Scagnoli, U. Staub, Y. Bodenthin, R.A. de Souza, M. García-Fernández, to their product. It is closely related toroidal currents are of relevance M. Garganourakis, A.T. Boothroyd, to the anapole moment, which in this for many phenomena, ranging from D. Prabhakaran and S.W. Lovesey, particular case originates from the multiferroicity to superconductivity Science 332, 696-698 (2 011). toroidal orbital currents of the cobalt [4], while the combination of chirality [5] S. Mühlbauer, B. Binz, F. Jonietz, atoms, as depicted in Figure 18. The and magnetism could be a viable C. Pfl eiderer, A. Rosch, A. Neubauer, R. Georgii and P. Böni, Science 323, 915- more usual orbital currents generating route to store information in protected 919 (2009). the magnetic moment detected magnetic structures like skyrmions [5].

Principal publication and authors A NOVEL SOLUTION FOR THE T.-T. Lu (a), T.-C. Weng (b) and W.-F. Liaw (c), Angew. Chem. 53, 11562 DETERMINATION OF NO OXIDATION (2014). (a) Department of Chemistry, Chung STATE IN Fe-NO COMPLEXES Yuan Christian University (Taiwan) (b) SLAC National Accelerator Laboratory (USA)

– (c) Department of Chemistry, National To understand how the metal iron or NO . With regard to the utilisation Tsing Hua University (Taiwan) utilises its intrinsic reactivity and of valence-to-core X-ray emission redox propensity for the stabilisation spectroscopy (V2C XES) for quantifying of nitric oxide in the form of Fe-NO small-molecule activation such as complexes, an extensive study of N2 in iron-dinitrogen complexes, we the electronic structure of Fe-NO developed V2C XES as a spectroscopic complexes was carried out using a ruler for the quantitative determination variety of spectroscopic methods of NO oxidation state in Fe-NO [1-4]. The comparable energy levels complexes, which was circumvented by of NO π* orbitals and Fe 3d orbitals, invention of Enemark-Feltham notation however, complicate the Fe-NO bonding [5]. interaction and puzzle the quantitative assignment of the oxidation state of Fe- XES displays the dipole-allowed bound nitric oxide as NO+, NO radical, transitions from metal p orbitals to metal

The European Synchrotron 19 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

References 1s orbital aft er ionisation of a metal V2C transition peaks at ~7098 eV, [1] T.-T. Lu et al., Inorg. Chem. 50, 5396 1s electron. The XES valence-to-core ~7103 eV, and ~7110 eV exhibited by (2011). region, in particular, reveals the ligand nitrosyl iron complexes [(NO)Fe(SPh) ]–, [2] J.L. Hess, C.H. Hsieh, S.M. Brothers, 3  – – M.B. Hall and M.Y. Darensbourg, J. Am. np/ns metal 1s transitions, which [(NO)2Fe(SPh)2] , and [(NO)2Fe(OPh)2] . Chem. Soc. 133, 20426 (2011). are enhanced by ~5% contribution of Presumably, the two additional V2C [3] S. Ye, F. Neese, J. Am. Chem. Soc. 132, Fe 4p orbitals into the valence orbitals. transition peaks in the lower energy 3646 (2010). II 2– XES spectra of complexes [Fe (SPh)4] , region result from the incorporation of [4] J.E.M.N. Klein et al., Angew. Chem. 53, [(NO)Fe(SPh) ]–, [(NO) Fe(SPh) ]–, and nitric oxide. 1790 (2014). 3 2 2 – [5] J.H. Enemark and R.D. Feltham, [(NO)2Fe(OPh)2] were collected at Coord. Chem. Rev. 13, 339 (1974). beamline ID26. As shown in Figure 19, Theoretical calculation was further a broad transition peak at 7107.5 pursued to verify the nature of these eV is observed in the normalised Fe spectral transitions. DFT calculated K-valence spectrum of V2C XES spectra of all Fe-NO complexes II 2– complex [Fe (SPh)4] , resemble the experimental V2C as opposed to features. DFT calculation provides three distinctive a computational validation for the * dominant contribution from NO σ2s and σ2p orbitals, respectively, to these Fig. 19: Comparison XES V2C features (Figure 20). That is, of the normalised V2C XES probes the energy diff erence Fe K-valence XES between the σ * and σ orbitals of spectra of complexes 2s 2p II 2– Fe-bound nitric oxide. To test the [Fe (SPh)4] (blue), – [(NO)Fe(SPh)3] (red), application of this energy diff erence, [(NO) Fe(SPh) ]– (black), 2 2 ∆Eσ2s*–σ2p of NO, to identify the oxidation – and [(NO)2Fe(OPh)2] state of nitric oxide, DFT calculation of (purple). + - ∆Eσ2s*–σ2p of NO , NO radical and NO was carried out to establish the linear

relationship between ∆Eσ2s*–σ2p of NO Fig. 20: DFT calculated and its oxidation state. The correlation valence-to-core XES between ∆Eσ2s*–σ2p of NO and its spectra of complexes oxidation state follows the equation: – [(NO)Fe(SPh)3] (top), ∆E = -0.550 × (oxidation state [(NO) Fe(SPh) ]– σ2s*–σ2p 2 2 of NO) + 5.079. Negative correlation (middle), and – between ∆E of NO and its [(NO)2Fe(OPh)2] σ2s*–σ2p (bottom) and the oxidation state suggests that the corresponding addition of electrons into NO π* orbitals molecular orbitals for weakens N-O bonding interaction, each of the transitions. raises the energy level of the σ orbital, Calculated peaks are 2p * displayed in the vertical while reducing that of the σ2s orbital, lines. therefore resulting in an increase in the energy separation between these two orbitals.

Based on the ∆Eσ2s*–σ2p exhibited by – Fe-NO complexes [(NO)Fe(SPh)3] , – – [(NO)2Fe(SPh)2] , and [(NO)2Fe(OPh)2] , the oxidation state of these nitrosyl iron complexes are –0.58±0.18, –0.77±0.18, and –0.95±0.18, respectively. This fi nding suggests that binding of nitric oxide toward non-heme iron, for the sake of the storage and transport of nitric oxide, aff ords Fe-NO complexes and stabilises NO via partial reduction. We aim to expand this spectroscopic method to quantify NO oxidation state in a variety of Fe-NO complexes and to bring insights into Fe-NO bonding interaction modulated by its geometric structure and supporting ligands in the near future.

20 ESRF ELECTRONIC STRUCTURE aND maGNETISm

3+ Principal Publication and authors ABSENCE OF Ce SITES IN J.-D. Cafun (a), K.O. Kvashnina (a), E. Casals (b,c) , V.F. Puntes (b,c) and CHEMICALLY ACTIVE COLLOIDAL CERIA P. Glatzel (a), ACS Nano 7, 10726–10732 (2013). NANOPARTICLES (a) ESRF (b) Institut Català de Nanotecnologia (ICN), Bellatera (Spain) (c) Universitat Autònoma de Barcelona Materials based on ceria (CeO2) lifetime broadening [1]. In the case of (UAB), Bellaterra (Spain) have a capacity to store and release rare earth systems, HERFD-XAS reveals oxygen in their intrinsic n-type the fi ne structure of the chemically fl uorite structure. This provides them active 4f orbitals and thus provides a with unique redox properties that are direct, in situ probe of the Ce chemical widely used and off ers potential for state (Figure 21). promising applications in the areas of catalysis, fuel cells, and biomedicine. A simple ionic picture has been shown

Developing tailored CeO2 materials for to be insuffi cient to describe inner-shell industrial applications involves a clear spectroscopic data of ceria and a multi- understanding of the chemical activity confi guration approach is necessary. of CeO2. This ultimately depends on The electronic structure of Ce in bulk the electronic structure, which can CeO2 with formal valence state of IV can be controlled at the nanoscale by the be described as a mixture of f0 and f1L morphology. contributions (where L denotes a hole in a ligand orbital) thus considering orbital

Reducing the size of CeO2 to the mixing and sharing of electron density nanoscale enhances the chemical between Ce and O. Such formalism activity. The generally advocated implies that the system remains in a mechanism states that particle singlet spin state (S = 0) independent miniaturisation results in an increase of the f0 and f1L mixing ratio because in the surface to bulk ratio and the Ce f electron is spin paired with a thus produces a signifi cantly non- ligand electron. The resulting spectral stoichiometric CeO2-d owing to surface shape is very diff erent from a system oxygen vacancies. Electrons left behind with two distinct f0 (Ce4+) and f1 (Ce3+) by O-vacancy formation are believed sites where the spectrum presents to reduce Ce4+ to Ce3+ ions, resulting the sum of the two sites. Oxygen in a higher level of catalytically active defi cient CeO2-d may be described Ce3+ ions at the surface that are responsible for the increased chemical activity of nanoceria when compared to bulk ceria. This ionic description of the catalytic activity is commonly supported by experiments performed under conditions that are very diff erent from the environment of the working catalyst and may even change the chemical state of ceria (e.g. high vacuum). The electronic structure of ceria nanoparticles critically depends on the reaction environment, thus experiments need to be performed under in situ / operando conditions to fully reproduce the environment under which activity is observed.

Fig. 21: a) Experimental setup for We monitored the electronic structure HERFD-XAS of colloidal nanoparticles. of ceria nanoparticles using high energy The suspension is circulated and the spectra resolution fl uorescence-detected hard are recorded on a free liquid jet. X-ray absorption spectroscopy (HERFD- b) Comparison between conventional total XAS) at beamline ID26. This technique fl uorescence yield (TFY) with HERFD Ce L3-edge spectra in bulk CeO2. The electronic allows spectral features to be observed transitions are assigned to the spectral beyond the limitation of the L3 edge features.

The European Synchrotron 21 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

Fig. 22: Spectral changes in ceria nanoparticles (NPs) with average size of 3 nm aft er synthesis in ethanol (blue) and in the presence

of H2O2 (red). The inset shows a magnifi cation of the 3+ 2p → 4f transitions and a comparison to Ce in Ce(NO3)3.

nanoparticles show rapid ( < 1 minute) and signifi cant changes in the HERFD- XAS (Figure 22) that are accompanied by an increase in the Ce-O inter-atomic distance. The pre-edge peak remains as one strong resonance, leading us to conclude that Ce is in a singlet state by an increased f1L contribution in over the whole course of reaction. The presence of strong orbital mixing and changing intensity ratio between the an unchanged singlet spin state. Such two strong 5d absorption peaks can be a description, however, contradicts the rationalised as a change in the f0 and generally accepted view of a Ce3+ site f1L contributions to the ground state with a local, unpaired spin moment in electronic structure. chemically active nanoceria [2]. Thus, the picture of a delocalised We tested the catalytic activity of electron distribution emerges where References the ceria NPs by monitoring the the entire NP participates in the [1] K. Hämäläinen, D.P. Siddons, decomposition of hydrogen peroxide reaction as opposed to single Ce3+ sites.

J.B. Hastings and L.E. Berman, Phys. Rev. (H2O2). This reaction mimics the The challenge now is to fi nd a quantum Lett. 67, 2850-2853 (1991). activity of catalase, a common chemical theoretical description of the [2] F. Esch, S. Fabris, L. Zhou, T. Montini, enzyme in organisms that balances electron structure that accounts for C. Africh, P. Fornasiero, G. Comelli and R. Rosei, Science 309, 752-755 (2005). the level of reactive oxygen species the particle morphology and is able to and thus protects cells. The 3 nm model X-ray absorption data.

Principal publication and authors R. Bjornsson (a,b), F.A. Lima (a,c), HERFD XAS REVEALS A Mo(III) IN THE T. Spatzal (d,e), T. Weyhermüller (a), P. Glatzel (f), E. Bill (a), O. Einsle(d), NITROGENASE ACTIVE SITE F. Neese (a) and S. DeBeer (a,g). Chem. Sci. 5, 3096–3103 (2014). (a) Max-Planck-Institut für Chemische The conversion of inert nitrogen to Energiekonversion, Mülheim and der ammonia represents a process that Ruhr (Germany) (b) Present address: Science Institute, is fundamental for all life on earth. University of Iceland, Reykjavik (Iceland) While both plants and animals (c) Present address: Centro Nacional require bioavailable ammonia for de Pesquisa em Energia e Materiais, incorporation into proteins and DNA, Brazilian Synchrotron Light Laboratory – neither plants nor animals possess LNLS, Campinas (Brazil) (d) Institute for Biochemistry, Albert- enzymes that are capable of eff ecting Ludwigs-Universität Freiburg (Germany) this conversion. As such, we rely on (e) Present address: Howard Hughes diazotrophs (bacteria and archaea that Medical Institute and Division of are able to fi x atmospheric nitrogen Chemistry and Chemical Engineering, gas using the enzyme nitrogenase) as California Institute of Technology, Pasadena (USA) the biological source of ammonia. The (f) ESRF enzyme active site that is responsible (g) Department of Chemistry and for catalysis is an MoFe7S9C cluster, Chemical Biology, Cornell University, known as FeMoco (Figure 23). The Ithaca (USA) long-time enigmatic cluster was only recently revealed to contain a central carbon atom, and many details of the electronic structure remain unknown [1]. However, understanding the iron Fig. 23: The FeMoco site oxidation state, the molybdenum of nitrogenase.

22 ESRF ELECTRONIC STRUCTURE aND maGNETISm

oxidation states, and the spin coupling assignment and to also show that the References within this complex biological cofactor Mo(III) is in an unusual spin-coupled [1] K.M. Lancaster, M. Roemelt, P. Ettenhuber, Y. Hu, M.W. Ribbe, is essential for understanding the confi guration. This fi nding represents F. Neese, U. Bergmann and S. DeBeer, mechanism and is a prerequisite for the fi rst example of Mo(III) in biology. Science 334, 974-977 (2011). rational bio-inspired catalytic design. The implications that this electronic [2] F.A. Lima, R. Björnsson, confi guration has for nitrogenase T. Weyhermüller, P. Chandrasekaran, Recently, using high-energy resolution reactivity are part of ongoing research P. Glatzel, F. Neese and S. DeBeer, Phys. Chem. Chem. Phys. 15, 20911 (2013). fl uorescence detected Mo K-edge X-ray eff orts in our laboratory. absorption spectroscopy (HERFD XAS) at beamline ID26, we were able to show that the nitrogenase contains a Mo(III) centre in the FeMoco active site (Figure 24). Previous Mo K-edge studies were unable to make a clear assignment of the Mo oxidation state in this cluster, due in large part to the signifi cant core hole lifetime broadening at the Mo K-edge. By using HERFD XAS at beamline ID26, we were able to overcome many of the limitations of standard XaS measurements.

Furthermore, by using a time- dependent density functional approach Fig. 24: A comparison of experimental and to calculate the Mo K-edges [2], calculated HERFD Mo K-edge XAS data for we were able to verify the Mo(III) FeMoco and reference model complexes.

Principal publication and authors WHAT DRIVES OSCILLATIONS IN CO S.J.A. Figueroa (a,b) and M.A. Newton (a), J. Catal. 312, 69 OXIDATION BY O2 OVER RH/AL2O3 (2014). (a) ESRF CATALYSTS? (b) Now at Brazilian Synchrotron Light Source (LNLS), Campinas (Brazil)

Spontaneous oscillatory behaviour To resolve this situation a 5wt%Rh/ in heterogeneous catalysis is of al2O3 catalyst was investigated considerable interest from fundamental during CO oxidation catalysis using and applied points of view; anticipating synchronous application of Rh K and controlling such behaviour is edge dispersive EXAFS (at beamline important to process engineering, as it ID24), diff use refl ectance infrared may be highly problematic to process spectroscopy (DRIFTS) and mass control and effi ciency. spectrometry (MS), at sub-second time resolution. [3] Previous studies of oscillations in

CO oxidation over Rh/Al2O3 systems Figure 25a shows CO2 production have resulted in diff ering opinions as at varying temperatures when a to their source. Shanks and Bailey [1] stoichiometric (2CO:O2) fl ow of concluded that oscillations arose from reactive gas is passed over the Rh/ temperature fl uctuations within the al2O3 catalyst. Rapid variations in reactor, whereas Ioannides et al. [2] CO2 evolution that vary in period/ preferred redox processes within the Rh magnitude are observed. Figure 25b phase as their preferred explanation, shows (I) DRIFTS during the oscillatory though no technique directly sensitive behavior at 433K and (II) the diff erence to the oxidation state of the Rh was between the maximum and minimum applied. levels of CO2 production. Linear, bridged and geminal carbonyls are observed but no relation to the oscillatory

The European Synchrotron 23 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

Fig. 25: a) Oscillations in CO2 (MS) as a function of temperature. b) DRIFTS showing (I) carbonyl speciation present during oscillatory behaviour at 433K, and (II) the diff erence

in spectra taken at points of maximal and minimal levels of CO2 production. c) Temporal variation in overall infrared refl ectivity.

behaviour observed. These species do surface temperature, or changes in show that Rh is present in at least two the refl ectivity of the Rh arising from forms, reduced nanoparticles (Rh(CO) a change from metallic nanoparticles I and Rh2(CO)) and Rh (CO)2 centres, but to more oxidic entities. A precise cannot inform us as to whether redox correlation between these variations (Rh(0) to Rh (3+)) bi-stabilities are and the temperature of the sample directing these events. can be established. However, this does not rule out variations (0 to +3) Figure 25c shows changes in infrared in the oxidation state of the Rh as the refl ectivity (DRIFTS). Here we do underlying source of the oscillations. observe variations that precisely match those observed in MS. These changes Figure 26 shows Rh K edge XAFS may have two origins: variations in obtained at 0 seconds for each temperature. The XAFS envelope is diff erent at each temperature refl ecting changes in the balance between Rh in a I metallic state and (Rh (CO)2). However, the temporal variation of the Rh K edge position (I), the Rh white line (II) – both very sensitive to the Rh0/Rh3+ ratio- and a second XANES feature (III) that indexes the Rh(0) nanoparticulate state (Figure 26b) again show no

relationship to the oscillatory CO2 production.

As such, neither the carbonyl functionality (DRIFTS) nor the oxidation state of the Rh (XAFS) are the sources of the oscillations. Furthermore, the relationship between the temperature Fig. 26: a) Rh K edge XAFS at t=0 seconds of the sample (Figure 25) and the within the CO oxidation experiment as production of CO2 shows that this a function of temperature. b) Temporal cannot be the cause of the oscillations variation observed at the three points but rather a result of them. Therefore we indicated in (a) within the XAFS envelope. have to search for another explanation that is invisible to both the DRIFTS and the Rh K edge EXAFS.

24 ESRF ELECTRONIC STRUCTURE aND maGNETISm

Previous studies [5] have shown onset and magnitude as a function of References [1] B.H. Shanks and J.E. Bailey, J. Catal. that, alongside adsorbing molecular temperature, the oscillations in CO2 110, 197 (1988). CO, in the regimes of Rh particle size production we observe. [2] T. Ioannides, A.M. Efstathiou, and temperature investigated here, Z.L. Zhang and X.E. Verykios, J. Catal. Rh nanoparticles also dissociate CO Therefore, the evidence points to 156, 265 (1995). to yield adsorbed atomic carbon; a a unique mechanism based upon [3] M.A. Newton, Topics in Catalysis 52, speciation invisible to infrared and CO dissociation. This leads to an 1410 (2009). [4] V. Matolin, M.H. Elyakhloufi , K. Masek Rh K edge XAFS. Also, the probability accumulation and periodic ignition of and E. Gillet, Catal. Letts. 21, 175 (1993). of CO dissociation as a function of adsorbed carbon as the source of the temperature matches, in terms of oscillatory phenomena.

RATE CAPABILITY ENHANCEMENT BY Principal publication and authors N. Kiziltas-Yavuz (a), A. Bhaskar (a), D. Dixon (a), M. Yavuz (a,b), ru DOPING IN HIGH VOLTAGE LITHIUM- K. Nikolowski (a,c), L. Lu (d), R. Eichel (e,f) and H. Ehrenberg (a,b), ION BATTERY CATHODE MATERIAL J. Power Sources 267, 533-541 (2014). (a) Karlsruhe Institute of Technology LiNi Mn O (KIT), Institute for Applied Materials 0.5 1.5 4 - Energy Storage Systems (IAM-ESS), Eggenstein-Leopoldshafen (Germany) (b) Helmholtz Institute Ulm (HIU) Lithium-ion batteries are the most vehicles. The application of Li-ion Electrochemical Energy Storage, common energy storage systems batteries in electromobility demands Karlsruhe (Germany) commercialised so far for applications high volumetric as well as gravimetric (c) Fraunhofer Institute for Ceramic in electric vehicles or hybrid electric energy and power densities which Technologies and Systems IKTS, Dresden (Germany) (d) National University of Singapore, Dept. of Mechanical Engineering Fig. 27: a) Cycle (Singapore) number vs. (e) Forschungszentrum Jülich, Institut discharge capacity für Energie- und Klimaforschung (IEK-9) plots at C/2 charge- (Germany) discharge rate. (f) RWTH Aachen University, Institut für b) Cycle number Physikalische Chemie (Germany) vs. discharge capacity plots at C/2 charge rate and varied discharge rates for Ru-doped and un-doped LNMO samples cycled between 3.5-5.0 V at room temperature.

The European Synchrotron 25 SCIENTIFIC HIGHLIGHTS ELECTRONIC STRUCTURE aND maGNETISm

References are factors oft en determined by the work, we have proved that doping of [1] C.M. Julien and A. Mauger, Ionics. 19, positive electrode or cathode materials the LNMO with Ru further increases the 951-988 (2013). [1]. Nowadays, most of the commercial rate capability and cycling stability. [2] A. Bhaskar, D. Mikhailova, N. Kiziltas- Yavuz, K. Nikolowski, S. Oswald, lithium-ion batteries employ layered The Ru doped spinel (LNRMO) with the N.N. Bramnik and H. Ehrenberg, Li(Ni,Co,Mn)O2 and spinel-type composition LiNi0.4ru0.05Mn1.5O4 was Progress in Solid State Chemistry 1, LiMn2O4 (working at ~4 V) as cathode prepared by citric acid-assisted sol-gel 1-21 (2014). and graphite as anode. However, the method with heat treatment at 1000°C spinel type cathode materials possess for the fi nal compound. The material high cycling stability in comparison has an initial capacity of ~139 mAh to the layered counterpart. Spinels g-1 and capacity retention of 84% aft er

have a general formula AB2O4, where 300 cycles at C/2 charging-discharging the oxygen atoms occupy 32e site rate between 3.5-5.0 V vs. Li/Li+. The – (Fd3m space group) and form a face- achievable discharge capacity at 20C centred cubic close packing. The A for a charging rate of C/2 is ~136 mAh cations occupy tetrahedral 8a site, g-1 (~98% of the capacity delivered at whereas the B cations are located in C/2). These results are summarised octahedral 16d site. The unoccupied in Figure 27. The Ru doping inside octahedral 16c site together with the spinel material was confi rmed by the 8a tetrahedral site forms a three XaS studies carried out at beamline dimensional pathway for the lithium- BM23. From the XANES analysis of ion diff usion in these materials. Ni- the Ru K edge data, Ru was found to doped Mn spinel with the composition be in +4 oxidation state. The EXAFS

LiNi0.5Mn1.5O4 (LNMO) is considered as analysis shows that Ru has a nearest highly promising cathode material due coordination bond distance of 2.02 Å to its large reversible capacity obtained and 2.94 Å with oxygen and transition at a high operating voltage of around metal Ni or Mn. This conclusively 4.7 V where the reversible Ni2+  Ni4+ proves that the Ru exists inside the redox reactions take place [2]. In this spinel lattice (Figure 28).

Fig. 28: a) Atomic cluster of

LiNi0.4ru0.05Mn1.5O4 spinel (Red: oxygen; purple: Ni, Mn and Ru, black: Li). b) The Ru K-edge XANES of LNRMO

compared to the reference RuO2, reveals that the oxidation state of Ru is +4. The Fourier transform of the EXAFS clearly shows a strong second peak confi rming Ru bonding to the transition metal (Mn/ Ni).

Principal publication and authors LOCAL FIVEFOLD SYMMETRY IN LIQUID A. Di Cicco (a), F. Iesari (a), S. De Panfi lis (b), M. Celino (c), S. Giusepponi (c) and A. Filipponi (d), AND UNDERCOOLED Ni Phys. Rev. B 89, 060102(R) (2014). (a) Università di Camerino (Italy) (b) IIT, Roma (Italy) Perfect crystal structures can contain might form liquids with local fi ve-fold (c) ENEA Casaccia (Italy) pyramids, cubes, or hexagons, but not symmetry, in the form of icosahedral (d) Università dell’Aquila (Italy) pentagons. The fi ve-fold symmetry structures. The presence of such close- of a pentagon is indeed impossible packed fi vefold confi gurations, favoured to replicate over and over in space to by energetic considerations, was make a conventional crystal. Most considered as a possible explanation solid metals actually form closely for the peculiar undercooling properties packed cubic crystals, but in the 1950s, of liquid metals. Crystal nucleation is calculations showed that metals hampered by the positive liquid-crystal such as copper, gold, silver, and lead interfacial energy, resulting in the

26 ESRF ELECTRONIC STRUCTURE aND maGNETISm

possibility of bringing a liquid sample Fig. 29: X-ray absorption into a metastable (undercooled) molten temperature scans of the Ni sample. The change in state below the melting temperature. the absorption coeffi cient However, fi ve-fold ordering in metals at E = 8.338 keV is plotted is not entirely unprecedented. with respect to the solid Quasicrystals are metal alloys that Ni absorption level during have a peculiar semi-crystalline fi ve- repeated thermal cycles. Inset: temperature dependence of fold symmetry that never quite repeats the Ni nucleation rate from the itself. The determination of the local cooling scan. structure in liquid matter is quite a challenging task. Most techniques (X-ray and neutron diff raction) can only fi nd the relative positions of two atoms at a time, which makes it diffi cult to pin Fig. 30: EXaFS extracted signals down the fi ve-fold symmetry prediction. for diff erent phases of Ni. On the other hand, X-ray absorption spectroscopy (XAS) is a technique particularly sensitive to short range ordering and able to determine also three-body distributions. In this work, we used the ESRF XAS beamline [1] (now BM23) to perform XAS measurements combined with complementary techniques, such as fi xed energy X-ray absorption temperature scans and X-ray diff raction. Reaching extremely high temperatures, we have investigated the phase at 1493 K. The extracted signals short-range structure in solid and liquid are compared in Figure 30. Ni, following a previous study on liquid Cu [2] and other metals. Samples were Using a sophisticated data-analysis prepared from a submicrometric Ni and technique, known as Reverse Monte alumina (Al2O3) powder mixtures with Carlo (RMC), coupled with GNXAS a 1:20 mass ratio that were suspended multiple-scattering calculations, it in alcohol, fi ltered, and pressed has been possible to create a model into 100-μm-thick, 13-mm-diameter structure for the liquid reproducing pellets. The furnace consisted of a our experimental data with some 130-mm-diameter cylindrical Pyrex constraints. Moreover, ab initio glass vessel with a suitable window for computer molecular-dynamics (MD) X-rays. The pellets were placed inside simulations provided further insights the crucible of the furnace and the heat to the determination of the local treatments were performed under high- structure. From both RMC and MD vacuum conditions (P = 10-5 mbar). confi gurations we were able to calculate Temperature was monitored using a diff erent physical quantities, such as high-temperature pyrometer. X-ray bond angle distribution functions. absorption at fi xed energy as a function Two diff erent approaches have been of temperature is shown in Figure 29. used to analyse the local geometry: Upon cooling, the absorption remains the common-neighbour analysis stable about 300 K below the melting (CNA) and the calculation of spherical point, showing the deep undercooling harmonic invariants, giving consistent properties of small metal particles. results. The fractions of fi vefold nearly The relatively broader decrease in icosahedral confi gurations found were absorption is associated with the in the range 14-18% for Ni in the progressive nucleation of the broad liquid undercooled phase and 11-14% mass droplet distribution in the sample. for liquid Ni just above the melting Analysis of this part of the scan using temperature. The emerging picture for References previous methods allows us to extract the liquid structure is that of a mixture [1] A. Filipponi, M. Borowski, the Ni nucleation rate in a relatively of nearly icosahedral structures D.T. Bowron, S. Ansell, S. De Panfi lis, wide temperature range (Figure 29, embedded in a disordered network A. Di Cicco and J.P. Itié, Rev. Sci. Instr. 71, 2422 (2000). inset). High quality EXAFS spectra mainly composed of fragments of [2] A. Di Cicco, A. Trapananti, S. Faggioni have been collected in the solid phase highly distorted icosahedra, structures and A. Filipponi, Phys. Rev. Lett. 91, at 1493 K, in the stable liquid phase at reminiscent of the crystalline phase 135505 (2003). 1733 K and in the undercooled liquid and other confi gurations.

The European Synchrotron 27 SCIENTIFIC HIGHLIGHTS Highlights 2014

Soft condensed matter

This chapter presents a selection of (J. Moeller et al., Phys. Rev. Lett. 112, 2686, 2014). X-ray photon correlation articles related to Soft Matter Science 028101, 2014), nanoscale structure spectroscopy has made new inroads

published during the past year. A of Si/SiO2/organics interfaces (H.-G. into the probing of atomic dynamics particular emphasis in this edition Steinrueck et al. ACS Nano 8, 12676– in glasses (B. Ruta et al., Nature is on the self-assembly that is at the 12681, 2014 ), etc. Communications 5, 3939, 2014), origin of many fascinating features a publication summarised in the exhibited by soft matter systems. From On the technical front, very impressive chapter on Dynamics and extreme simple surfactant molecules to highly progress has been made during the conditions. At ID13, a pioneering serial tailored complex organic molecules year. The upgrade beamline project protein crystallography experiment and biomolecules, scattering UPBL9a was successfully completed was performed by an international experiments once again have proved at ID02 and the new beamline team (P. Nogly et al., IUCr J., 2, doi: to be unique for unravelling the reopened for Users in July with 10.1107/S2052252514026487,(2015)). complex pathways and mechanisms unique capabilities for ultra low angle Furthermore, ID13 has received a of self-assembly processes. Indeed, scattering. The upgrade beamline was state-of-the-art Eiger 4M detector advanced modelling and computer officially inaugurated during the ESRF from Dectris that is expected simulations were essential for the Science Advisory Committee meeting to revolutionise nanodiffraction extraction of quantitative information on the 06 November. The sister project experiments. from these scattering experiments, UPBL9b at ID09 is progressing well, which again proved to be a winning the combined time-resolved X-ray Also during the year, ID09B had a combination in the investigation emission spectroscopy and wide- successful beamline review in May. of complex systems. Such studies, angle scattering setup has been The scientific staff of the Structure of driven primarily out of fundamental commissioned. New focusing optics Soft Matter Group have relocated to curiosity, are now approaching closer and an area detector are expected the Science Building, sharing the third to the engineering of smart materials to be in place by 2016, when the floor with the Large Scale Structures and devices pertinent to modern beamline will become fully dedicated Group of the ILL. The Partnership for technology. Needless to say that only to time-resolved experiments. A new Soft Condensed Matter (PSCM) has a small fraction of the exciting results diffractometer for liquid scattering attracted several collaborative partners published during the year made it into will be installed at the soft interfaces from academia and their contracts are this chapter. To mention a few notable station of ID10. In addition, a notable currently being finalised. New services omissions: the millisecond dynamics step has been made in the three- offered to soft matter industrial of protein upload in multicompartment dimensional analysis of highly- customers using techniques such as lipid cubic nanocarriers (B. Angelov et organised surface structures by SAXS, WAXS and microdiffraction al., ACS Nano 8, 5216, 2014); pressure means of grazing-incidence small- are yielding good dividends. Finally, induced re-entrant phase separation angle scattering (E.A. Chavez Pandura the local organisation for the next in concentrated protein solutions et al., ACS Appl. Mater. Interfaces 6, International Soft Matter Conference in 2016 (ISMC2016) is advancing well. This meeting, under the auspices of the SOFTCOMP network of excellence will be jointly organised by the ESRF and ILL together with Grenoble University, CEA and CNRS.

T. Narayanan

Inside view of the 34 m detector tube at ID02. The detector wagon at the far end travels along the rails.

28 ESRF SOFT CONDENSED maTTER

Principal publication and authors FROM SPHERES TO WORMS: G.V. Jensen (a), R. Lund (b), J. Gummel (c), T. Narayanan (c) and J.S. Pedersen (a), Angew. Chem. Int. Ed. MONITORING TRANSFORMATION 53, 11524–11528 (2014). (a) Department of Chemistry and KINETICS OF SURFACTANT MICELLES Interdisciplinary Nanoscience Center (iNANO), Aarhus University (Denmark) USING TIME-RESOLVED SAXS (b) Department of Chemistry, University of Oslo (Norway) (c) ESRF Since their discovery about a century This makes it exceedingly diffi cult to ago, surfactant micelles have been monitor the structural evolution in real a topic of great interest in both time, due to the lack of methods with academic and technological research. both nanostructural resolution and Surfactant molecules consist of distinct millisecond temporal resolution. hydrophobic and hydrophilic parts that, in water, lead to aggregation into Here we use the high brilliance of the a wide variety of interesting and highly ID02 beamline to monitor the kinetics dynamic nanostructures. The shape of the transformation from spherical to and size of the micelles depend on worm-like micelles induced by adding the type of surfactant and conditions sodium chloride (NaCl) to an aqueous such as temperature and solvent solution of the archetypical anionic composition. Due to the widespread surfactant, sodium dodecyl sulphate applications ranging from detergents (SDS). In aqueous solution SDS forms and additives in daily products to small spherical or ellipsoidal micelles advanced applications in biotechnology due to the repulsion between the and nanotechnologies [1], insight into charged head groups which favours their fundamental behaviour is still a large surface curvature. By adding highly relevant. salt, the repulsions are screened, and a smaller curvature is favoured, Equilibrium structures of surfactant leading to a transition into fl exible, micelles have been studied extensively, “worm-like” cylinders (Figure 31a). but the pathways of transitions Using a stopped-fl ow apparatus for between diff erent morphologies are still rapid mixing, combined with the not well understood. Micelle kinetics small-angle X-ray scattering (SAXS) usually takes place in the millisecond setup at ID02, we were able to follow and even sub-millisecond regime. the structural details of the process

Fig. 31: a) Transition from spherical to worm-like SDS micelles upon addition of salt. b) Experimental setup where the X-ray pulse is synchronised with the sample mixing. Using a stroboscopic-type sampling, millisecond time resolution was obtained. c) Scattering curves: Intensity vs. modulus of the momentum transfer vector, Q, as a function of time. The data are described using a model for core-shell ellipsoids and worm-like cylinders (lines).

The European Synchrotron 29 SCIENTIFIC HIGHLIGHTS SOFT CONDENSED maTTER

Fig. 32: Results of the quantitative model fi t analysis showing the time dependence of a) the fraction of surfactant in ellipsoidal micelles, and b) the average contour length of the worms. The hatched area indicates the resolution limit of the experiment. c) Schematic illustration of formation of “worm-like” micelles.

over time. The time gap between data by growth in the contour length of frames was limited by the detector the worms. Interestingly, the growth read-out time, and a stroboscopic- approximately follows a linear law

type sampling was employed to Lave ∼ t (Figure 32b). This resembles capture the rapid kinetic process step-wise growth polymerisation (Figure 31b). By repeating the mixing reactions where a linear relationship sequence several times, varying the between mean molecular weight and delay between the time of mixing and time is observed [3]. In such reactions, collection of the fi rst data frame, we the polymers grow through addition obtained sampling with only a few reactions at the chain ends, indicating milliseconds increment. that a similar mechanism applies here. This suggests a growth mechanism of Figure 31c depicts the scattering micelle fusion at the cylinder ends, curves obtained aft er mixing a 1% where the curvature and surface energy SDS solution with 2 M NaCl (1:1). is highest (Figure 32c). At a lower NaCl The intensity rapidly increases, concentration of 1 M, the repulsion accompanied by a Q-2 behaviour at between the surfactant head groups is low Q at later times, demonstrating less well screened. The micelles do not formation of worm-like micelles. To undergo a transition into worms, but quantify the results, the data were a slight elongation (within ~10 ms) is analysed using a model consisting observed (Figure 32c). of a linear combination of ellipsoidal and worm-like core-shell particles In conclusion, by time-resolved SAXS, [2]. The result, in terms of fraction we have been able to shed new light of surfactant in ellipsoidal micelles on an old classical surfactant system References [1] D. Attwood and A.T. Florence, and contour length of the worm-like and obtain insight into morphological Surfactant systems, their chemistry, micelles, is plotted in Figure 32. transition of micelles. Through a pharmacy and biology, Chapman and quantitative analysis, we deduced that Hall, New York (1983). The fraction of ellipsoids approximately the kinetic pathway of the transition [2] J.S. Pedersen and P. Schurtenberger, follows a single exponential decay, from spherical to worm-like micelles Macromolecules 29, 7602 (1996). [3] P.J. Flory, Principles of polymer ϕell(t) ~ exp(–t/tE), with a time involves successive fusion events, chemistry, Cornell University Press constant of about tE = 10 ms bearing similarities to classical (1953). (Figure 32a). This decay is followed polymerisation reactions.

30 ESRF SOFT CONDENSED maTTER

Principal publication and authors NON-EQUILIBRIUM COLLECTIVE T. Reusch (a), D.D. Mai (a), M. Osterhoff (a), D. Khakhulin (b), M. Wulff (b) and T. Salditt (a), Physical DYNAMICS IN PHOTO-EXCITED LIPID Review Letters 111, 268101 (2013). (a) Institut für Röntgenphysik, Georg- MEMBRANES August-Universität Göttingen (Germany) (b) ESRF

All living systems are intrinsically out of as a function of time delay, from a few thermal equilibrium. Their metabolism picoseconds to several microseconds, results in a multitude of local force before the procedure was repeated to centres on a molecular scale. ‘Active’ gain suffi cient statistics. Figure 33 molecular transport in cells, motor summarises the experiment proteins, and shape transformations schematically. of biological membranes are just some of the most prominent examples. From the data, we extracted the From a fundamental point of view, the temporal evolution of the height understanding of how a local energy correlation functions describing release may change the properties of the lipid bilayer undulations aft er a soft biomolecular assembly remains excitation (Figure 34). The results elusive in many cases. We wanted to indicate that pulsed laser illumination investigate how a system with long- even at quite moderate peak intensities range non-covalent interactions relaxes of about 105 W/cm2 leads to signifi cant back to the equilibrium state aft er it changes of the in plane membrane has been driven out of equilibrium by correlation length by up to 50% as external forces and the extent to which well as the excitation of transient collective modes of many interacting conformal undulation modes of a biomolecules are involved. well defi ned lateral wavelength. The

For the important case of membranes, considered to be Nature’s most important interface, non-equilibrium fl uctuations can be expected to be driven by active proteins or external forces. We have now shown that photo excitation of fl uorophor-labelled lipids in a membrane can lead to a very characteristic collective response and have studied the pathway and time scales on which structural dynamics evolves.

In the experiment, we monitored the shape fl uctuations in a stack of phospholipid bilayers mainly composed of the model lipid DOPC (1,2-dioleoyl-sn-glycero-3- phosphatidylcholine), deposited on a Fig. 33: a) Schematic of the laser quartz surface in solution. To excite the pump/X-ray probe experiment at beamline system, a short picosecond laser pulse ID09B on the multilamellar lipid stack in was used, matched in wavelength to the fl uid phase. The temporal evolution of the absorption band of fl uorescently membrane undulations was studied in response labelled lipids (Texas-red), which to a short pulse excitation, by recording the diff use scattering pattern and line-shape were mixed into the membranes, as analysis of the individual lamellar refl ections. routinely used for optical fl uorescence b) Example of a lamellar diff raction pattern experiments. Aft er energy uptake, the with primary beam (PB), specular beam (SB), system’s response was probed by well and the two fi rst lamellar diff raction orders. defi ned picosecond X-ray pulses. The c) From the two-dimensional distribution, the temporal evolution of the lateral and vertical characteristic diff raction pattern of correlation functions are extracted. d) Sample the membrane stack (non-specular chamber for oriented Texas-Red labelled lipid diff use scattering) was then recorded multilayers fully immersed in solution.

The European Synchrotron 31 SCIENTIFIC HIGHLIGHTS SOFT CONDENSED maTTER

observed phenomena evolve on nano- For compensation, the membranes to microsecond timescales aft er optical buckle, opening up a faster relaxation excitation, and can be described in mechanism than the thermal expansion terms of a modulation instability in the of the entire system. The results may lipid multilamellar stack. have important implications for many other optical fl uorescence experiments, In simple terms, the energy uptake in which it is tacitly assumed that at the molecular level leads to non- the supra-molecular structural state thermalised local vibrations which can of the sample remains unchanged, be regarded as “local heat”, before the apart perhaps from more or less long-range undulation modes have negligible heat generation. Aft er all, the equilibrated. In such a transient state, excitation of embedded fl uorophores the lipid molecules tend to expand their in biomolecular assemblies is a very inter-molecular distances, but on short common tool in biomolecular and life time scales this motion is quenched. science. Similar collective dynamic eff ects may thus occur in many photo Fig. 34: a) Evolution labelled soft matter sample systems, of the diff use eff ectively driving the structure away scattering intensity from thermal equilibrium. as a function of time aft er excitation. The Importantly, the non-equilibrium increase refl ects the transient increase eff ects do not directly manifest in membrane themselves on the molecular length undulations. scale, but rather on mesoscopic b) A characteristic scales via collective mechanisms. In undulation mode the present example, local molecular with a wavelength band centred energy uptake by absorption of photons around 130 nm in the membrane leads to peculiar appears, reaching long-range mesoscopic changes its maximum about observable over microseconds before 1 µs aft er excitation. the sample has reached equilibrium again. The energy dissipation pathway involves a characteristic sequence of modes at length scales of a few hundred nanometres. We speculate that this may be a quite general phenomenon, not only applicable to optical excitation, but also to other excitation mechanisms and local force centres occurring in active states of biomolecular and soft matter.

Principal publication and authors TWO-DIMENSIONAL BINARY S.-H. Lim (a), H.-S. Jang (a), J.-M. Ha (a), T.-H. Kim (b), P. Kwasniewski (c), T. Narayanan (c), SUPERLATTICE OF A SINGLE-WALLED K. S. Jin (d) and S.-M. Choi (a), Angew. Chem. Int. Ed. 53, 12548 (2014). CARBON NANOTUBE/CYLINDRICAL- (a) Department of Nuclear and Quantum Engineering, KAIST (Rep. of Korea) MICELLAR SYSTEM (b) Korea Atomic Energy Research Institute (Rep. of Korea) (c) ESRF Binary or multicomponent properties through synergetic (d) University of Science and Technology (Rep. of Korea) superstructures of nanoparticles, are of coupling between diff erent types of great interest for potential applications nanoparticles. Exciting progress has such as optical and electronic devices, been made in the fabrication of binary sensing, and energy storage as well spherical-nanoparticle superlattices as their own scientifi c merit [1]. Their with various symmetries by using an synthesis may provide new emerging interplay of entropy and van der Waals,

32 ESRF SOFT CONDENSED maTTER

electrostatic, and other interactions. Fig. 35: a) Azimuthally However, systematic experimental averaged SAXS intensities of C E /water (45:55, studies on the formation of binary 12 5 top), p-SWNT/C12E5/water superlattices by mixtures of two (15:45:55, bottom). diff erent types of 1-D nanoparticles have been very rare, and the reported structures have been limited to simple small-angle neutron scattering liquid-crystalline structures. Here, a experiment. The coupled peak- highly-ordered and highly-aligned position ratios and the origin binary superlattice of 1-D nano-objects of new peaks in the SAXS is demonstrated. intensity indicate that the mixture forms a highly ordered One approach to the synthesis of intercalated hexagonal binary highly-ordered binary 1-D nanoparticle superlattice of p-SWNTs and superlattices is to fi rst fi nd conditions C12E5 cylinders, in which a under which 1-D nanoparticles of hexagonal array of p-SWNTs one type self-assemble into a highly is embedded in a honeycomb ordered superlattice, and then to add lattice of C12E5 cylinders. another type of 1-D nanoparticles into the preformed superlattice so that To unequivocally identify the second type of 1-D nanoparticles the structure, we carried out can collectively self-assemble with additional SaXS measurements the fi rst type of 1-D nanoparticles. at beamline ID02 for the Hydrophilically functionalised single- mixture under an oscillatory walled nanotubes (p-SWNTs [2]) with shear (Figure 36 a,b). The a diameter of 5 nm were added to a p-SWNT/C12E5/water loaded in solution of the nonionic surfactant a temperature-controlled shear cell was

C12E5 in water with a fi xed 12 C E5/ slowly cooled down from the isotropic water mixing ratio (45:55 by weight) phase (30°C) to the intercalated at which the C12E5/water solution hexagonal phase (19°C) under an shows a highly ordered hexagonal oscillatory shear with a shear stress phase (hexagonally packed cylindrical of 500 Pa and an oscillation frequency micelles with a diameter of 4.44 nm) of 5 Hz. The 2-D SAXS patterns at low temperature. Small-angle X-ray measured along the radial direction are scattering (SAXS) measurements from highly anisotropic; all the scattering the C12E5/water (45:55) and p-SWNT/ peaks were found along the direction C12E5/water (10:45:55) systems perpendicular to the fl ow direction. The performed at beamline ID02 (Figure 35) 2-D SAXS pattern measured along the show that a new sharp peak appeared tangential direction shows clear six-fold at q = 0.715 nm-1, which has a ratio of symmetries. These results unarguably -1 1:√3 with the peak at q=1.242 nm . confi rm that the p-SWNT/C12E5/water In fact, all the SAXS peaks of p-SWNT/ indeed forms the intercalated hexagonal

C12E5/water are directly coupled to binary superlattice (Figure 36c). When each other with specifi c peak-position p-SWNTs are added to the hexagonally ratios (1:√3:2:√7:3:√12:√21). The packed C12E5 cylindrical-micellar new scattering peaks in the p-SWNT/ system, some C12E5 cylinders are

C12E5/water system arise from replaced with p-SWNTs in such a way correlations between p-SWNTs, which that the free-volume entropies for was confi rmed by a contrast-matched both p-SWNTs and C12E5 cylinders are

Fig. 36: SAXS Intensities of p-SWNT/

C12E5/water aligned by an oscillatory shear, measured along the radial (a) and tangential (b) directions. c) Schematic illustration of the intercalated hexagonal

binary superlattice of p-SWNT/C12E5/ water.

The European Synchrotron 33 SCIENTIFIC HIGHLIGHTS SOFT CONDENSED maTTER

References maximised, resulting in the intercalated SWNT superstructure with hexagonal [1] J.J. Urban, D.V. Talapin, hexagonal binary superlattice. The symmetry. The understanding obtained E.V. Shevchenko, C.R. Kagan and alignment was maintained for more in this study may provide new insight C.B. Murray, Nat. Mater. 6, 115 – 121 (2007). than 24 hours. Although there have for engineering highly-ordered binary [2] T.H. Kim, C. Doe, S.R. Kline and been many examples of highly superlattices of 1-D nanoparticles S.M. Choi, Adv. Mater. 19, 929 – 933 oriented SWNT systems, most of with new synergetic functionalities (2007). which had no specifi c symmetry in such as metallic, semiconducting, and the cluster of SWNTs, this structure is magnetic nanorods, which are of great the fi rst example of a highly oriented current interest.

Principal publication and authors M.A.J. Gillissen (a,b), SOLVENT EFFECTS PROMOTE TRIPLE M.M.E. Koenigs (a,b), A.J.H. Spiering (a), J.A.J.M. Vekemans (a), HELIX FORMATION IN AMPHIPHILIC A.R.A. Palmans (a,b), I.K. Voets (a,b) and E.W. Meijer (a,b), J. Am. Chem. Soc. 136, DISCOTICS 336-343 (2014). (a) Laboratory of Macromolecular and Organic Chemistry, TU Eindhoven (The Self-assembling synthetic molecules via π−π interactions. In pure IPA, the Netherlands) in water is of great interest for the hydrophobic eff ect is absent and only (b) Institute of Complex Molecular development of functional soft matter π−π interactions are operative, resulting Systems, TU Eindhoven (The and fi nds applications in regenerative in one-dimensional helical fi bres Netherlands) medicine and sensors. Defi ning rules (Figure 37, left ). for eff ective self-assembly in water with the ability to control the shape The behaviour of the discotic and function of the supramolecular amphiphile was investigated as a structures formed is, however, less function of solvent composition in

trivial. A delicate balance between IPA–H2O mixtures and temperature. hydrophobic eff ects and other non- Remarkably, three distinct structural covalent interactions is oft en required regimes were detected with all but not easy to achieve. techniques (Figure 38). SAXS showed that a 1D helical fi bre was present With the help of small-angle X-ray at high IPA concentration, evidenced scattering (SAXS) at beamline ID02, in by a scattering curve typical for long combination with circular dichroism elongated objects with a homogenous (CD) spectroscopy and cryogenic electron length density (regime III). The transmission electron microscopy addition of water induced the formation (cryo-TEM), we have now unravelled of small, bundled aggregates, showing how the self-assembly behaviour a signature scattering curve for core– in water of an amphiphilic discotic shell assemblies with a low aspect ratio molecule (Figure 37, middle) is aff ected (regime II). At high water content, long, by the presence of isopropanol (IPA), one-dimensional core–shell assemblies added to enhance its solubility. The are formed, consisting of 3-4 fi bres in molecular structure of the discotic the cross-section. Because the discotics amphiphile is designed to form bundled have a fl exible propeller shape and

structures in H2O (Figure 37, right) are never completely fl at, the most and incorporates both hydrophilic/ likely conformation is a triple helix of hydrophobic peripheral side chains helices. The results obtained by SAXS and a core that strongly aggregates measurements were corroborated by cryo-TEM measurements: short fi bres were observed for regime II and long fi bres for regimes I and III. CD spectroscopy showed strong Cotton eff ects in all three regimes, confi rming the presence of helical aggregates.

Fig. 37: Self-assembly of an amphiphilic discotic into (left ) one-dimensional helical fi bres at low water content and (right) triple helical structures at high water content.

34 ESRF SOFT CONDENSED maTTER

Fig. 38: a-c) Schematic depiction of the three regimes in self-assembly of the discotic -4 amphiphile in IPA–H2O. CD-spectra of c = 3x10 M and T = 0°C for: d) fIPA = 0.25; -4 e) fIPA = 0.5; f) fIPA = 1; Cryo-TEM images at c = 3x10 M for g) fIPA = 0.25; ) fIPA = 0.75; i) fIPA = 1; SAXS profi les (symbols) and form factor fi ts (lines) in IPA–H2O mixtures at -4 c = 3×10 M and T = 20°C for: j) fIPA = 0.25; k) fIPA = 0.5; l) fIPA = 1.

The most noteworthy fi nding of this The above fi ndings demonstrate research is that the boundaries of the that the self-assembly of the discotic three regimes coincide with the two amphiphile discussed here in IPA– extrema (φmax and φmin) that are present H2O mixtures is intricately linked to in the excess enthalpy of mixing the mixing behaviour of the IPA–H2O phase diagram of IPA–H2O mixtures solvent mixture. This fi nding is highly [1]. This suggests that the solvent important since the nanostructures mixing thermodynamics drive the obtained upon self-assembly in observed morphological transitions, water can be greatly infl uenced by as also suggested by Trappe and co- the preparation method. Frequently, workers for the phase behaviour of samples are prepared by the direct poly(N-isopropylacrylamide) in IPA– injection into water of solutions

H2O mixtures [2]. To confi rm that the of small self-assembling units in a boundary between regime II and III is “good” solvent such as an alcohol or indeed coupled to excess enthalpy of THF. Understanding how the solvent mixing, we performed a temperature- composition infl uences the aggregate dependent experiment at φIPA = 0.95. structure helps us to unravel the At 80°C, φmax is well below 0.95 while pathways by which the fi nal structures at 0°C φ shift s to values larger are formed in water. max References than 0.95. This means that cooling a [1] a) F. Franks and D.J.G. Ives, Q. Rev. solution of the discotic amphiphile in Chem. Soc. 20, 1 (1966);

φIPA = 0.95 should induce a change in b) T. Katayama, Kagaku Kogaku the nature of the aggregates typical for Ronbun. 26, 361 (1962); c) R.F. Lama, B.C.Y. Lu, J. Chem. Eng. Data 10, 216 regime III to an aggregate typical for (1965). regime II. Indeed, the CD spectrum at [2] I. Bischofb erger, D.C.E. Calzolari, 10°C is typical for regime III, while it is P. De Los Rios, I. Jelezarov and V. Trappe, typical for regime II at 0°C. Scientifi c Reports 4, 4377 (2014).

The European Synchrotron 35 SCIENTIFIC HIGHLIGHTS SOFT CONDENSED maTTER

Principal publication and authors F. Liu (a), M. Prehm (b), X.B. Zeng (a), SKELETAL CUBIC, LAMELLAR, AND C. Tschierske (b) and G. Ungar (a), J. Am. Chem. Soc. 136, 6846−6849 (2014). (a) Department of Materials Science and RIBBON PHASES OF BUNDLED Engineering, University of Sheffi eld (UK) (a) Institute of Chemistry, Organic THERMOTROPIC BOLAPOLYPHILES Chemistry, Martin-Luther-University Halle-Wittenberg, Halle (Germany) Ordered periodic structures can self- In the work highlighted here, a series assemble on several diff erent length of T-shaped polyphilic molecules scales. Crystals form by ordered packing composed of a rigid linear aromatic of atoms, i.e. units on the 0.1 nm scale. core with a polar glycerol group at Block copolymers can form regular each end and one “swallow-tail” lattices of the diff erent polymer blocks partially fl uorinated side-chain of typically of 10-100 nm size and diff erent lengths n (see Figure 39) were colloidal particles, such as polystyrene synthesised and their LC phases were beads or inorganic glassy spheres (such investigated by small-angle grazing as in opal), can form “colloidal crystals” incidence X-ray scattering (GISAXS) at with periodicities between a fraction beamline BM28, as well as by other of a µm to several µm. The most techniques. prominent periodic self-assembled systems in the 1-10 nm range can be Similar molecules with smaller side- classifi ed as liquid crystals (LC) in a groups show a remarkable series of broader sense. They contain molecules honeycomb-like LC phases with the typically combining two or more aromatic rods as the “wax” and the poorly compatible chemical groups side-chains as “honey”. However, (amphiphilic or polyphilic) as well as here we fi nd that, as the side-groups combining both fl exible and rigid units increase in size, the honeycomb cells of rod-, disk-, or wedge-like shape. This open up, resulting in a structure of expanding fi eld of soft matter science alternating rigid aromatic and fl exible has produced some very intriguing fl uoroalkyl layers (Figure 40g). As structures, growing in complexity year temperature is raised, the fl exible aft er year. The challenge is to learn how side-chains expand faster than the to design molecules that will assemble rigid cores, and the growing mismatch into pre-determined superstructures, results in the aromatic layers (green) which could be used in molecular being shredded into parallel ribbons electronic and optical devices, sensors, (symmetry c2mm). Interestingly, the actuators, membranes and other smart molecules are highly aligned along the molecular materials. ribbon axis, forming a nematic LC phase self-confi ned within these 2-3 nm wide channels (Figure 40h). Fig. 39: Structure of the T-shaped polyphilic molecules. As the temperature is increased further, birefringence is lost (Figure 40c) - as the structure becomes cubic (Ia 3d, Figure 39). This phase is of the cubic type known as “double gyroid” bicontinuous, also found in some lipid-water systems, block copolymers and some LC systems. It consists of two infi nite interpenetrating networks that never touch. However, unlike in

Fig. 40: (a-c) Texture as seen between crossed polarisers in the (a) - lamellar phase at 105°C; (b) ribbon phase at 131°C and (c) Cubic Ia 3d phase at 140°C. Scale bar: 50 µm. (d-f): GISAXS patterns from a thin fi lm in the (d) lamellar, (e) ribbon and (f) cubic phase. (g-i): Histogrammatic electron density maps with added schematic molecules for (g) lamellar, (h) ribbon, and (i) cubic phase. Discrete solid colours are allocated to three diff erent density ranges, purple for the highest, green for the middle and red for the lowest; they correspond roughly to chemical group colours in the chemical formula.

36 ESRF SOFT CONDENSED maTTER

any other bicontinuous phase, the need no alignment. For alignment is molecular rods are parallel to the oft en hard to achieve and is required segments of the networks. Also, the in the alternative anisotropic systems distance between the network junctions such as columnar LCs. This work also is exactly two molecular lengths and is shows how a single appropriately fi x e d . designed organic compound can display molecular layers, wires and This new mode of molecular self- networks just by changing temperature. assembly promises potential 3-D If required, each of these could be semiconductors for applications such stabilised and fi xed by chemical cross- as organic photovoltaic materials that linking.

CRYSTAL STRUCTURE OF CONJUGATED Principal publication and authors A.A.Y. Guilbert (a), J.M. Frost (a), T. Agostinelli (a,b), E. Pires (c), POLYMERS: A COMBINED SCATTERING S. Lilliu (c,d), J.E. Macdonald (c) and J. Nelson (a), Chemistry of Materials 26, AND ATOMISTIC SIMULATION 1226-1233 (2014). (a) Centre for Plastic Electronics, Department of Physics, Blackett APPROACH Laboratory, Imperial College London (UK) (b) Plastic Logic Ltd., Cambridge (UK) Organic photovoltaics (OPV) is of the side chains attached to the (c) School of Physics and Astronomy, a promising technology with conjugated backbones to solubilise the Cardiff University (UK) the potential to deliver a cheap, polymer and side chain attachment (d) Masdar Institute of Science and Technology, Abu Dhabi (United Arab lightweight, fl exible and large area can strongly infl uence polymer Emirates) solution for niche applications in packing, and thus infl uence both electricity generation. In polymer OPV the polymer crystallisation and the devices, light is harvested by a low- polymer:fullerene phase separation. bandgap conjugated polymer with an absorption band closely matching the We studied a number of solar spectrum. Upon photo-excitation cyclopentadithiophene-co-benzothia- of the polymer, a coulombically bound diazole polymer derivatives diff ering in electron-hole pair, called an exciton, the heteroatom (carbon or silicon, see is formed. Exciton dissociation into Figure 41) that links the side chains free charge carriers occurs when a to the cyclopentadithiophene unit second material with higher electron and in the structure of the side chains affi nity, typically a fullerene derivative, (2-ethylhexyl, 3,7-dimethyloctyl and is blended with the polymer to form hexadecyl). Previous studies had shown the so-called bulk-heterojunction. that the photocurrent produced by a This blend can be solution processed, solar cell could be greatly increased by leading to thin fi lms around 100 nm, substituting the carbon heteroatom and subsequently sandwiched between with a silicon atom or by processing the two electrodes with diff erent work carbon analogue with additives. This functions to complete the device. eff ect had been assigned to a higher tendency of the silicon-substituted Device performance is strongly aff ected polymer to crystallise leading to more by the microstructure of the polymer- pronounced phase separation in the fullerene blends, which in turn is blend. infl uenced by the processing conditions and the chemical structure of the blend components. For instance, the nature

Fig. 41: GIWAXS 2D diff raction pattern of a) poly[2,1,3-benzothiadiazole-4,7- diyl-[4,4-bis(2-ethylhexyl)-4H-cyclopenta[2, 1-b:3,4-b’]-dithiophene-2,6-diyl]] (C-PCPDTBT) and b) poly(4,4-dioctyldithieno(3,2-b:2’,3’-d)silole-2,6-diyl-alt- (2,1,3-benzo-thiadiazole)-4,7-diyl) (Si-PCPDTBT). Adapted with permission from a.a.y. Guilbert et al., Chemistry of Materials 26, 1226-1233 (2014). Copyright (2014) American Chemical Society.

The European Synchrotron 37 SCIENTIFIC HIGHLIGHTS SOFT CONDENSED maTTER

Grazing-incidence wide-angle X-ray cyclopentadithiophene and scattering (GIWAXS) off ers us the benzothiadiazole units are more likely possibility to investigate the crystalline and that the side chains of the silicon structure of such conjugated polymer analogue are more fl exible. We built in thin fi lms. Figure 41 displays the a crystal structure with polymers in GIWAXS pattern for the carbon and the trans confi gurations, a π−π stack silicon analogue taken with a MAR SX- featuring a two-fold axis parallel to 165 area detector at BM28 – the XMaS the main chain axis simply translated beamline. The silicon analogue, which to create the lamellar stack (See displays a better defi ned diff raction Figure 42). Then, we scanned the energy pattern, shows a higher tendency to surface of this structure while varying crystallise than the carbon analogue. the π−π stacking and lamellar stacking From the patterns, we notice that distances (for each confi guration, the the crystals of the two analogues are side chains are relaxed by performing oriented diff erently with respect to the a MD simulation at 300 K in a NPT substrate. By substituting the carbon ensemble, see Figure 42). No diff erence atom with a silicon atom, the π−π can be observed for the carbon and stacking distance (the distance between silicon, but both display two minima. the fl at, conjugated backbones of the The π−π stacking and lamellar stacking polymers when stacked on top of one distances of one of the minima agrees another cofacially) decreases at the very well with the distances extracted cost of increasing the lamellar stacking from the GIWAXS pattern of the silicon distance (the lateral distance between analogue. The other minimum displays polymer backbones, arranged in a a lamellar stacking in agreement sheet). This is illustrated in Figure 42. with the carbon analogue but the π−π stacking is overestimated. In further To bridge the gap between chemical studies of structures with longer structure and crystal structure, we and less bulky side chains, we found combine GIWAXS with molecular only one minimum corresponding to dynamics simulations (MD). the short π−π stacking distance and From the quantum chemical long lamellar distance and in good calculations, it appears that cis and agreement with previous reports. trans conformations between the We propose that the greater side chain fl exibility of the silicon bridged polymer enables that material to fi nd the minimum with the shorter π−π stacking distance, which should enable improved hole transport, leading to the higher hole mobilities that have been observed and to easier photocurrent generation in a solar cell. These studies confi rm that structural characterisation using GIWAXS combined with molecular modelling can help to explain diff erence in performance of organic electronic materials in terms of their chemical structure and packing tendency.

Fig. 42: Structure used for potential energy surface calculations showing the π−π stack and the lamellar stack. Calculated potential energy surface (PES) for C-PCPDTBT. The step size is 0.25 Å for the π−π stacking and 0.5 Å for the lamellar stacking. The PES is not shown for Si-PCPDTBT since no noticeable diff erences can be observed. Adapted with permission from a.a.y. Guilbert et al., Chemistry of Materials 26, 1226-1233 (2014). Copyright (2014) American Chemical Society.

38 ESRF SOFT CONDENSED maTTER

Principle publication and authors INVESTIGA TING THE LOCAL STRUCTURE C. Gutt (a), L. Grodd (a), E. Mikayelyan (a), U. Pietsch (a), R. J. Kline (b) and S. Grigorian (a), OF A π−π CONJUGATED NETWORK B Y J. Phys. Chem. Lett. 5, 2335–2339 (2014). X-RAY NANODIFFRACTION AND CROSS (a) Fachbereich Physik, Universität Siegen (Germany) CORRELATION (b) Materials Science and Engineering Division, National Institute of Standards and Technology, Gaithersburg (USA) Poly(3-hexylthiophene) (P3HT) is nanocrystals. The broadening of one of the most widely used the peaks confi rms the size of the semicrystalline polymers in organic nanocrystals of 5−10 nm as deduced electronics. Cast P3HT thin fi lms are by the Debye−Scherrer formula. Due to typical semicrystalline where well- the small spot size of 150 nm, the X-ray ordered crystalline nanodomains diff raction pattern becomes sensitive with sizes of about 5-10 nm are to the local crystalline structure and embedded in a disordered host matrix. the local angular or orientational order Considerable eff ort has been made of the crystallites can be traced by the to understand the charge transport use of X-ray cross-correlation analysis mechanism in terms of the size and (XCCA) [2]. orientation of these nanodomains and their degree of crystallinity. It was In comparison to integrated profi les, found that the P3HT fi lms, with edge- which do not show any pronounced on orientation of thiophene moieties spatial variation over the whole with π−π stacking direction along the nanomesh, the correlation functions substrate plane, are a favourable way elucidate the variety of local structures of achieving high charge mobility in existing in P3HT. organic fi eld-eff ect transistor (OFET) devices [1]. At certain spots, the angular correlations with 2-fold symmetry (Figure 44a) The majority of structural and pronounced 4-fold symmetries investigations of organic thin fi lms (Figure 44b) are found, while other has been performed using the grazing positions show no sign of orientational incidence X-ray diff raction (GIXD) variation with fl at correlation functions technique, where the large footprint C(Δ) (Figure 44c). Despite 2-fold accounts for spatial averaging over symmetry being a typical motif in the more than ten thousand crystalline arrangement of P3HT crystals (even nanodomains. This drawback of GIXD on larger length scales), the 4-fold in terms of sample averaging has been symmetry oscillations provide an overcome by using an X-ray nanobeam enhanced local degree of ordering and in transmission geometry (Figure 43) cannot be resolved in conventional at the ID13 nanofocus beamline. We GIXD measurements. employed nanobeam X-ray diff raction using an X-ray spot size of 150 nm to At each spot on the sample, both the investigate the local structure of P3HT magnitude and the direction (the phase) thin fi lms. were deduced for the L = 2 Fourier component of the scattering intensity. P3HT drop-cast fi lms on rectangularly shaped grids with 15 nm thick Si3N4 membranes (Dune Sciences, Inc.) were measured by using a mesh scan Fig. 43: Schematic technique with a step size of 180 nm. setup of the From this we obtained 121 scattering X-ray scattering patterns. experiment employing transmission The (020) diff raction ring associated geometry. with the π−π stacking of the thiophenes is the most prominent signal visible at q = 16 nm−1 and originates from the edge-on orientation of the

The European Synchrotron 39 SCIENTIFIC HIGHLIGHTS SOFT CONDENSED maTTER

Fig. 44: Correlation functions C(Δ) at four diff erent positions on the sample showing key aspect for π−π conjugated polymers diff erent symmetries: (a) 2-fold, (b) 4-fold, is the question of connectivity through (c) fl at correlation at the (020) Bragg peak the nanoscale complex network. position q = 16 nm−1, (d) Stream lines Figure 44d visualises a form of (white) of the Fourier component L = 2. connectivity by displaying streamlines through the diff erent angular order maps. The streamlines indicate how distant parts of the sample can be connected along the orientational order direction. Figure 44d shows that for L = 2 Fourier component it is diffi cult to connect over long distances when angular disorder is present. The P3HT system possesses a complex interplay between order and disorder, and the understanding of such interconnectivity remains a manifold task.

Our X-ray study reveals the complex nanoscale structure of the crystals in P3HT where spatially resolved The arrows in Figure 44a have length investigation provides strong variations associated with the magnitude of the of the structural order on a local scale. Fourier component and direction given The diff raction patterns confi rm that a by the phase. The colour of the pixels high degree of angular order can exist represent the magnitude binned into in the polymer network within an X-ray three intervals: strong (brown), medium spot size of 150 nm. The XCCA technique (green) and weak (blue), respectively. allows us to determine local features on the scale of the working device, We observe a strong spatial variation avoiding the averaging procedures. without gradual transition from spot The present approach investigating the References to spot in the correlation functions. The ordering and interconnectivity of the [1] H. Sirringhaus et al., Nature 401, 685 overall appearance of C(Δ) is diff erent π−π conjugated network at the 150 nm (1999). [2] P. Wochner et al., PNAS 106, 11511 at each spot, indicating that the level has particular potential for the (2009). structural motifs are short-range with optimisation of organic electronic length scales below the spot size. One devices.

Principal publication and authors R. Angelini (a,b), E. Zaccarelli (c,b), GLASS–GLASS TRANSITION DURING F.A. de Melo Marques (d), M. Sztucki (e), A. Fluerasu (f), G. Ruocco (b,d) and AGEING OF A COLLOIDAL CLAY B. Ruzicka (a,b), Nat. Commun.5, 4049 (2014). (a) CNR-IPCF, UOS Roma (Italy) (b) Physics Department, Sapienza University, Roma (Italy) Colloidal suspensions are characterised measurements and Monte Carlo (MC) (c) CNR-ISC, UOS Sapienza, Roma (Italy) by a variety of microscopic interactions, simulations, we provide evidence of a (d) Center for Life NanoScience, IIT@ which generate unconventional phase spontaneous glass–glass transition in a Sapienza, Roma (Italy) diagrams encompassing fl uid, gel and colloidal system consisting of aqueous (e) ESRF (f) Brookhaven National Laboratory, glassy states and off er the possibility Laponite suspensions. Upton, New York (USA) to study new phase and/or state transitions. Glass–glass transitions Figure 45 shows a dilution experiment in particular have rarely been found, that allowed us to distinguish between especially at ambient conditions. Here, two diff erent glasses: the fi rst glass through a combination of dilution melts upon addition of water within experiments, X-ray photon correlation three days, i-j, while the second glass spectroscopy (XPCS) at beamline no longer melts upon addition of water ID10, small-angle X-ray scattering aft er three days, k-l. The former is (SAXS) at beamline ID02, rheological dominated by repulsion, a Wigner glass

40 ESRF SOFT CONDENSED maTTER

[1], while for the latter an additional (inset of Figure 45n) and a moderate mechanism must have been involved, increase of the low Q signal, that is, of probably due to attraction [2]. These the compressibility. results suggest that a previously unreported transition occurs in such To gain information on the Laponite glass at weight concentrations macroscopic diff erences between the

Cw = 3.0% around a critical time tc ≈ 3 two glasses, we performed oscillatory days. rheological measurements at diff erent

times before and aft er tc. Interestingly, To further investigate this transition, a signifi cant change of the storage XPCS measurements were performed on modulus G’(ν) is found (Figure 45o): rejuvenated samples and the intensity for times below 3 days G’(ν) ≈103 Pa, autocorrelation functions shown in however, for times above 3 days, G’(ν) Figure 45m have been fi tted through is roughly one order of magnitude β 2 4 the expression g2(Q,t)-1=B∙exp[-(t/t) ] higher (G’(ν) ≈10 Pa) as found in where t is an ‘eff ective’ relaxation systems with depletion interactions time whose distribution is measured when passing from a repulsive to by the exponent β. Two diff erent an attractive glass. This result thus behaviours were again observed upon supports the existence of two glasses varying the rejuvenation time, tr: with signifi cantly diff erent elastic samples rejuvenated before tc show properties and suggests a crucial role a stretched behaviour (β < 1) while of attraction for the occurrence of the samples rejuvenated aft er tc exhibit a spontaneous glass–glass transition. compressed behaviour (β > 1) (inset of Figure 45m) [3]. To clarify the nature of the attraction, we have carried out Monte Carlo To complement these observations, simulations of a simple model of we also investigated the structure of Laponite platelets that combines both rejuvenated samples. Measurements repulsive and attractive ingredients, of static structure factors SM(Q) are see Figure 46. The simulated static shown in Figure 45n: two diff erent structure factor Ssim(Q) of platelets behaviours were found for samples interacting via electrostatic repulsion rejuvenated before tc (black curves) and and a long-range orientational aft er tc (red curves) showing a shift of attraction at fi xed Cw=3.0% on varying the main peak to larger wave vectors attraction strength mimics the eff ect

Fig. 45: (a–l) Photographic time sequence of a dilution experiment for Laponite suspensions at concentration Cw = 3.0% as described in the text. (m) XPCS intensity autocorrelation functions and (n) Static structure factors for rejuvenated aqueous Laponite suspensions at the indicated rejuvenation times tr. In the insets, the β exponent as a function of Q and the peak position as a function of tr are shown. (o) Frequency dependence of the storage modulus G’(ν) at the indicated waiting times tw.

The European Synchrotron 41 SCIENTIFIC HIGHLIGHTS SOFT CONDENSED maTTER

References of time taking place in the ageing of All of the above observations, providing [1] B. Ruzicka et al., Phys. Rev. Lett. 104, Laponite suspensions, producing a information on the microscopic 085701 (2010). qualitative similar behaviour for the structure (SAXS) and dynamics (XPCS) [2] B. Ruzicka et al., Nature Mater. 10, 56 (2011). structure factors as in the experiments. as well as on the macroscopic properties [3] R. Angelini et al., Soft Matter 9, (dilution experiments and rheological 10955 (2013). measurements) and on the microscopic interactions (MC simulations), indicate that two diff erent glassy states are distinguishable with evolving waiting time: the fi rst is dominated by long- range screened Coulombic repulsion (Wigner glass), while the second is Fig. 46: Simulated static structure factor stabilised by orientational attractions of Laponite platelets (disconnected house of cards - DHOC

at fi xed Cw = 3.0% with glass), occurring aft er a much longer varying attraction time. strength that mimics the eff ect of time in the ageing of Laponite suspensions.

Principal publication and authors M. Koifman Khristosov (a), FORMATION OF CURVED MICROMETRE- L. Kabalah-Amitai (a), M. Burghammer (b,c), A. Katsman (a) SIZED SINGLE CRYSTALS and B. Pokroy (a), ACS Nano, 8, 4747– 4753 (2014). (a) Technion Israel Institute of One remarkable outcome of biogenic size of the single crystal growing within Technology (Israel) crystal growth is the formation of it. The mechanism we employed for (b) ESRF (c) Ghent University (Belgium) unfaceted sculptured single crystals the formation of such droplets was the with rounded shapes and even crystals dewetting phenomenon in nanometric that are highly porous. This is in contrast thin fi lms. To induce dewetting into to crystal growth by classical methods, liquid droplets, a thin fi lm needs to be which produce facets dictated by the brought to its melting point. A eutectic atomic structure and minimisation of system such as gold-germanium the surface energy, leading to specifi c seemed an ideal choice for this purpose low surface energy facets. Drawing owing to the relatively low temperature inspiration from nature in regard to at which the liquid state can be achieved

the growth of unfaceted single crystals, (TE = 361°C). If a eutectic fi lm of Au-Ge we demonstrate that single crystals is deposited on a non-wetting surface

of gold can indeed also exhibit curved such as SiO2, it dewets on reaching the surfaces. This occurs during crystal eutectic temperature (Figure 47a) and growth without the need for any top- produces liquid droplets within which down fabrication steps. These single Au single crystals form (Figure 47b). crystals grow from a confi ned volume of a droplet of a eutectic composition Energy dispersive spectroscopic (EDS) melt, which forms via the dewetting of examination of the single crystals nanometric thin fi lms. We can control proved that they are gold. To confi rm crystal curvature by controlling the that these crystals are indeed single process environment, including several throughout their entire volume, we parameters, such as the contact angle prepared cross-sections of these and the curvature of the drops by droplets using a focused ion beam changing the surface tension of the (FIB) (Figure 47c). This enabled us to liquid drop during crystal growth. observe that each crystal was faceted inside the droplet and possessed a By controlling the shape and size of the curved surface confi ned by the droplet’s droplet, we aim to tailor the shape and surface. Moreover, the 3D curvature of

42 ESRF SOFT CONDENSED maTTER

Fig. 47: HrSEM micrographs of Au-Ge droplets aft er dewetting. a) A large area view of the dewetted surface revealing several droplets. b) A side view of a curved gold single crystal within a micro droplet. c) A droplet cross-section obtained by FIB reveals facets inside the droplet but a curved shape on the surface.

these single crystals was identical to eutectic melt, we have demonstrated that of the droplets. a method for creating a single curved crystal that grows in the confi ned To obtain a crystallographic proof that space of its own melt, replicating the these crystals were single, we used 3D shape of the droplet. sub-micrometre scanning synchrotron diff ractometry at beamline ID13, on a FIB-sectioned crystal of gold (Figure 48a). At each point of the scanned area (1793 scanning points) the same thermal diff use scattering pattern of a bulk single crystal was observed (Figure 48b). We also verifi ed this by averaging all the 1793 diff raction images from the diff erent locations of the crystal and found that this integrated image remained identical to any one of the individual diff raction images (Figure 48c-d). Furthermore, a single crystal Kossel line pattern emerging from Bragg-type diff raction of gold X-ray fl uorescence radiation generated by the primary beam was clearly visible in the averaged pattern (Figure 48e).

The energetic mechanism that permits Fig. 48: a) Cross-section of a FIB- the formation of such curved single prepared droplet that was used for the crystals is that the free energy of a high-resolution X-ray scan. b) A binary mask derived from mapping of average liquid/solid gold interface is yet lower intensity by simple thresholding in each than the free energy of vacuum/solid diff raction pattern, representing the X-ray gold interface. That is why the gold fl uorescence background of the scanned crystals favour the energetic curved gold crystal. c) Average diff raction pattern surfaces within the droplet, rather of all corrected images of the area in (b) (1793 diff ractions). Nano-beam than breaking out of the droplet and Kossel lines can be observed due to the exposing their facets. Au-Fluorescence. d) Corrected (spatial distortion and fl at fi eld) diff raction pattern In conclusion, using a relatively simple of scan point which was taken from the method such as the dewetting from thin centre of the single crystal, marked with a circle on (b), for reference. e) High pass fi lms, and exploiting thermodynamic fi ltered version of the average image process such as crystallisation from an enhancing the Kossel-lines.

The European Synchrotron 43 SCIENTIFIC HIGHLIGHTS Highlights 2014

X-ray Imaging

The articles presented in this chapter illustrate the broad range of scientific topics encountered at the X-ray imaging beamlines, from medicine and life science to palaeontology, nanowires, stardust or historical artefacts. All the examples share a common property: the heterogeneity (and often complexity) of the samples. However, they usually differ in terms of size, composition and structure. While ID17, the biomedical beamline, offers a beam large enough to image objects the size of organs (for example Figure 49, imaging the complete human knee), the NINA project (nano-imaging and nano-analysis, upgrade beamline project UPBL4) provides two new beamlines, ID16A and ID16B, designed to deliver the ultimate nanometric resolution. Accordingly, the five X-ray Fig. 49: Axial section of a full human knee in vitro imaged at ID17 using imaging beamlines offer our users a propagation-based phase contrast imaging. The image reveals an unprecedented visualisation of the various soft tissue structures (from [1]). very wide range of accessible lateral resolution and corollary sample sizes, as well as an extended set of imaging and analytical tools. fluorescence analysis and magnified 2015 and the first XRD users are phase tomography, are now regularly scheduled for June 2015. In parallel This year, the NINA project has reached being put to good use, either on their to these developments, discussions fruition. ID16B, the nano-analysis own or in combination during a single have been initiated to consider the beamline, welcomed its first users in experiment. The beamline will add future of the infrared beamline and the April. Two beam modes can be used: three major capabilities in 2015: a possibility of reaching the nanometre pink beam and monochromatic beam high energy focusing mirror operating scale. and various techniques are available at 34 keV, ptychographic imaging for using a nanobeam focused down ultimate resolution and cryogenic Finally, at the opposite end of the to around 50 × 50 nm2: 2D and 3D sample preservation. beam size scale, the paleontology fluorescence, 2D and 3D diffraction, facility project aims at improving 3D phase contrast imaging and 2D At an intermediate submicrometric X-ray coherent imaging capabilities X-ray excited optical emission (XEOL). scale, ID21, the spectromicroscopy for large samples. Important The very first papers based on these beamline, offers complementary tools instrumental developments are pioneering experiments have recently to the NINA beamlines with an X-ray underway at ID17 and ID19 to enlarge been accepted. As an example, microscope in the tender X-ray domain the field of view and simultaneously Laforce et al. reports, the 3D nano (2-9 keV) for 2D X-ray fluorescence increase the lateral resolution. fluorescence imaging of meteoric (XRF) and XAS, a FTIR microscope for A new multi-modal monochromator nanoparticles in the journal Analytical molecular mapping, including a side (Bragg, Laue and soon multi- Chemistry [2]. Nano-X-ray absorption branch that should soon provide users layer) has been installed at ID19, spectroscopy (XAS), one important with a stable and easy to use instrument covering an energy range as large as deliverable of the beamline project, for 2D µXRF and µXRD mapping. This 18-200 keV. Different developments still requires additional development is currently the main project at ID21, in optics and detectors are on-going and commissioning, but should be which is being implemented without to further increase the field of view, available in 2015. At ID16A, the nano- any impact on the regular operation and the general versatility of the imaging beamline, the first users of the main branch. In 2014, major various imaging configurations. At arrived in October. They were very achievements included producing the ID17, a new Bragg-Bragg coherence- pleased to test a focused beam as first monochromatic beam in July preserving monochromator has been small as 35 nm vertical by 20 nm 2014 (fixed energy at 8.5 keV) and the commissioned and it now allows horizontal at 17 keV with a flux as first focused beam in November 2014 the full exploitation of the available high as 1012 photons/s. The two (0.9 µm vertical × 2.0 µm horizontal). energy range (40-100 keV) with a flat main imaging configurations, X-ray Commissioning will continue in March and uniform incoming field.

44 ESRF X-RAY IMAGING

Going to ultra-fast imaging is another major objective for the coming years, in particular within Phase II of the upgrade programme. Feasibility tests at ID15 and ID19 have already demonstrated that the pulsed time- structure of the ESRF’s synchrotron light can be used to capture snapshot images of ultra-fast processes such as crack propagation, shown in Figure 50. Using the isolated X-ray flash from a single-bunch of electrons in the storage ring, which is as short as 140 ps, allows access to a completly new time domain for hard X-ray imaging [3].

M. Cotte

References Fig. 50: Time series recording crack [1] A. Horng, E. Brun, A. Mittone, S. Gasilov, L. Weber, T. Geith, S. Adam-Neumair, propagation in a glass plate initiated S.D. Auweter, A. Bravin, M.F. Reiser and P. Coan, Cartilage and soft tissue imaging using X-rays: by an accelerated bolt. The pictures propagation-based phase-contrast computed tomography of the human knee in comparison were acquired via single-bunch with clinical imaging techniques and histology, Investigative Radiology 49, 627-34 (2014). imaging at beamline ID19 (from [3]). [2] B. Laforce et al., Nanoscopic X-ray fluorescence imaging of meteoritic particles and diamond inclusions, Analytical Chemistry 86, 12369–12374 (2014). [3] A. Rack et al., Exploiting coherence for realtime studies by single-bunch imaging, Journal of Synchrotron Radiation 21, 815-818 (2014).

Impact of manganese on primary Principal publication and authors A. Daoust (a,b), Y. Saoudi (a,b), J. Brocard (a,b), N. Collomb (a,b), hippocampal neurons in rodents C. Batandier (c), M. Bisbal (a,b), M. Salomé (d), A. Andrieux (a,b), S. Bohic (a,b,d) and E.L. Barbier (a,b), X-ray fluorescence (XRF) imaging was We have addressed the impact of Hippocampus 24, 598–610 (2014). used to map the spatial distribution Mn2+ on cells in culture: mouse (a) Inserm, U836, Grenoble (France) of manganese in neurons to help in an hippocampal neurons (HN) and (b) Université Joseph Fourier, Grenoble Institut des Neurosciences (France) assessment of the safety of manganese Neuro2a neuroblastoma cells (N2a, a (c) Laboratoire de Bioénergétique enhanced magnetic resonance imaging neuron-like cell) using several different Fondamentale et Appliquée, Grenoble (MEMRI). approaches. Here we summarise the (France) main results obtained in HN cells. (d) ESRF Manganese is an increasingly Acknowledgements important contrast agent for MRI, XRF microscopy, carried out at the ID21 The authors acknowledge the support of which is now widely used to image the and ID22NI (now at ID16B) beamlines, the Grenoble MRI Facility (IRMaGe) and internal body structures of rodents and showed that HN cells not exposed to the ESRF for provision of synchrotron 2+ radiation beamtime (respectively, non-human primates. Mn is an MRI exogenous MnCl2 naturally contain contrast agent [1] and, as a divalent 5 µg/g of Mn, a result in line with a exp. MD-458, MD-548). AD received a stipend from the Région Rhône-Alpes – 2+ ion, Mn can substitute calcium previous study [4]. A 24h exposure Cluster HVN. (Ca2+) by entering active neurons to a Mn2+ concentration leading to through Ca2+ channels [2] and can 50% lethality (100 µM) yielded an participate in axonal transport [3]. The intracellular Mn concentration of 248 unique ability of Mn2+ to image brain µg/g for HN cells. cytoarchitecture and connectivity makes MEMRI particularly effective for In control HN and HN exposed to studying neurodegenerative diseases. moderate concentration of Mn2+ However, at high doses, manganese is (20 µM), phosphorus (P) and zinc (Zn, toxic for neurons and glial cells. not detected for N2a) were mostly

The European Synchrotron 45 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

Fig. 51: Map of metal distribution in hippocampal neurons obtained by SR-XRF. Left : Two-dimensional elemental distribution in HN 24 h aft er adding diff erent concentrations of Mn2+ to the culture medium: 0 (control), 20, and 100 µM of Mn2+. Cell contours appear as a white dotted line. The colour scale is in mg of element per g of dry weight. Right: Normalised XRF spectra, averaged over the entire map. The green line corresponds to the baseline and the red line to the fi t.

Fig. 52: Impact of Mn2+ on the morphology of

hippocampal neurons. 2+ Aft er 48 h in culture, of Mn below LC50 (90 µM) shortened HN were exposed to neuritic length and decreased Mn2+ for 24 h. Four mitochondria velocity aft er 24 h concentrations of Mn2+ of incubation (Figure 52). Similar were used: 0 (control), concentrations of Mn2+ also facilitated 20, 50, and 100 µM. Immunofl uorescence the opening of the mitochondrial labelling: neuronal fi brillar permeability transition pore in isolated actin (red), tubulin (green) mitochondria from rat brains. and nucleus (blue). The reduction in neuritic length 2+ This study reports a number of new as Mn concentration 2+ increases is readily visible. observations on the impact of Mn on cultured HN. When cells are exposed to a concentration of Mn2+ below the LC50, Mn is preferentially localised found in the cell nucleus (Figure 51). around the nucleus, most likely in the Aft er adding a high concentration Golgi apparatus or in the endoplasmic of Mn2+ to the culture medium (100 reticulum, and in discrete spots in µM), the distributions of P and Zn neurites. When neurons are exposed became much more heterogeneous. to a concentration of Mn2+ above the In control HN, Ca and Fe were barely 50% lethality level, the distribution detectable whereas, aft er adding a of Mn becomes more diff use within moderate concentration of Mn2+ to the cell, suggesting a mechanism the culture medium (20 µM), Ca and of cell detoxifi cation. Exposure of Fe were clearly visible and localised cultured HN to Mn2+ also induced a in the perinuclear, Mn2+ rich, region. shortening of neurites, and a decrease 2+ References At a higher Mn concentration of mitochondria velocity. Altogether, [1] A.P. Koretsky and A.C. Silva, NMR (100 µM), Ca and Fe were redistributed these data support the use of primary Biomed 17, 527–531 (2004). throughout the cytoplasm. Note that neuronal cultures as a model to [2] K. Narita, F. Kawasaki and H. Kita, the distribution of Zn and Cu in HN did study Mn2+-induced toxicity in the Brain res 510, 289–295 (1990). not change upon addition of Mn2+. brain. Future experiments should be [3] R.G. Pautler, NMR Biomed. 17, 595–601 (2004). performed on co-cultures of astrocytes [4] A. Daoust, E.L. Barbier and S. Bohic, Immunofl uorescence studies of HN and neurons to better mimic brain NeuroImage 64, 10-18 (2013). neurons revealed that concentrations anatomical conditions.

46 ESRF X-RAY IMAGING

Principal publication and authors A.J. Westphal (a), R.M. Stroud (b), Chemistry and structure of H.A. Bechtel (c), F.E. Brenker (d), A.L. Butterworth (a), G.J. Flynn (e), D.R. Frank (f), Z. Gainsforth (a), contemporary interstellar J.K. Hillier (g), Frank Postberg (g), A.S. Simionovici (h), V.J. Sterken (i,j,k,l), dust particles collected by the L.R. Nittler (m), C. Allen (n), D. Anderson (a), A. Ansari (o), Stardust spacecraft S. Bajt (p), Ron K. Bastien (f), N. Bassim (b), J. Bridges (q), D.E. Brownlee (r), M. Burchell (s), M. Burghammer (v), H. Changela (u), The interstellar region contains At ID13, analyses by synchrotron P. Cloetens (v), A.M. Davis (w), several percent of the whole mass of XRF microscopy of the elemental R. Doll (x), C. Floss (x), E. Grün (y), the galaxy. New stars and planetary composition of eight candidate impact P.R. Heck (o), P. Hoppe (z), systems will be generated from this so- features extracted from the Stardust B. Hudson (aa), J. Huth (z), called interstellar matter, (Figure 53). Interstellar Dust Collector (SIDC) were A. Kearsley (bb), A.J. King (w), B. Lai (cc), J. Leitner (z), L. Lemelle (dd), It contains the basic ingredients of all performed, coupled with scanning A. Leonard (x), H. Leroux (ee), known space objects including our own XRD [1,3]. Six of the features were R. Lettieri (a), W. Marchant (a), Earth. Samples of this unique extra- unambiguous tracks, and two were R. Ogliore (ff), W.J. Ong (x), solar system matter were brought back crater-like features. Five of the tracks M.C. Price (s), S.A. Sandford (gg), to Earth during NASA’s stardust mission are so-called “midnight” tracks - that J.-A. Sans Tresseras(v), S. Schmitz (d), T. Schoonjans (t), K. Schreiber (x), in order to study it in the best-suited is, they had trajectories consistent G. Silversmit (t), V.A. Solé (v), laboratories around the world including with an origin either in the interstellar R. Srama (hh), F. Stadermann (x),†, the most powerful synchrotron sources. dust stream or as secondaries from T. Stephan (w), J. Stodolna (a), impacts on the Sample Return S. Sutton (cc), M. Trieloff (g), Capsule (SRC). Synchrotron XRD P. Tsou (ii), T. Tyliszczak (c), B. Vekemans (t), L. Vincze (t), analyses of these stardust interstellar J. Von Korff (a), N. Wordsworth (jj), preliminary examination (ISPE) D. Zevin (a) and M.E. Zolensky (n), 30714 candidates [3] revealed that two of Stardust@home dusters (kk), Science 15, these particles contain crystalline 786-791 (2014). materials: the terminal particle of (a) Space Sciences Laboratory, University of California at Berkeley (USA) track 30 contains olivine and spinel, (b) Materials Science and Technology and the terminal particle of track Division, Naval Research Laboratory, 34 contains olivine. The terminal Washington (USA) particle of track 30, Orion, shows (c) Advanced Light Source, Lawrence elemental abundances, normalised to Berkeley Laboratory, Berkeley (USA) (d) Geoscience Institute, Goethe Fe, that are close to chondritic type CI University Frankfurt, (Germany) values, and a complex, fine-grained (e) State University of New York at structure (Figure 54). The terminal Plattsburgh (USA) particle of track 34, Hylabrook, (f) Jacobs Technology/ESCG, NASA Fig. 53: Infrared image of the dust rich shows abundances that deviate Johnson Space Center (JSC), Houston horsehead nebula, a star forming region (USA) (image credit NASA, ESA, HST). strongly from CI, but shows little fine (g) Institut für Geowissenschaften, structure and is nearly homogenous. University of Heidelberg (Germany) The stardust spacecraft collected dust An additional track, with a trajectory (h) Institut des Sciences de la Terre, Observatoire des Sciences de l’Univers de particles from the interstellar medium Grenoble (France) over several months. The micrometre (i) Institut für Raumfahrtsysteme (IRS), to sub-micrometre sized, fast flying University of Stuttgart (Germany) grains were collected within Aerogel (j) IGEP, TU Braunschweig (Germany) tiles, a transparent, high purity solid (k) Max Planck Institut für Kernphysik, Heidelberg (Germany) silicon oxide glass foam, developed to (l) International Space Sciences Institute, slow down and capture these grains Bern (Switzerland) with minimum damage. Thin slices of (m) Carnegie Institution of Washington Aerogel containing impact tracks were (USA) cut from the trays. The tiny particles (n) Astromaterials Research and Exploration Science, NASA JSC, Houston were studied in situ, in the Aerogel, (USA) applying nano synchrotron X-ray (o) Field Museum of Natural History, fluorescence (XRF) and diffraction Chicago (USA) (XRD) techniques at the nanofocussing (p) Deutsches Elektronen-Synchrotron, beamline ID13 [1], and the ID22NI Hamburg (Germany) (q) Space Research Centre, University of nanoprobe (now at to ID16B) and ID22 Fig. 54: Tricolour RGB = (Fe, Ni, Ca) high- Leicester (UK) microprobe (now closed) beamlines resolution (100 nm per pixel) map of Orion (r) Department of Astronomy, University [2]. (reproduced with permission from [1]). of Washington, Seattle (USA) (s) University of Kent, Canterbury (UK) (t) University of Ghent (Belgium) The European Synchrotron 47 SCIENTIFIC HIGHLIGHTS X-RAY IMAGING

(u) University of New Mexico, Albuquerque consistent with secondary ejecta from Ni-bearing particle and contained (USA) (v) ESRF an impact on one of the spacecraft about 33 fg of distributed high-Z (w) University of Chicago (USA) solar panels, contains abundant elements (Z > 12). The measurements (x) Washington University, St. Louis (USA) Ce and Zn. This is consistent with at ID22NI show that Orion is a highly (y) Max-Planck-Institut für Kernphysik, the known composition of the glass heterogeneous Fe-bearing particle and Heidelberg (Germany) covering the solar panel. contains about 59 fg of heavy elements (z) Max-Planck-Institut für Chemie, Mainz (Germany) located in hundred nanometre phases, (aa) 615 William Street, Apt 405, Midland, In addition eight interstellar candidate forming an irregular mantle that Ontario (Canada) impact features were studied by hard surrounds a low-Z core. (bb) Natural History Museum, London (UK) X-ray, quantitative, XRF elemental (cc) Advanced Photon Source, Argonne imaging and XRD on the ID22NI To date, only a small number of the National Laboratory, Lemont (USA) (dd) LGL-TPE, Ecole Normale Superieure de nanoprobe and ID22 microprobe particles captured by the Stardust Lyon (France) beamlines [2]. Three features were Interstellar Dust Collector and (ee) University Lille 1 (France) unambiguous impact tracks, and the returned to Earth have been extracted (ff) University of Hawai’i at Manoa, other five were identified as possible, and analysed. Of these, seven have Honolulu (USA) but not definite, impact features. features consistent with an origin (gg) NASA Ames Research Center, Moffett Field (USA) Overall, an absolute quantification in the contemporary interstellar (hh) IRS, University Stuttgart (Germany) of elemental abundances in the dust stream. The interstellar dust (ii) Jet Propulsion Laboratory, Pasadena 15 ≤ Z ≤ 30 range was produced by candidates are readily distinguished (USA) means of corrections of the beam from debris impacts on the basis of (jj) Wexbury, Farthing Green Lane, Stoke parameters, reference materials, and elemental composition and/or impact Poges, South Buckinghamshire (UK) (kk) Worldwide. List of individual dusters fundamental atomic parameters. trajectory. The seven candidate is at http://stardustathome.ssl.berkeley. Seven features were ruled out as interstellar particles are diverse edu/sciencedusters. interstellar dust candidates (ISDC) in elemental composition, crystal based on compositional arguments. structure, and size. The presence of References [1] F.E. Brenker, A. Westphal, L. Vincze et One of the three tracks, previously crystalline grains and multiple iron- al., Meteoritics & Planetary Science 49, studied at ID13, contained two bearing phases, including sulfides, 1594–1611 (2014). physically separated, micrometre- in some particles, indicates that [2] A.S. Simionovici, L. Lemelle et al., sized terminal particles, the most individual interstellar particles differ Meteoritics & Planetary Science 49, promising ISDCs, Orion (see above) from any other representative model 1612–1625 (2014). [3] Z. Gainsforth et al., Meteoritics & and Sirius. We found that the Sirius of interstellar dust inferred from Planetary Science 49, 1645–1665 (2014). particle was a fairly homogenous astronomical observations and theory.

Principal publication and authors B. Enders (a), M. Dierolf (a), R obUST LENSLESS IMAGING WITH A P. Cloetens (b), M. Stockmar (a), F. Pfeiffer (a) and P. Thibault (a,c), Appl. high-flux instrument Phys. Lett. 104, 171104 (2014). (a) Physics Department, Technische Universität München, Garching In the last decade, coherent diffractive scan point [1,2]. Unlike traditional (Germany) imaging (CDI) has shown great promise X-ray microscopy or other CDI (b) ESRF as a dose-efficient and high-resolution techniques, ptychography can deliver (c) Current address: Physics & Astronomy complement to traditional (lens- quantitative amplitude and phase Department, University College London based) X-ray microscopy. In contrast images of extended specimen beyond (UK) to conventional microscopy where the resolution limit of lenses. an objective lensing system catches the diffracted X-ray photons to form However, the benefits of diffraction a direct image on the detector, the imaging usually come at the cost X-ray photons are allowed to freely of requiring “clean” diffraction data propagate to the detector forming for the algorithms to work properly. a diffraction image. A computer Commonly this requirement is met algorithm is used to reconstruct an by strong spectral and spatial filtering image of the specimen’s transmission for beam purity in combination with a from the traces left in such an indirect photon counting, pixel-array detector measurement. Ptychography is a CDI for an ideal measurement. While the variant where the specimen is scanned first measure drastically reduces the transversally through the X-ray beam, available photon flux, the latter cannot collecting a diffraction image at each distinguish single photon events at

48 ESRF X-RaY ImaGING

high fl ux. Ironically, this very need Our work proves that mere adjustments to reduce fl ux translates directly into in the data analysis steps can lead lower achievable resolution [3]. to dramatic improvements in image quality. With its newly demonstrated An experiment was carried out at robustness, ptychography will the nano-imaging endstation of the be widely applicable with relaxed former ID22 beamline (now ID16- experimental requirements, paving the NI) to investigate the robustness of way to high-throughput, high-fl ux and ptychography in a confi guration that high-resolution diff ractive imaging. allows for high-fl ux experiments, under conditions that were previously considered insuffi cient: a broad- bandwidth X-ray beam focused by a pair of crossed multilayer-coated Kirkpatrick-Baez mirrors and an integrating scintillator-based FReLoN detector (Figure 55). As can be seen in the diff raction image (D in Figure 55), the acquired diff raction data was degraded by a large halo around the central beam (bluish in the image). The high dynamic range in the diff raction signal caused this degradation by amplifying even faint eff ects like diff use air scattering or a point spread in the Fig. 55: Schematic of the experimental scintillator coupled to the CCD. setup (not to scale) used for ptychography at the former ID22NI endstation. The X-ray Consequently, conventional beam enters through an aperture (A) and is ptychographic algorithms struggled focussed by two multilayer Kirkpatrick-Baez when reconstructing sample mirrors (B). The sample (C) is placed close to the focus and the diff raction patterns are transmission (Figure 56b) and recorded downstream by a CCD coupled to a illuminating wavefront (Figure 56a) scintillator (D). To avoid artifacts, the sample from these data. The illumination is scanned along a circular pattern (E). contains speckly noise and the phase image of the spoke-pattern of the sample is blurred.

Next, the mixed states approach [4] was applied, which is one of the latest algorithmic developments for the treatment of ptychographic data. It permits a number of concurrent modes to form the illumination and thereby emulates some defi ciencies of the detector and the beam, e.g. partial coherence. These modifi cations alone Fig. 56: Conventional allowed reconstruction of the sample in reconstruction (left ) high detail and quality (Figure 56d) and compared with mixed- states approach the main mode (Figure 56c) now shows (right). Conventional the shape expected for the illumination reconstruction: (a) modulus in the sample plane. of the retrieved illumination wavefront (in a.u.) and (b) phase (in rad) Furthermore, an analysis of the of the reconstructed sample transmission function. Mixed state reconstruction: (c) References illumination modes revealed that the showing the reconstruction of the main [1] P. Thibault et al., Science 321, 379 algorithm not only corrects for lower mode of the illumination and (d) phase of (2008). coherence and detector point spread the sample transmission function. (e)-(g) an [2] M. Dierolf et al., Nature 467, 436 (Figure 56, e and f), but also attempts to assortment of other modes reconstructed (2010). correct for an inhomogeneous response in the Illumination, where (e) and (f) can be [3] M. Howells et al., Journal of Electron associated with the detector point spread Spectroscopy and Related Phenomena of the four detector quadrants in function or partial coherence while (g) is 170, 4 (2009). algorithmically less constrained parts related to the inhomogeneous response of [4] P. Thibault and A. Menzel, Nature of the illumination frame (Figure 56g). the four quadrants of the detector. 494, 68 (2013).

The European Synchrotron 49 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

Principal publication and authors J. Segura-Ruiz (a, b), REVEALING PHASE SEPARATION IN G. Martinez-Criado (a), C. Denker (c), J. Malindretos (c), A. Rizzi (c), Nano Lett. 14, 1300−1305 (2014). SINGLE InGaN NANOWIRES (a) ESRF (b) Present address: Institut Laue- Langevin, Grenoble (France) Semiconductor materials are evolving xN has an absorption edge that can be (c) IV. Physikalisches Institut, Georg- toward the development of more complex modifi ed between 0.7 to 3.4 eV. This August-Universität Göttingen, alloys and nanoscale heterostructures. range covers most of the solar spectrum, (Germany) Unexpected properties and improved making this material an excellent functionalities can be achieved with the candidate for the development of high combination of these two approaches. effi ciency multijunction solar cells and

For instance, the ternary alloy InxGa1− light-emitting diodes. Nanometric-sized structures such as nanowires confer Fig. 57: (a) Multi-technique additional advantageous properties that setup at the nanoimaging can enhance the performance of the station ID22NI (now at beamline nanodevices. For instance, nanowire- ID16B). (b) SEM image (left ), In (centre) and Ga (right) Kα XrF shaped semiconductors absorb light in maps of representative single a much more effi cient manner [1], and as-grown nanowires on the Si allow the possibility to build core−shell substrate taken from the top. structures that further improve the performance of solar cells [2], and light- emitting devices [3].

Characterisation of these complex nanoscale structures requires the combination of several techniques, preferably with a high spatial resolution to avoid the averaging that could hide local details. In this work, we have used the hard X-ray nanoprobe at beamline ID22NI (now at ID16B) to characterise

self-assembled single InxGa1−xN nanowires using a multi-technique Fig. 58: (a) InGaN (002) XRD refl ection (open circles) approach and nanometric resolution. acquired at the bottom (a) and top (b) of a single Figure 57a shows a schematic view nanowire along with the best of the setup at the X-ray nanoprobe, multi-Gaussian fi ts (solid which integrates X-ray fl uorescence line). Diff erent Gaussian contributions are plotted with (XRF), X-ray diff raction (XRD) and X-ray dotted lines. The insets show absorption spectroscopy techniques. the corresponding CCD images. This multi-technique approach (c) XLD spectra around the Ga allows the simultaneous study of the K-edge measured at the top composition, long- and short-range (squares) and bottom (circles) of a single nanowire, and the order of the samples, with a very calculated spectrum (triangles) high lateral spatial resolution. As- for wurtzite GaN. grown and dispersed nanowires placed onto thin membranes were scanned with both, pink (ΔE/E ~ 10−2, fl ux of ~ 5 × 1012 ph/s, 55 × 55 nm2 spot size) and monochromatic (ΔE/E ~ 10−4, fl ux of ~ 5 × 1010 ph/s, 105 × 110 nm2 spot size) beams. XRF maps of the nanowires show an axial and radial elemental heterogeneity with Ga accumulation in the bottom and in the outer region of the nanowire, examples are displayed in Figure 57b.

Nano-XRD and nano-XRF maps were performed simultaneously on single dispersed nanowires, using a 28 keV

50 ESRF X-RAY IMAGING

monochromatic beam. Figure 58 in Figure 58c. Both XLD spectra of the References shows the XRD spectra of the InGaN nanowire exhibit the strong anisotropy [1] J. Wallentin, N. Anttu, D. Asoli, M. Huffman, I. Åberg, M.H. Magnusson, (002) reflection acquired at the bottom of the X-ray absorption, typical for the G. Siefer, P. Fuss-Kailuweit, F. Dimroth, (a) and top (b) of a single nanowire. wurtzite GaN matrix. This confirms B. Witzigmann, H.Q. Xu, L. Samuelson, A closer inspection of the (002) that despite the elemental modulation, K. Deppert and M.T. Borgström, Science reflection shows at least three Gaussian the tetrahedral order around the Ga 339, 1057−1060 (2013). contributions to this XRD peak at absorbing atoms remains along the [2] M.D. Kelzenberg, S.W. Boettcher, J.A. Petykiewicz, D.B. Turner-Evans, the bottom, and single Gaussian-like nanowires. M.C. Putnam, E.L. Warren, J.M. Spurgeon, XRD peak at the top. These different R.M. Briggs, N.S. Lewis and H.A. Atwater, contributions to the (002) diffraction In conclusion, a large phase separation Nat. Mater. 9, 239−244 (2010). peak are observed even in the original at the bottom, and a single phase at the [3] G. Martínez-Criado, A. Homs, B. Alen, CCD images, shown in the insets of top of the nanowires were revealed. We J.A. Sans, J. Segura-Ruiz, A. Molina-Sánchez, J. Susini, J. Yoo and Figure 58a and b. Finally, X-ray linear anticipate that this methodology will G.-C. Yi, Nano Lett. 12, 5829−5834 dichroism (XLD) spectra around the Ga contribute to a greater understanding (2012). K-edge measured at the top and bottom of the underlying growth concepts, not of a single nanowire, and the calculated only of nanowires, but also of many spectrum for wurtzite GaN are shown nanostructures in materials science.

Principal publication and authors Tuning the oxygen doping level A. Pagliero (a), L. Mino (b), E. Borfecchia (b), M. Truccato (a), in the Bi-2212 superconductor by A. Agostino (b), L. Pascale (b), E. Enrico (c), N. De Leo (c), X-ray nano-beam irradiation C. Lamberti (b, d), and Gema Martínez-Criado (e), Nano Lett. 14, 1583 (2014). The search for higher speed and lower that 12.4 keV X-rays are able to induce (a) Depart. of Physics, NIS Centre of power consumption has driven the persistent changes in the transition Excellence, University of Turin, Torino semiconductor industry to produce temperature (T ) of La CuO (Italy) c 2 4+y (b) Depart. of Chemistry, NIS Centre of increasingly smaller integrated superconducting cuprates by means of Excellence, University of Turin, Torino circuits. This trend has stimulated photon-induced displacement of light (Italy) deep innovation in the lithographic atoms in their crystal structure [2]. (c) INRIM, National Institute of techniques based on the use of short Metrological Research, Torino (Italy) wavelength radiation. Nevertheless, In our study, we have explored the (d) Southern Federal University, Rostov- on-Don (Russia) so far, hard X-rays-based lithography application of X-ray direct-writing (e) ESRF with l ≈ 0.1 nm has been hampered lithography to the Bi2Sr2CaCu2O8+d by problems in the fabrication of (Bi-2212) superconductor, with suitable masks with both sufficiently typical Tc ≈ 80-90 K. As Bi-2212 high contrast in their absorbing power crystals below Tc are able to emit and and small enough features. Another sense THz radiation [3], an obvious appealing possibility is represented by application could be the production of direct-writing techniques allowing the these devices. The possibility of direct- modification of substrate properties writing X-ray lithography is even more without any photoresist. Focused- interesting because of the expected electron-beam- and focused-ion- increasing availability of hard X-ray beam-induced deposition and etching, sources with extreme brilliance and along with the scanning-probe power, such as X-ray free electron lithography, are the most frequently lasers (XFELs). used techniques, and all enabling nanometric spatial resolution [1]. A few examples of X-ray direct-writing- Fig. 59: a) Scheme of the lithography have been reported experimental setup used at beamline recently, but in all cases patterning was ID22NI (now at ID16B). b) SEM limited to organic or organometallic micrograph of the typical chip used materials. Since these processes for electrical measurements of the are essentially based on molecular Bi-2212 micro-crystals. Current and voltage electrodes used for the photodissociation, the limitation four-probe technique are labelled as to this kind of materials seemed I+, I–, V+ and V–. The arrow indicates to be intrinsic to the use of X-rays. the nanobeam direction during Nevertheless, it has also been shown irradiation.

The European Synchrotron 51 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

References In particular, we investigated the eff ect measured by the four-probe method [1] S.J. Randolph et al., Crit. Rev. Solid of localised hard X-ray irradiation (Figure 59b) aft er each irradiation State Mat. Sci. 31, 55 (2006). of Bi-2212 micro-crystals using the session. We have clearly proven that [2] N. Poccia et al., Nat. Mater. 10, 733 2 (2011). 117 × 116 nm nanoprobe available irradiating this material by 17 keV [3] L. Ozyuzer et al., Science 318, 1291 at the ID22NI beamline (now at X-rays (doses in the order of 1012 Gy) (2007). ID16B) (Figure 59a). The structural results in an elongation of the c-axis [4] I. Piñera et al., Phys. Stat. Sol. A 204, modifi cations induced by the X-ray and in an increase of both the critical 2279 (2007). nanobeam were monitored using nano- temperature and the normal state XRD, while the corresponding changes resistivity (Figure 60). This highlights in the electrical properties were a modifi cation in the material’s oxygen doping level, moving from an over- Fig. 60: a) Four-probes doped state towards a progressively ab-plane resistivity vs. temperature curves for more under-doped state. Bi-2212 micro-crystals in as-grown conditions and Moreover, modelling our diff erent aft er increasingly higher experimental conditions using the X-ray doses. The insets fi nite element method (FEM) allowed us report a magnifi cation of the transition region to exclude heating induced by the X-ray nanobeam as the major cause for the and the Tc vs. resistivity dependence at room change in the doping level. Conversely, temperature. a signifi cant role could be played by b) Dependence of the the photoelectrons acting as a possible c-axis length and of the Tc on the X-ray dose. source of knock-on for the O atoms in the BiO layers, due to their low displacement threshold energy, which has already been predicted for the O atoms in the CuO chains of YBCO [4].

Our results suggest that the use of hard X-rays as a novel direct- writing, photoresist-free lithographic process is an attractive possibility to overcome the current limitations in the fabrication of miniaturised devices with potential nanometric resolution and 3D capability.

Principal publications and authors THE in vivo RADIOSENSITISING EFFECTS I. Miladi (a), C. Alric (a), S. Dufort (b), P. Mowat (a), A. Dutour (c) , C. Mandon (a), G. Laurent (d), OF NANOPARTICLE-BASED MRI E. Bräuer-Krisch (e), N. Herath (f), J.L. Coll (g), M. Dutreix (f), F. Lux (a), CONTRAST AGENTS R. Bazzi (d), C. Billotey (a), M. Janier (a), P. Perriat (h), G. Le Duc (e), S. Roux (d) and O. Tillement (a). Small 10, 1116- Radiotherapy is commonly used in use of radiosensitisers composed of 1124 (2014); cancer treatment, either alone or heavy elements (high atomic number, G. Le Duc (e), S. Roux (d), A. Paruta Tuarez (a), S. Dufort (b), in combination with surgery and/ Z), combined with a preferential E. Brauer-Krisch (e), A. Marais (a), or chemotherapy, since X-rays accumulation in the tumour tissues, C. Truillet (a), L. Sancey (a), P. Perriat (h), induce lethal alterations of the has been proposed in order to increase F. Lux (a) and O. Tillement (a), Cancer tumour. However, the sparing of the the ratio between the dose deposited Nanotechnology 5:4 (2014). surrounding healthy tissue remains a in the tumour and the dose deposited (a) Institut Lumière Matière, Université de Lyon (France) challenge because the similarities in in the surrounding healthy tissues (b)Nano-H S.A.S, Saint Quentin-Fallavier the atomic composition of the tumour [1]. In this context, gadolinium plays (France) and the surrounding tissues do not an important role because of its high (c) Centre Léon Bérard, Lyon (France) favour a preferential absorption of the relaxivity value and because of its (d) Institut UTINAM, Université de therapeutic beam. For this reason, the relatively high atomic number value Franche-Comté, Besançon (France)

52 ESRF X-RaY ImaGING

(Z = 64), highlighting its potential for nanoparticles (when compared to (e) ESRF MRI guided radiation therapy [2]. chelates) within the tumour allows a (f) Institut Curie, Orsay (France) radiotherapy session to be carried out (g) Institut Albert Bonniot, Grenoble (France) Studies using gold nanoparticles when the ratio between the gadolinium (h) MATEIS, INSA, Lyon (France) coated with gadolinium chelates content of the tumour versus the have already been carried out and surrounding healthy tissues is optimal. the design of the particles optimised This time delay can be assessed to allow both MRI monitoring (owing by MRI prior to the radiotherapy. to the gadolinium chelates) and Finally, both gadolinium and gold/ radiosensitisation (due to the gold gadolinium based nanoparticles were core, Z = 79). The delay between the proven to enhance the lifespan of rats administration of the nanoparticles bearing glioma when compared to and the radiotherapy session was radiotherapy alone. determined from the data collected by MRI (Figure 61) prior to irradiation. Although better results are expected by refi ning the biodistribution of the nanoparticles inside the tumour, the lifespan of the rats bearing glioma was enhanced by combining microbeam radiation therapy (MRT) and nanoparticle injection resulting in lifespans extended by 473%, which compares favourably to the 222% obtained when MRT was used without nanoparticles.

The potential for image guided radiotherapy of gadolinium based Fig. 61: Temporal evolution of the MRI nanoparticles was compared to signal in tumour (red squares) and in an routinely used gadolinium chelates, equivalent surface in normal tissue in by using a glioma model implanted in the left hemisphere (black circles) aft er intravenous injection of gadolinium chelate the rat brain. Viewed by MRI imaging, coated gold nanoparticles. The grey area a high and relatively constant represents the time interval during which concentration was conserved by the the diff erence of signal intensities is the tumour while the nanoparticles were highest. rapidly cleared from healthy tissues. This is far better than for gadolinium chelates which are rapidly eliminated from the tumour. Another advantage of this protocol is that the same injection can be used for both MRI and radiotherapy. When radiotherapy was carried out at beamline ID17 using identical experimental conditions, the combination of gadolinium nanoparticles and radiotherapy allowed the median survival time of rats bearing glioma to be doubled when compared to the use of only gadolinium chelates. These results are shown in Figure 62. Fig. 62: Survival curves of 9LGS bearing To conclude, two kinds of gadolinium rats treated by MRT only (black curve, n = based nanoparticles were used both 15 rats), treated by MRT 20 minutes aft er References in MRI and radiotherapy in order to injection of DOTAREM® (blue curve, n = 16 [1] J.F. Hainfeld, D.N. Slatkin and rats), and of AGuIX nanoparticles (red curve, develop the concept of image guided H.M. Smilowitz, Phys. Med. Biol.49, 309- n = 8 rats). The irradiation was performed radiotherapy, i.e. in order to use the 315 (2004). 10 days aft er tumour implantation. [2] D. Francis, G.M. Richards, same contrast media on images as A. Forouzannia, M.P. Mehta and a dose media during radiotherapy. D. Khuntia. Expert. Opin. Pharmacother. The long time of residence of the 10, 2171-2180 (2009).

The European Synchrotron 53 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

Principal publication and authors S. Layachi (a), L. Porra (b,c), WHY LUNG VENTILATION IS UNEVENLY G. Albu (d), N. Trouillet (e), H. Suhonen (c), F. Peták (f), H. Sevestre (e), P. Suortti (b), DISTRIBUTED IN ASTHMA: NEW A. Sovijärvi (g), W. Habre (d) and S. Bayat (a), J Appl Physiol 115, 1057- INSIGHT FROM SYNCHROTRON K-EDGE 1064 (2013). (a) Université de Picardie Jules Verne and SUBTRACTION IMAGING Amiens University Hospital (France) (b) Department of Physics, University of Helsinki (Finland) (c) ESRF Asthma, the most common chronic abnormal respiratory mechanics and (d) Geneva Children’s Hospital, University disease in children, is characterised by gas exchange, but also because it can Hospitals of Geneva, and Geneva chronic infl ammation of th e airways. signifi cantly impact the distribution University (Switzerland) Exposure to an allergen in sensitised of inhaled drugs commonly used in (e) Department of Pathology, Amiens University Hospital (France) asthmatics triggers the expression the treatment of asthma. We have (f) Department of Medical Physics and release of mediators including previously shown that both the route and Informatics, University of Szeged cytokines, chemokines, adhesion (inhaled vs. blood-born) and nature (Hungary) molecules, infl ammatory enzymes (allergic vs. non-specifi c) of airway (g) Department of Clinical Physiology and receptors by infl ammatory cells, challenge play a signifi cant role in and Nuclear Medicine, Helsinki University Central Hospital (Finland) which leads to bronchial smooth determining the heterogeneity of muscle constriction and an increased airway constriction [2 ] . However, resistance of the bronchi to air fl ow. the exact mechanisms causing such Allergen exposure also causes transient striking regional diff erences in lung increases in eosinophils, T-cells, mast ventilation in allergic asthma remain cells and neutrophils in the lung and elusive. infl ammatory cell infi ltration of airway walls. In the Brown Norway rat, a strain We studied the relationship between genetically predisposed to allergic the spatial distribution of the cellular hypersensitivity, allergen provocation eff ectors involved in asthma and produces patchy infl ammatory the emergence of defects in regional cellular infi ltrates in the walls of both ventilation following allergen exposure. airways and pulmonary blood vessels. For this purpose, we used in vivo high- Eosinophils are thought to play a central resolution K-edge subtraction (KES) role in allergic infl ammation in asthma, synchrotron imaging at the ID17 however, the relationship between beamline, during xenon inhalation in eosinophilic infi ltration and regional ovalbumin (OVA)-sensitised Brown- lung function is poorly understood. Norway rats at baseline and repeatedly following OVA challenge. Bronchoconstriction in asthmatic subjects leads to a signifi cant The KES technique allows simultaneous heterogeneity in the distribution of imaging of the airway structure inhaled air within the lung, and the and quantitative measurements of appearance of ventilation defects the regional tissue density and the [1]. This phenomenon is important concentration of inhaled Xe gas [2]. not only because it contributes to This technique allowed us to measure

Fig. 63: High resolution K-edge subtraction images for one representative animal. Upper row: quantitative Xe-images. Following OVA, transient clustered areas of poor ventilation emerge (blue). Lower row: tissue-density KES images.

54 ESRF X-RaY ImaGING

both airway narrowing and the Fig. 64: Eosinophil cellular distribution of inhaled Xe gas within the densities in airway wall, vessel wall and parenchyma, pulmonary alveoli. Histological slices within and outside the of the lung anatomically matched to ventilation defects, defi ned the computed-tomography images based on the regional were stained and digitised. Total cells, concentration of Xe. CD68+ and eosinophils were counted Data are mean ± SE in 6 sensitised and challenged in the walls of bronchi and vessels rabbits and 3 controls. *: randomly selected within and outside p < 0.05 vs. outside of VD, of ventilation defects, delineated based #: p < 0.05 vs. control. on the Xe-KES images. Following OVA challenge, we observed transient airway narrowing and closure [3]. This ventilation defects (Figure 63), and led to the theory that the appearance increases in airway resistance and of ventilation defects may emerge tissue elastance. Eosinophil (Figure 64), from the complex dynamic system CD68+ and total cell counts were behaviour of the lung. signifi cantly higher in the bronchial and vascular walls within vs. outside Our fi ndings further the understanding of the ventilation defects. Furthermore, of uneven ventilation and show that airway constriction during OVA- in asthma, airway infl ammation induced bronchoconstriction was is locally heterogeneous, and that correlated with eosinophil and total cell this phenomenon is involved in densities in the airway walls within the determining the heterogeneity of poorly-ventilated zones. regional airway constriction and that of regional ventilation following The bronchial tree is a branching exposure to an allergen. Moreover, References structure leading to terminal acini the global response to an allergen [1] A.R. Sovijarvi, L. Poyhonen, (clusters of alveoli) where gas exchange that causes central airway narrowing L. Kellomaki and A. Muittari, Thorax 37, 516-520 (1982). is strongly correlated to airway wall occurs. Previously, computational [2] S. Bayat, L. Porra, G. Albu, S. Layachi, modelling studies have shown that infi ltration by infl ammatory cells such F. Petak, P. Suortti, Z. Hantos, due to feedback between regional tidal as eosinophils. These fi ndings bring A.R.A. Sovijarvi and W. Habre, Am J expansion and airway constriction, signifi cant new insight to the current Respir Crit Care Med 181, A5029 (2010). even a uniform constriction of airway understanding of the emergence of [3] J.G. Venegas, T. Winkler, G. Musch, M.F. Vidal Melo, D. Layfi eld, heterogeneous airway constriction, smooth muscle above a critical level N. Tgavalekos, A.J. Fischman, can lead to self-organised emergence and have important implications for R.J. Callahan, G. Bellani, and R.S. Harris. of patchy areas of ventilation defects the treatment strategies for allergic Nature 434, 777-782 (2005). resulting from clustered peripheral asthma.

MORPHOLOGICAL EVOLUTION OF Principal publication and authors B. Bayerlein (a), P. Zaslansky (b), Y. Dauphin (c), A. Rack (d), P. Fratzl (a) THE GROWING MOLLUSC SHELL and I. Zlotnikov (a), Nature Materials 13,1102–1107 (2014). REVEALED BY X-RAY COMPUTED (a) Department of Biomaterials, Max Planck Institute of Colloids and MICROTOMOGRAPHY Interfaces, Potsdam (Germany) (b) Charité Universitätsmedizin BSRT and Julius-Wolff Institute, Berlin (Germany) Living organisms form complex despite a relatively narrow range (c) Micropaléontologie, UFR TEB, mineralised biocomposites that of constituent inorganic materials. Université P. & M. Curie, Paris (France) perform a variety of essential Hence, a signifi cant eff ort has been (d) ESRF functions. These biomaterials are oft en directed to study the process of multifunctional, being responsible biomineralisation – to understand for not only structural support how organisms assimilate elements and mechanical strength, but also from the environment in which they provide optical, magnetic or sensing live and incorporate them into living capabilities. This remarkable diversity tissues. Many studies have focused on in functionality is accomplished the biological and chemical aspects

The European Synchrotron 55 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

of the interaction between inorganic perpendicular to the layer as a measure precursors and macromolecular of time. components resulting in controlled growth of diff erent biominerals. In The thickness of the organic contrast, thermodynamic constraints interprismatic matrix surrounding that govern many aspects of the calcitic prisms matches well the microstructure formation and resolving power accessible at ID19, morphology of the mineralised tissues the microtomography beamline. are much less understood. To study the shell geometry, a combination of absorption- and Mollusc shells are widely used as phase-contrast enhanced imaging model systems to investigate the modes was used to resolve the biomineralisation process of calcium organic/calcite prism boundaries and

carbonate (CaCO3) by living organisms. to reconstruct the microstructure of For example, the shell of Pinna nobilis the prismatic layer. The analysis of has two juxtaposed mineralised layers the 3D microtomography data was with completely diff erent properties performed in the scope of classical and therefore, provides two diff erent grain growth and coarsening theories examples of calcium carbonate while considering the growth as a two- assemblies within the same organism: dimensional problem where the time a calcitic layer exhibits an elongated parameter was replaced by the direction prismatic microstructure and an of the structural evolution [2]. We aragonitic layer exhibits a ‘brick-and- focused on mean fi eld considerations, mortar’ nacreous microstructure. In where the growth kinetics of a single the prismatic layer, the mineral prism prism is described by an average building blocks are surrounded by behaviour of the entire prismatic layer a 1 µm think organic interprismatic and topological considerations of space matrix defi ning the shape of individual fi lling. prisms. The biomineralisation process of the entire prismatic layers is initiated Figure 65a shows a representative on the organic layer covering the microtomography cross section, shell (periostracum) and propagates exhibiting a honeycomb-like towards the inside of the shell by the microstructure in the plane secretion of a mixture containing the perpendicular to the growth direction precursors for calcite formation [1]. of the prismatic layer. A representative The microstructure of the layer, a 3D rendering of a segment of the few millimetres in thickness, evolves layer, reconstructed from a continuous as an assembly of growing prisms set of 2D sections similar to the one and the interprismatic matrix as a presented in Figure 65a, is shown function of the growing thickness of in Figure 65b. These data clearly the layer. Hence, the growth history demonstrate that the average area of the prismatic layer is recorded (cross-section) of the prisms increases in its microstructure and can be during growth, in an eff ort to decrease accessed by an appropriate imaging the total interface area between the technique, interpreting the coordinate prisms and the adjoining organic matrix. In complete agreement with the predictions from classical grain growth theories based on mean fi eld and on topological considerations, some of the prisms are seen to shrink and disappear while others grow in their

Fig. 65: a) Representative 2D microtomography section obtained perpendicular to the growth direction of the prismatic layer. b) Representative 3D microtomography section of the prismatic layer of Pinna nobilis. Growing prisms are colour-coded blue, shrinking prisms are colour-coded purple.

56 ESRF X-RaY ImaGING

cross-sectional area (purple and blue prismatic layer in P. nobilis using prisms in Figure 65b, respectively). high-resolution microtomography and Moreover, the data reveal that the based on the comparison with well- grain size distribution renormalised by known theories of grain growth is a its mean value is time invariant during fundamental step towards a better microstructural evolution, so that the understanding of various aspects References [1] F. Marin, P. Narayanappa and coarsening process can be considered of biomineralisation. The results S. Montreuil, in Molecular statistically self-similar. show how the living organism uses Biomineralization Vol. 52, W.E.G. Müller thermodynamic boundary conditions (Ed.), Springer Verlag, 353-395 (2011). This quantitative description of the to defi ne the morphology of its calcite [2] H.V. Atkinson, Acta Metall. 36, 469- structural development of the calcitic shell. 491 (1988).

COARSENING AND BREAK-UP REVEALED Principal publication and authors D. Bouttes (a), E. Gouillart (b), E. Boller (c), D. Dalmas (b) and BY X-RAY MICROTOMOGRAPHY OF D. Vandembroucq (a), Physical Review Letters, 112, 245701 (2014). PHASE-SEPARATED GLASSES (a) Laboratoire PMMH, UMR 7636 CNRS/ESPCI/University Paris 6 UPMC/ University Paris 7 Diderot (France) (b) Surface du Verre et Interfaces, In their molten state, some silicate Aft er image processing, we were able UMR125 CNRS/Saint-Gobain, glasses spontaneously phase-separate: to observe the intricate geometry of Aubervilliers (France) they form domains of diff erent one of the two phases (Figure 67). At (c) ESRF compositions [1, 2]. Such domains the beginning of the heat treatment, can take diff erent shapes, such as the system is bicontinuous, which droplets or interconnected channels. means that there is only one domain The physics is then very similar to an for each phase, and these domains emulsion of oil and water, with the are interpenetrating. The geometry microstructure gradually coarsening is similar to a sponge, where there is to reduce the interfacial energy. In only one “solid” part, that is contiguous the case of a silicate glass, the liquid (otherwise it would collapse), and only state is obtained at temperatures one “void” part (otherwise the water generally higher than 700°C, and by could not fi ll it all). In fact, it is possible cooling it below this temperature one to produce a “glass sponge” by leaching can “freeze” the system. The domains one of the two separated phases with formed in the liquid state then become acid, which, for example, is used to a microstructure in the solid, and can be produce porous glass membranes. used to tune some physical properties. Using X-ray tomography at ID19, we During the heat treatment, we were were able to visualise the formation able to measure how the typical size of this microstructure in a barium borosilicate glass in three dimensions (3D), and observe the eff ect of the time spent in liquid state.

With X-ray tomography, one reconstructs 3D images of the Fig. 66: Experimental absorption of a sample from several setup at ID19. The hundreds of radiographs. The high sample, a 2 mm cylinder, intensity fl ux at ID19, together with a is placed in an alumina crucible mounted on a setup suited for ultrafast acquisition, rotation stage. A furnace, allows fast scans. Moreover, a dedicated mounted on a vertical furnace is used such that the same translation stage, is sample can be heated for some time at lowered over the sample a given temperature then quenched to for the heat treatments. In this image the X-rays be scanned, then heated again, and so would be arriving from on (Figure 66). the left .

The European Synchrotron 57 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

References of the domains increases in order to scale, which itself grows with time. The [1] A.J. Bray, Theory of phase-ordering reduce the interfacial area where the experiments confi rmed this theoretical kinetics. Advances in Physics, 43(3), two immiscible phases are in contact. prediction, up to a limit: the coarsening 357-459 (1994). [2] O.V. Mazurin and E.A. Porai-Koshits General predictions suggest that this is slowed down by the fragmentation of (Eds.). Phase separation in glass, Elsevier coarsening should be self-similar. the less viscous phase into disconnected (1984). This means that the geometry of the droplets. This important fragmentation domains is statistically the same at mechanism was identifi ed for the any time provided that all lengths fi rst time thanks to the possibility of are divided by a characteristic length following the same sample at diff erent heat treatment times.

The experimental setup at ID19 has recently been improved with a new camera and rotation stage. The acquisition time has been reduced from a few minutes to a few seconds to obtain 3D images with micrometric resolution. This has triggered new studies, which are ongoing, to investigate the origin of the unexpected fragmentation. It is now possible to follow in situ the coarsening (without quenching to room temperature). In particular, the dynamics of evolution of domain shapes are considered, and point towards hydrodynamics eff ects. Such understanding of microstructure formation could lead to new ways of building tailored microstructures in phase-separated glasses.

Fig. 67: 3D rendering of the barium-rich phase during coarsening. The largest domain, in green, is percolating. Other domains, in purple, are isolated droplets. The shades of colour represent the mean curvature value. The lateral size of the cubes is 140 µm.

Principal publication and authors DIRECT THREE-DIMENSIONAL IMAGING Y. Cheng (a), H. Suhonen (b,c), L. Helfen (a,b), J. Li (a,d), F. Xu (a,b), OF POLYMER–WATER INTERFACES M. Grunze (a,d), P.A. Levkin (a,d) and T. Baumbach (a), Soft Matter 10, 2982- 2990 (2014). BY NANOSCALE HARD X-RAY PHASE (a) Karlsruhe Institute of Technology (Germany) TOMOGRAPHY (b) ESRF (c) Presently at: University of Helsinki (Finland) Surfaces with special wettability on lubricant-infused porous polymer (d) Department of Applied Physical Chemistry, University of Heidelberg can exhibit unique properties like substrates [1] were found to have (Germany) self-cleaning or cell repellency and particularly good anti-bacterial and therefore attract much research marine anti-biofouling properties. interest in the quest to improve coating Visualisation of such surfaces and technology. Superhydrophobic and their interfaces to water or aqueous superoleophobic surfaces, in particular, solutions have been attempted via have been studied intensively in recent confocal laser-scanning and via cryo- years due to their extraordinary liquid FIB/SEM microscopy but artefacts due repellent properties. Slippery liquid- to the required staining or freezing infused porous surfaces (SLIPS) based processes cannot be excluded.

58 ESRF X-RaY ImaGING

Fig. 68: Cross-section of the microporous BMA-EDMA surface on a glass substrate at an eff ective pixel size of 50 nm. Insets (a) and (b) are enlargements of the white dashed selections. Red colour labels A, B and C denote polymer globules, air pockets and dispersed globules, respectively. Air was trapped by the structured BMA-EDMA polymer. The coexistence of the Wenzel state (blue arrows) and Cassie–Baxter state (green arrows) was observed. A volume of 0.1×0.05×0.01 mm3 was rendered in 3D via grey level segmentation, see Figure 69a.

To better understand such surfaces in their natural state, we investigated their structure on the micro- and nanoscale by three-dimensional (3D) hard X-ray phase tomography. To have a representative interface area, several mm thick samples were needed and therefore hard X-rays were used. Since water, lubricant (oil) and polymers provide weak absorption Fig. 69: a) Volume contrast, we used holotomography at rendering of the local the nano-imaging endstation ID22NI mass density of the porous BMA-EDMA (now at ID16) for propagation-based surface, voxel size 50 nm. phase imaging of such interfaces. Blue and green arrows Multilayer-coated Kirkpatrick–Baez indicate Wenzel and (KB) optics provided a tightly focused Cassie–Baxter states, spot (50×50 nm2 H×V, FWHM) of respectively. b) Rendering of the slippery BMA-EDMA monochromatic X-rays (E = 29.6 keV, surface with a voxel size of -2 ∆E/E ≈10 ). The sample was placed 0.32 µm. downstream of the focus in the slightly diverging beam, and, by varying the source-specimen and specimen- detector distances, the geometrical magnifi cation M and thus the spatial resolution of the imaging setup could be adjusted. Fresnel diff raction patterns acquired with diff erent M and eff ective propagation distances D in this so-called projection microscopy set-up allow retrieval of the phase shift introduced by the specimen with a high sensitivity that allows us to distinguish between water, polymer and oil aft er globules of the porous structure can be tomographic reconstruction of the distinguished. The visualisation of the specimen’s electron density. structure (see also Figure 69) allows us to identify regions which coexist As the material for our study, we chose and follow two existing models that porous poly(butyl methacrylate-co- explain the hydrophobicity of a porous ethylene dimethacrylate) (BMA-EDMA) surface, namely the Wenzel state [3], surfaces, as described in [2]. The where the increased surface area of the substrates of 5×25 mm2 surface area rough interface is responsible for its were imaged inside a polypropylene hydrophobicity or the Cassie–Baxter tube of 8 mm diameter that was state [4] where gas trapped between fi lled with water. Each scan at all four the rough surface and water favours propagation distances took around 45 hydrophobicity. minutes. Figure 69 shows 3D visualisations The 2D distribution of the electron of such interfaces at two diff erent density on a tomographic cross- spatial resolutions in the nm-scale section through the substrate is shown (a) and on the (sub-)µm scale. For in Figure 68. Water, air and polymer the oil-infused surface in (b), such 3D

The European Synchrotron 59 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

data sets can be used to determine with hard X-ray phase-contrast features such as the infusion quality nanotomography. The method or the spatial fl uctuation of the provides isotropic volumetric References thickness of the lubricant – here structural characterisation of complex [1] T.-S. Wong, S.H. Kang, S.K.Y. Tang, E.J. Smythe, B.D. Hatton, A. Grinthal perfl uoropolyalkylether (PFPE) oil – interfaces as well as bulk materials and J. Aizenberg, Nature 477, 443–447 e.g., due to the anisotropic pressure from sub-µm to nm scales whilst (2011). imposed by the surrounding water preserving the ambient conditions and [2] P.A. Levkin, F. Svec, J.M.J. Fréchet, environment. representative specimen sizes. These Adv. Funct. Mater. 19, 1993-1998 advantages are of central importance (2009). [3] R.N. Wenzel, Ind. Eng. Chem. 28, Summarising, we have demonstrated to help understand the correlations 988–994 (1936). non-invasive direct 3D imaging between surface properties and their [4] A.B.D. Cassie and S. Baxter, Trans. of surface–water interfaces on physical micro- or nanostructures Faraday Soc. 40, 546–551 (1944). hydrophobic porous BMA-EDMA which are needed to identify present and slippery BMA-EDMA surfaces limitations of such coatings.

Principal publication and authors HOW CAN LOCAL ENVIRONMENTAL A. Accardo (a), V. Shalabaeva (a), M. Cotte (b), M. Burghammer (b), R. Krahne (a), C. Riekel (b) and CONDITIONS INFLUENCE THE S. Dante (a), Langmuir 30, 3191−3198 (2014). SECONDARY STRUCTURE OF AN (a) Istituto Italiano di Tecnologia, Genova (Italy) AMYLOID BETA PEPTIDE? (b) ESRF

Extracellular amyloid plaques are a oleoylphospatidylcholine) and POPS hallmark material for the study of (1-palmitoyl-2-oleoylphospatidylserine). Alzheimer’s disease. The neurotoxicity of their main components, the amyloid The investigation was supported by β peptides (Aβs), may be mediated by a multi-technique approach based their direct interaction with the neural on a combination of micro-raman lipid membrane. In order to shed (μRaman), synchrotron radiation more light on conformational and (SR) micro-Fourier-transform infrared structural processes of amyloidosis spectroscopy (μFTIR) at ID21 and involving membranes, we focused micro X-ray diff raction (μXRD) at ID13. our attention on the Aβ(25−35) and In order to enhance the sensitivity of aβ(1-42) species in the presence of these probes, we deposited droplets a phospholipid mixture that mimics with Aβ and phospholipid suspensions the phospholipid composition of the on highly hydrophilic nanostructured neural membranes, i.e., a lipid mixture substrates. Ring-like solid residues were made of POPC (1-palmitoyl-2- formed almost immediately due to the strong heterogeneous evaporation rate on hydrophilic substrates (Figure 70a,b). The aid of micro- fabricated supports for structural studies had already been successfully exploited in previous experiments [1]. As detected by µFTIR, the outer rim of aβ-25-35 residue (Figure 71a,b) was characterised by β-sheet material

Fig. 70: a,b) SEM micrographs of Aβ(25−35) ring- like residues; c) Sketch of an evaporating droplet on a hydrophilic substrate. The evaporation rate is stronger near the contact line, provoking a radial convective motion of the solute from the centre toward the contact line.

60 ESRF X-RaY ImaGING

(peak at 1634 cm−1), while the inner zone (Figure 71c) contained peptide in α-helical conformation (1657 cm−1). The high evaporation fl ux at the triple contact line resulted in an accumulation of solute while the inner regions were depleted (Figure 70c). Therefore, the convective fl ow and the enhanced peptide concentration in proximity to the external rim, have been shown to result in self-assembly of β-amyloid fi brils. At the inner rim, the lower amount of material was not suffi cient for the α/β transition and the peptide remained arrested in a α-helical conformation. The aβ(25−35)/POPC/POPS mixture showed, on the contrary, the exclusive presence of β-sheet material without any trace of α-helices indicating that the presence of phospholipids plays an active role in the fi brillation process, as already suggested by previous studies [2]. As a control, μFTIR experiments on aβ(1−42) in the presence and absence of POPC/POPS revealed in both cases the presence of parallel β-sheet material. Complementary information on the crystalline fractions of the lipid and peptide system was obtained by micro small and wide angle X-ray scattering (μSAXS/WAXS). The pure aβ(1−42) pattern (Figure 71d) showed β-sheet features with a 4.2 Å peak characteristic of the distance between hydrogen-bonded strands and a 10 Å peak due to the β-sheet stacking. In contrast, the Aβ(25−35) rim showed only a very weak signature of a peak in the range of 10 Å, suggesting stacking disorder (Figure 71e) and was further Fig. 71: a,b,c) FTIR spectra recorded from confi rmed by a more detailed analysis several regions of the Aβ(25−35) residue of the peak profi le by fi tting Gaussian dried on highly hydrophilic BaF2 surfaces. functions (Figure 71g). A composite Typical μXRD mesh scan and relative peak with d ~4.7/4.1 Å (Figure 71g) azimuthal averages of the Aβ(1−42) (D) agreed with the possible coexistence and aβ(25−35) (e) droplet rims. (f) XRD of a β-phase (4.1 Å) together with an mesh-scan of the Aβ(25−35) solid rim. (g) α-phase (~4.7 Å). Peak fi tting of the Aβ(25−35) XRD pattern of Figure 71e. Based on these results, we were able to demonstrate the presence of an α/β transition from the internal rim of the α-helical structures and introduced coff ee-ring residues to the external rim the presence of pure β-type material. in the pure Aβ(25−35) case, probably An integrative μXRD analysis of the due to the convective fl ow inside the same residues confi rmed the μFTIR droplet sitting on highly hydrophilic experiments. Further work could help References substrates that enhances the local to clarify the role of enzymes, such as [1] A. Accardo, E. Di Fabrizio, T. Limongi, G. Marinaro and C. Riekel, Journal of concentration of the protein at the acetylcholinesterase, and of diff erent Synchrotron Radiation 21, 643-653 external edge of the drying drop. In lipid systems in the amyloidosis of (2014). contrast, the presence of POPC/POPS the analysed peptides and of other Aβ [2] A.M. Sheikh and A. Nagai, FEBS lipids in the protein did not result in fragments. Journal 278, 634−642 (2011).

The European Synchrotron 61 SCIENTIFIC HIGHLIGHTS X-RaY ImaGING

Authors E. Pouyet (a,b), M.Cotte (a,c), Micro-FTIr STUDY OF THE POLYMER A.Nevin (d) and D. Saviello (e). (a) ESRF (b) ARC-Nucléart - CEA/Grenoble DEGRADATION IN DESIGN OBJECTS (France) (c) Laboratoire de Physique des Solides, UMR 8502 Université Paris Sud, Orsay Recent developments in analytical fragments from historical Italian design (France) chemistry are being put to good objects from the 1960’s [ 2] (Figure 72). (d) Consiglio Nazionale delle Ricerche use in the study of artwork. Some – Istituto di Fotonica e Nano tecnologie analyses can be carried out directly (CNR-IFN), Dipartimento di Fisica, Fourier transform infrared Politecnico di Milano (Italy) on the objects, while in some cases, spectroscopy (FTIR) is well suited to (e) Dipartimento di Chimica Materiali micro-fragments can be sampled, the characterisation of such materials e Ingegneria Chimica, Politecnico di allowing a more accurate analysis since it probes molecular groups Milano (Italy) of the material’s composition, at the and allows the disappearance of the micrometre scale. Care is needed in th e reactive groups and the formation of preparation of these tiny fragments degradation products to be followed. In- to reveal as much detail as possible of depth analysis of such phenomena has their intrinsic structure (usually multi- already been reported, but limited by layer) and to preserve the chemical an in-depth resolution of about 100 µm and physical properties of the sample. [ 3]. The goal of the present study was New preparation strategies have been to study transversal sections of the recently developed at beamline ID21 for polymers with synchrotron-based thin sections from painting fragments micro-FTIR (at ID21), in order to locate [1]. We report here their application the diff erent components constituting to the study of the degradation of the polymers, at the micrometre acrylonitrile-butadiene-styrene (ABS) scale. This required the preparation polymer in historical Italian design of thin-sections (2 µm), sliced objects [2] . perpendicularly to the block surface. The sample fragment, smaller than ABS has been widely used over the 1 mm3 and representative of the entire last 50 years as a plastic material stratigraphy, was clamped between two in industrial applications. It is sheets of polycarbonate or polyether present in many design objects and ether ketone foil, PEEK, (3 × 6 x 1 mm3). household items such as telephones, Thin-sections of diff erent thicknesses domestic appliances, lamps, musical were obtained using a microtome, at instruments, some of which are room temperature. now being exhibited in international museums. However, as observed with This method produced very thin paintings, ABS suff ers from photo- samples with high surface quality while ageing, which leads to the alteration avoiding the use of embedding media of its physical properties, chemical and polishing or grinding; it solves modifi cations such as crosslinking the risk of chemical and analytical and chain scission, colour changes interference, both for IR and X-ray (yellowing and discolouration) and analyses. worsening of mechanical performance. The objective of this study was to On such a section, both the internal characterise and assess the depth pristine polymer and the superfi cial of alteration in terms of chemical products formed during photoageing modifi cations on two types of samples: can be identifi ed and localised, in a artifi cially photo-aged ABS blocks and single FTIR map. The visible image (Figure 73a) and UV fl uorescence images (Figure 73b) highlight a luminescent upper layer of about 45 µm on the surface of an artifi cially aged block. In Figure 73c, three FTIR

Fig. 72: Italian design objects made of ABS. (left ) Grillo (Marco Zanuso and Richard Saper for Siemens), 1965; (right) E63 (Umberto Riva for Francesconi) 1963, Private Collection.

62 ESRF X-RaY ImaGING

References [1] E. Pouyet et al., Analytica Chimica Acta 822, 51-59 (2014). [2] D. Saviello et al., Analytica Chimica Acta 843, 59-72 (2014). [3] J.G. Bokria and S. Schlick, Polymer 43, 3239-3246 (2002).

spectra from three diff erent depths are Fig. 73: SR µFTIR study of the oxidation shown: 40, 194 and 268 µm. When of an ABS model fi lm (sample exposed to UV-irradiation for 1000 hours); a) Visible comparing surface and bulk areas, the light microscope image of a 2 µm thin- spectra from the most external parts section, b) UV light microscope images of of the sample show a clear increase of 1000 hour aged ABS sample on the left , the signal relative to C = O stretching and of pure resin thin-section (2 µm) on the in the region 1800-1600 cm1 and the right c) FT-IR spectra acquired at diff erent positions corresponding to three diff erent appearance of an intense and broad depths from the sample surface and one signal relative to O-H stretching spectrum characteristic of the absorption centred at 3436 cm-1. These bands of an embedding resin, prepared as a 2 µm were attributed to ketones, aldehydes, thin-section. esters, α,β-unsaturated carbonyl, carboxylic acids and alcohols. The depth of oxidation in samples was also evaluated, for samples aged for diff erent amounts of time. µFTIR highlighted a selective and progressive gradient of degradation, greatest at the surface of the sample and which decreases in- depth. The oxidation compounds tend to form a passivation layer, limiting further oxidation of the bulk ABS [ 2].

To illustrate the strength of this sample preparation protocol, Figure 73c includes the spectrum of a standard embedding resin (Historesin, Leica), showing strong overlaps with FTIR bands characteristic of the chemical bonds formed during photo-ageing. Identifi cation and precise localisation of these degradation compounds would have been diffi cult or even impossible using a standard resin embedding protocol.

The European Synchrotron 63 SCIENTIFIC HIGHLIGHTS Highlights 2014

Dynamics and extreme conditions

This year’s highlights in dynamics and The technical design report for the and Utrecht University. A BMBF extreme conditions include the study move of ID09A to ID15B has been grant under the leadership of the of fundamental phenomena such as completed, and the current planning Bayerisches Geoinstitut Bayreuth dynamics in glasses and clathrates, foresees the closure of the high- aims at the further development of and magnetism at very high pressures pressure branch of ID09 at the end high-pressure sample environment beyond 200 GPa. This chapter also of 2015. ID15B will be available for in materials at extreme conditions comprises work on the synthesis user operation from scheduling period at ID18. Further support for ID18 has and in situ characterisation of novel II/2016 onwards. been received in the framework of materials at extreme conditions, the the Helmholtz Russia Joint Research investigation of correlated electron New capabilities at ID27 comprise Group (“Sapphire ultra optics for systems, and unique insights into the a Soller slit system interfaced to synchrotron radiation”; FZ Jülich) structure and dynamical phenomena the CO2 laser heating bench for and the XNAP collaboration for the of Earth’s interior. In addition, the the collection of high quality X-ray development of a two-dimensional chapter hosts a contribution on the diffraction patterns at high pressures counting pixel X-ray detector based atomic motion in network glasses and temperatures, and the full remote on silicon avalanche photodiodes. A revealed by coherent X-rays from control of the rotational-tomography further BMBF grant, led by the Goethe- beamline ID10A. Experiments Paris-Edinburgh cell (rotoPEC) for Universität Frankfurt am Main, performed at the Dynamics and diffraction and imaging applications. supports the installation of the high- Extreme Conditions Group’s beamlines throughput diffractometer for diffuse resulted in about 100 publications in At the Nuclear Resonance Beamline scattering studies as a side-station peer reviewed journals with 20 papers (ID18), the portfolio of available to beamline ID28. A second grant in high-impact journals. Mössbauer isotopes was expanded for ID28 (ANR-DPG grant; Goethe- to include the Ru-99 resonance (at Universität and ESRF) is dedicated Besides user operation, the beamline 89 keV) and the Os-187 resonance (at to the understanding of the role of teams and support groups were 9.8 keV). The synchrotron Mössbauer dynamic disorder in carbonates. highly active in developing and source (SMS) setup took up nearly Finally, ID27 profits from the ANR implementing new instruments and half of the available beamtime in Blanc grant MOFLEX “Structure and facilities. 2014, focusing on investigations of dynamics of simple molecular fluids magnetic and electronic properties under extreme conditions of pressure The technical design report for under extreme conditions, including and temperature“, coordinated by the hard X-ray microscope (HXRM) double sided laser heating capabilities. IMPMC Paris and CEA Bruyères. project at beamline ID06 has been completed, and the project has The new inelastic X-ray scattering This year’s achievements would not entered its implementation phase. In beamline, ID20, has reached all its have been possible without a vibrant parallel, proof-of-concept experiments target performances, and has been user community and the continuous were conducted, and first results 100% available for user operation support and effort of all the other were recently published in Nature since 1 March 2014. As a consequence, ESRF divisions. The advent of the Communications (H. Simons et al., the UPBL6 project was officially closed ESRF Upgrade Phase II provides a Nat. Commun. 6, 6098 (2015), doi: on 31 December 2014. further stimulus for a sustained 10.1038/ncomms7098). R&D programme. The significantly The high-throughput diffuse higher photon flux density will allow The large volume press (LVP at ID06) scattering diffractometer project, studies to be carried out on much continued to operate at 30% of the a side-station of beamline ID28, smaller sample volumes at much standard allocation time for an ESRF was officially launched in February higher pressures and temperatures, beamline. Deformation experiments 2014, and is well under way. First and on much shorter time scales. are now possible up to 10 GPa and commissioning experiments are The increased coherence of the X-ray 1700 K, with strain rates of 10-6 expected to take place towards the beam at high energies will allow new to 10-4 sec-1 and 35% maximum end of 2015. techniques to be used at extreme total strain. Two-stage X-ray conditions such as imaging methods transparent anvils were successfully External funding has become a using phase contrast, coherent commissioned, allowing fully open significant complement to the diffraction, and ptychography. 2D solid-angle diffraction up to investment and staff capacity of For further details, please consult 20 GPa and 2000 K. Furthermore, the group. The HXRM project at the chapter on matter at extreme new detection and data acquisition ID06 profits from collaboration with pressures and temperatures in the schemes are under development. Technical University of Denmark recently published Orange Book. (DTU), Norwegian University of Science and Technology (NTNU) M. Krisch

64 ESRF DYNamICS aND EXTREmE CONDITIONS

Principal publication and authors FUNDAMENTAL PHENOMENA A.I. Chumakov (a), G. Monaco (b, a), A. Fontana (b, c), A. Bosak (a), R.P. Hermann (d, e), D. Bessas (d, e, a), B. Wehinger (a), W.A. Crichton (a), roLE oF DISorDEr In THE THErMoDynAMIcS AnD M. Krisch (a), R. Rüff er (a), G. Baldi (f), AToMIc DynAMIcS oF gLASSES G. Carini Jr. (g), G. Carini (h), G. D’Angelo (h), E. Gilioli (f), G. Tripodo (h), M. Zanatta (i, c), In the late 1950s, the low-temperature vibrational excitations in the ~1 THz B. Winkler (j), V. Milman (k), K. Refson (l), M. T. Dove (m), N. Dubrovinskaia (n), thermal conductivity of glasses was frequency range of the density of states L. Dubrovinsky (n), R. Keding (o) and found to be several orders of magnitude (DOS) g(E) [2]. In the reduced DOS Y.Z. Yue (p), Physical Review Letters 112, smaller and the heat capacity presentation, g(E)/E2, these states form 025502 (2014). several times bigger [1] than those of a distinct peak (the so-called ‘boson’ (a) ESRF crystals. The anomalously low thermal peak) at a position where the reduced (b) University of Trento (Italy) (c) University of Rome (Italy) conductivity was soon successfully DOS of the corresponding crystals (d) Jülich Centre for Neutron Science and explained by phonon scattering is nearly fl at. This excess of heat Peter Grünberg Institut (Germany) caused by disorder. However, the capacity and the corresponding excess (e) University of Liege (Belgium) understanding of the excess of the heat of vibrational states are universally (f) IMEM-CNR, Parma (Italy) capacity demanded longer studies. observed for all glasses and with all (g) IPCF-CNR, Messina (Italy) (h) University of Messina (Italy) relevant experimental techniques. (i) University of Perugia (Italy) The excess of the heat capacity in Nevertheless, the results of numerous (j) Goethe-University, Frankfurt glasses at ~10 K was found to be studies still do not converge to a (Germany) associated with a similar excess of the unifi ed answer on how disorder causes (k) Accelrys, Cambridge (UK) (l) STFC Rutherford Appleton Laboratory, (UK) (m) Queen Mary University of London (UK) (n) University of Bayreuth (Germany) (o) Max Planck Institute for the Science Fig. 74: The density of states g(E) (a–c), the reduced density of states g(E)/E2 (d–f), and the of Light, Erlangen (Germany) heat capacity c /T3 (g–i) for various glassy and crystalline polymorphs of SiO . The left P 2 (p) Aalborg University (Denmark) panels (a, d, g) compare the most common glassy and crystalline polymorphs and reveal a noticeable diff erence in the displayed properties. The middle and the right panels compare the glassy and crystalline polymorphs with matched densities, namely, the low-density (b, e, h) and high-density (c, f, i) polymorphs. They demonstrate that the atomic dynamics and thermodynamics of the glassy and crystalline polymorphs with matched densities do not diff er much from each other. From Ref. [3].

The European Synchrotron 65 SCIENTIFIC HIGHLIGHTS DYNAMICS AND EXTREME CONDITIONS

these features. Recent experiments The measured DOSs revealed that SiO2 conducted at beamlines ID18 and glasses have the same number of low- ID28 have now revealed that disorder energy states and the same number

actually is not responsible. of excess states as SiO2 crystals. For the most common glass, these states

The studies were performed on SiO2, are located at lower energy than for which exists in various glassy and the typical crystal and provide the crystalline polymorphs with different higher heat capacity (left column of structures and densities. The density Figure 74).

(ρ) of the most common SiO2 glass (ambient silica glass, ρ = 2.20 g / cm3) For glasses and crystals with matched is smaller than the density of the densities (middle and right columns most common crystal (α-quartz, of Figure 74), the excess states are 3 ρ = 2.65 g/cm ). However, SiO2 also located in the same energy region exists in a lower density crystalline and provide the same heat capacity. form (α-cristobalite, ρ = 2.29 g/cm3). Furthermore, the boson peak of glasses The density of this crystal is nearly the appears as a smooth counterpart of the same as the density of the ambient van Hove acoustic singularity of the

glass. Furthermore, SiO2 exists also corresponding crystals. in a higher-density glassy form ρ 3 Thus, being responsible for the References (densified silica glass, = 2.67 g/cm ). [1] R.C. Zeller, R.O. Pohl, Phys. Rev. B 4, The density of this glass is nearly the anomalous thermal conductivity of 2029 (1971). same as the density of α-quartz. Thus, glasses, disorder does not greatly affect [2] U. Buchenau, N. Nücker and SiO2 offers a suitable 2×2 matrix of the frequency spectrum of the atomic A.J. Dianoux, Phys. Rev. Lett. 53, 2316 (dis)order versus density, which allows vibrations and the heat capacity. The (1984). usually lower heat capacity of glasses [3] A.I. Chumakov and G. Monaco, J. for a model-independent evaluation of Non-Cryst. Solids 407, 126-132 (2015). the relative role played by structural is caused by their usually lower density. disorder and density.

Loc alISATIon of propagating phonons in a Principal publication and authors S. Pailhès (a), H. Euchner (b), perfectly crystalline solid V.M. Giordano (a,f), R. Debord (a), A. Assy (c), S. Gomès (c), A. Bosak (d), D. Machon (a), The energy loss due to thermal oscillators and the observed S. Paschen (e), M. de Boissieu (f), Phys. Rev. Lett. 113, 025506 (2014). conversion processes has triggered the suppression of the thermal conductivity (a) Institute of Light and Matter, search for new functional materials was then rationalised by considering UMR5306 Université Lyon 1-CNRS, allowing for more efficient energy the inelastic resonant scattering of Université de Lyon, Villeurbanne consumption. Promising candidates heat carrying phonons with these (France) are ultra-low thermal conductivity oscillators. However, in 2008, a study of (b) Institute of Materials Science and Technology, Vienna University of materials that can be made use of in phonons in skutterudites demonstrated Technology (Austria) thermoelectric converters for waste heat that this intuitive picture of resonant (c) Centre de Thermique de Lyon recovery as well as in thermal barrier scattering by loosely bound guest CETHIL UMR5008, UCBL, INSA de Lyon, coatings. In non-metallic crystalline atoms is not applicable, putting into CNRS (France) solids, heat is mainly carried by atomic question the current understanding of (d) ESRF (e) Institute of Solid State Physics, motion, in the form of propagating the anomalous heat properties in these Vienna University of Technology waves. In a quantum mechanical materials [1]. (Austria) picture, the atomic vibrations can be (f) SIMAP, UJF, CNRS, INP Grenoble, St described by means of the free motion We have investigated the mechanism Martin d’Hères (France) of bosonic quasiparticles, known as at the origin of the low thermal phonons. conductivity in cage based materials by mapping the phonon states in the Cage-like structures encapsulating simplest representative intermetallic

guest atoms or molecules such clathrate, Ba8Si46 (see Figure 75), as clathrates, skutterudites, using Inelastic X-ray Scattering (IXS) at β-pyrochlore and zeolites are known beamline ID28. Phonons propagating to exhibit anomalously low thermal along the main crystallographic conductivities. The encaged atoms were directions were studied and the formerly considered as independent corresponding IXS cross sections

66 ESRF DYNamICS aND EXTREmE CONDITIONS

Fig. 75: a) Ba8Si46 microcrystal grown at 5 GPa and 1000°C. b) Representation of

face sharing BaSi20 and BaSi24 cages as building units of type-I clathrates. Atoms are represented by their atomic displacement ellipsoids obtained from single crystal X-ray diff raction. were simulated by ab initio density Fig. 76: a-h) Example functional theory (DFT) calculations. of constant Q-scans for photons propagating In Figure 76, raw data of constant along the 001 direction. Q-scans, along the [100] direction, are Orange lines are fi ts of shown together with the corresponding the data. Blue profi les simulated IXS spectra, convoluted with correspond to the ab the instrumental resolution function. A inito simulated IXS spectra, convoluted surprisingly good agreement between with the instrumental the experimental and the simulated resolution function. energy profi les can be observed for both The blue vertical lines the position and the energy width of indicate the simulated the phonons. For phonon wave vectors phonon positions without the convolution. close to the Brillouin zone centre (q –> 0), all phonon spectra consist of a main peak at energies distinctly below the optical (op) guest excitations, which stems from the acoustic (ac) branch and follows a linear dispersion. No energy broadening was needed to fi t the acoustic and optic phonon profi les. The main eff ect that was observed in our experimental data was an abrupt decrease of the intensities of the acoustic phonons for wave vectors higher than a critical wave vector, qc, concomitant with an increase of the intensities of the guest phonons. This was accompanied by a strong bending of the dispersion relation of the acoustic excitations, with its slope approaching zero. qc was interpreted as the limit of the acoustic regime above which the acoustic character of the modes vanishes. The abrupt changes of the acoustic intensities were remarkably well reproduced by the ab lifetime and their propagative nature, initio simulation. until a critical wave vector is reached: at this point the character of the acoustic In conclusion, the fi lling of the phonons changes, resulting in modes cages with guest atoms leads to with an increasing contribution of the suppression of host propagative the guest atoms. A careful theoretical phonons, which are highly effi cient analysis provides evidence that the at carrying heat, over a wide range of guest modes are not propagative. energies and wave vectors in phase These modes release their energy space. The guest-host interaction can locally in an isotropic manner, leading be understood as a purely harmonic to a change from a propagative to a References [1] M. Marek Koza, M.R. Johnson, process, i.e. it is not dominated diff usive energy transport mechanism, R. Viennois, H. Mutka, L. Girard and by phonon-phonon scattering. The which may, in turn, be responsible for D. Ravot, Nature Materials 7, 805 acoustic phonons preserve a very long the low lattice thermal conductivity. (2008).

The European Synchrotron 67 SCIENTIFIC HIGHLIGHTS DYNamICS aND EXTREmE CONDITIONS

Principal publication and authors AToMIc MoTIon In nETWorK gLASSES rEVEALED By B. Ruta (a), G. Baldi (b), Y. Chushkin (a), B. Ruffl é (c), coHErEnT X-rAyS L. Cristofolini (d), A. Fontana (e), M. Zanatta (f) and F. Nazzani (g), Nat. Commun. 5, 3939 (2014). The remarkable properties of glasses important progress has been made in (a) ESRF are related to their out-of-equilibrium the understanding of the dynamics of (b) IMEM-CNR Institute, Parma (Italy) state in which relaxations occur on their high temperature phases, the so (c) Université Montpellier 2, Laboratoire time scales that were believed to be too called glass-forming liquid phase, a Charles Coulomb UMR 5221 (France) (d) Physics and Earth Sciences long to be observed. Thanks to the use microscopic theory of glasses is still Department, Parma University (Italy) of coherent X-rays, it has been possible missing due to the extreme diffi culty (e) Dipartimento di Fisica, Trento to investigate the atomic motion of a of probing the structural relaxation of University (Italy) silicate glass and discover the existence a glass far below the glass transition (f) Dipartimento di Fisica e Geologia, of unexpected atomic rearrangements temperature T , both with experiments Universitá di Perugia (Italy) g (g) Department of Physics, University of even in the deep glassy state, contrary and simulations [1]. Fribourg (Switzerland) to the common belief of a fully arrested state. These results modify our Thanks to the instrumental conception of the glassy state, suggest developments in the collection of more experiments, and ask for a new sparse scattering signals [2] and to microscopic theory of glass dynamics. an increased fl ux and coherence of X-ray beams, X-ray photon correlation Glasses are essential materials in spectroscopy (XPCS) has recently present day science and technology. emerged as a very powerful technique The optimisation of the properties of able to follow the evolution of the materials for technological applications dynamics at the atomic length scale in is oft en the result of a trial-and-error crystalline and amorphous materials approach, as a basic understanding of [3]. many of their properties is still lacking. One of the challenges for glasses is We used XPCS at ID10 to follow the that their intrinsic non-equilibrium temporal and temperature evolution nature poses formidable challenges of the relaxation dynamics for the both for experiments and theory. While fi rst time in a prototypical network glass, a sodium silicate, at both the mesoscopic and atomic length scales. Fig. 77: Arrhenius plot We revealed the existence of collective of the relaxation times measured with XPCS at atomic motion in the glassy state. This -1 is in contradiction with the common Qmax = 1.53 Å (coloured symbols) by cooling from idea of an ultraslow arrested out-of- the ultraviscous liquid equilibrium state: upon cooling the with diff erent cooling system from the melt to the glassy rates R. Black symbols are macroscopic data taken state, the structural relaxation time from the literature. The strongly departs from the supercooled dashed lines are Arrhenius liquid behaviour and displays a weak fi ts while the full line is the temperature dependence characterised macroscopic prediction of by relaxation times of about 103 s even the TNM model [1]. in the deep glassy state (Figure 77). This fast atomic motion is accompanied by the absence of any detectable physical ageing, not even in the glass transition region, in disagreement with macroscopic studies and theoretical Fig. 78: Wave vector approaches (Figure 77) [1]. Indeed, dependence of the relaxation current phenomenological models for times (circles) measured in a relaxation in glasses, such as the TNM silicate glass cooled down to model [1], assume that the physical ambient temperature with a ageing takes place on a timescale rate of 100 K/s from the high temperature melt. The data comparable to the structural relaxation are reported together with the time, which is usually the case for corresponding static structure macroscopic observables. In contrast to factor (squares). macroscopic studies, the microscopic dynamics remains stationary in the glassy state, even in those cases where

68 ESRF DYNAMICS AND EXTREME CONDITIONS

the annealing time is much longer than region, at lower scattering vectors, the structural relaxation time derived and corresponding to the mesoscopic from a simple extrapolation of the scale (Figure 78). While the first is supercooled liquid relaxation times. reminiscent of the de Gennes-like These results indicate that a different narrowing, also observed in coherent physical mechanism, unpredicted by neutron scattering experiments previous experimental and theoretical of supercooled liquids, the second studies, is responsible for the relaxation corresponds to the characteristic dynamics at the microscopic level. distance between the sodium diffusion channels in the network, in agreement Some insight into this novel dynamics with a high temperature simulation in References can be gained by inspecting its molten silicates. [1] L. Berthier and G. Biroli. Rev. Mod. dependence on the scattering vector. Phys. 83, 587-645 (2011). The relaxation time (t) shows a This observed complex microscopic [2] Y. Chushkin, C. Caronna and A. Madsen, J. Appl. Cryst. 45, 807-813 broad peak with two contributions: dynamics calls for a reexamination (2012). a shoulder close to the position of of phenomenological models and [3] B. Ruta et al. Phys. Rev. Lett. 109, the maximum of the static structure highlights the need for a new 165701 (2012). factor; a maximum in the pre-peak microscopy theory for network glasses.

C orrelATED ELECTRON SYSTEMS

Hyperfine splitting and room-temperature Principal publication and authors I. Sergueev (a), L. Dubrovinsky (b), ferromagnetism of Ni at multimegabar pressure M. Ekholm (c), O.Yu. Vekilova (d), A.I. Chumakov (e), M. Zajac (e), V. Potapkin (e,b), I. Kantor (e), High pressure generally leads to the did not predict the critical pressure S. Bornemann (f), H. Ebert (f), S.I. Simak (d), I.A. Abrikosov (d) and at which the magnetic moment is suppression of ordered magnetism in R. Rüffer (e), Physical Review Letters the solid state. The critical pressure quenched. The calculations show that 111, 157601 (2013). where magnetism disappears depends the ferromagnetic state is persistent (a) Deutsches Elektronen-Synchrotron, on the material. Here we observed up to 300 GPa at 0 K [4] and the Curie D-22607 Hamburg (Germany) ferromagnetism in Ni metal at room temperature is expected to decrease to (b) Bayerisches Geoinstitut, Universität Bayreuth (Germany) room temperature at about 160 GPa temperature up to 260 GPa, the highest (c) Swedish e-Science Research Centre pressure where magnetism has been [5] or at 300 GPa [6]. (SeRC), Department of Physics, reported. Chemistry and Biology (IFM), Linköping In this work, the magnetic properties University (Sweden) The critical pressure at which a material of Ni metal under high pressure were (d) Department of Physics, Chemistry and Biology (IFM), Linköping University investigated by nuclear forward becomes non-magnetic is of importance (Sweden) for the fundamental understanding scattering (NFS), using the 67.4 keV (e) ESRF of the origin of its magnetism. In nuclear transition in 61Ni. The (f) Department of Chemistry, Ludwig- particular, this is important for simple measurements performed at beamline Maximilians-Universität München 3d-transition metals Fe, Co and Ni ID18 revealed the presence of magnetic (Germany) due to their technological use. Fe and hyperfine splitting of Ni metal at Co become nonmagnetic above 16 room temperature up to 260 GPa (see [1] and 120 [2] GPa, respectively. In Figure 79), which confirm that Ni stays these cases, however, the collapse of ferromagnetic up to this pressure. ferromagnetism is associated with The magnetic hyperfine field shows a structural transitions of the crystal peculiar pressure dependence, reported structure. In contrast for Ni, no in Figure 80. It slowly increases up to structural transition is observed at its maximum value around 225 GPa least up to 200 GPa [3]. Thus, Ni is a and then decreases towards higher good candidate to study solely the pressures. This result seems to be in effect of pressure on the magnetic contradiction to the observed and properties. It was shown recently by theoretically predicted continuous X-ray absorption spectroscopy that decrease of the magnetic moment, Ni stays ferromagnetic up to 200 GPa which is supposed to be proportional [3]. Theoretical calculation so far to the magnetic hyperfine field.

The European Synchrotron 69 SCIENTIFIC HIGHLIGHTS DYNamICS aND EXTREmE CONDITIONS

The theoretical explanation of this behaviour reveals the importance of relativistic eff ects. In fact, the hyperfi ne fi eld consists of two main contributions of diff erent signs: a positive contribution due to the Fermi contact interaction and a negative contribution due to the orbital magnetic moment of the 3d electrons induced by spin- orbit coupling. The magnitude of both Fig. 79: Time evolution contributions decreases with pressure, of the nuclear forward similar to the magnetic moment but scattering for Ni with diff erent slopes. Thus, the sum of measured at room the two contributions, that determines temperature and diff erent the total hyperfi ne fi eld, increases up pressures. The solid lines show the best fi t to the to 100-180 GPa, reproducing well the data. The oscillation experimentally observed behaviour. period of the signal is inversely proportional to This work demonstrates a the magnetic hyperfi ne breakthrough in the technique that splitting. enables magnetism to be studied in Ni containing compounds at high and ultra-high pressures. The effi ciency of the method signifi cantly benefi ts from the application of high pressure because of the hardening of the material and, consequently, a substantial increase in the Lamb-Mössbauer factor. Thus, the highest pressure observed in this work is not limited by the NFS technique but by the mechanical stability of the diamond anvil cells used. Recent advances in high pressure technology [7] as well as th e decrease of the X-ray source size with the future upgrade of the ESRF would allow the technique to study Ni and Ni containing compounds at pressures even above 500 GPa.

Fig. 80: Pressure dependence of the hyperfi ne magnetic fi eld in Ni from experiment and theory, calculated using fully relativistic ab initio calculations. Positive and negative contributions to the calculated fi eld are shown as scalar relativistic calculations (mainly due to the Fermi contact interaction) and 3d orbital contribution.

References [1] O. Mathon et al., Phys. Rev. Lett. 93, 255503 (2004). [2] R. Torchio et al., Phys. Rev. B 84, 060403 (2011). [3] R. Torchio et al., Phys. Rev. Lett. 107, 237202 (2011). [4] T. Jarlborg, Physica C 385, 513 (2003). [5] S. Shang et al., J. Appl. Phys. 108, 123514 (2010). [6] V. G. Tissen, Sov. Phys. Solid State 25, 11 (1983). [7] L. Dubrovinsky et al., Nature Communications 3, 1163 (2012).

70 ESRF DYNamICS aND EXTREmE CONDITIONS

MATERIALS SYNTHESIS

Principal publication and authors DIVErSE cHEMISTry In THE MAgnESIUM/cArBon T.A. Strobel (a), O.O. Kurakevych (a, b), SySTEM D.Y. Kim (a), Y. Le Godec (b), W. Crichton (c), J. Guignard (c), N. Guignot (d), G.D. Cody (a) and A.R. Oganov (e-h), Inorg.Chem. 4, 53, Magnesium and carbon, both of which A bright golden-yellow powder of this 7020−7027 (2014). form numerous compounds with other new Mg-C compound was recovered (a) Geophysical Laboratory, Carnegie elements, have a remarkably low from high-pressure, high-temperature Institution of Washington, Washington affi nity for one another. At ambient experiments (Figure 81). Mass DC (USA) pressure, only the reaction of Mg or spectrometry analysis of the hydrolysis (b) Institut de Minéralogie, de Physique des Matériaux et de Cosmochimie, MgO with hydrocarbons leads to the products showed allene, C3H4, UPMC,CNRS, Paris (France) 4- formation of metastable carbides with indicative of the presence of C3 anions (c) ESRF reasonable yields, and pure Mg and within the structure. In addition, 13C (d) SOLEIL, Gif-sur-Yvette (France) carbon do not react to form stable NMR analysis confi rmed the presence (e) Department of Geosciences, SUNY, compounds at any temperature. With of two structurally distinct carbon Stony Brook, NY (USA) (f) Institute for Advanced Computational experimental data obtained at beamline atoms in an approximate 1:2 ratio, Science, SUNY, Stony Brook, NY (USA) ID06LVP, we have demonstrated and Raman spectra indicated carbon (g) Moscow Institute of Physics and that thermodynamically stable Mg-C stretching modes from [C=C=C]. The Technology, Dolgoprudny City (Russian compounds are indeed possible under crystal structure –clearly distinct from Federation) high-pressure conditions (above 5 GPa previously reported compounds– was, (h) Northwestern Polytechnical University, Xi’an (China) and 1300 K). The new compound however, impossible to solve using the synthesised, β-Mg2C3, crystallises in available powder X-ray diff raction data. the monoclinic C2/m space group, and 4- contains rare allene-derived C3 anions To resolve the structure of this new that are isoelectronic with CO2. magnesium carbide, two important steps were taken. First, potential

Mg2C3 structures were predicted using USPEX, an ab initio structural evolution algorithm. Second, in situ high-pressure synthesis was performed at beamline ID06 using the recently-commissioned large-volume press. The in situ synthesis allowed the nearly phase-pure synthesis of the new compound, and high-resolution X-ray diff raction patterns obtained allowed conclusive comparison with theoretical models.

Remarkably, the high-

pressure Mg2C3 structure predicted via USPEX was a perfect match to the experimental X-ray diff raction data (Figure 82). Thus, the

structure of β-Mg2C3 was solved. Unlike α-Mg2C3,

Fig. 82: In situ diff raction 13 Fig. 81: C MAS NMR spectrum of β-Mg2C3 data collected during recovered from 9 GPa showing two unique HPHT synthesis at ID06. carbon environments. Inset: optical Crystal structure of microscope picture of the golden-yellow 4- β-Mg2C3 showing C3 ions powder. aligned along the c-axis.

The European Synchrotron 71 SCIENTIFIC HIGHLIGHTS DYNamICS aND EXTREmE CONDITIONS

which contains alternating layers Finally, these results indicate that the 4− of C3 chains oriented in opposite Mg-C system should be completely 4− directions, all C3 chains within revised under high pressure [2]. β-Mg2C3 are nearly aligned along the The phase behaviour in this system crystallographic c-axis. is important for the production of structurally-perfect diamonds with Aft er exploring pressure as an high growth rates as compared to other additional dimension for the chemistry known metal-carbon systems, and also of the Mg−C system, four magnesium for the synthesis of unique nitrogen- References carbides are now known: (1) tetragonal free and boron-doped semiconductive [1] O.O. Kurakevych, T.A. Strobel, MgC2, (2) orthorhombic α-Mg2C3, (3) diamonds. With these new results, we D.Y. Kim and G.D. Cody, Angew. Chem. monoclinic β-Mg2C3, and (4) cubic now have a much deeper understanding Int. Ed. 52, 8930-8933 (2013). Mg C [1]. Taking into account that, at of Mg-C thermodynamics under [2] O.O. Kurakevych, Y. Le Godec, 2 T.A. Strobel, D.Y. Kim, W.A. Crichton ambient pressure and at pressures up to extreme conditions. and J. Guignard, J. Phys. Chem. C 118, ~5 GPa, the elements do not interact at 8128–8133 (2014). any temperature, it is quite astonishing to observe such rich chemistry.

A nEW crySTAL PHASE oF FErrocEnE Principal publication and authors D. Paliwoda (a,b), K. Kowalska (a), M. Hanfl and (b) and A. Katrusiak (a), J. Phys. Chem. Lett. 4, 4032–4037 (2013). In 1951 the serendipitous has been determined independently in (a) Faculty of Chemistry, Adam discovery of ferrocene, FeC10H10, several laboratories, but its molecular Mickiewicz University, Poznan (Poland) bis(cyclopentadienyl)iron(II), the fi rst conformation and interactions (b) ESRF and one of the most stable metallocenes, remained unclear. Under normal ignited the ‘explosion’ of organometallic conditions, ferrocene polymorph I chemistry [1,2]. Since then, thousands is monoclinic with staggered metallocenes and their derivatives cyclopentadienyl rings, i.e. their have been synthesised and used as torsion angles t equal to 36º + n·72º catalysts in chemical laboratories (t is defi ned as torsion C-cCp···cCp’-C’, and in technological processes [3,4]. where unprimed and primed symbols The proposed sandwich-like structure refer to two Cp-rings of one molecule, of the ferrocene molecule, consisting cCp and cCp’ are the rings centroids, of two parallel cyclopentadienyl (Cp) and n is an integer). The staggered rings with Fe2+ ion conformation is energetically sandwiched in-between, unfavoured and high-resolution was soon confi rmed by neutron-diff raction measurements spectroscopic and X-ray revealed structural models with Cp- diff raction studies. The rings disordered in two or three sites structure of ferrocene (Figure 83). Below 164 K, ferrocene I transforms into the triclinic form II, where the Cp-rings of two independent molecules are each distorted by about 9º off the energetically favoured eclipsed conformation (t = 0º, Figure 83). Dunitz et al. [5] crystallised Fig. 83: Schematic ferrocene below 98 K and obtained its illustration of the orthorhombic polymorph III with the conformers of ferrocene Cp-rings eclipsed at t equal to 0º. These molecule and their potential results suggested that the unfavoured energy function (E ). The p staggered conformation in ferrocene I torsion angle t describes the mutual orientation of is promoted by intermolecular the cyclopentadienyl rings interactions. However, under normal (Cp), as defi ned in the text. conditions, these crystal cohesion The dotted green and blue forces are too weak to completely lines in the drawing of the stabilise the molecular rings at the molecule in phase I indicate the positions of disordered unfavourable staggered conformation. Cp-rings. The intermolecular interactions of

72 ESRF DYNamICS aND EXTREmE CONDITIONS

Fig. 84: a) Unit-cell dimensions of ferrocene phases I and I’ compressed up to 11.6 GPa. The vertical dashed line marks the transition between phases I and I’ at 3.24 GPa. The fi lled circles represent laboratory data, crossed circles and empty circles indicate the compression and decompression runs measured at beamline ID09A, respectively; b) structural formula of the molecule in the staggered conformation, used in most chemistry textbooks; c) space- fi lling model of the ferrocene molecule in phase I; and d) molecules of ferrocene located at the inversion centres in the monoclinic polymorph I’. ferrocene I have been enhanced by high pressure in our single-crystal X-ray diff raction experiment performed at beamline ID09A. We found that ferrocene I undergoes an isostructural phase transition at 3.24 GPa (Figure 84), preserving the crystal unit cell and symmetry of space group energetically-favoured conformation.

P21/n. Therefore, this new high-pressure We were also able to explain the U-turn phase of ferrocene has been denoted as shaped compression of the unit-cell phase I’. In its structure, the molecules parameter b, by the compensation of are ‘pushed’ by intermolecular forces the shortest intermolecular distances into the staggered conformation and in the compressed structure. However, the disorder of Cp-rings is eliminated. several intriguing observations This result is consistent with the could be made about the measured occurrence of the average staggered data. The discontinuities in unit-cell conformers unfavoured by 4 kJmol-1 in parameters occur when angle β, obtuse ferrocene I. In contrast, this pressure- in ferrocene I, passes through 90º and induced phase transition is in many it becomes acute in phase I’. The β = 90º respects diff erent to most other order- coincides with the point of structural disorder phase transitions. Usually in transformations when the fi rst and high-temperature phases molecules second shortest intermolecular or ions are disordered, in two or more contacts between the centres of equivalent or nearly equivalent sites. molecules in ferrocene I become equal On lowering temperature, the disorder and then reversed in phase I’. In these References in the structure is eliminated when Cp- respects, the structure of ferrocene [1] T.J. Kealy and P.L. Pauson, Nature 168, 1039−1040 (1951). rings reside in only one of the sites. phases I and I’ resembles a mechanical [2] P.L. Pauson, J. Organomet. Chem. Thus, below 164 K the disordered box with meshing gears regulated by 637−639, 3−6 (2001). ferrocene I transforms into ferrocene II their positions and distances. Our study [3] H. Werner, Angew. Chem. Int. Ed. 51, with two independant molecules shows that a lot of information is still 6052−6058 (2012). ordered. We were able to show that high missing for a complete understanding [4] G.B. Kauff man, J. Chem. Educ. 60, 185−186 (1983). pressure has a very diff erent eff ect on the of the basic properties of prototypical [5] P. Seiler and J.D. Dunitz, Acta crystal structure, and that in ferrocene I’ compounds at their molecular level, Crystallogr. Sect B 38, 1741−1745 it does not favour any of the disordered and that high pressure studies provide a (1982). sites, but the average and the least new perspective for their investigation.

(n ) nE : A HIgH PrESSUrE VAn DEr WAALS InSErTIon 2 6 7 Principal publication and authors coMPoUnD T. Plisson (a), G. Weck (a) and P. Loubeyre (a), Phys. Rev. Lett. 113, 025702 (2014). Through increased pressure, two van der Waals compounds. The (a) CEA, DAM, DIF, Arpajon (France) diff erent atomic or molecular species eff ect of pressure is to modify the that interact via weak van der Waals microscopic interactions and stabilise forces can auto-organise to form new structures. Remarkably, none of stoichiometric compounds, called these van der Waals compounds have

The European Synchrotron 73 SCIENTIFIC HIGHLIGHTS DYNamICS aND EXTREmE CONDITIONS

been observed at low temperatures in of a clathrate-like insertion compound. cryocrystals, which makes pressure a It is the fi rst van der Waals clathrate powerful driving force for self-assembly ever observed and it could open a route of molecular compounds. for the development of a new family of host-guest systems in high pressure Most of the compounds that have molecular systems. been observed so far under pressure,

such as NeHe2, Ar(H2)2, CH4(H2)2, To unveil the existence of this Xe(O2)2 crystallise in the form of Laves stoichiometric compound, we phases, i.e. analogous to intermetallic performed a detailed measurement

compounds and colloidal systems, the of the N2-Ne binary phase diagram, stability of which is explained in terms at ambient temperature, by studying of binary crystals of hard sphere like 16 diff erent concentrations, as shown particles. However, the van der Waals in Figure 85. The fl uid mixtures were forces can also be anisotropic which, loaded in a diamond anvil cell. A combined with packing constraints, ruby gauge was used to measure the may lead to more exotic structures. For pressure of the sample. The diff erent

example, in pure N2, the quadrupole- phases were identifi ed by Raman quadrupole interaction between the spectroscopy of the vibration modes of

N2 molecules is responsible for a rich the N2 molecule. polymorphism under pressure. The sample was compressed to several In this study, we have gigaPascals, and the solidifi cation discovered a new type of was observed visually through the van der Waals compound diamond anvils, using a microscope.

in the N2-Ne system at high The pressure at which the solid crystal pressure. This compound disappears in the liquid is the melting has the stoichiometry pressure for a given concentration (red

(N2)6Ne7, with the structure dots in Figure 85). The particular form of the binary phase diagram reveals the presence of a stoichiometric compound Fig. 85: Binary phase diagram at 54 mol% concentration of neon, of the N2-Ne mixture at 296 K, plotted as pressure vs Ne corresponding to the formula (N2)6Ne7. concentration. S1 and S2 are At this pressure, congruent melting was the pure solids and F is the observed and a solid with hexagonal fl uid phase. Inset: Photograph structure was identifi ed, as shown in of a single crystal of (N ) Ne 2 6 7 the inset of Figure 85. surrounded by a solid-solid phase separation, in the diamond anvil cell. The structure of this compound was determined by single-crystal angular dispersive X-ray diff raction at beamline

Fig. 86: Crystal structure of the (N2)6Ne7 compound. Neon and nitrogen are represented by black and blue spheres respectively. The red planes show the surfaces supported by the

centres of the N2 molecules.

74 ESRF DYNAMICS AND EXTREME CONDITIONS

ID09A. The synthesis of a high quality Such an exotic structure could also single crystal, and the use of a high be used to stabilise new compounds brilliance, monochromatic synchrotron synthesised at high pressure. For beam optimised for diffraction at example, pure nitrogen was shown extreme conditions using diamond anvil to polymerise at high pressure [1]. cells was essential to the determination This polymeric single-bonded form of the structure of (N2)6Ne7. The is known to be a dense high-energy diffraction signal was composed of material that could be used as fuel, 225 reflections, which were indexed explosive or propellant. But it has to find that the structure is hexagonal not yet been recovered at ambient with symmetry R-3m. The positions of pressure. Constraining the geometry of the atoms were determined by direct the nitrogen molecules in such van der methods and refined. The structure is Waals compounds could change the shown in Figure 86 The N2 molecules conditions of polymerisation and the surround the Ne atoms, forming cages recovery of the polymeric compound with pentagonal faces, similar to the at lower pressure. It could also serve structures of clathrates. The fact that as a pathway for the observation of van der Waals interactions could other allotropic structures predicted References drive the self-organisation of such a theoretically, among which is an [1] M.I. Eremets, A.G. Gavriliuk, I.A.Trojan, D.A. Dzivenko and R. Boehler, complex structure was unexpected. intriguing diamondoid structure formed Nat. Mater. 3, 558 (2004). The prediction of this structure could by N10 cage units [2]. The compression [2] X. Wang et al, Phys. Rev. Lett. 109, be a challenging test for first principle of this compound to the megabar range 175502 (2012). calculations. is underway.

E arTH AND PLANETARY SCIENCE Principal publication and authors D. Andrault (a), G. Pesce (a), M.A. Bouhifd (a), N. Bolfan-Casanova (a), J.-M. Hénot (a) Melting of subducted basalt at the core-mantle and M. Mezouar (b), Science 344, 892- boundary 895 (2014). (a) Laboratoire Magmas et Volcans, Seismological studies have In search of an alternative, researchers Université Blaise Pascal, CNRS, IRD, demonstrated the complexity of the explored the melting properties of mid- Clermont-Ferrand (France) (b) ESRF lowest 150-300 km-thick mantle oceanic ridge basalt (MORB), which can layer, called the D” layer, situated just reach the lowermost mantle following above the core-mantle boundary (CMB) subduction of oceanic crust. In the [1] (see Figure 87). The anomalies lower mantle, the MORB presents four detected could arise from various coexisting phases: free silica, the Mg types of chemical or mineralogical and Ca bearing silicate perovskites heterogeneities including phase and a cage-type structure like calcium- transitions in mantle minerals, ferrite. This mineralogy is radically chemical reactions with the outer core, different from the bulk mantle and, presence of an ancient (crystallised) therefore, the melting properties should basal magma ocean and/or the deep also be different. Due to the number descent of subducted slabs. Partial of chemical elements, melting occurs melting of the bulk mantle is often progressively from the solidus, where advocated to explain the seismic the liquid contains a high concentration features, but it cannot explain the non- of incompatible elements, to the ubiquitous character of the ultra-low liquidus. The difference between solidus velocity zones (ULVZ). Partial melting and liquidus can be of several 100 K at would require a very hot core, higher the CMB pressure of 135 GPa. than 4150 K [2,3], which is more than 1000 K above the mantle adiabat The difficulty of the experiments was extrapolated to the CMB. Moreover, the therefore to determine the onset of hypothesis of a very hot core would be melting using a laser heated diamond at odds with the current explanation anvil cell at pressures of the CMB. for the existence of a geodynamo over The sample behaviour was followed prolonged geological periods. continuously using in situ X-ray

The European Synchrotron 75 SCIENTIFIC HIGHLIGHTS DYNamICS aND EXTREmE CONDITIONS

diff raction at beamline ID27. Upon pressure of ~80 GPa (corresponding heating, we observed a sequence of to a mantle depth of ~1900 km), events: crystallisation of the high- the solidus temperature of ~3200 K pressure phases; continuous grain is similar to that of the bulk mantle growth; and then the sudden apparition (Figure 88). With pressure increased to of much larger spots on the CCD 135 GPa, the MORB solidus increases detector. These spots moved quickly to ~3800 K, whereas that of the bulk and eventually disappeared, and mantle increases to ~4150 K. Thus, the randomly reappeared, as a function of onset of melting in the MORB occurs at time. The fast grain rotation indicates ~350 K below the mantle solidus. partial melting of the sample. At a The fl attening of the MORB solidus could be linked to a change in liquid composition with pressure. There are three indications of an increase of the

SiO2 content in the liquid with pressure during partial melting of MORB. First, the MORB solidus curve always plots 1000-1500 K below the lowest melting

curve for pure CaSiO3, MgSiO3 or SiO2 phases (Figure 88). Although SiO is the Fig. 87: 2 Location of most refractory phase at low pressures, the D” layer, it becomes the least refractory above between the 60 GPa. Second, we observe the total Earth’s core disappearance of the SiO2 peaks above and lower the solidus for pressures higher than mantle. 70 GPa, which indicates samples under-

saturated in SiO2 compared to the liquid composition. Third, the mineralogical Fig. 88: The MORB solidus and chemical analyses performed on and liquidus compared with the recovered samples (Figure 89) melting curves of the bulk show higher SiO content and presence (chondritic) mantle. At the 2 core-mantle boundary, the of the high-pressure polymorph of SiO2 MORB solidus reaches 3800 (seifertite), while the surrounded rim of (±150) K. Its curvature is much material is composed of bridgmanite, more pronounced than that of a which is the MgO-bearing phase of the chondritic-type mantle [ 2 ] , which MORB lithology. induces a lower melting point for the MORB at 135 GPa, the core-mantle boundary pressure. Following the hypothesis that the ULVZ The MORB solidus also plots of the lowermost mantle are associated 1000-1500 K below the melting with partial melting, our experimental curves of pure SiO and MgSiO , 2 3 results provide explanations for the as determined from shock wave experiments. seismological features. The SiO2-rich liquid that is generated by MORB partial melting could either remain trapped in the subducted slab, producing the seismic anomalies, or solidify aft er reaction with the surrounding Mg- bearing mantle, which would result in a remixing of the subducted basalt in the lowermost mantle.

Fig. 89: Phase relations in the sample References recovered aft er partial melting at 120 GPa. [1] T. Lay, Q. Williams and E.J. Garnero, Mineralogical content (lower frames for Nature 392, 461-468 (1998). Bridgmanite (Pv, left ) and SiO2-seifertite [2] D. Andrault, et al., Earth and (right)) and Si chemical map (upper frame) Planetary Science Letters 304, 251-259 were measured on the recovered samples (2011). by X-ray diff raction and scanning electron [3] G. Fiquet, et al., Science 329, 1516- microscopy. They show higher SiO2 content 1518 (2010). in the sample region corresponding to the quenched liquid.

76 ESRF DYNamICS aND EXTREmE CONDITIONS

BUoyAnT MAgMAS TrIggEr SUPErVoLcAno ErUPTIonS Principal publication and authors W.J. Malfait (a), R. Seifert (a), S. Petitgirard (b), J.-P. Perrillat (c), M. Mezouar (b), T. Ota (d), Supervolcanoes are considered a SiO2-rich magmatic liquids that fed E. Nakamura (d), P. Lerch (e) and rare but serious threat due to their supereruptions. C. Sanchez-Valle (a), Nature Geosciences destructive capacity only comparable 7, 122-125 (2014). (a) Institute for Geochemistry and to giant impacts [1]. Well-known We conducted experiments at beamline Petrology, ETH Zurich (Switzerland) supervolcanoes are located in the ID27 to determine the density of dry (b) ESRF Yellowstone Caldera (USA), Lake and hydrous rhyolitic/granitic liquids (c) Laboratoire de Géologie de Lyon, Université Claude Bernard Lyon 1 and Toba (Indonesia) and Lake Taupo (78 wt% SiO2) using a Paris-Edinburgh (New Zealand), where the last press and a synchrotron X-ray Ecole Normal Supérieure de Lyon, CNRS UMR 5276, Villeurbanne (France) supereruption occurred 26500 years absorption technique. Finely-powdered (d) Institute for Study of the Earth’s ago. Supervolcano eruptions are synthetic rhyolitic glasses with various Interior, Okayama University (Japan) explosive and eject more than 450 H2O contents (0, 4.5 and 7.7 wt% H2O) (e) Paul Scherrer Institute, Villigen km3 (or 1015 kg) of magma (e.g. La were loaded into natural diamond (Switzerland) Garita, Yellowstone) compared to the cylinders and subjected to pressures (P) mere 10 km3 supplied by the largest and temperatures (T) up to 3.6 GPa and conventional volcanic eruptions (e.g., 1950 K. Aft er compression, the samples Pinatubo in 1991) [2]. Moreover, were heated until complete melting supereruptions have a major impact on was confi rmed by X-ray diff raction. global climate as the large volumes of The liquid density was determined in emitted ashes and volcanic gasses (e.g. situ from the X-ray absorption contrast

SO2) raise the Earth’s albedo and could between the sample and the diamond decrease the global temperature by up cylinder used as container (Figure 90). to 10°C for over a decade. The experimental densities were implemented in a model for the density In spite of their devastating impact, of dry and hydrous granitic liquids the triggering mechanisms for at crustal and shallow upper mantle supereruptions have remained elusive conditions (< 100 km depth) that allows until now. Processes occurring in the fi rst evaluation of the buoyancy conventional volcanic system are indeed not scalable to the much larger magma chambers beneath supervolcanoes (∼ 5-10 km thick and up to 100 km wide). Because supervolcanoes are located in areas with high heat fl ux, the crust around their magma chambers is relatively hot and ductile. As a result, overpressure due to a change in volume, e.g. by magma recharge, volatile saturation or mush rejuvenation, is dissipated more effi ciently in the plastic supervolcano magma chambers than in the more rigid magma chambers of conventional volcanoes [3]. Supervolcanoes therefore do not erupt very oft en (~1/100000 years on average, [1]). Fig. 90: a) X-ray absorbance scan through the sample assembly (i.e., cylindrical In an alternative mechanism, diamond capsule and sample) as a function magma buoyancy in large magma of the X-ray beam position in the horizontal chambers potentially creates an direction (in mm) at 1.02 GPa - 1735 K (dotted line). The density (ρ ) is derived overpressure suffi cient for an eruption PT from the attenuation coeffi cient of the [3]. Contrary to the other plausible sample (μPT), obtained from the fi t (blue triggers, the buoyancy overpressure line) to the measured absorbance profi le. does not decrease upon wall rock The strong absorption near 45.6 mm deformation and expansion of the indicates the position of Pt-bearing magma chamber. However, the pressure-temperature calibrants touching the outside of the diamond capsule. b) X-ray viability of this mechanism could not absorbance map of the gap between the be evaluated until now due to the lack anvils, reconstituted from horizontal scans of appropriate density data for the collected at diff erent vertical positions.

The European Synchrotron 77 SCIENTIFIC HIGHLIGHTS DYNamICS aND EXTREmE CONDITIONS

References overpressure in supervolcano magma chamber (Figure 91b) were computed [1] S. Self and S. Blake, Elements 4, 41 chambers. with our liquid density model and (2008). literature data for relevant mineral [2] B.G. Mason, D.M. Pyle and C. Oppenheimer, Bull. Volcanol. 66, 735 For each melt composition and magma and volatile phases. Our results show (2004). chamber conditions (i.e. thickness, that for magma chambers thicker [3] A.M. Jellinek and D.J.A. DePaolo, Bull. degree of crystallisation and P-T), than 5 km, if there are no volatiles, the Volcanol. 65, 363-381 (2003). the density profi les (Figure 91a) and buoyancy overpressure in the roof of [4] O. Bachmann and G.W. Bergantz, J. overpressures at the roof of the magma the magma chamber is lower than the Petrol. 45, 1565 (2004). minimal critical overpressure (10 MPa) required to break the cap rock and to initiate the eruption. Therefore, magma buoyancy provides a background overpressure that can progressively increase by the addition of heat and/ or magma during the evolution of a supervolcano and that primes the magma chamber for another (small) trigger or may ultimately lead to a supereruption by itself. The present results thus show that external triggers are not a priori required to start an eruption and that buoyancy alone may trigger a supereruption.

Fig. 91: a) Density profi les of a supervolcano magma chamber consisting of a 2 km thick granitic melt layer on top of 8 km of crystal mush (40 or 60% crystallinity) [4] and density contrast (∆ρ) with the host rock. b) Buoyancy overpressure as a function of magma chamber thickness for a range of plausible chamber conditions; the red area denotes the overpressure for the present-day magma chamber conditions at Yellowstone (3.6-8 km thick, 68-90% crystallinity). In magma chambers thicker than 5 km, the buoyancy derived overpressure may exceed the critical overpressure (10-40 MPa) required for an eruption, even in the absence of volatile saturation.

gIAnT PrESSUrE-InDUcED VoLUME coLLAPSE In THE

PyrITE MInErAL MnS2

Magnetism plays an important role and electronic Coulomb repulsion to in determining the properties of the determine the size of the magnetic minerals that make up Earth’s crust and moment. High pressure can thus mantle [1]. This is due to the ubiquity make a high-spin (e.g., S = 5/2) state of transition metals (notably Fe) with unstable with respect to a low-spin unpaired electrons. Under geological state (e.g. S = 1/2). The primary driving pressures, these may undergo abrupt force for this eff ect is an increase in Δ volume collapses due to spin-state as the metal–ligand bond distances transitions, where the number of decrease. Many simple materials show unpaired electrons is sharply reduced. pressure-induced spin state transitions, This occurs when the crystal fi eld and volume collapses of ΔV ≈ 5% are splitting Δ, competes with Hund’s rule regarded as notable. Understanding

78 ESRF DYNamICS aND EXTREmE CONDITIONS

distortion, with neighbouring Mn Principal publication and authors atoms forming short bonds along one- S.A.J. Kimber (a), A. Salamat (a,b), S.R. Evans (a,c), H.O. Jeschke (d), dimensional chains of edge-sharing K. Muthukumar (d), M. Tomić (d), MnS6 octahedra. Density functional F. Salvat-Pujol (d), R. Valenti (d), theory calculations show that this M.V. Kaisheva (e), I. Zizak (f) and structure is low spin S = 1/2, and that T. Chatterji (g), Proc. Nat. Acad. Sci., the remaining spin per site is occupied 111, 5106 (2014). (a) ESRF in chemical bonding. Magnetism is thus (b) Lyman Laboratory of Physics, completely quenched in this phase. Harvard University, Cambridge (US A) (c) Departement für Chemie Not only is the size of the volume und Biochemie, Universität Bern (Switzerland) collapse in MnS2 much larger than that (d) Institut für Theoretische Physik, found in other minerals, it therefore Goethe-Universität Frankfurt, Frankfurt also occurs by a new mechanism. am Main (German y) Traditionally, the electronic and (e) School of Chemistry, University of magnetic properties of oxides and Edinburgh (UK) sulphides have been explained using (f) Helmholtz-Zentrum Berlin für Materialien und Energie, Berliner a so-called ‘single-ion’ model [3]. This Elektronenspeicherring-Gesellschaft takes into account the spin, charge and für Synchrotronstrahlung-II, Berlin orbital degeneracies of isolated cations. (Germany) The results presented here go beyond (g) Institut Max von Laue-Paul Langevin, this simple picture, and show that Grenoble (France) molecule-like units can spontaneously emerge under pressure in otherwise ordinary continuous lattice materials.

In MnS2, the unpaired electrons on neighbouring cations form chemically Fig. 92: a) Ambient pressure cubic structure bonded dimers, and it is this ‘squeezing of MnS2, showing the fcc lattice of Mn sites (blue) and highlighting the valence out’ of magnetism which stabilises the eg orbitals of one cation. b) Pressure giant increase in density of the new dependence of diff raction data (ID09A) phase. Although its scarcity means through the volume collapse, showing the that MnS is geologically insignifi cant, formation of a disordered intermediate. 2 similar phenomena might be found in the mechanisms of such magnetically more abundant minerals containing 3+ driven transitions, and predicting the iso-electronic cations such as Fe . high-pressure structures, is ultimately of great importance for modelling the Earth›s mantle in particular.

Using diamond anvil cells at ID09A and ID27, we discovered that a simple cubic manganese mineral (Hauerite,

MnS2) undergoes a truly giant 22% shrinkage of unit cell volume above 12 GPa. Instead of the expected transition between crystalline high and low-spin phases, compression induces the formation of a disordered intermediate (Figure 92). As we were unable to solve the high pressure structure from this data, we made use of recently developed ab initio structure searching methods implemented in the Fig. 93: a) High-pressure structure of USPEX package [2], which predicted MnS2 as predicted using USPEX and as a new monoclinic polymorph. Using refi ned against the ID27 diff raction data References the recently commissioned online CO2 shown in b) at 20 GPa aft er laser heating. [1] R.E. Cohen, I.I. Mazin, Science, 275, laser heating at ID27, we were able to c) Perspective view of the high pressure 654 (1997). recrystallise the sample at 20 GPa. structure showing the formation of short [2) C.W. Glass, A.R. Oganov and dimer bonds between neighbouring N. Hansen, Comput. Phys. Commun, Indexing and Rietveld refi nement Mn cations in the edge-sharing chains. 175, 713 (2006). confi rmed the predicted structure This viewpoint is perpendicular to that [3] J.B. Goodenough, Phys. Rev., 100, (Figure 93). This contains a notable shown in a). 564 (1955).

The European Synchrotron 79 SCIENTIFIC HIGHLIGHTS Highlights 2014

S trUCTURAL Biology

With two new end-stations also provide access to in situ (i.e. in in the Protein Data Bank remains constructed and commissioned as crystallisation plate) screening and very high and covers a broad range part of the UPBL10 Upgrade project data collection. of structural biology research. The entering user operation, 2014 has highlights presented here illustrate seen both a major evolution in While the UPBL10 project is nearing that membrane proteins, either facilities available for Structural its end, the long-term evolution of transporters or receptors, remain Biology at the ESRF and the coming the facilities available for structural major targets of the research carried to fruition of much hard work, biology will continue. Indeed, the out at the Structural Biology Group’s spread over a number of years1. recently published ‘Orange Book’ beamlines as do studies aimed at MASSIF-1 [ID30A-1] and MASSIF-3 contains many ideas for future understanding the molecular basis [ID30A-3] provide new options for facilities and experiments that will of disease. Concerning the latter, experiments in structural biology at be enabled by the ultra brilliant X-ray a particular highlight in 2014 was ESRF. beams that will be produced following the elucidation, by Cusack and the Phase II Upgrade of ESRF. colleagues, of the crystal structures Based around two ESRF-developed of the polymerases of bat-specific devices, the RoboDiff, which acts Our beamlines will also evolve in the influenza A and human influenza B both as sample changer and medium term. Even as this piece is in complex with single stranded viral goniometer and a high capacity being written the mirrors making RNA promoter. Comparison of the dewar (HCD) that is capable of up the focussing optics of ID23-2 two structures provides an atomic- holding 240 SPINE standard sample are being replaced. This should give level model of the mechanism of holders, MASSIF-1 supplies for a a new lease of life to this pioneering action of flu polymerases and may new paradigm in structural biology: microfocus end-station prior to the provide new insight into the design completely automatic, hands-off completion of a technical design of anti-influenza drugs. The articles screening and/or data collection. report (TDR) aimed at producing, in here also emphasise the increasing Following a successful pilot study, 2016, an upgraded ID23-2 with a use of techniques (small-angle X-ray such as service is now available to all beamsize 1 µm, or less, in diameter scattering, atomic force microscopy, our users and is proving extremely at the sample position. Additionally, in cristallo UV-visible or Raman popular. MASSIF-3 is a fixed-energy ideas for improving the functionality spectroscopy) complementary to end-station providing a highly and X-ray beam characteristics on the X-ray crystallography, a trend that is intense (~2 x 1013 photons / second) two MAD beamlines ID23-1 and ID29 sure to grow in the coming years. microfocus (~15 µm diameter) X-ray are actively being considered. The beam at the sample position. This provision at the ESRF of techniques Finally, following a year’s sabbatical end-station is therefore ideal for complementary to macromolecular leave Sean McSweeney and Elspeth the collection of diffraction data crystallography continues to Gordon officially resigned their from microcrystals of biological evolve: upgrade of the sample positions at the ESRF in September macromolecules and, once the Eiger environment at the ID29-S Cryobench 2014. As Head of the Structural 4M detector associated with the microspectroscopy facility is planned Biology Group from 2000 to 2013, end-station is operating at its full while the BioSAXS beamline BM29 Sean’s contribution to the success potential (750 Hz frame rate), will be will soon benefit from the installation of the ESRF as an experimental an excellent vehicle for experiments of a new, more user-friendly on-line resource for structural biology exploiting the burgeoning technique HPLC set-up. This, coupled with a was extraordinary as was that of of synchrotron serial crystallography multi-technique sample environment, Elspeth in making use of synchrotron (SSX). will greatly improve the information radiation an indispensible technique available during BioSAXS experiments. in the development of new The last element of the UPBL10 project pharmaceuticals. I’m sure that all will be put into place very shortly A strength of the Structural Biology readers of ESRF Highlights 2014 with the coming on stream of ID30B. Group has been its ability to continue will want to join us in wishing Sean This fully-tuneable end-station will to operate existing experimental and Elspeth every success in their offer both a high photon flux (~1013 facilities while at the same time new ventures and in overcoming the photons/sec) and a variable spot constructing and commissioning challenges that lie ahead. size (20 to 200 µm2) at the sample new ones. The past year is no position. In its final configuration, exception to this rule. Indeed, the G. Leonard foreseen for mid-2015, ID30B will number of ESRF-based depositions and C. Mueller Dieckmann

1 Although space does not allow acknowledgement of all those involved in the UPBL10 project, we would like to give a special mention here to technical staff of the Structural Biology Group (H. Caserotto, F. Dobias, T. Giraud, N. Guichard, M. Lentini, J. Soudarin, J. Surr), the staff of the EMBL Grenoble Outstation Synchrotron Instrumentation Group, the ESRF Beamline Control Unit (M. Guijarro, A. Beteva), the ESRF Data Analysis Unit (O. Svensson) and the staff of ESRF ISDD, particularly P. Theveneau and W. Schmid.

80 ESRF STRUCTURaL BIOLOGY

Principal publications and authors STRUCTURAL INSIGHTS INTO VIRAL RNA A. Pfl ug, D. Guilligay, S. Reich and S. Cusack, Nature 516, 355–360 (2014); and S. Reich et al., Nature 516, 361–366 SYNTHESIS BY INFLUENZA POLYMERASE (2014). European Molecular Biology Laboratory, Grenoble Outstation (France) Infl uenza virus causes a highly snatching mechanism whereby the contagious respiratory illness (the ‘fl u’) 5’ cap-structure is cleaved from host to which everybody on the planet is cell mRNAs by an endonuclease and susceptible. Viruses that infect humans stitched onto the beginning of the viral can have three diff erent origins. transcript. Replication produces full- Seasonal fl u causes annual epidemics length copies of the genomic vRNA and and has a high public health and proceeds via a complimentary cRNA economic cost. The estimated global intermediate. For many years we have infection rate is 5%–10% in adults and been trying to obtain a high resolution 20%–30% in children and about 250 crystal structure of the infl uenza 000 to 500 000 seasonal fl u-related polymerase in order to understand the deaths occur each year worldwide, distinct mechanisms of RNA synthesis mainly amongst the very young, the and to enable structure-based drug elderly or the chronically ill. Bird fl u design targeting this essential viral is caused by highly pathogenic avian replication machine. strains such as H5N1 and H7N9 (water birds are the natural reservoir for the To overcome the problem of production virus), which through point mutations, of suffi cient amounts of pure and active are able to infect humans with up recombinant polymerase, we used a to 60% mortality, but as yet do not novel baculovirus vector, developed transmit between humans. Pandemic by I. Berger (EMBL), in which the three fl u occurs when novel human strains polymerase subunits are expressed as a are generated by recombination fusion protein and post-translationally between avian and mammalian cleaved by a protease into separate strains, oft en via an intermediate pieces. Bat-specifi c infl uenza A host such as the pig. This occurred in (FluA) and human infl uenza B (FluB) 1918, 1957, 1968 and, most recently, polymerases were then co-crystallised in 2009. Pandemics are unpredictable with the vRNA promoter and the and potentially devastating. While structures solved at 2.7 Å resolution vaccines can provide good protection using data collected at beamline against infl uenza, there is a need for ID23-1. These are the fi rst structures eff ective anti-infl uenza drugs to treat of a complete polymerase from any severe cases and in the event of a negative strand RNA virus. dangerous pandemic where no vaccine is available. The three polymerase subunits, PA, PB1 and PB2 mutually stabilise each The infl uenza virus genome comprises other through extensive and intricate eight segments of negative sense, interfaces (Figure 94). PB1 has a single stranded viral RNA (vRNA). Each canonical polymerase fold, with a segment is packaged in an individual large, enclosed catalytic and RNA ribonucleoprotein particle (RNP). binding cavity connected to solvent RNPs contain one heterotrimeric RNA- by channels which allow nucleotide dependent RNA polymerase, which is bound to the conserved 3’ and 5’ extremities of the vRNA (the ‘promoter’) and multiple copies of nucleoprotein. The viral polymerase both transcribes and replicates each genome segment. Transcription, which produces viral mRNA, occurs by a unique cap-

Fig. 94: Front and back views of the complete infl uenza polymerase with PA shown in green, PB1 in cyan and PB2 in red. The 3’ and 5’ ends of the viral RNA are shown in yellow and violet respectively.

The European Synchrotron 81 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

triphosphate (NTP) entrance, template and replication. In the FluA structure, entrance and product exit. The an ordered priming loop suggests that domains involved in cap-snatching, the infl uenza might initiate unprimed PA endonuclease domain and the PB2 template replication by a similar cap-binding domain, form protrusions mechanism to that seen for hepatitis which face each other across a solvent C and Dengue virus polymerases. channel. Ten nucleotides at the extreme Interestingly, the FluA and FluB 5’ end of the promoter form a compact structures diff er in the orientation stem-loop structure (‘hook’) which is of the PB2 cap-binding domain, bound in a pocket formed by PB1 and suggesting a mechanism for cap- PA. The next four 5’ end nucleotides dependent transcription. In the FluA form canonical base-pairs with the structure, the cap-binding site faces distal part of the promoter 3’ end. The the PA endonuclease active site, a 3’ end proximal nucleotides interact confi guration compatible with cap- with all three subunits and are directed snatching. In the FluB structure, the towards the template entrance to the cap-binding domain has rotated in polymerase active site. situ by 70° and RNA-like residual extra electron density descends part way Comparing the FluA and FluB from the cap binding site towards the polymerase structures gives PB1 active site. Modelling this as RNA, mechanistic insight into transcription its prolongation is compatible with the primer strand in known structures of other RNA polymerases bound to primer-template duplexes (Figure 95). The tight binding of the 5’ hook of the viral promoter to the polymerase is compatible with the role of this element in allosterically activating polymerase functions and in stalling transcription at the 5’ proximal oligo-U stretch on the vRNA leading to poly(A)-tail generation on the viral mRNA.

These structures not only lay the basis for an atomic-level mechanistic Fig. 95: Schematic diagram showing a model of how understanding of the multiple functions cap-snatching (left ) and of infl uenza polymerase but will also cap-dependent priming of shed light on host-specifi c polymerase transcription (right) works by in variants and open new opportunities situ rotation of the cap-binding for anti-infl uenza drug design. domain (orange).

Principal publication and authors LIGHT-INDUCED OPENING OF A H. Takala (a, b), A. Björling (b), O. Berntsson (b), H. Lehtivuori (a), S. Niebling (b), M. Hoernke (b), PHYTOCHROME I. Kosheleva (c), R. Henning (c), A. Menzel (d), J.A. Ihalainen (a) and S. Westenhoff (b), Nature 509, 245-248 All life on Earth is based on sunlight, The structures of several phytochromes (2014). and diff erent organisms have have been solved in their resting state, (a) University of Jyväskylä, Jyväskylä developed multiple systems to detect but the structural details of their light (Finland) and react to their light environment. activation have remained elusive (b) University of Gothenburg (Sweden) Plants, bacteria, and several micro- [1]. Now, in a study conducted at the (c) University of Chicago (U.S.) (d) Paul Scherrer Institut, Villigen organisms use light-sensitive proteins ESRF, the Swiss Light Source (SLS) (Switzerland) called phytochromes to control cellular and Advanced Photon Source (APS), functions in response to the changes we have assembled structural data in light. For example, a red/far-red on the activation mechanism of the light-absorbing bacterial phytochrome phytochrome sensory module. These from Deinococcus radiodurans controls data provide exciting new insights into pigment production in response to light. how light activates the protein. Our

82 ESRF STRUCTURaL BIOLOGY

Fig. 96: Diff erence signals seen in light-induced SAXS (blue) and WAXS (red) scattering patterns from solutions of the phytochrome studied. The diff erence signals were calculated by subtracting the X-ray solution scattering pattern of the resting state from the solution scattering pattern of the active state and thus indicate structural changes between the states. fi ndings provide critical information for understanding these light-sensing proteins and set the basis for novel applications, e.g., in agriculture.

In the studies described here, we aimed to reveal the structural details of a bacterial phytochrome from D. radiodurans. To attain this, we compared X-ray solution scattering data of the resting and activated forms of the protein. The data involved small X-ray scattering (SAXS) data from the ESrF beamline BM29 and time-resolved Fig. 97: solution X-ray data measured at SLS Crystal and APS in two states (Figure 96). This structures of the was achieved by switching the sample resting and between the two states using fl ashes active state of the of diff erent wavelengths of red light phytochrome fragment. while the data were being collected. The chromophore is shown in red and the These SAXS data indicated large tongue region of the protein is in green. The resting state was obtained by growing the conformational changes in the protein crystals in complete darkness; the active upon activation. state crystals needed sequential red light illumination to appear. Aft er collecting the promising X-ray solution scattering data described the rest of the phytochrome protein, above, we verifi ed the exact structural and fi nally to the next proteins in the changes between the states using signalling cascade. X-ray crystallography. We managed to crystallise the photosensory module in Our results have led to a comprehensive two states: resting (Pr, red-absorbing) picture of the light-induced activation and active (Pfr, far-red absorbing) of the phytochrome. First, red light and collected diff raction data at ESRF causes small conformational changes beamline ID23-1. In both states the in the chromophore (red in Figure 97), phytochrome appeared as a parallel which are amplifi ed and transmitted dimer, as shown in Figure 97. However, to the rest of the protein. The tongue the conformation of the dimers diff ered region of the phytochrome rearranges considerably between the states: and shortens, which forces changes in the resting state adopts a “closed” the relative orientations of the rigid confi guration, the active state an “open” protein domains, leading to “opening” confi guration in which the gap between of the entire dimer structure. In this way the two subunits making up the dimer a tiny change induced by absorption is opened by several nanometres. of a photon is amplifi ed to a dramatic conformational rearrangement of the In the structure of this phytochrome, protein. an evolutionally highly conserved amino acid sequence forms a tongue- This fi nding sets the ground for the like structure (green in Figure 97). development of phytochrome-based This tongue was shown to form a applications and further studies in β-sheet structure in the resting state. the fi eld of phytochrome research Remarkably, the tongue adopted [2]. Our next goal is to decipher how a completely diff erent secondary the structural signal changes the structure in the activated state, taking biochemical activity of the output up a shorter, α-helical conformation. region of the protein, which ignites References [1] A.T. Ulijasz and R.D. Vierstra, Curr The resulting shortening of the tongue signalling cascades within the cell. Opin Plant Biol. 14, 498-506 (2011). by ~2.5 angstroms explain how the Naturally, this requires collecting more [2] A.W. Baker and K.T. Forest, Nature. light-induced changes are relayed to structural data on phytochromes. 509, 174-5 (2014).

The European Synchrotron 83 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

Principal publication and authors J.T. Grotwinkel, K. Wild, B. Segnitz and THE “strong arM” OF SRP68 SHAPES I. Sinning, Science 344, 101-104 (2014). Heidelberg University Biochemistry Center (BZH), Heidelberg (Germany) THE SIGNAL RECOGNITION PARTICLE

Targeted protein transport is a vital to a three-way junction ([2], for review process for all cells. A priori targeting see [3]). While this binding site is quite can occur while the protein is being distant from the Alu domain and the synthesised (co-translational) or signal sequence binding protein SRP54, thereaft er (post-translational) and SRP68 was found to be essential each pathway depends on a dedicated for translation-arrest and protein machinery. In eukaryotic cells, translocation. We have now determined membrane proteins and secreted the X-ray structures of the human and proteins are co-translationally targeted a fungal SRP68-RBD domain both to the endoplasmic reticulum (ER) alone and in the context of the SRP S by the universally conserved signal domain. Data were collected on the recognition particle (SRP) [1]. SRP tuneable beamlines ID23-1 and ID29 recognises an N-terminal hydrophobic and initial structure determination signal sequence within the nascent was performed by multi-wavelength polypeptide chain, stalls the translation anomalous dispersion (MAD) using process, and upon interaction with selenomethionine labelled protein. The the SRP receptor (SR), the ribosome– high brilliance of the beamlines was nascent chain complex (RNC) is docked essential for obtaining superior data onto the translocation channel (the Sec quality from the small and weakly translocon) within the ER membrane. diff racting crystals of the protein-RNA Here, protein biosynthesis resumes and complex. the nascent chain is directly inserted into the ER membrane or translocates SRP68-RBD forms an unusual RNA- into the lumen of the ER. Eukaryotic SRP binding module resembling the purely is a complex ribonucleoprotein particle α-helical tetratricopeptide repeat (TPR) consisting of six proteins assembled fold (Figure 98b). The RBD exposes a on a SRP RNA of 300 nucleotides highly positively charged surface, which (Figure 98a). SRP can be divided into tightly binds to an RNA three-way two domains, the S domain, where junction. While previous footprinting the signal sequence binds, and the Alu experiments accurately mapped the domain, responsible for elongation binding site on the SRP RNA, the arrest. In recent years, the molecular consequences of the interaction were details of most SRP components have unexpected. SRP68-RBD remodels been determined with the beamlines at the SRP RNA on two levels and the ESRF playing a major role. However, this remodelling elegantly explains it was only this year that structural previous structural and functional information on SRP68, one of the data. First, binding to the three-way largest, eukaryote specifi c components, junction induces a kink in S domain became available. SRP RNA. When the new structure is fi tted into the cryo-EM structure of the SRP68 forms a heterodimer with SRP72 entire eukaryotic SRP/RNC complex, and the SRP68 RNA-binding domain the “kinked” S domain is necessary to (RBD) has been previously shown correctly explain the RNA density and within the context of the RNC to bind to establish an important RNA-RNA contact between SRP and the ribosome (Figure 99a). Without SRP68 present to induce this contact, the Alu

Fig. 98: Co-translational protein targeting by the signal recognition particle (SRP). a) Human SRP bound to the RNC. b) Ternary complex of SRP S domain including SRP19, SRP68-RBD, and SRP RNA

84 ESRF STRUCTURaL BIOLOGY

Fig. 99: SRP RNA remodelling by SRP68-RBD. a) SRP68-RBD induces a kink in SRP RNA necessary for contacting the ribosome. b) SRP68-RBD inserts an ARM into the major groove of SRP RNA and opens the 5f-loop. domain could not, as already described conserved unpaired nucleotides. It is 30 years ago, function in translation- these nucleotides that form the contact arrest. Second, one α-helix of SRP68- with ribosomal RNA (Figure 99a). A RBD is inserted into the major groove corresponding RNA-bulge in bacterial References [1] P. Grudnik, G. Bange and I. Sinning, of SRP RNA in the 5f-loop region next SRP RNA is essential for SRP RNA Biol Chem 390, 775-782 (2009). to the three-way junction (Figures 98b mediated GTPase activation of the [2] M. Halic et al., Nature 427, 808-814 and 99b). The major groove in RNA- SRP GTPases present in SRP and its SR (2004). helices is narrow and deep, and it needs [4]. GTPase activation is essential for [3] K. Wild, M. Halic, I. Sinning and a “strong arm” to open it. For this to driving the SRP cycle of co-translational R. Beckmann, Nat Struct Mol Biol 11, 1049-1053 (2004). happen, an arginine-rich motif (ARM) targeting. Whether this mechanism [4] G. Bange and I. Sinning, Nat Struct is inserted into the groove, and the also applies to the eukaryotic SRP Mol Biol 20, 776-780 (2013). 5f-loop is remodelled by bulging-out system remains to be seen.

A PROMISCUOUS INTERMEDIATE Principal publication and authors C. Sayou (a), M. Monniaux (a), M.H. Nanao (b,c), E. Moyroud (a), UNDERLIES THE EVOLUTION OF SINGLE S.F. Brockington (d), E. Thévenon (a), H. Chahtane (a), N. Warthmann (e), COPY TRANSCRIPTION FACTORS M. Melkonian (f), Y. Zhang (g), G.K.-S. Wong (g,h), D. Weigel (e), F. Parcy (a,i) and R. Dumas (a), Science A class of proteins known as one copy of the gene. In this scenario, 343, 645-8 (2014). transcription factors (TFs) is central to the organism will continue to survive (a) Laboratoire de Physiologie Cellulaire et Végétale (LPCV ), Grenoble (France) the evolution of organisms. TFs exert because it has a working copy of (b) European Molecular Biology their control on the cell by binding both the DNA and TF, which frees the Laboratory (EMBL), Grenoble (France) specifi cally to DNA sequences and second copy of the TF to evolve new (c) Unit of Virus Host-Cell Interactions, causing a cascade of downstream binding specifi cities. However, in cases Université Grenoble Alpes–EMBL–CNRS, events to occur. A key question in where only a single copy of the gene UMI 3265, Grenoble (France) (d) Department of Plant Sciences, the evolution of TFs is how they exists in the cell, it is less obvious how University of Cambridge (UK) can evolve new DNA specifi cities. changes can occur without deleterious (e) Department of Molecular Biology, This is complicated by the fact that eff ects. In this work, LEAFY, a generally Max Planck Institute for Developmental altering the target DNA sequence or single copy master regulator of fl oral Biology, Tübingen (Germany) the specifi city of the transcription development, was used as a model (f) Botanisches Institut, Lehrstuhl I, Universität zu Köln, Biozentrum Köln factor will in most cases change the system. A key discovery which enabled (Germany) binding properties of the TF towards this research was the identifi cation of (g) Beijing Genomics Institute (BGI)– its original target DNA sequence, ancient forms of LEAFY from hornworts Shenzhen (China) which will in turn prevent the normal and algae. The DNA binding preferences (h) Department of Biological Sciences, downstream events from occurring. of these proteins and other previously Department of Medicine, University of Alberta (Canada) A simple solution to this problem is described forms from angiosperms and (i) Centre for Molecular Medicine the duplication of genes, followed by gymnosperms were determined using and Therapeutics, Child and Family the evolution of new properties (most a technique called SELEX (Systematic Research Institute, University of British notably DNA binding specifi city) of Evolution of Ligands by EXponential Columbia (Canada)

The European Synchrotron 85 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

enrichment). From the data obtained, from type I/II in that a central region in in combination with classical the recognised DNA is absent. Using biochemical binding assays, a “family SELEX, biochemical binding assays and tree” of DNA binding specifi city could the type I and type II LEAFY structures be assembled. The family tree showed as a guide, it was shown that a region in that there are at least three diff erent the LEAFY protein of type I/II allows for types of DNA recognition sites: types binding to sequences with this central I, II and III. region (Figure 101). When this region is mutated (as in the hornwort and algal Using the microfocus macromolecular LEAFYs which bind type III sequences), crystallography beamline ID23-2 [1] LEAFY binds to type III sequences. and a special data collection method With a molecular understanding to reduce radiation damage, the of how, through evolution, LEAFY crystal structure of one of these family can bind to diff erent sequences, the members (from moss) was determined, original question of how diff erent DNA in complex with its preferred (type specifi cities in a single copy TF could II) DNA sequence. When compared be better addressed. Upon further to an earlier structure of LEAFY in analysis, the Hornwort LEAFY gene was complex with a type I sequence [2], this shown to have structural characteristics structure revealed that an additional of LEAFYs that should allow it to bind contact is made between the protein both type I/II and type III sequences. and DNA, and provided a molecular Biochemical assays were used to explanation for the altered DNA confi rm this hypothesis experimentally specifi city compared to other LEAFY and showed that this LEAFY is indeed forms (Figure 100). promiscuous in its binding specifi city. This seemingly simple fi nding has Equally important was the elucidation profound consequences for our of the molecular determinants of understanding of how TFs could evolve primitive LEAFY binding to the type III diff erent DNA specifi cities. In particular, motifs. Type III motifs diff er dramatically such an intermediate state could have acted as an ancient evolutionary crossroad through which LEAFY Fig. 100: Structural developed into specialised “modern” basis for type forms. The reasons for the diversity of II specifi city modern forms of LEAFY remain elusive, (PpLFY1-DBD), in comparison with and are the subject of ongoing research. type I (AtLFY- DBD). a) Overall structure of the moss LEAFY dimer bound to a type II DNA sequence. b) Superimposition of the structures of angiosperm (AtLFY-DBD, in grey) with its cognate DNA (orange) and moss (moss-DBD, in pink) with its cognate DNA (green).

Fig. 101: a) Moss (PpLFY1)-LEAFY (in red and pink) bound to DNA (in green, except the black 3-bp spacer). Interactions between monomers (involving α helices α1 and α7) are References shown with dashed lines. [1] D. Flot et al., J Synchrotron Radiat. b) Modelled type III binding with 17, 107-18 (2010). DNA shown in blue. The dashed [2] C. Hamès et al., EMBO J. 27, 2628-37 vertical line denotes the centre of the (2008). pseudopalindromic DNA sequence.

86 ESRF STRUCTURaL BIOLOGY

Principal publication and authors A TRANSPORTER CAUGHT IN THE ACT L. Malinauskaite (a), M. Quick (b,c), L. Reinhard (a,d), J.A. Lyons (a), H. Yano (b,e), J.A. Javitch (b,c), P. Nissen (a), Nat Struct Mol Biol. 21, Secondary transporters are membrane and the end of the forward transport 1006-12 (2014). proteins that use the energy stored in cycle. However, the mechanism of (a) Aarhus University (Denmark) electrochemical gradients (typically H+ intracellular sodium release - where (b) Columbia University College of or Na+) to drive antiport or symport transport becomes coupled to the Na+ Physicians and Surgeons, New York of specifi c solutes. Their function is gradient - has remained elusive. (USA) (c) New York State Psychiatric Institute, generally based on an “alternating New York (USA) access” mechanism [ 1 ] where the We studied another bacterial model of (d) Current address: Karolinska Institutet protein switches the orientation the NSS family, the multihydrophobic Stockholm (Sweden) and Karolinska between outward- and inward-facing amino acid transporter MhsT from Institutet, Hamburg (Germany) states. This mechanism allows a net Bacillus halodurans. Two crystal forms (e) Current address: National Institutes of Health, Baltimore (USA) transport and accumulation of solute of this protein with bound L-tryptophan determined by the driving gradient. and sodium ions (diff raction data Over the last 10 years, a wealth of collected at ESrF beamline ID23-2 and structures of secondary transporters at Diamond Light Source) revealed the has emerged. However, information as long-sought occluded, inward-facing to how transporters actually switch state where sodium and substrate are sides and release the driving ion is, in primed for release to the cytoplasmic fact, scarce. environment and highlights both the fi rst steps of solvation of the driving Classical examples of secondary sodium ion and the structural changes transporters include the that have switched the transporter neurotransmitter-sodium symporters from an occluded outward- to an (NSSs) that perform an active re- occluded inward-facing state. A unique uptake of molecules such as serotonin, feature was observed: an unwinding dopamine and γ-aminobutyric acid at the cytoplasmic end of the fi ft h (GABA). The signalling between transmembrane helix (TM5) allows neurons is achieved by the release of solvent from the cytoplasm to enter neurotransmitters into the synaptic the driving-ion binding site, which also cleft from the presynaptic neuron changes from a trigonal pyramidal that activates receptors and signal to an octahedral coordination with a transmission in the postsynaptic cytoplasmic water molecule as the neuron. Based on the steep Na+ gradient sixth ligand (Figure 102). maintained by the Na+, K+-ATPase, the NSS family transporters perform an active re-uptake to terminate the signal and recycle the neurotransmitters to the presynaptic neuron. It is of particular interest to study the structure-function relationships of this family as they are implicated in various psychological and neurological disorders such as epilepsy, depression and schizophrenia. NSSs are also the key targets for pharmaceuticals including antipsychotics and antidepressants, and for psychostimulants such as cocaine and amphetamine.

Several crystal structures have been obtained of a bacterial member of the Fig. 102: The MhsT inward-facing occluded NSS family - the amino acid transporter structure. The electrostatic surface is shown LeuT from Aquifex aeolicus. These have for MhsT. The close-up with a transparent captured the outward-open, sodium- surface shows a solvent-fi lled, cytoplasmic cavity that reaches the Na2 binding site. bound state [ 2 ] , the occluded outward- Associated detergent molecules are shown facing, substrate-bound state [3 ] , and as ball-and-sticks, Na+ as green spheres the inward-open apo state [2 ] and and water molecules as small red spheres provide snapshots of the beginning (insert).

The European Synchrotron 87 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

References The forward transport reaction can membrane potential. Aft er the release [1] O. Jardetzky, Nature 211, 969-70 now be described for the NSS family of sodium the intracellular cavity opens (1966). (Figure 103): sodium binding at two for complete substrate release to occur [2] H. Krishnamurthy and E. Gouaux, Nature 481, 469-74 (2012). sites stabilises the outward-open state [2 ] . Finally, the transporter switches [3] A. Yamashita et al., Nature 437, 215- [2 ] , and allows substrate binding, back to an outward-open state 23 (2005). which stimulates the transporter through sodium-free return states, the to occlude the substrate binding structures of which are as yet unknown. sites [3 ] . Subsequent closure of an extracellular, hydrophobic vestibule This new structure in the NSS cycle – triggers unwinding, at the intracellular caught in the act of exposing sodium to side, of TM5. This provides access of the intracellular environment - provides cytoplasmic water to the sodium site. novel insight into NSS function and, This initiates sodium release to the perhaps, into how sodium-dependent low-sodium intracellular environment, transporters generally operate, and possibly stimulated by a negative brings us closer to a complete picture of their transport mechanism and how this is regulated.

Fig. 103: Transport mechanism of the NSS family. Na+ binding stabilises the outward-open state (1) and allows the substrate to bind (dashed arrow), triggering substrate site occlusion (2). Closure of a hydrophobic, extracellular vestibule (EV) stabilises the transition to an inward-facing state where TM5 unwinding provides a solvation pathway for the Na2 driving site (3). The helix unwinding provides an opportunity for Na+ to escape to the intracellular low-sodium environment (4). Na+ release from Na2 allows TM5 to reassume a helical structure and TM1a to swing out to release the substrate with Na+ from the Na1 site in the inward-open state. Finally, the transporter switches back to the outward- open state via as yet unknown sodium-free return states (5-6). A scaff old domain of the transporter is shown in dark grey, a bundle domain in pink, half channel openings as grey lines, solvated regions by light blue shadows, TM1 as red and TM5 as cyan cylinders, Na+ ions as green, and amino acid substrate in orange spheres. The membrane is indicated by straight, black lines.

Principal publication and authors STRUCTURE OF A MAMMALIAN G. Hassaine (a,b), C. Deluz (a), L. Grasso (a), R. Wyss (a), M.B. Tol (a), R. Hovius (a), A. Graff (c), SEROTONIN RECEPTOR REVEALED BY H. Stahlberg (c), T. Tomizaki (d), A. Desmyter (e), C. Moreau (f), X-RAY CRYSTALLOGRAPHY X.-D. Li (g), F. Poitevin (h), H. Vogel (a) and H. Nury (a,f), Nature 512, 276-81 (2014). The complex mental capacities and signals. These receptors, targets (a) LCPPM, EPFL, Lausanne (Switzerland) motor behaviours of mammals rely of many psycho-active and (b) Present address: Theranyx, Marseille (France) on the evolution of nervous systems therapeutic compounds, function as (c) CCIN, Biozentrum, University of Basel in which ligand-gated ion channels allosteric signal transducers whose (Switzerland) facilitate fast intercellular signalling. conformational transitions, regulated (d) SLS, PSI, Villigen (Switzerland) The Cys-loop receptor family by neurotransmitter binding, operate (e) AFMB, CNRS & Université Aix- encompasses pentameric ion channels the gating of a channel that let ions fl ow Marseille, Marseille (France) (f) Institut de Biologie Structurale, such as the nicotinic acetylcholine through the plasma membrane. Each Grenoble (France) and serotonin 5-HT3 (resp. GABA of their fi ve subunits contains three (g) LBR, PSI, Villigen (Switzerland) and Glycine) receptors which are domains: a β-sheet-rich extracellular (h) DSM, Institut Pasteur & CNRS, Paris cation- (resp. anion-) selective and domain comprising the ligand-binding (France) promote excitatory (resp. inhibitory) site, an α-helical transmembrane

88 ESRF STRUCTURaL BIOLOGY

Fig. 104: a) View parallel to the membrane plane of the crystal structure of the split 5-HT3A receptor in complex with VHH15 in cartoon representation. Two out of the fi ve VHH15 molecules observed are shown in pale green. The ion (pictured in red) permeation pathway is shown as a continuous solvent accessible volume (grey). b) Rotated view of the interface between two neighbouring extracellular domains and the solvent accessible volume (pink) representing the serotonin binding site. One VHH15 molecule (hidden in the main panel) is also displayed in cartoon and surface representation, highlighting the extensive interaction between its loops and the ligand- binding site. c) Close up of the transmembrane domain pore.

Fig. 105: a) View of one subunit of the split 5-HT3A receptor shown in cartoon representation. b) The intracellular domain. Key residues are shown as spheres. c) Same view, in surface representation, highlighting the vestibule and the two constricted ion exit pathways.

domain made of helices named M1 to M4, and an intracellular domain, a major determinant of gating kinetics and channel conductance. Even though structural knowledge of these receptors is rapidly increasing (see e.g. [1-2]), the latter domain, specifi c to metazoa, is absent in crystal structures derived thus far.

To further document the mechanistic details of the functional processes of pentameric ligand-gated ion channels, high-resolution structures of full-length mammalian receptors have to be solved. In an eff ort toward this goal, we crystallised a mouse homopentameric 5-HT3A receptor that retains a large In these crystals, the 5-HT3 receptor part of its intracellular domain. The full- subunits are arranged in pseudo length receptor was expressed in HEK fi ve-fold symmetry around a central 293 cells [3]. First attempts yielded ion pathway perpendicular to the crystals diff racting X-rays only to 20 membrane plane (Figure 104a). The Å resolution and limited proteolysis extracellular domain reveals the experiments improved this limit to detailed anatomy of the serotonin 7 Å. We eventually found that the binding-site capped by a VHH15 Llama-derived single-chain antibody bound radially to the subunit interface VHH15 formed a stable complex with (Figure 104b). The transmembrane the 5-HT3 receptor and crystals of this domain delimits an aqueous pore complex diff racted to 3.5 Å resolution lined by the M2 helices (Figure 104c), at beamline ID29. whose state (open or closed) cannot be

The European Synchrotron 89 SCIENTIFIC HIGHLIGHTS STRUCTURAL BIOLOGY

unambiguously defined without more into ion permeation mechanism dynamical information. through the intracellular vestibule. The missing digested segment, between A portion of the intracellular MX and MA, contains interaction domain structure is revealed in the sites for proteins modulating activity, crystal structure (Figure 105a): assembly, trafficking and clustering. the post-M3 loop and the short MX helix clamp the helix MA, the latter The structure of the split 5-HT3 a continuous extension of M4, and receptor shows a number of interesting create a vestibule exposing charged features. Firstly, it reveals part of the residues on its inner surface that were intracellular domain, thus expanding the identified as determinants of single structural basis for understanding the channel conductance by mutagenesis operating mechanism of mammalian studies (Figure 105b). Here, lateral Cys-loop receptors. Secondly, our ion pathways that might be formed finding that a particular nanobody References between MA helices are blocked by the binds at the subunits’ extracellular [1] L. Sauguet et al., PNAS 111, 966-71 post-M3 loop and, further down, the MA interface opens the possibility to (2014). helical hydrophobic bundle tightens, specifically target Cys-loop receptor [2] P. Miller and A.R. Aricescu, Nature thus also constricting a possible exit subtypes: a crucial challenge in 512, 270-5 (2014). pathway for ions (Figure 105c). Again, developing new active compounds and [3] G. Hassaine et al., Biochim. Biophys. Acta 1828, 2544-52 (2013). a more dynamical view of the receptor in determining the crystal structures of would be needed to gain further insight pentameric ion channels.

Principal publication and authors S trUCTURAL BASIS FOR NUCLEAR G.N. Maertens (a), N.J. Cook (b), W. Wang (c), S. Hare (a), S.S. Gupta (a), I. Öztop (c), K.-E. Lee (d), import of splicing factors V.E. Pye (b), O. Cosnefroy (b), A.P. Snijders (b), V.N. KewalRamani (b), A. Fassati (e), A. Engelman (c) and The nuclear/cytoplasmic trafficking nuclear/cytoplasmic trafficking are P. Cherepanov (a,b), PNAS 111, 2728– machinery that is essential for currently being evaluated as cancer 2733 (2014). metazoan development and therapeutics. (a) Division of Infectious Diseases, Imperial College London (UK) homeostasis, is often deregulated (b) Clare Hall Laboratories, Cancer in cancer and can be hijacked by One fundamentally important type Research UK, London Research Institute, pathogens, such as HIV. The transport of a nuclear localisation signal, Potters Bar (UK) of macromolecules between cytoplasm comprising sequences rich in Arg-Ser (c) Department of Cancer Immunology and nucleus is orchestrated by nuclear dipeptides (known as RS domains), and AIDS, Dana–Farber Cancer Institute, Boston (USA) import and export receptors. Referred belongs to the family of Ser/Arg- (d) HIV Drug Resistance Program, to as importins and exportins, rich (SR) proteins. These nuclear National Cancer Institute, Frederick (USA) these proteins bind their cargos and proteins also contain RNA recognition (e) Wohl Virion Centre and Medical translocate them across the nuclear motif (RRM) domains and play Research Council Centre for Medical and pore complex. This process is regulated essential roles in pre-mRNA splicing Molecular Virology, Division of Infection and Immunity, University College London by Ran, a small GTPase, which and 3´-processing, participating (UK) partitions between cytoplasm and in transcription regulation, mRNA nucleus in the predominantly GDP- transport, translation and nonsense- and GTP- bound form, respectively. mediated decay. The canonical Importins associate with their cargos splicing factors ASF/SF2 and SC35 in the cytoplasm, and the competitive are among the best-characterised SR binding to RanGTP induces them to proteins. RS domains are processively release their cargos in the nucleus. phosphorylated on their Ser residues Most nuclear import/export receptors by a set of dedicated kinases. This belong to the b-karyopherin family of phosphorylation is thought to trigger proteins of which 22 members encoded import of SR proteins into the nucleus, in the human genome. Typically, where it is subsequently required for b-karyopherins bind their cargos the spliceosome assembly. During pre- directly, recognising a linear nuclear mRNA splicing and 3´-processing, RS localisation or export signal, and/or a domains are involved in establishing specific tertiary/quaternary structural networks of protein-protein feature. Small molecules targeting interactions.

90 ESRF STRUCTURaL BIOLOGY

Fig. 106: Overview of the crystal structures of unliganded Tnpo3 (a), Tnpo3-RanGTP (b) and Tnpo3-ASF/SF2 (c) complexes. The structures were refi ned to 3.0, 3.4 and 2.6 Å resolution, respectively. Tnpo3 is shown as cartoons, while ASF/SF2 and Ran are shown in space-fi lling mode. Tnpo3 is coloured green, except the R-helix, which is shown in orange; Switch I (SI) and Switch II (SII) regions of Ran are shown in magenta and brown, respectively, with the remainder of the Ran structure shown in yellow; RRM2 and RS domains of ASF/SF2 are shown in blue and magenta, respectively.

The human β-karyopherin Tnpo3 was shown to recognise the RS domains of ASF/SF2 and SC35 and execute their nuclear import. Likewise, Mtr10, the yeast version of Tnpo3, has long been known to facilitate nuclear import of splicing factors, but no atomic structural information on any of the Tnpo3 orthologs has been reported to date. Using data collected at beamlines ID14-4 and ID23-1, we determined crystal structures of the human β-karyopherin in its three key function states: as a free protein, and as complexes with RanGTP and ASF/ SF2 (Figure 106).

Tnpo3 is a highly fl exible molecule, comprised of 20 HEAT repeats. The Fig. 107: Details of the structure of protein adopts an open circular/toroidal Tnpo3-ASF/SF2. a) Overview with both shape, with its N- and C-terminal protein chains displayed as cartoons. The arches facing each other (Figure 106a). RRM2 and RS domains of ASF/SF2 and the The structure of the Tnpo3-ASF/SF2 R-helix of Tnpo3 are indicated. b) Details of the interface between Tnpo3 and the RS complex revealed a tripartite protein- domain seen in the structure of ASF/SF2. protein interface (Figure 106c, 107a), with the most important contacts involving the RS domain of the cargo. In its mode of binding to Tnpo3, the RS engaging RanGTP in the nucleus. A domain is reminiscent of Velcro tape, close comparison of the structures of and its structure is then determined Tnpo3-ASF/SF2 and of Tnpo3-RanGTP by the underlying Tnpo3 surface. (Figure 106b) revealed that the binding Therefore, the interface is highly platforms for RanGTP and ASF/SF2 adaptable, aiding the β-karyopherin overlap on the concave surface of to accommodate highly degenerate RS Tnpo3, with the Switch I region of Ran domains. The inner helix of Tnpo3 HEAT obstructing the approach of the RS repeat 15 (R-helix) projects an Arg side domain to the R-helix. Collectively, chain at each turn towards the concave these observations provide the surface of Tnpo3 and contributes to structural basis for the dissociation of the phospho-Ser binding platform Tnpo3-cargo complexes via competitive (Figure 107b). The phosphorylated RS binding of RanGTP upon nuclear entry. domain forms a tight network of salt bridges with complementary Tnpo3 residues on and around the R-helix (Figure 107b).

The fl exibility of Tnpo3 may allow it to bind a wide range of SR proteins and, furthermore, aid in their release upon

The European Synchrotron 91 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

Principal publication and authors A. Frick (a), U. Kosinska-Eriksson (a), T HE CRYSTAL STRUCTURE OF HUMAN F. de Mattia (b), F. Öberg (a), K. Hedfalk (a), R. Neutze (a), W. de Grip (c), P.M.T. Deen (b) and AQUAPORIN 2 S. Törnroth-Horsefi eld (a, d), PNAS 111, 6305-6310 (2014) (a) Department of Chemistry and Maintenance of body water A striking diff erence between AQP2 Molecular Biology, University of homeostasis is an essential part and previous structures of mammalian Gothenburg (Sweden) of human physiology. A key event aquaporins is the signifi cant fl exibility (b) Department of Physiology, Radboud University Medical Center, Nijmegen in this is the kidney’s ability to of the C-terminus. In particular, the (The Netherlands) regulate urine volume in response to C-terminal helix, which in previous (c) Department of Biochemistry, dehydration. This takes place in the aquaporin structures occupies a Radboud Institute for Molecular Life collecting duct through the action of conserved position at the cytoplasmic Sciences, Radboud University Medical aquaporin 2 (AQP2), a membrane- interface, adopts four distinct new Center, Nijmegen (The Netherlands) (d) Department of Biochemistry and bound water channel, in the apical conformations (Figure 109a). In Structural Biology, Centre for Molecular membrane of the principal cells. Upon one monomer, the C-terminal helix Protein Science, Lund University dehydration, AQP2 is moved from interacts with the cytoplasmic interface (Sweden) intracellular storage vesicles to the of a symmetry-related AQP2 molecule apical membrane, thereby increasing (Figure 109b), thereby shielding four its water permeability. This process, leucines aligned on the same side of the which is known as traffi cking, involves C-terminal helix from the cytoplasm. the pituitary hormone vasopressin, These leucine residues are responsible which stimulates phosphorylation for the interaction between AQP2 and of the AQP2 C-terminus, thereby LIP5, a regulatory protein involved triggering its translocation. Defective in targeting AQP2 for lysosomal traffi cking of human AQP2 leads to degradation. The corresponding nephrogenic diabetes insipidus (NDI), residues are also involved in the a water balance disorder characterised interaction between AQP0 and by large urine volumes (up to 20 L/day) calmodulin [3]. We therefore suggest and subsequent dehydration [1]. that exposed hydrophobic residues on the C-terminal helix are a common We have solved the crystal structure motif for protein-protein interactions of human AQP2 at 2.75 Å resolution amongst mammalian aquaporins. using data collected at ID14-4. As seen in previous aquaporin An unexpected fi nd in the AQP2 structure structures, this reveals a tetrameric was the presence of two divalent cations fold with each monomer comprising per tetramer (Figure 109c), assigned six transmembrane helices and two as cadmium due to its presence during half-membrane spanning helices crystallisation, but likely to be replaced constituting a seventh pseudo- by calcium in vivo. The fi rst binding transmembrane segment (Figure 108). site is located at the cytoplasmic The water-translocating channel interface between monomers a and D passes through the middle of each and causes a rearrangement of loop D, monomer in which water is transported thereby exposing the binding pocket in a single fi le while excluding ions and in which the C-terminal helix from the protons [2]. symmetry-related monomer binds (Figure 109b). The second calcium ion binds at the C-terminus and positions the C-terminal helix for this interaction. It thus seems that calcium ion binding may very well be important for protein- protein interactions involving the AQP2 C-terminus.

Fig. 108: Structure of the AQP2 tetramer viewed in the plane of the membrane (a) and from the cytoplasm (b).

92 ESRF STRUCTURaL BIOLOGY

Due to its role in NDI, AQP2 is the most clinically studied member of the aquaporin family. Over fi ft y diff erent AQP2 point mutations have been identifi ed in patients suff ering from NDI, most of which cause endoplasmic reticulum (ER)-retention. By obtaining the structure of AQP2, we now have a structural framework to help us understand why these mutations cause disease thereby opening the possibility of fi nding new pharmacological treatments. Furthermore, the crystal structure provides a platform for exploring membrane protein cellular sorting in general, including how this is guided by protein-protein interactions and how misfolded proteins are recognised by the ER quality control.

References [1] P.M.T. Deen, M.A. Verdijk, N.V. Knoers, B. Wieringa, L.A. Monnens, C.H. van Os and B.A. van Oost, Science 264, 92-95 (1994). Fig. 109: Structural details of human [2] U. Kosinska Eriksson, G. Fischer, AQP2. a) Overlay of four monomers R. Friemann, G. Enkavi, E. Tajkorshid showing the conformational variability and R. Neutze, Science 340, 1346-1349 of the C-terminus. b) Interaction between (2013) the C-terminal helix and the cytoplasmic [3] S.L. Reichow, D.M. Clemens, interface of symmetry-related AQP2 J.A. Freites, K.L. Nemeth-Calahan, molecules. c) Two cadmium ions can be M. Heyden, D.J. Tobias, J. E. Hall and seen per tetramer (Cd1 and Cd2) and are T. Gonen, Nat. Struct. Biol. 20, 1085- likely to represent calcium ion binding sites 1092 (2013). in vivo.

STRUCTURAL ELUCIDATION OF MADS Principal publication and authors S. Puranik (a), S. Acajjaoui (a), DOMAIN TRANSCRIPTION FACTOR S. Conn (b), L. Costa (a), V. Conn (b), A. Vial (a), R. Marcellin (a,c), R. Melzer (d), E. Brown (a), D. Hart (e), G. Theißen (d), SEPALLATA3 C.S. Silva (f), F. Parcy (f), R. Dumas (f), M. Nanao (g,h) and C. Zubieta (a,f,i), The Plant Cell 26, 3606-3615 (2014). Virtually all terrestrial habitats are the seed. Flowering plants are not only (a) ESRF dominated by angiosperms, or fl owering the dominant type of land plants, but (b) SA Pathology and the University of South Australia, Adelaide (Australia) plants. However, angiosperms are a are also the primary source of food (c) Faculté des Sciences de Montpellier relatively recent step in evolution. Their and habitat for all animals, including (France) success in colonising new habitats and humans. The evolution of fl owers has (d) Friedrich Schiller University, Jena superseding other species is due to the been the subject of speculation from (Germany) advent of a novel reproductive structure the time of Charles Darwin who termed (e) Université Grenoble Alpes, CNRS, Integrated Structural Biology Grenoble, – the fl ower. Flowers are composed of the dominant rise and diversifi cation UVHCI, Unité Mixte Internationale 3265, four diff erent organs: the sepals which of fl owering plants as “an abominable UMS 3518 Grenoble (France) protect the developing bud, the petals mystery” due to the lack of a smooth (f) LPCV, Grenoble (France) which are the showy organs that act transition from non-fl owering to (g) EMBL, Grenoble (France) as pollinator attractors, the stamen fl owering plants in the fossil record. (h) UVHCI, Université Grenoble Alpes- EMBL-CNRS, Grenoble (France) which are the male organs containing With the sequencing of multiple (i) CEA, DSV, iRTSV, Grenoble (France) the pollen and the carpels which are the genomes from basal angiosperms, female organs and contain the ovules. gymnosperms and higher fl owering The fl ower unites the male and female plants, certain gene families that organs into one structure and encloses play a central role in the development

The European Synchrotron 93 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

and evolution of the fl ower have been gene expression and/or repression [2]. identifi ed. Our research focuses on one Some of our recent work investigated such family of high-level regulators the molecular mechanisms of SEP3 called the MADS transcription factor function. Using a combination of (TF) family, that orchestrates fl ower structural techniques, including X-ray development. We are interested crystallography and atomic force in understanding the molecular microscopy (AFM), we determined mechanisms of the MADS family and how SEP3 is able to form diff erent how these proteins are able to control protein-protein complexes and directly complex reproductive functions. demonstrated DNA looping following SEP3 tetramer formation. Unlike animals which have only a few MADS family members, plants have The 2.5 Å resolution crystal structure greatly expanded and diversifi ed the of the SEP3 oligomerisation domain family during evolution. MADS TFs was determined from diff raction data play signifi cant roles in all aspects obtained at beamline ID14-4. This is of plant life from seed germination the fi rst structure for a plant MADS to fruit and fl ower development. One transcription factor and demonstrates of the most central members of the novel dimerisation and tetramerisation MADS family is SEPALLATA3 (SEP3). interfaces formed by amphipathic SEP3 is able to form over 50 diff erent α−helices. Figure 110 shows that complexes with other MADS proteins the protein monomer folds into two and has roles in virtually all aspects amphipathic α−helices separated by of plant reproduction including the a rigid kink region, that the dimer is establishment of the fl oral meristem, formed via hydrophobic interactions the development of all fl oral from helix 1 and helix 2 and that the organs and fruit ripening tetramerisation interface is more [1]. SEP3-containing protein limited, relying on hydrophobic complexes are postulated to residues contributed by the C-terminal bind DNA at multiple binding portion of helix 2. Mutagenesis studies sites resulting in DNA loops that have confi rmed the importance of these are important for downstream residues to the tetramerisation of the protein. To study the full-length protein, AFM was performed in collaboration with F. Comin and L. Costa of the ESRF Surface Science Laboratory. These Fig. 110: SEP3 tetramer experiments directly demonstrate depicted as a cartoon with SEP3-DNA binding. Figure 111 shows each monomer coloured a SEP3 tetramer binding two adjacent diff erently. For each monomer, the N and C DNA-binding sites leading to the termini are labelled. formation of a DNA loop. This ability is lost when the C-terminal portion of helix 2 is deleted.

These results provide the fi rst structural basis for MADS domain transcription factor function. The novel dimerisation and tetramerisation interfaces present two independent protein-protein interaction surfaces, allowing the proper binding of many diff erent partners for both dimer and tetramer formation. We are currently developing and testing predictive models for MADS domain TF function based on our structural studies References Fig. 111: AFM results showing full length of SEP3 using an integrated biophysical [1] R.G.H Immink, I.A. Tonaco, S Folter, SEP3 in complex with a 1-kb DNA fragment and genetic approach. As a long-term A Shchennikova, A.DJ. Dijk, J. Busscher- containing two SEP3 binding sites. The goal, these studies will help us to tune Lange and J.W. Borst, Genome Biology inset clearly shows that SEP3 tetramer 10, R24 (2009). binding to two adjacent DNA-binding sites the function of MADS TFs in plants and [2] R. Melzer, W. Verelst and G. Theissen, (indicated by arrows) leads to formation of to alter diff erent reproductive pathways Nucleic Acids Res. 37, 144-157 (2009). a DNA loop. and structures in a predictable manner.

94 ESRF STRUCTURAL BIOLOGY

Principal publication and authors Tracking the bacterial immune S.G. Wong (a), and A. Dessen (a,b), Nature Comm. 5, 4917 (2014). (a) Institut de Biologie Structurale (IBS), system CNRS, CEA, Univ. Grenoble (France) (b) Brazilian National Laboratory for Biosciences (LNBio), Campinas (Brazil) Alpha-2-macroglobulins (A2Ms) are thioester bond without engendering plasma proteins that trap and inhibit conformational modification of the a broad range of proteases and are macroglobulin. Interestingly, the major components of the eukaryotic thioester motif is stabilised by a innate immune system. Their main structural lock provided by Tyr1175, a functions include clearing pathogenic residue which is present only in A2Ms or parasitic proteases from circulation, of bacterial species. The necessity for as well as blocking proteases that this ‘lock’, which guarantees that the participate in inflammatory and blood thioester remains entrenched (and clotting events. A2Ms are composed protected from hydrolysis) within the of ~180 kDa subunits that generally macroglobulin in the absence of a associate into oligomeric forms. The target protease, is potentially linked human form, for example, is a tetramer to the fact that bacterial A2Ms are (thus, a 720 kDa protein). Despite it monomeric, and thus structurally being believed that these molecules simplistic, when compared to their existed uniquely in metazoans, A2M- oligomeric eukaryotic counterparts. like proteins were recently identified Human A2M, for example, entraps in pathogenically invasive bacteria as target proteases in a central cage-like well as in species that colonise plants, structure formed by its four monomers. fish, and insects [1], suggesting that In StA2M a Y1175G mutation opens bacteria could also require ‘protection’ this lock. Here, the thioester region is from proteases secreted by the target unprotected, the covalent thioester host or by other organisms competing bond is broken, and the macroglobulin for nutrients or space. is unable to trap proteases. Our structural and mechanistic studies thus Bacterial A2Ms are monomeric not only pinpoint how bacteria can molecules located in the periplasm inhibit protease attack during infection of the cell, thus right underneath or colonisation, but also indicate that the outer membrane, where they are bacteria have developed a rudimentary believed to provide protection to the innate immune system whose cell upon membrane breach [2]. This mechanism mimics that of eukaryotes. occurs by trapping external proteases through a covalent interaction with a thioester bond buried within a hydrophobic environment inside the inactive macroglobulin. In the test tube, activation of the thioester can be mimicked by a small molecule called methylamine. Our paper reports the crystal structures and the characterisation of the alpha-2 macroglobulin from the human pathogen Salmonella typhimurium Fig. 112: Schematic arrangement of the (StA2M) in different states of thioester S. typhimurium A2M, which comprises over activation. Diffraction data were 1600 residues that fold mostly into b-sheet- collected at beamlines ID29 and ID23-2, carrying domains. The only exception is the and solution scattering data at beamline thioester domain (TED), which is mostly helical (blue), and harbours the thioester BM29. The crystal structures revealed bond that attacks target proteases. thirteen domains (Figure 112), whose arrangement displays high similarity to proteins involved in eukaryotic immune defense, such as complement References [1] A. Budd, S. Blandin, E. Levashina and C3. The crystal structure of (StA2M) in T.J. Gibson, Genome Biol 5, R38 (2004). complex with methylamine confirms [2] N. Doan and G.W. Gettins, J Biol that the small molecule can attack the Chem 283, 28747-28756 (2008).

The European Synchrotron 95 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

Principle publication and authors M. Szczepek (a,b), F. Beyrière (a), CRYSTAL STRUCTURE OF A COMMON K.P. Hofmann (a), M. Elgeti (a), R. Kazmin (a), A. Rose (a), F.J. Bartl (a), D. von Stetten (c), GPCR-BINDING INTERFACE FOR M. Heck (a), M.E. Sommer (a), P.W. Hildebrand (a) and P. Scheerer (a,b), G-PROTEIN AND ARRESTIN Nature Communications 5, 4801 (2014). (a) Institute of Medical Physics and Biophysics, Charité - The large family of G-protein-coupled GDP/GTP exchange in the Gα-subunit Universitätsmedizin Berlin (Germany) receptors (GPCRs) controls various (1). Another binding partner is (b) Institute of Medical Physics and Biophysics, Group Protein sensory and physiological responses arrestin, whose canonical role is to X-ray Crystallography and and is thus a major focus of research stop signalling by blocking receptor Signal Transduction, Charité - and of interest for medical drug coupling to G-proteins, or to initiate Universitätsmedizin Berlin (Germany) targets. These receptors bind a variety arrestin-dependent signalling by (c) ESRF of ligands (e.g. light-depending retinal, bringing together other signalling hormones, opioids, odorants, peptides, proteins. In our studies, we have neurotransmitters) and communicate addressed the question of how arrestins such extracellular signals into the employ the cytoplasmic crevice within cell by coupling to intracellular the 7 transmembrane (TM) bundle heterotrimeric G-protei ns (Gαβγ) and of the active receptor for the high- arrestins (Figure 113) [1]. In G-protein- affi nity interaction necessary to block signalling, the Gα C-terminus (GαCT) G-protein binding or stimulate arrestin binds to a cytoplasmic crevice mediated signalling [4,5]. This work formed in the agonist-bound active was motivated by our observation of receptor [2,3], eventually catalysing a consensus sequence motif, (E/D)

Fig. 113: Overview of rhodopsin signal transduction and deactivation. (1) Dark-state bovine rhodopsin with inverse agonist 11-cis-retinal. (2) Light-activation and formation of metarhodopsin II with agonist all-trans-retinal. (3) Coupling to the G-protein. (4) Phosphorylation of the activated receptor by its specifi c kinase. (5) Arrestin binding to the phosphorylated receptor. Inset (left ): Sequence alignment of a proposed common sequence motif: rod photoreceptor arrestin (Arr1; residues 68–77), cone photoreceptor arrestin (Arr4; 63–72), β-arrestin-1 (Arr2; 64–73), β-arrestin-2 (Arr3; 65–74), wildtype

GtαCT (340–349), high-affi nity variant GtαCT-HA (340–349), Gsα (385–394). Inset (right): View of the cytoplasmic receptor crevice from the crystal structures depicted as ribbon-stick

diagrams of a), R*- GtαCT (coloured in black-blue, PDB entry 3DQB, [2]) and b) R*- ArrFL-1 (coloured in orange-purple, PDB entry 4PXF, [6]).

96 ESRF STRUCTURAL BIOLOGY

x(I/L)xxxGL (Figure 113, inset left), To verify the functional relevance of which is conserved in GtaCT of the Gi/ the observed peptide binding to the Gt-family and the “finger loop” (ArrFL) receptor crevice, we applied three from the central crest region of all four different functional assays to native subgroups of arrestin [6]. disc membranes filled with rhodopsin. These functional experiments showed We have used protein X-ray similar R* binding properties of crystallography, Fourier-transform- Gi/Gt and ArrFL peptides. The key infrared spectroscopy, functional player, Arg135, essential for receptor competition assays and UV-Vis activation and G-protein-coupling, has absorption spectroscopy on solutions also a decisive role in arrestin binding (“Extra-Meta II assay”) and on (Figure 113, right). However, significant crystals (at beamline ID29S – The differences between ArrFL and GteaCT ar Cryobench) to gain insight into how seen at the rim of the receptor crevice. an active receptor binds to ArrFL. In ArrFL-1, the interaction with TM5/6 Firstly, we co-crystallised the active is partially replaced by interactions with receptor rhodopsin R* (as light- TM7/H8 (specifically to the conserved activated metarhodopsin II or as the NPxxY(x)5,6F motif) and directed towards active form of the apo-protein opsin, the intracellular loop 2 (CL2). Ops* [1,2,3,6]) with three different ArrFL peptides derived from rod Overall, the crystal structure indicates a photoreceptor arrestin, b-arrestin-1 general mechanism of arrestin binding and b-arrestin-2. The crystal structure and the remarkable homology seen of active rhodopsin with a peptide between GtaCT and ArrFL suggests References analogue derived from ArrFL-1 (rod that the (E/D)x(I/L)xxxGL motif evolved [1] K.P. Hofmann et al., Trends Biochem photoreceptor arrestin), was solved in both G-protein and arrestin to Sci. 34, 540-552 (2009). [2] P. Scheerer et al., Nature 455, 497- at 2.75 Å resolution using diffraction allow recognition and interaction with 502 (2008). data collected at beamline ID29. ArrFL the same binding crevice presented [3] H.-W. Choe et al., Nature 471, 651- binds to the cytoplasmic crevice of R* by the active GPCR [6]. Despite 655 (2011). with a similar reverse turn structure these similarities, we also identified [4] A.K. Shukla et al., Nature 497, 137- as observed for the C-terminal G aCT pronounced differences, namely a 141 (2013). t [5] Y.J. Kim et al., Nature 497, 142-146 [2]. In both cases, a hydrogen bond is larger hydrophobic interaction surface (2013). formed between the ArrFL-1 reverse in the case of Gt,aCT which may [6] M. Szczepek et al., Nature turn and Arg135 from the highly contribute to the distinct biological Communications 5, 4801 (2014). conserved E(D)RY motif of R*. functions of G-proteins and arrestin.

Insight into an intermediate step Principal publication and authors S. Bui (a), D. von Stetten (b), in O -dependent catalysis P.G. Jambrina (c), T. Prangé (d), 2 N. Colloc’h (e), D. De Sanctis (b), A . Royant (b,f), E. Rosta (c) and R.A. Steiner (a), Angew.Chem.Int.Ed. The unambiguous identification Uricase (UOX) is the archetypal 53, 13710–13714 (2014). of reaction intermediates is one cofactor-independent oxidase [1]. It (a) Randall Division of Cell and Molecular of the most challenging tasks in catalyses the breakdown of uric acid Biophysics, King’s College London (UK) structural enzymology. Although to 5-hydroxyisourate. A recombinant (b) ESRF (c) Dept. of Chemistry, King’s College peroxide intermediates are likely to be version of UOX finds therapeutic London (UK) formed in many reactions involving application to aid uric acid clearance (d) LCRB, UMR 8015-Université Paris molecular oxygen (O2), their structural in paediatric and adult cancer patients Descartes-CNRS (France) identification has so far proven elusive. and for gout treatment. To investigate (e) ISTCT, UMR 6301–UCBN–CNRS–CEA- Here, X-ray crystallography combined the step involving dioxygen in UOX- Normandie Université (France) (f) IBS, Grenoble (France) with online in-crystallo Raman mediated catalysis, we employed spectroscopy at beamline ID29 and 9-methyl uric acid (MUA). When radiation damage effects have provided anaerobic UOX:MUA co-crystals are direct evidence for a C5(S)-(hydro) exposed to O2, electron density maps at peroxide intermediate in a cofactor-free 1.3 Å resolution unambiguously reveal oxidase reaction that also allowed the that MUA converts into its C5(S)-peroxo visualisation of the reactive enzyme- derivative (5-PMUA) (Figure 114a). 3 substrate-O2 complex. 5-PMUA is sp hybridised at carbon C5.

The European Synchrotron 97 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

Fig. 114: a) The UOX: 5-PMUA complex formed by reacting UOX:MUA crystals grown in anaerobic conditions with air. Electron density is shown for 5-PMUA. Hydrogen bonds involving the peroxo atoms are shown by broken lines. b) In-crystallo raman spectra of apo-UOX (grey), anaerobic UOX:MUA (green), UOX:5- PMUA (brown), and UOX:H5-PMUA (gold). H5- PMUA (H for heavy) refers 18 to MUA reacted with O2.

The peroxo Op1-Op2 bond is 1.47 Å 600 cm-1 exhibits changes sensitive long, whilst the length of the C5-Op1 to the chemistry of the oxygenation bond refi nes to 1.51 Å. reaction (Figure 114b). Upon MUA peroxidation, a distinct band develops Non-resonant Raman spectra recorded at 605 cm-1 (brown trace), whilst the from crystals of substrate-free UOX, shoulder at 597 cm-1 for the UOX:MUA anaerobic UOX:MUA and UOX:5-PMUA complex (green) disappears. Neither complexes are very similar overall. band is present in the spectrum of However, the region centred around substrate-free UOX (grey) indicating that they arise from the bound organic molecules. Furthermore, the 605 cm-1 band could be selectively shift ed by carrying out MUA peroxidation with 18 -1 ~ -1 O2 (shift to 596 cm , Δν= –9 cm , gold), thus confi rming that this band specifi cally involves Raman modes with contributions from the peroxide oxygen atoms. Quantum mechanics/ molecular mechanics (QM/MM) calculations at the MP2/6-31+G* level of theory predict a band at 600 cm-1 (experimental 605 cm-1) for the 5-PMUA hydroperoxide resulting from a set of modes involving C5-Op1 bond stretching and C5-Op1-Op2 bending coupled to ring distortions. The theory further predicts a -9 cm-1 isotope shift (experimental -9 cm-1) for 5-PMUA featuring 18Op1 and 18Op2 peroxide oxygen atoms. Overall, the calculations agree remarkably well with the experiment identifying the 605 cm-1 5-PMUA Raman band as a ‘signature’ for its peroxide state.

Fig. 115: a) Dose-stamped snapshots for UOX:5-PMUA radiolysis. Electron density is shown for the organic moieties and solvent molecules in close proximity. b) Online in-crystallo Raman spectroscopy reveals the specifi c dose-dependent decrease of the 605 cm-1 5-PMUA ‘fi ngerprint band’. c) 5-PMUA decay is biphasic suggesting a mechanism of peroxide regeneration. 5-PMUA occupancies from crystallographic refi nement are shown as fi lled circles. Open circles are 5-PMUA occupancies estimated from the integration of the 605 cm-1 ram an band.

98 ESRF STRUCTURAL BIOLOGY

Exposure to X-rays can lead to specific Online Raman analysis shows that modifications in proteins and nucleic peroxide rupture causes the height of the acids, including bond rupture. We 605 cm-1 ‘signature band’ to selectively observed that the C5-Op1 bond is decrease in a dose-dependent manner susceptible to selective radiolysis at (Figure 115b), consistent with our very low X-ray doses. We tracked this QM/MM calculations that assign this quantitatively using on line Raman- band to the stretching of the C5-Op1 assisted crystallography by performing bond. The dose-dependent 5-PMUA multiple data collections interspersed occupancy from crystallographic by spectrophotometric measurements refinement at near-atomic resolution on a single UOX:5-PMUA crystal. The is in excellent agreement with its dose-dependent rupture of the C5-Op1 orthogonal estimation from the bond is accompanied by the loss of Sp3 integration of the 605 cm-1 Raman hybridisation at C5 leading to a planar signal (Figure 115c). 5-PMUA decay organic structure (Figure 115a 2.5- follows a biphasic profile. This is not 655 kGy). Concomitantly, a diatomic consistent with a simple peroxide molecule with a shorter bond length decomposition process. The observed (1.2 Å) than the original peroxide is decay can be rationalised assuming liberated and trapped above it. We a pathway of peroxide regeneration interpreted the elongated electron acting alongside its decomposition. We density as molecular oxygen as reactive have fitted this process kinetically. oxygen species are likely transiently formed in the process and are expected Our multi-technique investigation to rapidly convert to the more stable offers unambiguous evidence for a

O2 molecule. Dioxygen adopts a well- C5-peroxide intermediate in cofactor- defined position and orientation above free UOX catalysis. Peroxide radiolysis the flat organic molecule. The O2 followed by online Raman-assisted molecular axis lies parallel to the plane X-ray crystallography afforded at a distance of 3.15 Å and is rotated by exquisite insight into the elusive References 15° as compared to the original Op1- reactants’ configuration leading to [1] S. Fetzner and R.A. Steiner, Appl. Op2 peroxide bond (Figure 115a). the peroxo intermediate. The latter is Microbiol. Biotechnol. 86, 791-804 suggested to be formed by a radical (2010). recombination process.

Shaping the early event of prion Principal publication and authors R.N.N. Abskharon (a, b, c), formation G. Giachin (d, e), A. Wohlkonig (a, b), S.H. Soror (a, b, f), E. Pardon (a, b), G. Legname (d) and J. Steyaert (a, b), J Am. Chem. Soc. 136, 937-944 (2014). The conversion of the cellular prion makes its structural characterization (a) Structural Biology Brussels, Vrije protein (PrPC) into its misfolded and extremely difficult. The N-terminal Universiteit Brussel (Belgium) amyloidogenic isoform, denoted as region of PrPC features a palindromic (b) Structural Biology Research Center, prion or PrPSc, is the central event in sequence (AGAAAAGA) which can VIB, Brussels (Belgium) (c) National Institute of Oceanography prion diseases which affect humans and lead to the formation of neurotoxic and Fisheries (NIOF), Cairo (Egypt) animals. In prion biology, unravelling fibrils enriched in β-sheet motifs. (d) Department of Neuroscience, the molecular mechanisms leading Experimental evidence also supports Laboratory of Prion Biology, Scuola the conversion process whereby the idea that the palindromic sequence Internazionale Superiore di Studi α-helical motifs are replaced by β-sheet plays a critical role for prion generation Avanzati (SISSA), Trieste (Italy) (e) Present address: ESRF. secondary structures is of utmost and transmissibility. However, the (f) Center of Excellence, Helwan importance. intrinsic flexibility of this N-terminal Structure Biology Research, Faculty of region has hampered efforts to obtain Pharmacy, Helwan University, Cairo The structure of human PrPC consists atomic information on the structural (Egypt) of a disordered N-terminal part features of the palindromic sequence. (residue 23-127) and a structured C-terminal domain (residue 128- In this study, we used nanobodies 228) (Figure 116). The insoluble (Nb) to solve the first structures and heterogeneous nature of PrPSc of the full-length HuPrP and its

The European Synchrotron 99 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

Fig. 116: Structural conversion of the cellular prion protein (PrPC) to its pathological isoform, denoted as prion or PrPSc. PrPC is mostly involved in synaptic functions. PrPSc causes prion diseases such as Creutzfeldt-Jakob disease in humans which is characterised by spongiform neurodegeneration (arrow) in the brain.

C-terminal truncated version by X-ray features of the hydrophobic segment crystallography. Nb are small (15 kDa) from residues 117 to 128, which had and stable single-domain fragments remained unresolved in all the PrP harbouring the full antigen-binding structures published so far. capacity of the original heavy chain– only antibodies that naturally occur in In our X-ray structures, the sequence camelids. Collective eff orts of several including the palindromic motif laboratories have demonstrated arranges in a novel β-strand we denoted that Nbs are exquisite chaperones as β0 (residues 118–122), which folds for crystallising complex biological into a three-stranded antiparallel systems such as membrane proteins, β-sheet with β1 and β2 (Figure 117). transient multiprotein assemblies, The same structural arrangement was transient conformational states and observed in both crystal structures, intrinsically disordered proteins [1]. suggesting that it does not result from crystal packing but might have In complex with this Nb, the total major biological implications for prion amount of structured polypeptide conversion. Edges of regular β-sheets (125 amino acids of antibody and are inherently aggregation-prone 108 amino acids of the prion protein) because the motif for H-bonding with rises to 71% as against 52% for free any other β-strand is available. From HuPrP, thus providing a much better our structures, it appears that the starting point for crystallisation. The exposed edge of the short β1 strand high-resolution X-ray crystal structures observed in all previous PrP structures (obtained using data collected at beam- serves as an intra-molecular nucleus line BM30A) of two HuPrPs in complex for edge-to-edge β aggregation of part with a selective nanobody of the palindromic sequence to form (Nb484) revealed a novel the β0 strand. The propensity of the structural feature. While the palindromic sequence to engage in segment from residues 128 to such β-structures strongly indicates 225 shares a fold that is very that this motif mediates β-enrichment similar to the corresponding in the PrPC monomer as one of the early NMR HuPrP structures, the events in the PrPC to PrPSc conversion. binding of Nb484 to a region adjacent to the fi rst β-sheet Here, we show that Nb-assisted (β1) unveils key structural crystallography is a powerful tool for unveiling local structural features Fig. 117: Crystal structure of of intrinsically disordered proteins. C the human PrP in complex with Our data provide structural evidence a nanobody (Nb484) with the newly formed β-strands in the that the palindromic motif of HuPrPs palindromic region highlighted. is important as a dynamic site for

100 ESRF STRUCTURaL BIOLOGY

β-sheet structural conversion in prion prion formation. Moreover, crystals of formation. The structures we solved the full-length human PrP in complex also support the hypothesis that with Nb484 are amenable to soaking References the conserved palindromic sequence experiments to study the interactions [1] E. Pardon et al., Nat Protoc 9, 674-93 C mediates β-enrichment in the PrP of small molecules with the fl exible (2014). monomers as one of the early events in region of HuPrP [1].

A MOLECULAR CALIPER FOR METERED Principal publication and authors D.R. Boer (a,b), A. Freire-Rios (c), HORMONE RESPONSE W.A.M. van den Berg (c), T. Saaki (c), I.W. Manfi eld (d), S. Kepinski (d), I. Lopez-Vidrieo (e), Hormones trigger a variety of responses binding motif with a spacer in between J.M. Franco-Zorrilla (e), S.C. de Vries (c), R. Solano (e), D. Weijers (c) and to control growth, development the two motifs. M. Coll (a,b), Cell 156, 577-589 (2014). and physiology. Understanding the (a) Institiute for Research in molecular underpinnings of specifi city When comparing two highly divergent Biomedicine, Barcelona (Spain) is a key challenge in biology. The plant ARF proteins, we found them to (b) Institut de Biologia Molecular de signalling molecule auxin is central adopt almost identical structures Barcelona (Spain) (c) Laboratory of Biochemistry, for growth and development and (Figure 118). Likewise, when we Wageningen University (Netherlands) controls a wide range of processes, determined the DNA-binding preference (d) Faculty of Biological Sciences, including embryogenesis and fl ower of ARF monomer proteins, the intrinsic University of Leeds (UK) development [1]. Here, we report specifi city was the same. Thus, neither (e) Centro Nacional de Biotecnologia, a molecular mechanism for the overall structure, nor intrinsic DNA Madrid (Spain) generation of diversity and specifi city binding specifi city can explain the in auxin response. diff erent biological functions of these two ARF proteins. As the binding of ARF Auxin is recognised by nuclear receptors, dimers to complex motifs is not only promoting degradation of inhibitor determined by DNA-binding amino proteins and consequent activation of acids but also by protein-protein DNA-binding auxin response factors interactions, we tested the hypothesis (ARFs), that control transcription that ARFs could diff er in the spacing of thousands of genes to bring tolerated between the two DNA motifs. about a multitude of developmental We found that ARFs diff ered in this responses. In the past, we had shown property. While ARF1 could bind with that variations in the ARF family in high affi nity to complex elements with the higher plant Arabidopsis thaliana a spacer of 7 or 8 bases, ARF5 could largely determine auxin response also bind to motifs with spacing of 5, 6 specifi city [2]. For the present research, or 9 bases (Figure 119). Thus, our work our biological and structural biology reveals a mechanism for DNA-binding teams joined forces to determine the specifi city within a transcription factor structural underpinnings of functional family that is based upon proteins diversity within the ARF family. acting as “calipers”. The overall structure of dimers determines the Crystallographic analysis using setting of the caliper, and thus the beamlines ID14-1 and ID14-4 revealed distance between two sites. that ARF proteins bind DNA as dimers, and let us identify the essential protein- This work now off ers a new framework DNA contacts. We then used a range of for understanding specifi city in plant biochemical and in vivo methods to hormone response and opens many show that these protein-DNA contacts new avenues for future investigations. are indeed essential for DNA binding and Of particular interest would be function in mediating auxin response and that ARF protein dimerisation is Fig. 118: Structure of Arabidopsis thaliana critical for function as it allows high- ARF1 protein dimer in complex with DNA affi nity cooperative binding to complex oligonucleotide with two inverted binding DNA elements. These elements are motifs. residues are coloured according to composed of inverted repeats of a whether they are identical (blue) or diff erent (magenta) between ARF1 and ARF5.

The European Synchrotron 101 SCIENTIFIC HIGHLIGHTS STRUCTURaL BIOLOGY

References to discover: how a caliper-based [1] S. Vanneste and J. Friml, Cell 136, mechanisms is leveraged in living cells 1005-1016 (2009). to recognise specifi c genes, how the [2] E.H. Rademacher, A.S. Lokerse, A. Schlereth, C.I. Llavata-Peris, M. Bayer, ARF caliper can be tuned by co-factors M. Kientz, A. Freire-Rios, J.W. Borst, and, perhaps most importantly, how W. Lukowitz, G. Jürgens and D. Weijers, this caliper mechanism evolved to Dev. Cell 22, 211-222 (2012). increase the complexity of the auxin response. Fig. 119: Affi nity of DNA binding by ARF1 and ARF5 depends on the space between the two inverted binding motifs.

Principal publication and authors J.E. Burke (a,b), A.J. Inglis (a), STRUCTURES OF PI4KIIIβ COMPLEXES O. Perisic (a), G.R. Masson (a), S.H. McLaughlin (a), SHOW HOW THE ENZYME RECRUITS F. Rutaganira (c), K.M. Shokat (c), R.L. Williams (a), Science 344, 1035- Rab11 WITH ITS EFFECTORS 1038 (2014). (a) MRC Laboratory of Molecular Biology, Cambridge Biomedical Campus (UK) Four mammalian enzymes we have employed a combination (b) Department of Biochemistry and phosphorylate the lipid of X-ray crystallography carried out Microbiology, University of Victoria phosphatidylinositol to produce at beamlines ID23-2 and ID29 with (Canada) phosphatidylinositol 4-phosphate hydrogen-deuterium exchange mass (c) Howard Hughes Medical Institute and (PI4P). One of these enzymes, PI4KIIIβ, spectrometry (HDX-MS) carried out Department of Cellular and Molecular Pharmacology, UCSF, San Francisco is enlisted by a range of RNA viruses, at the MRC-LMB HDX-MS facility. The (USA) including hepatitis C virus and HDX-MS analysis of PI4KIIIβ identifi ed enteroviruses to form PI4P-enriched intrinsically disordered regions in the References viral replication organelles [1] and enzyme and removal of these regions [1] N. Altan-Bonnet and T. Balla, Trends several compounds inhibit enteroviral enabled crystallisation of the enzyme Biochem Sci 37, 293–302 (2012). replication by targeting cellular in a complex with Rab11. The structure PI4KIIIβ. PI4KIIIβ is also important in of the complex shows that PI4KIIIβ has the malarial life cycle, and inhibitors a catalytic domain that resembles the of the P. falciparum PI4KIIIβ are potent phosphatidylinositol 3-kinases (PI3Ks) anti-malarial agents. and that it interacts with Rab11 using a helical solenoid domain N-terminal In addition to producing PI4P, PI4KIIIβ to the kinase domain (Figure 120). is also required for the recruitment of These crystal contacts were confi rmed both the small GTPase Rab11 and its in solution by HDX-MS. downstream eff ectors. To understand the structure and dynamics of PI4KIIIβ, PI4KIIIβ interacts with Rab11 in a contact that includes a unique direct interaction with the Rab11 GTP nucleotide, and this contact leaves the switch regions of the Rab11 largely free to interact with Rab11 eff ectors (see inset, Figure 120). It is the nucleotide- dependent change in conformation of these switch regions that is essential for the signalling role of Rab11. This suggests that PI4KIIIβ might be able to bind Rab11 while it, in turn, binds to its eff ectors. This was confi rmed by a low-resolution structure of a complex of PI4KIIIβ with Rab11 and the dimeric Rab11-binding domain (RBD) of the FIP3 Rab11 eff ector. These structures Fig. 120: The structure of PI4KIIIβ may provide the basis for design of in complex with PI4KIIIβ inhibitors for anti-viral and Rab11. anti-malarial applications.

102 ESRF SCIENTIFIC HIGHLIGHTS Highlights 2014

S trUCTURE OF MATERIALS

The development of new technologies developments for enhanced detector should open a new chapter in the plays a fundamental role in meeting systems. structure of materials studies at the major societal challenges. New ESRF and make it possible to attract advanced materials and devices such The construction of upgrade an even greater community of users as rechargeable batteries, catalytic beamline project UPBL2, the high from both academia and industry. materials, etc. form part of this energy beamline for buried interface development. The complexity of these structures and materials processing, V. Honkimäki heterogeneous devices can only be started at ID31 in January 2014. studied adequately by a combination The lead hutches are now ready of experimental methods, in order and installation of the beamline to reveal the interplay between the components will soon commence. microscopic material properties and This long beamline with nanofocus the macroscopic device performance. capabilities extended to high photon The following structure of materials energies will carry out the first highlights demonstrate the utility of user experiments in autumn 2015. synchrotron X-ray characterisation The final beamline to be rebuilt in techniques for the study of both real the current Phase I of the Upgrade devices under operating conditions Programme is ID15. At the moment, and idealised model systems under the old lead hutches are to be precisely controlled environments. demolished, and by the end of 2015 the new hutches will be ready for The past year was again especially the high pressure diffraction branch busy for the Structure of Materials (relocation of ID09A) and for the Group. ID01 has now been material chemistry and engineering reconstructed as the upgrade materials science branch. Due to the beamline project UPBL1, Diffraction canting of the straight section, these imaging for nanoanalysis. The first two branches are independent. coherent diffraction nano-imaging user experiments were carried out at The new buildings constructed during the end of the year taking advantage the first phase of the upgrade resulted of the enhanced nanofocus and in an increase in space available for coherence capabilities of this long the structure of materials group. beamline. Through optimisation of In particular, an electrochemistry the source and the geometry of the laboratory, under the responsibility monochromator, an increase of the of the ID03 team but open to all coherent flux by a factor of 5 has been the users, is now operational in obtained right from the first startup the new Lab and Office Building of the beamline. Further optical (LOB). In particular, it houses a new elements will be added in 2015 glove box dedicated to studies of Li allowing versatile optimisation of the batteries partly funded by the BM26- total flux and beam size over a wide Dubble and BM1-Swiss Norwegian energy range. The high resolution beamlines. The LOB also hosts a powder diffraction beamline moved catalyst laboratory, providing new from ID31 to ID22 at the beginning of opportunities for materials chemistry 2014 and it returned to user operation studies at ESRF. Furthermore, a at the beginning of the summer. nanoscope instrument to exploit high The flux has increased by a factor energy nanofocus beams at ID11 will of 2.5 due to the reduced horizontal be ready later this year. divergence at a high-β sector and its energy range now reaches 80 keV. The beamline performance Later this year, ID22 users will be improvements already mentioned in able to benefit from the large area addition to better detectors constitute detector for complementary studies a revolutionary enhancement in and from the new high-resolution performance. This, together with diffractometer allowing future the development of side laboratories

The European Synchrotron 103 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

Principal publication and authors M.P. Boneschanscher (a), ARTIFICIAL GRAPHENE MADE OF W.H. Evers (b), J.J. Geuchies (a), T. Altantzis (c), B.Goris (c), SEMICONDUCTOR QUANTUM DOTS F.T. Rabouw (a), S.A.P. Van Rossum (a), Quantum dot superlattices are Oriented attachment is a specifi c H.S.J. van der Zant (b), structures that can be formed method for the preparation of such L.D.A. Siebbeles (b), through the self-assembly of colloidal superlattices from the nanocrystal G. Van Tendeloo (c), I. Swart (a), semiconductor nanocrystals, which building blocks. In oriented J. Hilhorst (d), A.V. Petukhov (a), are nanoparticles with size-dependent attachment, nanocrystals fuse S. Bals (c) and properties [1]. Many applications together through atomically matched D. Vanmaekelbergh (a), Science have been envisioned for quantum dot bond formation between two specifi c 344, 1377-1380 (2014). superlattices, ranging from solar cells facets. Precise control of nanocrystal (a) Debye Institute for and transistors to LEDs. The properties and ligand concentration as well as Nanomaterials Science, Utrecht of such systems depend both on the temperature enabled the synthesis University (The Netherlands) properties of the constituent particles of large area, defect free honeycomb (b) Opto-electronic Materials, and on the structure of the superlattice. superlattices. The formation of these Kavli Institute, Delft University of The novel eff ects related to the systems is remarkable, given that Technology (The Netherlands) formation of a superstructure remain several demanding conditions must (c) Electron Microscopy for Materials to be addressed experimentally. be fulfi lled. The nanocrystal building Science (EMAT), University of blocks must be nearly monodisperse in Antwerp (Belgium) One particular superlattice with very size and shape, and attachment should (d) ESRF interesting properties is the honeycomb occur with atomic accuracy, between superlattice [2]. It resembles the a geometrically defi ned subset of wonder material graphene, but instead nanocrystal facets. of having carbon atoms at the edges of its honeycomb, it has PbSe or The nanocrystal building blocks have the CdSe nanocrystals, which are fused shape of a truncated cube, approaching together into a single-crystalline sheet. that of a rhombicuboctahedron, Theory predicts this material to have implying that the nanocrystals are properties similar to those of graphene, terminated with {100}, {110} and such as very high charge carrier {111} facets. Characterising the mobility, which could be conveniently honeycomb superlattice with high- modifi ed by an electric fi eld. In addition, angle annular dark fi eld (HAADF) courtesy of its semiconductor building scanning transmission electron blocks, the electronic properties can microscopy (STEM) combined with be further tuned by the quantum dot electron diff raction reveals that the composition and size, making the <111> axis of the nanocrystals is “artifi cial graphene” very interesting for perpendicular to the substrate and that applications in the electronics industry the structure is single crystalline, having [3]. all nanocrystals aligned similarly (Figure 121a,b). There are three models which fulfi l this condition: a trigonal model, where the nanocrystals attach via {110} facets, a tetrahedral model where they attach via {111} facets and an octahedral model where they attach via {100} facets (see Figure 121c,d).

Fig. 121: Single-crystalline PbSe honeycomb structures created by means of oriented attachment. a) HAADF-STEM image of the honeycomb structure (bright on a dark background). The equilateral triangle shows the long-range ordering of the structure. Scale bar, 50 nm. b) High-resolution HAADF-STEM image showing that the 〈111〉 NC axes are perpendicular to the honeycomb plane and three of the 〈110〉 axes are perpendicular to the (c) and (d) models of the honeycomb structure, with truncated cubes as nanocrystals. The two inequivalent sites in the honeycomb lattice are indicated by yellow/red and blue/green NCs. c) and d) are the top and side view of the octahedral structure, respectively.

104 ESRF STRUCTURE OF maTERIaLS

These models appear similar from the top, but have a very diff erent 3-dimensional structure. The trigonal structure is fl at, exactly like graphene, whereas the tetrahedral and octahedral structures are buckled comparable to the silicene structure.

The HAADF-STEM images combined with the electron diff raction show that the <110> axes are perpendicular to the superstructure bonds, ruling out the possibility of the trigonal structure (where the <110> axes are parallel to the nanocrystal bonds). In addition, Fig. 122: a) GISAXS pattern of a honeycomb these images show increased scattering structure on top of a Si(100) substrate strength on the nanocrystal bonds, obtained at ID01. b) Line trace in the indicating a larger-than-average horizontal direction indicated in (a). Red lines mark diff raction peaks with thickness of the structure at these relative positions of 1 : √3 : 2 : √7 arising positions which originate from the out- from the hexagonal symmetry of the of-plane bonding of the honeycomb honeycomb lattice. structure. Scanning tunnelling microscopy (STM) measurements showed that the honeycomb structure bond length. The obtained bond length is indeed buckled. However in STM, of 95o ± 5o shows that the honeycomb the topography is a complicated structure is of the octahedral type, with convolution of the superlattice bonds between {100} facets. The bond topography and the tip shape, so no length was found to be 13% larger quantitative height data could be than the original nanocrystal diameter, extracted. A next-nearest neighbour indicating neck formation between the distance of nanocrystals at the same nanoparticles in the superlattice. height could reliably be extracted, which on average is 8.5 ± 0.8 nm. It has been predicted that honeycomb semiconductors of zinc blend To further study the long-range order, compounds (such as CdSe) show a truly grazing incidence small-angle X-ray new electronic band structure. The PbSe scattering (GISAXS) experiments were honeycomb structure is changed into performed at ID01 (Figure 122a,b). The CdSe via a cation exchange reaction, scattering pattern shows diff raction which has been confi rmed by means of rods arising from the hexagonal energy dispersive X-ray spectroscopy. symmetry of the honeycomb structure, Combining this with HAADF-STEM and with the expected relative positions of electron diff raction showed that the Se 1 : √3 : 2 : √7. The extracted hole-hole anion lattice is preserved with respect distance of the honeycomb structure, to the honeycomb periodicity. corresponding to the next-nearest neighbour (NNN) distance, is 8.5 nm, These results show that the honeycomb which is in agreement with the STM lattice, which is buckled and of measurements. octahedral symmetry, can be made on a liquid surface, and can subsequently To distinguish between the tetrahedral be transformed into a material of and octahedral model, electron interest via cation exchange. A next tomography was performed. Slices step in this research will be the in situ through the reconstructed tomogram measurements of the formation of this perpendicular to the substrate remarkable type of crystal. indicate that the nanocrystals in the honeycomb unit cell are located References on diff erent heights. Subsequently, [1] W.H. Evers et al., Nano Lett. 13, 2317 automated particle detection was (2013). [2] M.P. Boneschanscher et al., Science performed on a region of interest to 344, 1377 (2014). extract important parameters, such [3] E. Kalesaki et. al., Phys. Rev. X 4, as bond angle, bond length and NNN 011010 (2014).

The European Synchrotron 105 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

Principal publication and authors Z. Ristanović (a), J.P. Hofmann (a), CRYSTALLOGRAPHIC ARCHITECTURE OF U. Deka (a), T.U. Schülli (b), M. Rohnke (c), A.M. Beale (a) and B.M. Weckhuysen (a), Angew. Chem. Int. A SINGLE ZEOLITE ZSM-5 CRYSTAL Ed. 52, 13382 –13386 (2013). (a) Inorganic Chemistry and Catalysis, Utrecht University (The Netherlands) Zeolites represent a very important crystals have been used in numerous (b) ESRF class of natural and synthetic micro-spectroscopic studies due to their (c) Institute of Physical Chemistry, alumininosilicate materials with a wide well-defi ned structure and suitable size. Justus-Liebig-University Giessen range of applications including ion- They consist of at least 6 intergrown (Germany) exchange, adsorption and catalysis. subunits (Figure 123) that may diff er The size and topology of zeolite cages in their crystallographic orientation and channels is determined by a making the crystal properties highly highly crystalline 3-D porous network anisotropic. Therefore one of the that consists of the Si-O tetrahedra. intriguing questions related to the Substitution of al for Si atoms in structure and reactivity of zeolite ZSM- the crystalline structure leads to the 5 crystals is the precise orientation formation of acidic sites that can of the porous network in individual catalyse a broad range of chemical subunits and their impact on uptake reactions. This particular combination and diff usion of guest molecules. This of Al-related acidic sites and shape- question has been mostly studied selective properties defi ned by their using a number of optical microscopy narrow pores (< 0.8 nm) makes zeolite techniques providing only indirect catalysts a workhorse in numerous evidence of the crystallographic large-scale industrial processes. orientation of the subunits.

Zeolite ZSM-5 is one of the most Using highly focused X-rays at ID01, abundantly used solid acid catalysts we conducted a micro-X-ray diff raction in oil refi ning and petrochemical (µ-XRD) crystallographic imaging industries. To understand the catalytic study of a single zeolite ZSM-5 crystal. performance of this material, large Individual zeolite ZSM-5 crystals were coffi n-shape zeolite crystals (100 × mapped using an X-ray beam (8.5 keV) 20 × 20 µm3) have frequently been focused down to a size of 1 × 2 µm2. To investigated as model systems [1]. The minimise the footprint of the beam, we recorded higher order Bragg refl ections in the range of 2θ > 70° (Figure 123a). The a and b lattice parameters of zeolite ZSM-5 are almost identical, which enables reliable identifi cation of the crystalline phases based on distinct splitting of the higher order (16 0 0) and (0 16 0) refl ections. Using those refl ections, we further reconstructed diff raction intensity maps with lateral resolution of 2 µm and determined crystallographic orientation of the individual subunits. The results indicate 90° intergrowth structure of a zeolite ZSM-5 crystal, meaning that the top and bottom subunits have a 90° rotated crystal lattice as compared to the remaining subunits (Figure 123b). This implies that the direction of the straight and sinusoidal pores changes depending on the subunits, therefore having an impact on uneven uptake of guest molecules.

Fig. 123: a) Schematic of the µ-XrD imaging approach. b) Intergrowth structure of the zeolite ZSM-5 crystal and the crystallographic orientation of each subunit (colour-coded).

106 ESRF STRUCTURE OF maTERIaLS

The high angular and spatial resolution of the technique also enables subtle changes in the lattice parameters to be resolved. It was possible to map the distortion of the crystal lattice spatially along the crystal x and y axes (Figure 124). The results point towards a very small, but detectable 0.02 Å expansion of the unit cell from the centre of the crystal towards the edges. This is explained by a macroscopic gradient in the concentration of Al atoms and consequently the heterogeneous distribution of acidic sites. As a result of this gradient, inner regions of the zeolite crystal are Si-rich, while Al atoms mostly concentrate in the outer crystalline layer near the surface – a phenomenon that to the best of our knowledge has never been crystallographically substantiated (Figure 124b).

With the method demonstrated here, it is possible to look at the crystalline structure not only of a single zeolite crystal – but also at any advanced crystalline material where structural and chemical gradients play an important role. In addition, hard X-ray diff raction provides an ideal tool to study chemical processes in situ, where one could benefi t from the well- defi ned model systems, such as zeolite ZSM-5 crystals. Using a combination Fig. 124: a) Fine structure of the reconstructed single (x, y) pixel of µ-XrD and a wide range of micro- References diff ractograms, measured for the [1] L. Karwacki, M.H.F. Kox, spectroscopic techniques, it is now (16 0 0) and (0 16 0) Bragg refl ections. D.A. Matthijs de Winter, M.R. Drury, possible to elucidate the structure- b-e) Corresponding µ-XRD intensity maps J.D. Meeldijk, E. Stavitski, W. Schmidt, of the Bragg peaks integrated for the performance relationship for numerous M. Mertens, P. Cubillas, N. John, et al., 2θ regions as indicated in (a). processes that take place within Nat. Mater. 8, 959–965 (2009). functional materials.

GENERATION OF SURFACE STEPS ON Principal publication and authors O. Balmes (a), G. Prevot (b), X. Torrelles (c), E. Lundgren (d) and Pt(9 7 7) INDUCED BY THE CATALYTIC S. Ferrer (c,e), Journal of Catalysis 309, 33–37 (2014). OXIDATION OF CO (a) ESRF (b) Institut des NanoSciences de Paris, UMR CNRS 7588, Université Pierre et Atoms at surfaces have fewer This simple idea was proposed in Marie Curie-Paris 6 (France) neighbours than atoms in the bulk 1925 and since that time it has been (c) Institut de Ciència de Materials, CSIC, Bellaterra (Spain) (typically 9 vs 12) and consequently investigated extensively. Atoms at (d) Department of Synchrotron Radiation they have “unsaturated” chemical steps have been found to dissociate Research, Institute of Physics, Lund bonds. Atoms at surface steps (see molecules such as N2, CO or molecules University (Sweden) Figure 125) have even more “dangling” with C-C bonds; in most cases this (e) ALBA Light Source, Cerdanyola del bonds (seven neighbours only) and due has been determined using ultra-high Valles (Spain) to this, they might be key actors in vacuum techniques. However, chemical surface chemical reactions in catalysis. reactions are more complicated

The European Synchrotron 107 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

References than the adsorption and dissociation we were able to monitor the evolution [1] R. van Rijn et al., Rev, Sci. Instrum. process and many studies have set out of the concentration of steps at the 81, 014101 (2010). to question the simple picture of “more surface during the reaction. step atoms implies more reactivity”. This was the case for several reports We used a specially designed reactor

dealing with CO oxidation to CO2. [1] that allows the diff racted intensities and the reaction product to be measured We investigated the oxidation of CO at simultaneously while the reaction

beamline ID03 using a single crystal 2CO + O2 = 2CO2 was occurring. surface of platinum as the catalyst. The proportion of the reactants was This crystal has a periodical array of measured with the experimental

steps (see Figure 125), which generates variable R = PCO / PO2, the ratio of the specifi c diff raction features. partial pressures. R = 2 corresponds to the stoichiometric proportion and Figure 125b shows a portion of the R < 2, R > 2 to oxygen rich or oxygen diff racted intensity. The diff use lines defi cient proportions. The total

at H = 16 and 18 designated as CTR pressure PCO + PCO2 = 200 mbar and (crystal truncation rod) arise from the the reaction temperature was around

step array, whereas the intense peaks 490°C. The amount of CO2 produced (red colour) arise from the diff raction was monitored with a gas spectrometer. of bulk atoms, of no interest here. The diff use lines marked {111} indicate We found that, at a given temperature,

that the crystal surface also has regions the production of CO2 was maximum for with (111) symmetry, therefore not all R = 2 and that the maximum reactivity areas are stepped as depicted in the was associated with a maximum fi gure. Regions with (111) density of steps. Values of R other than

symmetry occur because 2 resulted in lower CO2 production and the 977 surface is not fewer steps in the Pt surface. Some of very stable and in some these observations are presented in places the steps bunch Figure 126. together generating (111) microfacets. These results illustrate that the morphology of the catalyst changes By measuring the with the reaction conditions. When intensities of the CTR and the reactivity is high, the number of of the {111} diff use lines, steps is also high, whereas when the reactivity is low, the number of steps decreases and the proportion of (111) microfacets increases. The catalyst is Fig. 125: a) Side view of the not a passive spectator of the reaction, 977 surface showing step and rather the contrary, it transforms terrace atoms. b) A portion of itself to optimise the reaction rate reciprocal space diff raction pattern. The diff use CTR lines by increasing its proportion of step arise from the monoatomic atoms. Thus, the atoms at the steps do steps and the {111} diff use appear to be a key ingredient for a high inclined line from (111) reaction rate. microfacets.

Fig. 126: Temporal variation of the proportion of the reactants (upper panel) and the corresponding changes in reciprocal space (lower panel). The intensity at H = –23 arises from steps and that at ~ –22.7 from (111) microfacets. At the start of the experiment (R = 2, stoichiometric mixture), the intensity from the monoatomic steps is pronounced (red indicates high intensity) and it fades out when the gas mixture is made CO rich, while the intensity from the facets that was initially low (green) clearly develops. When the gas mixture is set back to the stoichiometric proportion (near t = 5000 s), the intensity from the steps increases again (yellow segment), while that of the facets strongly decreases. Subsequent change to an oxygen rich mixture (t ≅ 10,000 s) results again in a decrease in the intensity from the steps in favour of the facets. Finally, the intensity from the steps reappears when the gas mixture is returned to the stoichiometric value.

108 ESRF STRUCTURE OF maTERIaLS

Principal publication and authors POSITIVE THERMAL EXPANSION OF F. Jean (a,b), T. Zou (c), N. Blanc (a,b), R. Felici (d), J. Coraux (a,b) and G. Renaud (c), Phys. Rev. B 88, 165406 GRAPHENE ON ITS SUBSTRATE (2013). (a) Université Grenoble Alpes, Inst. NEEL, Grenoble (France) Graphene (a single sheet of graphite) The transition between the two regimes (b) CNRS, Inst NEEL, Grenoble (France) is expected to exhibit unusual – graphene locked and unlocked on its (c) CEA-UJF, INAC, SP2M, Grenoble (France) mechanical properties related to its substrate – gives rise to a hysteretic (d) ESRF atomic thickness. In a free-standing change of graphene’s lattice parameter. form, graphene is predicted to have a Remarkably, aft er becoming unlocked negative thermal expansion below a few from the substrate, graphene locks 100 K, i.e. its lattice parameter should onto it again, adopting another, well- decrease as temperature increases defi ned epitaxial relationship. At (see ref. [1] and references therein). high temperature, this relationship This behaviour is associated with the approaches a coincidence of 21 carbon activation of out-of-plane vibrations atoms with 19 iridium atoms, which is when graphene is in a free-standing of high order, while at low temperature, form. Whether such vibrations are the relationship is of lower order, with still active when graphene rests on a 10 carbon atoms matching 9 iridium substrate is unknown. To shed some atoms (Figure 128). light on this matter, we have studied graphene grown in situ on iridium. Graphene is known to have a weak Fig. 127: Graphene lattice interaction with this substrate. parameter (aGr, left axis) as a function of temperature. The diff erent symbols refer We performed a series of in situ to diff erent samples prepared temperature-dependent high resolution under the same conditions. surface X-ray diff raction experiments, The lattice parameter of from 10 to 1300 K, at beamlines BM32 the substrate (aIr, right axis, same steps as left axis) as and ID03, with high quality graphene. a function of temperature The graphene was prepared in situ on is also displayed as a solid a (111)-terminated iridium crystal green line. The dotted green by chemical vapour deposition under is the same curve shift ed to ultra-high vacuum [2] to minimise the pass through the graphene data points. The cartoons contribution from contaminants. show the appearance/ disappearance of wrinkles We found that graphene on its in graphene, which merely substrate exhibits no negative thermal compresses and expands in expansion coeffi cient in the broad their absence. temperature range that we explored (Figure 127). The weak graphene- Fig. 128: Ratio of the graphene metal interaction indeed forces to iridium lattice parameters as a graphene’s lattice parameter to follow function of temperature. Dotted lines the temperature dependence of the highlight low- and high-order lattice metal lattice parameter. coincidences (commensurabilities).

The lattice properties of the two materials are distinctive, and this shows up as graphene becomes unlocked from the metal lattice at a certain threshold, corresponding to a situation when the in-plane compressive (upon cooling down) elastic energy overcomes the sum of the binding and bending energies of graphene on iridium. References Graphene forms linear delaminations: [1] Y. Magnin, G. D. Förster, F. Rabilloud, F. Calvo, A. Zappelli and C. Bichara. J. wrinkles several micrometres-long and Phys.: Condens. Matter 26, 185401 a few nanometres-high that can be (2014). observed with a microscope. [2] J. Coraux, A.T. N’Diaye, C. Busse and T. Michely. Nano Lett. 8, 565 (2008).

The European Synchrotron 109 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

Principal publication and authors T. Rosenthal (a), P. Urban (b), MULTINARY THERMOELECTRICS – A K. Nimmrich (a), L. Schenk (a), J. de Boor (c), C. Stiewe (c) and O. Oeckler (b), Chem. Mater. 26, CHALLENGE FOR CRYSTAL STRUCTURE 2567-2578 (2014). (a) Department of Chemistry, LMU DETERMINATION Munich (Germany) (b) Faculty of Chemistry and Mineralogy, Leipzig University (Germany) The conversion of thermal into X-ray fl uorescence and ∆f’(l) can be (c) German Aerospace Center, Cologne electrical energy is an important goal derived from it. The scattering contrast (Germany) for better energy effi ciency by means between neighbouring elements in the of waste-heat recovery. Modern periodic table can be enhanced by up to thermoelectric materials render 8 electrons [2]. this conversion possible and gained worldwide interest because of their Beamline ID11 is ideally suited to high- high ‘thermoelectric fi gures of merit’ resolution single crystal diff raction ZT, which depend on the thermal because of its stable beam and small (κ) and electrical (σ) conductivity, band pass. A fl uorescence detector the Seebeck coeffi cient (S) and the for the determination of dispersion temperature (T): ZT = (S2σ/κ)T. State- correction terms as well as a furnace of-the-art bulk materials oft en contain for high-temperature measurements a mixture of elements with very low are integral parts of the setup used scattering contrast like Pb and Bi or to refi ne the element distribution and

Ag, Sn, Sb and Te [1], which oft en share structural changes in Ge2Sb2Te 5 as atomic sites in the corresponding a function of temperature [3]. This crystal structures. These elements can material is well known as a fast phase- reliably be distinguished by the eff ect change material for data storage. Its of anomalous dispersion at absorption stable phase consists of tetradymite- edges, where the dispersion correction like distorted rocksalt-type slabs which term Δf’ is large and strongly modifi es are terminated by Te-atom layers the atomic form factor f(l,θ) = f 0(sinθ/l) and separated by van der Waals gaps + ∆f’(l) + i∆f’’(l). These correction (Figure 129). The element distribution terms can be determined directly at changes very little upon heating, but the beamline: ∆f’’(l) is calculated from instead the GeTe-like layers detach reversibly from the slab. In addition, this class of materials turned out to exhibit very promising thermoelectric properties which led to many substitution experiments to further enhance the effi ciency of energy conversion.

Ge~5Sb2( Te 0.13Se0.87)~8, for example, crystallises in a highly disordered rocksalt-type structure. TEM investigations revealed that the vacancies tend to order in parallel non- equidistant defect planes (111). This involves fourfold twinning and stacking disorder, which is also corroborated by diff use streaks in the diff raction patterns obtained at beamline ID11 (Figure 130). Semi-quantitative information on the real structure can be drawn by comparing the diff raction patterns of crystals with diff erent

compositions (GeTe)nSb2(SexTe 1-x)3

Fig. 129: Room-temperature structure of Ge2Sb2Te 5 (left ) with refi ned element distribution (top right); the principal trend of the dispersion correction terms is sketched (middle right); bottom right: SEM image of the crystal investigated.

110 ESRF STRUCTURE OF maTERIaLS

Fig. 130: Reciprocal lattice section (hhl) from a crystal of Ge~5Sb2( Te 0.13Se0.87)~8; note the diff use scattering running along <111>* (l/3 refl ections are marked with asterisks).

References and simulated patterns. The most refi nement on fi ve datasets (three [1] L.-D. Zhao, V.P. Dravid and promising thermoelectric properties at the K-edges of Sn, Sb and Te, one M.G. Kanatzidis, Energy Environ. Sci. 7, (ZT = 1.2 at 425°C) were found for further off -edge synchrotron and one 251-268 (2014). n = 7 and x = 0.2. Furthermore, the laboratory dataset). [2] S. Welzmiller, P. Urban, F. Fahrnbauer, L. Erra and O. Oeckler, J. refi nement of the atom distribution in Appl. Crystallogr. 46, 769-778 (2013). long-range ordered layered compounds In conclusion, the use of anomalous [3] P. Urban, M.N. Schneider, like Ge0.6Sn0.4Sb2Te 4 [4] yields clear dispersion is a unique way to reliably L. Erra, S. Welzmiller, F. Fahrnbauer and statements concerning the most determine structures of promising O. Oeckler, CrystEngComm 15, 4823- probable local structures in disordered thermoelectric materials that contain 4829 (2013). [4] S. Welzmiller, T. Rosenthal, P. Ganter, compounds. Occupancy factors on sites several elements with a low scattering L. Neudert, F. Fahrnbauer, P. Urban, shared by Sn, Sb and Te, impossible contrast. In situ investigations and C. Stiewe, J. de Boor and O. Oeckler, to distinguish using laboratory X-ray the refi nement of site occupancies of Dalton Trans. 43, 10529-10540 (2014). sources, could be determined in a joint mixed sites are easily possible.

DISCOVERY OF AN UNEXPECTED LI-ION Principal publication and authors X. Zhang (a), M. van Hulzen (a), D.P. Singh (a), A. Brownrigg (b), BATTERY CHARGING MECHANISM J.P. Wright (b), N.H. van Dijk (a) and M. Wagemaker (a), Nano Letters 14, 2279-2285 (2014). The revolution of personal electronics at high potential in electrode materials (a) Department of Radiation Science and Technology, Delft University of and rise of electronic mobility is made such as LiFePO4, making lithium-ion possible by the increasing energy batteries at least three times more Technology (The Netherlands) (b) ESRF densities of Li-ion batteries. Lithium, powerful than the traditional nickel- which is the third lightest element, cadmium and nickel metal hydride is a relatively small ion with a very varieties. One of the main restrictions strong tendency to give away its outer of today’s Li-ion batteries is the poor electron. As a consequence, large kinetics of the phase transitions taking numbers of lithium ions can be stored place in the electrodes upon Li-ion

The European Synchrotron 111 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

insertion and extraction. The impact of defects, particle size and charging rates on the phase transitions in Li- ion battery electrode materials have already been studied intensively, but in situ X-ray diff raction experiments at beamline ID11 have now allowed

us to study the chemistry of LiFePO4 batteries during very fast cycling rates for the fi rst time. This eff ectively made it possible to watch lithium atoms enter and exit the electrodes in real time. We have observed that at high charge and discharge rates, the electrode phase transitions can be bypassed – thereby enhancing the transport of lithium ions through otherwise poorly Li-ion conducting electrode material.

Electrode materials for Li-ion batteries

like LiFePO4 store lithium via a reversible fi rst-order phase transition

between LiFePO4 and FePO4, which is associated with a reduction in the a-axis lattice parameter. Diff raction during slow charging (5 hours) of the battery demonstrates this phase transition by the gradual disappearance of the

{200} refl ection of LiFePO4 and the Fig. 131: Top: 2D fi gure displaying the time appearance of the {200} refl ection evolution of the {200} refl ection during of FePO4, shown in the top part of slow charging (5 hours). Bottom: 2D fi gure Figure 131. In contrast, during fast displaying the time evolution of the {200} refl ection during fast charging (1 minute). charging (1 minute) the {200} refl ection gradually shift s from the Li-rich phase

position (LiFePO4) towards the Li-poor position (FePO4) showing a considerable intensity between the two end positions, indicating a diff erent phase transition mechanism. The intensity

distribution between the FePO4 and LiFePO4 {200} refl ections implies that the phase transition progresses via a distribution of lattice parameters,

gradually converting LiFePO4 to FePO4. This demonstrates that the dynamic fast charging conditions induce a solid solution transition rather than the thermodynamically predicted fi rst-order phase transition with a

Fig. 132: Model for the phase distribution

in LiFePO4 electrodes at high and low rates and upon relaxation. At a low rate of 5 hours (C/5) the electrode is near equilibrium and no signifi cant Li concentrations/overpotentials exist throughout the electrode. At high charge rate (60C corresponding to 1 minute charging) the potential gradient locally causes the bypassing of the fi rst-order phase transition and the appearance of the solid-solution compositions that are observed.

112 ESRF Structure of materials

phase coexistence of two well defined Interestingly, the observed bypassing individual phases. of the first-order phase transition is accompanied by a decrease in The in situ fast charging diffraction internal resistance of the battery. This indicates that only a small fraction of demonstrates that the solid-state the electrode is actively transforming, charge transport can be enhanced consistent with the existence of a Li- by the external charge conditions. ion concentration/potential gradient. Against intuition, this may indicate The consequence is that the electrode that charging and discharging at material localised at the gradient higher rates may be beneficial for a experiences an overpotential, absent battery’s efficiency and cycle lifetime. near equilibrium conditions, that drives In addition, we have learned that the the solid-solution transformation. This key to even faster charging is the Li- condition moves like a wave through the ion transport in the electrolyte located electrode as illustrated in Figure 132, in the pores of the electrode, which we transforming the complete electrode. can improve by smart design of the Interrupting the fast charging will allow electrode morphology. Currently, we the system to relax towards equilibrium are analysing the phase transitions in where, in contrast, the phase transition individual electrode grains in different appears to occur randomly throughout battery systems under varying the electrode. From previous research charge/discharge conditions, aiming [1], we know that the occurrence of to better understand the nucleation References a reacting phase front at high rates mechanisms, and the development of [1] D.P. Singh et al., Advanced Energy indicates that the Li-ion transport in the improved high-performance energy Materials 3, 572-578 (2013). electrodes limits faster (dis)charging. storage in batteries.

Mechanochemical synthesis Principal publication and authors I. Halasz (a), A. Puškarić (a), S.A.J. Kimber (b), P.J. Beldon (c), of pharmaceutical cocrystals A.M. Belenguer (c), F. Adams (d), V. Honkimäki (b), R.E. Dinnebier (d), monitored by real-time in situ B. Patel (c), W. Jones (c), V. Štrukil (e) and T. Friščić (e), Angewandte Chemie, powder X-ray diffraction International Edition 52, 11538-11541 (2013) (a) Ruđer Bošković Institute, Zagreb (Croatia) Mechanochemical synthesis, i.e. of solubility, solvolysis or thermal (b) ESRF chemical synthesis by milling or instability, inherent to other cocrystal (c) Department of Chemistry, University grinding, is a leading methodology discovery strategies. of Cambridge (UK) for environmentally-friendly, ‘green’ (d) Max-Planck-Institute for Solid State Research, Stuttgart (Germany) synthesis. However, one of the most Despite its broad synthetic potential (e) Department of Chemistry and established and valuable applications of and the widespread presence of milling FRQNT Centre for Green Chemistry and milling is screening for new solid forms in industrial processing, mechanisms of Catalysis, McGill University (Canada) of active pharmaceutical ingredients mechanochemical reactions are almost (APIs): cocrystals, polymorphs or salts unknown. This is most likely due to a [1]. Discovery of new cocrystals is one general inability to directly observe the of the principal pursuits in modern course of chemical reactions and phase drug development, as cocrystals transformations. Until recently, the only represent an exciting opportunity to means to follow changes in composition improve physicochemical properties of mechanochemical reactions has of the API (solubility, bioavailability, been by stepwise analysis, i.e. by tableting) and represent tremendous periodically interrupting the milling value in intellectual property process and analysing the reaction development. The application of mixture. Such an approach is, however, mechanochemistry in screening for very poorly informative or even cocrystals is based on the ability of misleading as the samples transform neat and liquid-assisted grinding (LAG) before analysis due to relaxation, to rapidly explore cocrystal formation recrystallisation, desolvation or simply without interference from problems moisture absorption.

The European Synchrotron 113 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

In collaboration with ESRF archetypal pharmaceutical cocrystal scientists, we have recently consisting of carbamazepine and developed the fi rst technique saccharin (Figure 133). Reaction for direct, real-time and in monitoring demonstrated their situ monitoring of structural amorphisation upon neat milling, and chemical transformations while milling in the presence of a during milling using catalytic liquid additive (liquid-assisted X-ray powder diff raction. grinding, LAG) led to rapid cocrystal This technique was fi rst formation observable by changes in introduced for monitoring the intensities of characteristic refl ections transformations of strongly for each phase. This revealed a diff racting inorganic oxides clear diff erence between neat and into metal-organic materials liquid-assisted approaches for solid- [2]. state molecular complexation, and demonstrated the unexpectedly high In continuation, we extend rate of mechanochemical reactions, the experimental design to leading to complete conversions in 2 monitor mechanochemical minutes (Figure 133). complexation of poorly scattering organic solids, We also explored the mechanochemical as encountered in the cocrystallisation of nicotinamide

formation of pharmaceutical (vitamin B3) with suberic acid, a model cocrystals. The required pharmaceutical system known to signal-to-noise ratio was encompass materials with diff erent achieved by performing the stoichiometric composition. In this experiment in a milling vessel system, in situ cocrystallisation made of non-crystalline, monitoring revealed complex, multi- poorly scattering Perspex step mechanisms involving short- (poly(methyl)metacrylate). lived and metastable materials not The fi rst application of observable by other means. Milling this new technique for cocrystallisation of nicotinamide monitoring pharmaceutical and suberic acid in a 2:1 respective cocrystallisation was on an stoichiometric ratio yielded the expected cocrystal over 40 min, but only through the formation of an intermediate cocrystal containing Fig. 133: a) Fragment of the crystal structure of the the two components in a 1:1 ratio. pharmaceutical cocrystal of Replacing the slower neat milling carbamazepine and saccharin. with much faster LAG revealed the b) Time-resolved X-ray formation of the fi nal product cocrystal diff ractogram for the LAG within about 5 minutes. Again, the cocrystallisation of saccharin and carbamazepine. c) reaction reaction involved the 1:1 cocrystal as course obtained via Rietveld an intermediate which, this time, lasted analysis. only for 3 minutes (Figure 134).

Monitoring of the LAG reaction also revealed another short-lived intermediate appearing about 1 minute

Fig. 134: Time-resolved diff ractogram for LAG cocrystallisation of suberic acid with nicotinamide, indicating a multi-step mechanism involving the formation of diff erent cocrystals.

114 ESRF STRUCTURE OF maTERIaLS

into milling. This novel intermediate cocrystals, immediately demonstrating has never previously been observed the unexpected speed of the process, in studies of this well-known model possibility of multi-step reaction pharmaceutical system and, although it mechanisms, as well as the formation could not be characterised structurally, of intermediate phases that have in situ monitoring now allowed it previously not been observed. The latter to be prepared under conventional is a particularly exciting development, References laboratory conditions. as it demonstrates the value of in situ [1] A. Delori, T. Friščić and W. Jones monitoring not only as a tool for the CrystEngComm 14, 2350-2362 (2012). [2] T. Friščić, I. Halasz, P. J. Beldon, The presented methodology provided academic investigation of reaction A. M. Belenguer, F. Adams, the fi rst in situ and real-time insight mechanisms, but also for the discovery S. A. J. Kimber, V. Honkimäki and into the course of mechanochemical of new, potentially commercially r. E. Dinnebier Nature Chem. 5, 66-73 formation of pharmaceutical valuable, solid forms of APIs. (2013).

TIME-RESOLVED X-RAY DIFFRACTION Principal publication and authors D. Ferri (a), M.A. Newton (b), WITH ENHANCED SENSITIVITY M. Di Michiel (b), G.L. Chiarello (c,d), S. Yoon (c) Y. Lu (c) and J. Andrieux (b,e), Angew. Chemie Int. Ed. 53, 8890–8894 X-ray diff raction (XRD) is a powerful time resolution is still an issue. High (2014). (a) Paul Scherrer Institut (Switzerland) method to determine the structure and energy XRD is suitable to circumvent (b) ESRF the structural changes of materials. An the inherent problems of XAS but is still (c) Empa, Swiss Federal Laboratories advantage of XRD is that it allows the limited by its intrinsic detection limit. for Materials Science and Technology observation of all XRD visible phases Metal nano-particles are visible by XRD (Switzerland) simultaneously within a complex only if dispersion is low and the metal (d) Present address: Università degli Studi di Milano (Italy) multiphase material. When using content high. Typically, detection is (e) Present address: Université de Lyon synchrotron radiation, the information limited to ca. 3 nm particle size for high (France) is obtained in a snapshot, i.e. with high loaded metals. time resolution. In the case of supported precious metal based catalysts, used at beamline ID15B, we have used for example to control the emissions high energy XRD in combination with of automotive engines, XRD provides a modulation approach and lock-in information on the precious metal amplifi cation based on phase sensitive particles and the metal oxide support(s). detection (PSD) [1] to follow redox Chemical reactions at the heart of processes on technologically relevant the abatement of pollutants (NOx, CO materials. A 2 wt% Pd/Ce0.6Zr0.4O2 and hydrocarbons) occur mainly on catalyst is subjected to repeated and the precious metal nano-particles. equally long pulses of CO and O2 at 573 K The function of the metal oxide is to provide stability against severe reaction conditions. For example, CeO2 is responsible for maintaining metal nano-particles in an oxidised state, thus retarding their growth which otherwise leads to loss of the catalytic function. Strongly X-ray absorbing materials like CeO2 are diffi cult to analyse by X-ray absorption spectroscopy (XAS) with high time resolution. On the one hand, the strong absorption of X-rays by the metal oxides hinders detection of the low loaded active metals (Pt, Pd, Rh) and fl uorescence based detection is oft en required. As a result, high

Fig. 135: Time-resolved XRD data. a) Colour map representation. b) Temporal responses of the XRD refl ections. c) Phase-resolved data.

The European Synchrotron 115 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

while XRD acquisition is performed set entails only an intensity change at 500 ms/pattern. A combined XRD/ with equally reversible behaviour. The DRIFTS/MS setup was used for this position of this sub-set of refl ections purpose [2]. Aft er averaging, the set coincides with the XRD pattern of fcc of time-resolved XRD patterns was Pd. Estimation of the crystallite size analysed by PSD, which provided a new by the Scherrer formula provides Pd dataset containing information only on particles of <2 nm that become visible the species that changed dynamically only aft er application of PSD. This size during the modulation experiment. estimate is supported by simulation of particles in the range 1-5 nm. The Figure 135 shows that the time- temporal behaviour of the two datasets resolved XRD data contain little or no demonstrates that evolution of the

information on any structural change structural change within Ce0.6Zr0.4O2 induced by the modulation experiment. and reduction-oxidation of the Pd Features that could be associated with component occur quasi simultaneously. Pd or PdO nano-phases are absent as Also, the function of Ce3+/Ce4+ becomes expected from the low metal loading clear in the appearance of the Pd(111) and the absence of long range order refl ection in the PSD data. This shows that are at the base of the peculiar that reduction of Pd is retarded until catalytic activity of such nano-entities. the very end of the reducing pulse as In marked contrast, the corresponding a result of the intimate contact of the

phase-resolved data obtained by Ce0.6Zr0.4O2 and Pd phases. PSD show two sub-sets of refl ections signifying that structural changes Therefore, we have demonstrated that occurred during the experiment. a greater detail of the redox kinetics of The fi rst one is characterised by a a complex material can be obtained by References diff erential profi le (both by intensity [1] D. Baurecht, U.P. Fringeli, Rev. Sci. the combination of high energy XRD, Instr. 72, 3782 (2001); D. Ferri, and 2θ variation) and is associated the modulation approach and PSD. M.A Newton, M. Di Michiel, S. Yoon, with expansion and contraction of the Subtle changes associated with metal G.L. Chiarello, V. Marchionni, Ce component upon reduction of Ce4+ nano-particles can be made visible S.K. Matam, M.H. Aguirre, (in CO) and oxidation of Ce3+ (in O ). using a bulk analysis method. Such A. Weidenkaff , F. Wen and J. Gieshoff , 2 PCCP 15, 8629 (2013). The very small (0.11%) and reversible enhanced insights should be achievable [2] M.A. Newton, M. Di Michiel, change of d spacing is associated in all systems that involve functional A. Kubacka and M. Fernàndez-Garcià, J. with the capacity of CeO2 to exchange materials where reversible cycling is Am. Chem. Soc. 132, 4540 (2010). oxygen with the environment under possible for one or more of the intrinsic reducing conditions. The second sub- process variables.

Principal publication and authors REVEALING RAPID DYNAMICS IN WELD W.U. Mirihanage (a), M. Di Michiel (b), A. Reiten (a), L. Arnberg (c), H.B. Dong (d) and R.H. Mathiesen (a), SOLIDIFICATION Acta Materialia 68, 159–168 (2014). (a) Department of Physics, Norwegian University of Science and Technology, Welding is considered the most In fusion welding, metals coalesce via Trondheim (Norway) important and economical material extremely rapid solidifi cation processes (b) ESRF processing route for joining metals. which are critical to the structural (c) Department of Materials Science and Engineering, Norwegian University Welds are employed extensively to integrity of the weld and oft en the origin of Science and Technology, Trondheim fabricate components and structures of catastrophic weld failures. Enhanced (Norway) ranging from tiny electronic chips to insight on weld solidifi cation is vital (d) Department of Engineering, massive ships and even for components to prevent losses of lives and property University of Leicester (UK) operated under critical and demanding [1], yet the inherent complexity, high conditions such as in nuclear reactors. temperatures and the rapidness of the welding processes make direct experimental investigations extremely demanding [2].

Polychromatic X-ray diff raction has Fig. 136: Schematic been used in time resolved studies experimental setup. of rapid solidifi cation microstructure

116 ESRF STRUCTURE OF maTERIaLS

evolution in situ during spot welding Fig. 137: a) Columnar at beamline ID15A. austenitic steel (c) crystal volume samples 2 mm thick were placed with evolution, c1-c5 the surface normal tilted 0.035 rad from denotes 5 diff erent the vertical axis towards an incident crystals, and favg presents the overall 50-150 keV white beam, and defi ned by solid volume fraction. upstream apertures to a cross section b) time-evolution of of 1 x 0.1 mm2 (H x V) at the sample. angular velocities Weld pools of ~5.0 mm diameter and for crystals c1-c5. <1.5 mm depth were formed in a few c) evolution of growth and tilting velocities seconds by focused radiation from for one representative halogen lamps. Solidifi cation was columnar crystal. initiated by powering off the lamps, and completed within 1.5 s. 2D data was collected in refl ection geometry at 1.0 kHz frame rate with a PCO. tips. As illustrated in dimax® CMOS equipped with an image Figure 137c, when intensifi er, see Figure 136. growth and tilting rates at the columnar In the chosen experimental geometry, tips are compared, scattered intensities stemmed from crystal growth the topmost ~140 µm of the weld pool, velocities and tilting with the most prominent growth mode velocities are found detected corresponding to columnar to be of the same dendritic grains growing in the surface- order of magnitude, near layer from the pool periphery especially in the towards the weld centre. In addition to earlier stages of crystal growth, a remarkable angular solidifi cation. Thus, motion of diff raction spots were solute pile-up ahead observed at the early growth stages, of the columnar front typically 3 - 4 orders of magnitude should be eff ectively larger than that associated with thermal removed not only by contraction of the solid at cooling strong convection, rates of 3-400 K/s. Grain rotation is but also by tilting, presumably driven by substantial melt such that growth convection in the pool. It is well known at this stage should be substantially that very strong convection currents diff erent from situations where only may arise during welding due to the melt convection and diff usion are extreme thermo-solutal conditions that present. apply, and a crude calculation indicate that fl ow velocities up to ~ 0.4 m/s can Experimental data both in terms of be expected in the fully molten pool for individual grain growth and motion are the case reported here. novel in the case of welding or similar rapid solidifi cation processing, and The data have been processed in will be very useful for benchmarking of terms of integrated diff raction spot more realistic simulation models. Grain intensities and angular positions to tilting has not been incorporated into allow for information to be extracted present simulation models, although on the growth of individual columnar its presence may be quite readily crystals, overall solid fraction recognised. Thus, the experimental evolution and accompanying grain results serve to stress that both melt rotation dynamics (Figure 137 a, b). fl ow and liquid to solid momentum Correlation between the temporal transfer have to be addressed and evolutions in integrated intensities, accounted for in future theoretical grain motions and 2D diff raction work and simulation models for spot areas pointed towards columnar welding. Furthermore, quite similar crystal tilting at their attaching roots experimental setups could be employed References in the future for fast in situ XrD studies [1] S.A. David and T. DebRoy, Science as the predominant mode of grain 257, 497 (1992). motion, and, by further data analysis, of real welding processes or other fast [2] M. Yonamura et al., J. App. Physics the tilting motion could be expressed in phase transformations under extreme 107, 013523 (2010). terms of metric velocities at the crystal conditions.

The European Synchrotron 117 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

Principal publication and authors A. Zorko (a, b) O. Adamopoulos (c), ON THE INHOMOGENEOUS M. Komelj (a), D. Arčon (a, d) and A. Lappas (c), Nat. Commun. 5, 3222 (2014). GROUND STATE OF A TRIANGULAR (a) Jožef Stefan Institute, Ljubljana (Slovenia) ANTIFERROMAGNET (b) EN—FIST Centre of Excellence, Ljubljana (Slovenia) (c) Institute of Electronic Structure and In condensed matter, phase separation The investigated compound α-NaMnO Laser, Foundation for Research and 2 breaking the translational symmetry is well known for its intriguing Technology—Hellas, Heraklion (Greece) (d) Faculty of Mathematics and Physics, of a Hamiltonian on a local scale is magnetic and structural properties [2]. University of Ljubljana (Slovenia) particularly intriguing, as it regularly It orders magnetically at TN = 45 K, but appears in chemically homogeneous diff use neutron magnetic scattering, systems [1]. This is related to some characteristic of short-range spin fundamental functional properties correlations, remains present down of materials, such as colossal to the lowest temperatures. The magnetoresistance of manganites, coexistence of such scattering with giant electrostriction of relaxors, and sharp magnetic Bragg peaks is a

possibly even high-Tc superconductivity clear indication of a magnetically [1]. According to a conventional inhomogeneous ground state intrinsic

paradigm, electronic charge plays to α-NaMnO2. Moreover, large and the leading role in promoting phase- highly anisotropic microstrains in the

separated states when competing crystal structure above TN suggest phases are present. Also, quenched structural inhomogeneities. Based on disorder – random deviations from the analysis of the neutron powder perfect system’s uniformity – is diff raction patterns, the microstrains generally required for the stabilisation were reported to grow with decreasing of such states. However, in spin systems temperature that should eventually that are geometrically frustrated, a result in a monoclinic-to-triclinic novel mechanism leading to spatial structural phase transition, occurring instabilities could be at play even simultaneously with the magnetic in the absence of charge degrees of ordering [2]. freedom and in the clean limit. In such spin lattices, geometrical constraints We have studied the alleged structural

prefer exotic degenerate ground states changes in α-NaMnO2 at beamline ID31 over conventionally ordered magnetic (now ID22). The enhanced resolution states. We have identifi ed a novel kind of of the synchrotron X-ray diff raction phase-separated state in the triangular compared to previous neutron studies

antiferromagnet α-NaMnO2, where the has revealed that a bulk structural interplay of geometrical frustration transition is impeded, in sharp contrast

and a structural instability of the to the isostructural CuMnO2 compound, lattice leads to a magnetostructural where such magnetic-order-induced inhomogeneity at the nanoscale. improper ferroelastic transition indeed

occurs at TN = 65 K (Figure 138). A characteristic symmetry-lowering

peak splitting is missing in α-NaMnO2, although the refl ections exhibit sizeable anisotropic broadening with decreasing temperature (Figure 138). Moreover, a peculiar asymmetric

line shape developed below TN that cannot be adequately modelled by the high-temperature monoclinic crystal structure (Figure 138). Regularly

Fig. 138: Upper panel: Bragg-peak splitting witnesses the

presence (CuMnO2) and absence (α-NaMnO2) of a structural phase transition. Lower panel: Rietveld fi t of the synchrotron

X-ray powder diff raction (XRD) profi le of α-NaMnO2 in the inhomogeneous phase, with the monoclinic (“m”) only and a combined monoclinic and triclinic (“m + t”) models.

118 ESRF STRUCTURE OF maTERIaLS

observed structural defects, such as twin boundaries and stacking faults, fail to account for the measured peak- shapes. Therefore, a more complex parameterisation of the structural model was required. A two-phase model incorporating the monoclinic and triclinic phase off ers a signifi cantly improved Rietveld refi nement (Figure 138). This implies the presence of randomly distributed nanoscale inhomogeneities, pertaining to the minority triclinic phase that grows at the expense of the majority monoclinic phase.

Such a nanoscale structural inhomogeneity is further supported by Fig. 139: Left : Illustration of the an electronic inhomogeneity. The latter magnetostructurally inhomogeneous has been witnessed in α-NaMnO2 by our ground state of α-NaMnO2. Right: Local local-probe magnetic investigations, structural deformation from the monoclinic incorporating both nuclear magnetic (m) to triclinic (t) symmetry pertinent to the defect (sphere) regions. resonance (NMR) and muon spin relaxation (mSR) techniques. These have unambiguously demonstrated the system’s tendency to remove a nanoscopic coexistence of two magnetic degeneracy associated diff erent magnetic environments with the geometrically frustrated that can be explained by a model of triangular spin lattice (Figure 139). triclinic magnetic defects in the form The associated magnetic-exchange of dispersed nano-sized islands, with energy gain due to the monoclinic-to- the magnetic order reversed at the triclinic distortion is compensated by interfaces with the monoclinic matrix the increase of the elastic energy. The (Figure 139). resulting near-degenerate monoclinic References [1] E. Dagotto, Science 309, 257–262 and triclinic structures, as suggested by (2005). We suggest that the newly discovered ab initio calculations, can then lead to [2] M. Giot et al., Phys. Rev. Lett. 99, magnetostructural phase separation nanoscale phase separation even in the 247211 (2007). in α-NaMnO2 is a consequence of limit of zero quenched disorder.

In situ MONITORING OF “JUMPING Principal publication and authors M.K. Panda (a), T. Runčevski (b), CRYSTALS” BY X-RAY POWDER S.C. Sahoo (a), A.A. Belik (c), N.K. Nath (d), R.E. Dinnebier (b) and DIFFRACTION P. Naumov (a), Nature Communications 5, 4811 (2014). (a) New York University Abu Dhabi, Abu Dhabi (United Arab Emirates) Certain crystals, when taken over a This extremely rare phenomenon of (b) Max Planck Institute for Solid State phase transition, are capable of self- “jumping crystals” was noticed in the Research, Stuttgart (Germany) actuation, jumping, exploding, bending 1980s and remained poorly understood (c) National Institute for Materials or changing their physical appearance, ever since. Etter and Siedle reported that Science, Tsukuba (Japan) thus presenting a novel platform when crystals of (phenylazophenyl) for clean and eff ective conversion of palladium hexafl uoroacetylacetonate thermal energy to mechanical work (PHA) are heated, they “literally fl y [1–3]. These thermosalient crystals, in off the hot stage” [4]. This impressive the near future, might also fi nd other dynamic eff ect is captured in applications, such as smart medical Figure 140. Although this system devices and implants, artifi cial muscles, could help to understand how ordered actuators, sensors, electromechanical matter responds to extreme internal devices, to name but a few. pressures and performs locomotion, the

The European Synchrotron 119 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

References relationships between the structural, gamma, which is accompanied by [1] Ž. Skoko, S. Zamir, P. Naumov and mechanistic and kinematic aspects of jumping of the crystals, and gamma– J. Bernstein, J. Am. Chem. Soc. 132, this system have remained unknown. to–beta, which is not thermosalient. 14191 (2010). [2] S.C. Sahoo, S.B. Sinha, M.S.R.N. Kiran, Determination of the crystal structure of U. Ramamurty, A. Dericioglu, M.C. Reddy When single crystals of PHA jump, they form gamma was crucial to understand and P. Naumov, J. Am. Chem. Soc. 135, increase their mosaicity and disintegrate, the jumping eff ect in PHA. This phase 13843 (2013). thwarting the crystal structure solution was found to exist in a triple mixture [3] S.C. Sahoo, M. Panda, N.K. Nath by single crystal diff raction. The with alpha and beta (Figure 141). and P. Naumov, J. Am. Chem. Soc. 135, 12241 (2013). macroscopic disintegration of single Moreover, the relative amounts of the [4] M.C. Etter and A.R. Siedle, J. Am. crystals is all but an obstacle for powder phases were found to change under Chem. Soc. 105, 641 (1983). diff raction (XRPD), and this method isothermal conditions. was used to investigate the system in greater detail. A laboratory XRPD study When dealing with crystal structures revealed as many as 5 polymorphs of metastable species, it is necessary between 98 and 388 K (a record for that the data are collected at the same the largest number of polymorphs of (or preferably lower) timescale relative an organometallic compound). Our to the lifespan of the metastable crystallographic, thermogravimetric, structures. Laboratory XRPD data, of microscopic and spectroscopic analyses suitable quality for ab initio crystal revealed that when heated, crystals of structure solution, requires hours for form alpha undergo two subsequent collection, thus conventional X-ray phase transformations: alpha–to– sources could not be applied to study this system. The high-resolution diff raction data collected at beamline ID31 (now ID22), with a time-resolution of 3 minutes per powder pattern, made this complicated system amenable to structural investigations (Figure 141).

Fig. 140: Snapshots (taken at frequency The crystal structures of forms alpha 103 s–1) of the kinematic eff ect of hopping of PHA crystals. and beta were already known [4]. The solution of the structure of form gamma provided us with a unique possibility to follow the structural changes on the alpha-gamma-beta phase transitions and to relate them to the thermosalient eff ect. During the phase transition accompanied by jumping, layers of molecules in form alpha slide atop each other, which results in slanting of the columns of stacked molecules that generates immense internal strain, which is released in form of forceful jumps of the crystals and phase transformation into form gamma. In the second phase transition, alternating layers of molecules in form gamma slide in opposite directions and interdigitate into the thermodynamically stable head-to-tail stacking pattern of form beta.

The detailed crystallographic study of this intriguing “jumping crystal” phenomenon was only possible by using Fig. 141: Rietveld plot of a three-phase high-intensity and high-resolution refi nement of PHA polymorphs used to synchrotron radiation. The PHA research detail the crystal structure of the gamma project once again demonstrated the form, which is relevant to the thermosalient behaviour. The crystal packing diagrams of necessity of acquiring high-resolution forms alpha, gamma and beta are given in diff raction data in frontier studies of the lower panels. novel materials.

120 ESRF STRUCTURE OF maTERIaLS

Principal publication and authors PUSHING CRYSTALLOGRAPHY ACROSS C. Tyrsted (a), N. Lock (a), K.M.Ø. Jensen (a,b), M. Christensen (a), E.D. Bøjesen (a), THE NUCLEATION BARRIER H. Emerich (c), G. Vaughan (d), S.J.L. Billinge (b) and B.B. Iversen (a), IUCr-J 1, 165-171 (2014). The century-old technique of X-ray (methanol-based solution at T = 275oC (a) Center for Materials Crystallography, diff raction continues to become and P = 250 bar), dissolved Zr4+ and Department of Chemistry, and iNANO, Aarhus University (Denmark) ever more important for materials y3+ species are transformed into (b) Department of Applied Physics characterisation, laying the foundation crystalline Zr1-xyxO2-d nanoparticles and Applied Mathematics, Columbia for many scientifi c and technological in a matter of minutes. Visually, the University, New York (USA) advances in academia and industry system may be observed to change (c) SNBL, ESRF alike. The rapid progress in X-ray from a translucent solution to an (d) ESRF diff raction methods has made it apparent gel-like structure upon possible for researchers to routinely nucleation, steadily densifying with identify known materials and solve prolonged hydrothermal processing. the structures of previously unknown Therefore, three diff erent structural compounds. The key requirement being stages likely exist during diff erent that the material exhibits a crystalline points in this process: 1) pre-nucleation long range ordering of the atomic molecular species, 2) a disordered lattice, the prerequisite for Bragg amorphous network and 3) an ordered diff raction. Digging into the world of crystal lattice. The experimental PDF’s nanoscience, the coherent atomic representing these diff erent structural domains of materials become ever stages are shown in Figure 142. Only smaller, challenging the condition for the third structural stage exhibits Bragg diff raction. This has meant that Bragg diff raction, the fi rst two stages crystallography was oft en disregarded giving rise to diff use scattering alone, in the structural characterisation of excluding the use of standard Bragg amorphous, liquid and extremely crystallography. Nonetheless, total nanocrystalline systems. However, this scattering clearly proves the existence perception is quickly changing, partly of distinct local atomic structuring in owing to the further development all three stages spanning further than and growing use of total scattering the fi rst immediate coordination shell. experiments and pair distribution Fig. 142: Local function (PDF) analysis [1]. atomic ordering as revealed by total In material science, it is a grand scattering. (d) Time- challenge to understand the atomic resolved view of ordering of materials during synthesis. the local structural region (0–10 A˚) of By knowing the mechanism of crystal the PDF. formation it may be possible to control it, thereby tailoring desired materials. This is especially a keen interest in the fi eld of solvothermal synthesis [2], where crystalline materials oft en form from dissolved chemical species or otherwise disordered gel systems. While the evolution of the growing crystalline particles themselves may be followed by diff raction, a barrier is met when trying to understand the local structures existing at the point of nucleation or the stage(s) preceding it.

In this study, total X-ray scattering in combination with X-ray absorption spectroscopy was used in an attempt Fig. 143: to further the understanding of the Structural stages formation of yttria-stabilised zirconia observed during yttria-stabilised zirconia nanoparticles during solvothermal formation. (a) Proposed solution species structure. (b) Amorphous network structure. processing. Under these conditions (c) Mature cubic crystalline structure of YSZ.

The European Synchrotron 121 SCIENTIFIC HIGHLIGHTS Structure of materials

The structures found to best describe far from the cubic lattice structure of these structural stages are shown in the end product, yet lacks the correct Figure 143. The local structure found long range topology. Over time, bond in the pre-nucleation stage is proposed reforming ensures that the local to consist of polymeric zirconia domains order within themselves species, formed during dissolution or and in relation to each other thereby present in the solid precursor from the forming the desired cubic lattice beginning. At the point of nucleation, structure. the data indicate that the polymeric species are rapidly fragmented into Numerous simplified models exist smaller species which cluster together to describe crystal nucleation and References and precipitate out as an amorphous growth phenomena. The present [1] T. Egami and S.J.L. Billinge. network. The atomic network of the study provides the first experimental Underneath the Bragg peaks: structural precipitate is not entirely randomised atomistic insight into the very complex analysis of complex materials. Vol. 16. despite the apparent chaotic nucleation chemical processes that take place Newnes (2012). [2] K.M. Jensen, C. Tyrsted, M. Bremholm process, but has a local ordering which during crystallisation, and the total and B.B. Iversen, ChemSusChem 7, may be described as monoclinic in scattering technique provides a new 1594-1611 (2014). nature. The local bonding environment approach to truly understanding of the precipitate is therefore not too nanomaterial formation.

Principal publication and authors Illuminating the physical process A.M. Beale (a,b,c), E.K. Gibson (b,c), M.G. O’Brien (a), S.D.M. Jacques (c,d), R.J. Cernik (d), M. di Michiel (e), of catalyst deactivation P.B. Cobden (f), O. Pirgon-Galin (f), L. van de Water (g), M.J. Watson (g) and B.M. Weckhuysen (a), J. Catal. 314, 94 Synchrotron-based X-ray diffraction which sulfur poisons these catalysts (2014). computed tomography (XRD-CT) is is still up for debate. It is thought that (a) Inorganic Chemistry and Catalysis, a technique developed at the ESRF sulfur, in the form of either H2S or Debye Institute for Nanomaterials for the purpose of revealing detailed similar compounds, accumulates on Science, Utrecht University (The Netherlands) chemical compositional information the surface of the metallic Cu species (b) Department of Chemistry, University in solid materials in 2 and 3 spatial blocking the active sites or else induces College London (UK) dimensions [1]. Much of this ‘chemical a phase transformation of the Cu (c) UK Catalysis Hub, Research Complex imaging’ has focused on the study of and ZnO into the respective sulfides at Harwell, Rutherford Appleton pre-shaped catalyst bodies often as a and oxy-sulfides. Since catalysts Laboratory, Didcot (UK) (d) School of Materials, University of function of time. Here we demonstrate are used in pre-shaped forms, i.e. Manchester (UK) how the technique has been used to as millimetre-sized catalyst bodies, (e) ESRF understand the importance of spatial to understand deactivation in real (f) Energy Research Centre of the variation during the process of sulfur- catalyst samples needs consideration Netherlands (ECN), Petten (The induced solid-state deactivation of low of the spatial perspective in order to Netherlands) (g) Johnson Matthey Technology Centre, temperature Cu/ZnO/Al2O3 water-gas assess macroscopic effects leading Billingham (UK) shift catalysts. to catalyst deactivation. XRD-CT is ideal for assessing crystalline phase Catalyst deactivation by sulfur distribution at the µm scale, whereas poisoning is a common problem for XAFS/XRD mapping can be used to a wide range of catalytic processes determine the relative importance of and which is likely to become more poisoning via sulfur adsorption vs. problematic as the chemical industry solid-state transformation of the active moves from relatively clean feeds, Cu/ZnO phase(s) into sulfide phases. such as natural gas and crude oil, to

alternative feedstocks, such as bitumen Sulfur, in the form of H2S, was added tars and some forms of biomass. Cu/ to the catalytic feed during the WGS ZnO catalysts represent the industrial reaction and the results of the XRD-CT standard for the production of methanol data recorded as a function of increasing and the water–gas shift (WGS) reaction. sulfur concentration are given in The catalysts are however very Figure 144. The initial crystallline susceptible to sulfur poisoning (ZnO Cu and ZnO phases disappear with

for example is often used as a ‘sulfur increasing [H2S] in the feed and both scrubber’). However the process by phases lose intensity and ‘shrink’

122 ESRF STRUCTURE OF maTERIaLS towards the core of the body. New sulfi de phases evolve on the outside, while the increasing [H2S] leads to a thicker sulfi de shell. Consequently, the thicker the sulfi de shell, the greater the loss in catalytic activity. In the sample poisoned using 500 ppm H2S, over 80% of the cylinder volume is converted into sulfi de-containing phases and almost all catalytic activity is lost.

Fig. 144: 2D XRD intensity colour maps for Whilst XRD-CT is able to identify the various crystalline phases (top) aft er two distribution of crystalline components concentrations of H2S (left ) were introduced over the 2D cross-section, it is in the WGS reactant feed. Right – colour insensitive to the presence of small scale map. amounts of adsorbed sulfur species on the surface of metallic Cu, which may block the active site. For this purpose, Fig. 145: Derived composition a chemical map derived from the fi tting data from the Cu-‘O’ bond distance map of a microtomed of the Cu-NN (nearest neighbour) cross-section of a Cu/ZnO/ contribution from XAFS spectroscopy al2O3 catalyst body treated on a section of the microtomed catalyst with 200 ppm H2S. Right – is given in Figure 145. The main Cu colour scale map. nearest neighbour component is < 2 Å at the core and > 2.1 Å towards the edge, consistent with the presence of CuO at the core and CuS at the edge.

There was no clear evidence for the amount of H2S used in this study, presence of either Cu-S or Zn-S phases the tendency for Cu and Zn to adsorb References in the centre of the sample [2]. S does mean that, over time, sulfi de [1] A.M. Beale, S.D.M. Jacques, phases will accrue in much the same E.K. Gibson and M. Di Michiel, Coord. The formation of sulfi de phases at the way as observed here. Clearly it is Chem. Rev. 277, 208 (2014). sample edges is highly detrimental to important to consider the spatial [2] A.M. Beale, E.K. Gibson, M.G. O’Brien, the catalytic activity of the sample component when trying to determine S.D.M. Jacques, R.J. Cernik, M. di Michiel, P.B. Cobden, O. Pirgon-Galin, since the catalyst deactivates before the structure-activity relationships in L. van de Water, M.J. Watson and all Cu/ZnO has been lost. Although real heterogeneous catalyst samples B.M. Weckhuysen, J. Catal. 314, 94 industrial Cu/ZnO/Al2O3 catalysts with imaging methods critical for such (2014). are unlikely to see anything like the purposes.

SHORT-RANGE CORRELATIONS IN Principal publication and authors A. Bosak (a), D. Chernyshov (b), M. Hoesch (c), P. Piekarz (d), MAGNETITE ABOVE THE VERWEY M. Le Tacon (e), M. Krisch (a), A. Kozłowski (f), A.M. Oleś (e,g) and TEMPERATURE K. Parlinski (g), Phys. Rev. X 4, 011040 (2014). (a) ESRF (b) SNBL, ESRF Magnetite, Fe3O4, is the fi rst magnetic is viewed as an ideal candidate for room- (c) Diamond Light Source, Didcot (UK) material discovered and utilised by temperature spintronic applications. It (d) Institute of Nuclear Physics, Polish mankind in Ancient Greece, yet it still crystalises in the inverse spinel cubic Academy of Sciences, Kraków (Poland) attracts attention due to its puzzling structure, with two types of Fe sites: the (e) Max-Planck-Institut für Festkörperforschung, Stuttgart properties. Discovered in the fi rst half tetrahedral A sites and the octahedral (Germany) of the twentieth century, the Verwey B ones. at TV = 124 K, a fi rst-order (f) AGH-University of Science and transition in magnetite [1] remains phase transition occurs as the electric Technology, Kraków (Poland) one of the most intriguing phenomena conductivity drops by two orders of (g) Marian Smoluchowski Institute of in solid-state physics. Magnetite is a magnitude with the simultaneous Physics, Jagellonian University, Kraków (Poland) ferrimagnet with an anomalously high change of the crystal structure from the

Curie temperature TC = 850 K. Hence, it cubic to monoclinic symmetry and with

The European Synchrotron 123 SCIENTIFIC HIGHLIGHTS STRUCTURE OF maTERIaLS

References spectacular anomalies in practically [1] E. J. W. Verwey, Nature 144, 327 all physical characteristics. The low- (1939). temperature structure of magnetite, [2] A. Faber, in Spectroscopy Techniques, B. Richards, H. Evans (Eds.), Oxford as deduced from recent studies, was University Press, New York, 1-34 (1999). identifi ed to be of monoclinic Cc space group symmetry, with a complex displacement pattern [2]. This suggests that electron localisation, responsible for the dramatically increased resistivity, appears in basic three-Fe- site units, called “trimerons”.

With the use of a state-of-the-art large area detector (PILATUS 6M), a detailed three-dimensional (3D) reciprocal space mapping can be performed, revealing an extremely rich diff use scattering pattern (Figure 146), inherited from the complex low-temperature structure below the Verwey transition. Even at room temperature, which is considerably higher than the 124 K transition temperature, electrons appear to be correlated over relatively large length scales (Figure 147). The characteristic length estimated from the width of diff use features varies from ~2 unit cells (u.c.) of the prototype

cubic structure at TV+2.5 K to a value Fig. 146: Isosurface representation of slightly larger than ~1 u.c. at room diff use scattering in magnetite slightly temperature. Thus, the ordering pattern above TV. Colour represents the distance cannot be reduced to the trimeron to the (000) node; diff use clouds in the features, but rather to complexes proximity of weak Bragg spots have been removed. The half-space above the H0L of trimerons. Therefore, our study plane has also been removed. White circles supports the polaron picture, and we mark strong Bragg refl ections in the H0L can state that its structure is in reality plane, and arrows denote HK0 and HK4 cuts much more complex than ever expected perpendicular to the image plane. previously. We were able to associate some characteristic experimental features with the topology of the Fermi surface of magnetite. While the structure of the low-temperature phase was frequently considered as a key to the understanding of the Verwey transition, our work indicates that the diff use scattering pattern above the transition actually encodes some key information.

Fig. 147: Diff use scattering in magnetite at variable temperature. Magnetite reciprocal space cuts HK0 (a) and HK4 (b) are shown at room temperature (RT),

slightly above (TV + d) and below (TV − d) the Verwey transition, with d ~ 2.5 K. Left -hand bottom panels: Sections perpendicular to the cell doubling direction. Right-hand bottom panels: sections parallel to the cell doubling direction. Selected regions of interest (see text) are shown in (c) together with the unit

cell vectors a* and b*. Cubic m3m (above TV) and tetragonal 4/mmm (below TV) Laue symmetries have been applied.

124 ESRF Enabling technologies Highlights 2014

By definition, enabling technologies crystallography beamlines. The third provide the means to generate article emphasises the impact of significant leaps in performance our long-standing effort to develop and capabilities of our facility and generic solutions for beamline therefore underpin every aspect of instrumentation. Similarly the fourth the ESRF’s scientific activity. In this article recalls the importance of context, 2014 saw the concomitant accurate modelling to optimise the execution of Phase I of the Upgrade performance of light converters that Programme and the development of can be integrated into high spatial Phase II. Together with the facility resolution X-ray detectors. The fifth operation, these two deliverables article describes the helium recovery represented a real challenge in network, which has a significant terms of resource management and financial impact on the use of liquid scheduling. On the one hand several helium at the beamlines, thus making UPBLs went through various stages it a sustainable operation. The last of completion, while on the other article describes the implementation hand an in-depth reflection on the of the new Internet addressing future scientific instrumentation scheme IPv6 for all institutes on and accelerator technology was the EPN Science Campus enabling carried out. The majority of the Internet connectivity compliant ESRF beamlines are expected to reap and performing with countries and immediate performance benefits laboratories having adopted this new from the improved source and in standard. many cases, novel instrumentation will be needed. The Orange Book R. Dimper and J. Susini provides a comprehensive overview of the ESRF’s strategy for developing future beamline instrumentation and infrastructure, driven by the new storage ring’s capacities.

A successful enabling technology programme is not limited to scientific instrumentation and in fact involves a very broad range of engineering areas. The content of this chapter on “Enabling Technologies” exemplifies this variety and touches a multitude of topics, all key in the current and future operation of the ESRF. The constantly growing sophistication of state-of- the-art instrumentation together with the very high expectations of our User Communities requires a holistic strategy involving many different factors that together determine the overall performance and reliability of the facility. All ESRF divisions, and in particular the two technical divisions, ISDD and TID, collaborate together to deploy state-of-the-art technologies taking into account the challenging trade-off between performance, cost, and maintenance effort.

The two first articles illustrate the impact that new cryo-cooling techniques and automation have on the new macromolecular

The European Synchrotron 125 HIGHLIGHTS 2014 ENaBLING TECHNOLOGIES

Principal publication and authors P. van der Linden, F. Dobias, H. Vitoux, HIGH PRESSURE COOLING OF U. Kapp, J. Jacobs, S. Mc Sweeney, C. Mueller-Dieckmann and P. Carpentier, J. Appl. Cryst. 47, 584 (2014). BIOLOGICAL CRYSTALS WITHOUT ESRF CRYOPROTECTION

Most diff raction data for biological crystals down to 77 K under 2 kbar crystals is recorded at cryogenic of helium pressure (Figure 148a) temperatures to reduce the secondary [1]. During this thermodynamical radiation damages and thereby extend process, the crystals are frozen in the life-time of samples in X-rays. their non-protected mother liquor However, crystals need to be cryo- which is directly transformed into protected prior to being cryo-cooled high density amorphous (HDA) to avoid solvent crystallisation and its ice under the eff ect of pressure, subsequent spoiling of the experimental avoiding water crystallisation and its results: the loss of diff raction orders destructive eff ects. Advantageously, and of diff raction quality. To a lesser the cryoprotectant-free HDA matrix extent, the cryo-protection process limits the augmentation of the also has its own undesirable eff ects: mosaicity which results from the (1) crystals are slightly destabilised LDA-ice expansion, eliminating the during the short soaking passage in unwanted eff ects that are associated the cryosolution and extremely fragile with the presence of cryo-agents, and crystals may dissolve regardless of the thus preserves crystals close to their cryo-agent type, (2) the vitrifi cation of native state. High pressure cooling has solvent into low density amorphous proven to be systematically applicable (LDA) ice at ambient pressure to all kinds of biological crystals produces a mechanical stress that regardless of the protein’s nature, the increases the mosaicity of crystals, solvent content or the crystallisation and (3) cryo-molecules may bind conditions. Moreover, the method has and disturb proteins possibly even recently been extended to cryopreserve at their active sites. The search for hydrated bacteria for imaging optimum cryoprotectant conditions experiments [2]. In addition, pressure which aim to limit these drawbacks is is a key thermodynamic parameter as so time-consuming for usually a poor important as temperature in studies improvement that crystallographers of macromolecules. Indeed, pressure oft en prefer to simply use glycerol. allows the energy landscape of proteins Therefore, we investigated alternative to be explored, providing access to cooling methods to eliminate the need diff erent conformational substates that for cryoprotection in an attempt to cannot be studied in any other manner circumvent this issue. High pressure (active site conformations or reaction cooling is one of the most eff ective of coordinates). Therefore, this technique these techniques and consists of fl ash- has potential to address many essential cooling cryoprotectant-free biological biological issues.

Fig. 148: a) Phase diagram of water. Unprotected crystals cooled at ambient pressure (AP-cooled) produce crystalline ice. At high pressure the solvent is transformed into HDA-ice. b) High pressure cooling installation at beamline ID23.

126 ESRF ENaBLING TECHNOLOGIES

Fig. 149: a) Ribbon representation of the structure of the HP-cooled protein FEA, largest structural displacements are represented in red (loop 132). b) Root mean square displacement of the HP-cooled FEA versus residue number. c) Crystallographic packing of FAE in surface representation (monomer/dimer).

Until now, users have been put off by this diffi cult approach reserved for specialists, which needs heavy instruments of low productivity to combine high pressures with cryogenic temperatures. We have designed an automated high pressure cooling instrument and developed a concomitant methodology and ancillary equipment, paying particular validated with a series of test crystals attention to the user-friendliness (lysozyme, insulin, thaumathin, and throughput (Figure 148b). FEA, proteinase-K and thermolysin). The computer controlled system is Following this, nearly a thousand theoretically capable of processing users’ crystals have been processed. 30 samples per hour. The throughput The resulting structures are relatively bottleneck now mainly depends on isomorphous to their counterparts in classical sample handling operations of the PDB, since the pressure induced crystallographers facilitated by a new structural changes are limited to fl exible pluggable SPINE compatible sample- loops at the proteins surfaces while their holder and its associated toolkit (i.e. structured cores are relatively pressure crystal fi shing, recovery and mounting/ insensitive. High pressure cooled FAE dismounting). Optionally, during the is a meaningful example, revealing a high pressure cooling procedure, the pressure induced modifi cation that pressurisation bench can be switched concerns only a single loop at the dimer to a noble gas circuit to produce xenon interface, but this simple modifi cation or krypton heavy atom derivatives results surprisingly in a superior for anomalous phasing purposes. In crystalline quality: a transition to a practice, the machine is capable of higher symmetry, a higher resolution, vitrifying nanolitres of aqueous solution lower B-factors and a lower mosaicity with only 5% glycerol for cryopreserving (Figure 149). dilute biological objects. Moreover, References the method is particularly eff ective The apparatus is installed in the [1] C.U. Kim, R. Kapfer and S.M. Gruner, for cooling biological crystals that are laboratory associated with beamline Acta Cryst. D61, 881 (2005). [2] E. Lima, Y. Chushkin, totally free of cryoprotectant since the ID23, and a high pressure cooling P. van der Linden, C. Un Kim, F. Zontone, ingredients of the mother liquor are service is now off ered to the ESRF P. Carpentier, S.M. Gruner, and P. Pernot, suffi cient to act as antifreeze agents. structural biology user community for Phys. Rev. E90, 042713 (2014). The methodology was successfully appropriate projects.

MXCuBE2 DEFINES A NEW PARADIGM IN Authors M. Oscarsson, M. Guijarro, A. Beteva, S. Monaco, O. Svensson, G. Leonard and MACROMOLECULAR CRYSTALLOGRAPHY D. de Sanctis. BEAMLINE CONTROL

Macromolecular crystallography (MX) robotics for sample handling, beamlines have been undergoing an development of high accuracy increasing evolution in automation, multiaxis diff ractometers, and faster with constant improvements in and more sensitive X-ray detectors

The European Synchrotron 127 HIGHLIGHTS 2014 ENaBLING TECHNOLOGIES

based on pixel technologies. This MXCuBE2 is built on a level of evolution has resulted in both an hardware abstraction that permits its increase in the productivity and adaptation to any kind of hardware an improvement in the quality of environment and its interfacing with diff raction data collected. Moreover, a variety of low-level control systems, the advent of microfocus beamlines thus making it compatible with most has made it possible to collect data components of any synchrotron- from microcrystals or from the best based MX facility. Like its predecessor, diff racting part of any crystal. However, MXCuBE2 off ers users the same developments in instrumentation have look-and-feel to its interface despite not been the only driving force behind controlling diff erent hardware, and it the increase in scientifi c productivity allows users to drive experiments from of MX beamlines: The simultaneous a single module that controls sample development of laboratory information handler, diff ractometer, detector and management systems (LIMS) [1], all other beamline components. In automatic pipelines for data collection this manner, it allows users to focus [2] and reduction [3], and soft ware on samples and experiments, without for beamline control has also been an having to have a specifi c knowledge essential contributor as it has allowed of how the hardware is going to carry researchers to perform increasingly out the instructions. Compared to complex experiments and to link that of MXCuBE, the appearance of these both to exhaustive metadata MXCuBE2 has been simplifi ed in order descriptions and to automated data to permit a closer interaction with the analysis. sample under study while, at the same time, providing a growing number of MXCuBE (Macromolecular new tools for experiment design and Crystallography Customised Beamline execution. In particular, MXCuBE2 Environment), is a beamline control provides an improved and more user- graphical user interface (GUI) for MX friendly environment for complicated experiments [4] that was initially multi-crystal/multi-position MX developed at the ESRF. The fi rst experiments performing them in a version of MXCuBE was released in modular, logical and automatic fashion. 2005, and since then it has become It also off ers both the possibility of the preferred data-acquisition performing routine experiments and soft ware for structural biologists at of accessing the advanced parameters many synchrotron sites, providing a and functionalities required for more platform for modern high-throughput complicated tasks through a single, MX that also allowed users to carry out unifi ed GUI. MXCuBE2 can thus be used experiments remotely. In particular, both to drive completely automatic the MXCuBE GUI provided the user data collection at the ID30A MASSIF with an intuitive means of interacting beamlines and, for example, in specifi c with beamline components by using experiments exploiting microbeams. graphical icons and visual indicators To facilitate the latter, MXCuBE2 allows rather than text-based interfaces. the storage of diff erent positions on a MXCuBE2 is a completely upgraded single sample that are then either used version of MXCuBE, its development for data collection with an X-ray beam and recent deployment has been of a defi ned size or to defi ne areas on driven by the new requirements of the sample that are to be characterised next-generation MX experiments at prior to full diff raction data collection. synchrotron sources. MXCuBE2 also facilitates experiments involving complex experimental phasing protocols, such as inverse beam geometry, interleaved multi- Fig. 150: (a) MXCuBE2 off ers improved interaction with the sample thus facilitating wavelength experiments and multi- advanced experiments, such as a helical positional data collection. The new scan in continuous mode (4dscan). tools available in MXCuBE2 also Diff raction data are collected while rotating include much-improved line and mesh and translating the crystal along the line scan protocols and, coupled with defi ned. (b) A new mesh drawing tool permits the drawing of a diff erent grid using improved motor control via ICEPAP the beam size as unit. Aft er analysis, scores modules, faster helical data collection are reported directly on screen. based on 4d scans [5] (Figure 150)

128 ESRF Enabling Technologies

and workflows for the collection of Ltd. Developments of the ISPyB References partial data sets from many samples LIMS [1] are keeping pace with the [1] S. Delagenière et al., Bioinformatics 27, 3186-92 (2011). contained on the same sample improved environment for experiment [2] S. Brockhauser, Acta Cryst. D68, support. execution that MXCuBE2 provides. As 975-984 (2012). the development of ISPyB will also [3] S. Monaco et al., J. Appl. Cryst. 46, MXCuBE2 is now developed as a joint soon be the subject of a collaboration 804-810 (2013). effort - the MXCuBE Collaboration - agreement between the ESRF and [4] J. Gabadinho et al., J. Synchrotron Rad. 17, 700-707 (2010). between the ESRF, other European other synchrotron sites, the future [5] D. de Sanctis et al., J. Synchrotron synchrotron sites [SOLEIL, EMBL@ of the MXCuBE/ISPyB combination Rad. 19, 455-461 (2012). PETRAIII, BESSY and MAXLAB] and as the software environment for MX the Cambridge-based computer experiments seems assured for many software company Global Phasing years to come.

Generic solutions ease the way Authors T. Mairs, R. Barrett, Y. Dabin, P. Fajardo, A. Goetz and P. Marion, representing for Phase I upgrade beamlines the many engineers and technicians of the ISDD who worked on the generic solutions. In 2009, the ESRF took on the contractors being identified early in ESRF challenge of designing, installing the development phase. and commissioning nine upgrade • The combined performance of the beamline projects (UPBL), renovating beamlines and accelerator can be four beamlines and moving three optimised due to the global view beamlines. Eighteen endstations of and know-how of the engineers and considerable complexity, in comparison technical staff over the whole ESRF with existing beamlines, were to be complex. completed in the six years of the upgrade Phase I using existing technical There are several minor disadvantages staff. A strategy of using standard and such as: generic building blocks to construct the • Longer design and production time beam delivery and detection systems for the first device of the generic yielded significant technical and family. financial advantages, leaving time to • Less freedom for beamline scientists develop ad hoc sample environment and engineers to choose their and manipulation instrumentation. individually-preferred and specifically tailored solutions. Advantages to generic solutions: • Technical staff are interchangeable Generic solutions at the ESRF between projects as the technology originated from the time of the adopted is similar leading to easier installation of the first beamlines in maintenance, lower ownership costs the 1990’s. Then, being the first third and quicker commissioning. generation synchrotron source, the • High performance is obtained thanks ESRF was obliged to develop a large to intense design and testing proportion of its instrumentation in- efforts invested to optimise and house as industrial suppliers had not characterise the first units of each enough experience to produce cutting- generic block, which then benefits all edge designs. This initial investment the devices using this block. in our in-house engineering and the • Design and development time early adoption of standard components is considerably shorter. A more such as primary and secondary slits, collaborative approach to design fluorescent screens, attenuators, work is fostered due to multiple use beamshutters, DPAP motor controllers, of the same technology. SPEC, TACO, Kohzu monochromators, • Cheaper and more efficient etc., allowed the first beamlines procurement due to increased to be installed and commissioned quantities of component building efficiently. As the source performance blocks with appropriate sub- improved and increased in power in

The European Synchrotron 129 HIGHLIGHTS 2014 ENaBLING TECHNOLOGIES

the late 1990’s, these initial, rather introduced for the particular beamline. basic, components, which were A typical example of this would be the essentially identical for all beamlines, ESRF UHV compatible beamviewers that were gradually replaced with higher have evolved from their initial concepts performance versions either for starting in 1993 to a stable design in standard devices (e.g. primary slits and 2012. These devices are particularly attenuators) or by generic variants that being used during the commissioning were tailored or adapted to individual phase of beamlines, and hence, large beamline requirements. quantities were foreseen for Phase 1 of the upgrade (Figure 151). a generic solution is considered to be an instrument that can be made up The beamviewers consist of the of building blocks which have already following building blocks: been used and tested beforehand with • Camera and focusing optics a number of relatively small changes • Scintillator and holder • Positioning device to bring the viewer into or out of the beam • Vacuum chamber • Soft ware to read, analyse and display the images

Basic imaging cameras are evolving quickly and the necessity to choose a standard camera for Phase 1 beamlines was identifi ed early on. An Ethernet GigE camera was chosen and implemented in a number of diff erent ESRF devices. The scintillator and its holder depend on the beam characteristics (white/pink/ monochromatic, size and position) and consequently each beamline is able to choose either a standard confi guration or one adapted to its needs. Two methods of positioning the device in the beam are available either pneumatically or using a stepper motor driven actuator. Integration of the device is specifi c to the beamline, Fig. 151: UHV Beam viewers. and hence, the vacuum chamber is designed on an ad hoc basis. The ESRF soft ware group is confronted daily with integrating images into beamline control so the beamviewers naturally benefi t from LIMA (Library for image acquisition) developed as a generic solution for reading cameras. a total of more than 50 devices from this family have been installed during the Phase 1 beamline programme.

The catalogue of generic solutions has continued to grow during Phase 1, which started with some high performance ESRF solutions ranging from opto-mechanics such as dynamic benders for KB mirror systems and precision jacks through more general mechanics such as hexapods and positioning technology (Figure 152). The generic approach extended to Fig. 152: Generic solutions. control instrumentation such as

130 ESRF ENaBLING TECHNOLOGIES

IcePAP motor controllers or MUSST transfocators, fi xed radius nano acquisition modules, Xray detectors focusing KB systems, white beam such as the FReLoN cameras and the mirror assemblies and multilayer fi rst generations of Maxipix devices. monochromators. and soft ware solutions with libraries • Mechanics: Improved hexapods. (PyFAI), applications ( MxCube, PyMCA), • Diagnostics and detectors: extended and complete toolboxes (TANGO). In range of focused beam viewers, new addition, all generic soft ware is open high resolution scintillator screens source and has been shared with a or CdTe based Maxipix detectors. large number of sites in Europe and the • Control electronics: continued rest of the world. improvements of IcePAP motor controllers and 100 MHz voltage-to- These generic solutions have been frequency converters. extensively used during the upgrade • Soft ware: TANGO, LIMA, MxCube, and are now complemented among PyFAI, PyMCA, e-motion, graphical others by: user interfaces, workfl ows, • Opto-mechanics: A range of white metadata and archive databases. beam and monochromatic beam

MODELLING OF THE MTF PERFORMANCE Authors F. Riva (a,b), P.A. Douissard (a), OF HIGH SPATIAL RESOLUTION X-RAY T. Martin (a) and C. Dujardin (b). (a) ESRF DETECTORS (b) ILM, Université Lyon 1-CNRS (France)

X-ray detectors for high spatial incident photons. Furthermore, the low resolution imaging are mainly based stopping power degrades the spatial on indirect detection. The detector resolution because of the energy consists of a converter screen made of a deposition spread in the converter scintillating material, light microscopy screen. optics and an imaging sensor (CCD or CMOS). The scintillating screen At the ESRF, GGG and LSO (lutetium converts part of the absorbed X-rays orthosilicate) scintillators are produced into a visible light image, which is by liquid phase epitaxy (LPE). projected onto the CCD by means of the Additionally, the development of new optics (Figure 153). The performance scintillating materials is currently of the detector is strongly infl uenced by being investigated: the goal is to the properties of the converter screen, improve the contradictory compromise which include X-ray absorption, spread between absorption and spatial of energy deposition, scintillation yield resolution, thereby increasing the and emission wavelength. detection effi ciency while keeping a good image contrast even at high To obtain detectors with micrometre energy. The development of a new and sub-micrometre spatial resolution, kind of scintillating thin (1-50 µm) single crystal fi lm (SCF) screen by LPE is time scintillators are required. consuming and expensive. Therefore, we developed At the ESRF, X-ray imaging applications a “simulation tool” able regularly make use of energies up to 120 keV. At relatively high energies, the effi ciency of the detector is reduced dramatically because of the weak absorption of the converter screen. Fig. 153: Schematic of a For example, at 20 keV a 5 µm thick detector for high spatial scintillator of gadolinium gallium resolution based on indirect garnet (GGG) absorbs only 10% of the detection.

The European Synchrotron 131 HIGHLIGHTS 2014 Enabling Technologies

References to predict the optimal combination of the scintillator. This means that part [1] S. Agostinelli et al., Nuclear of scintillating screen (in terms of of the visible image produced in the Instruments and Methods A 506, 250- thickness, composition and substrate) screen will be projected on the image 303 (2003). [2] H.H. Hopkins, Proc. R. Soc. Lond. A and visible light optics for different sensor as a defocused image, creating 231, 91-103 (1955). energy ranges. The aim of these a blur in the final image. To include this [3] Slanted Edge MTF: http://rsb.info.nih. simulations is to guide the choice of the effect, we applied analytical equations gov/ij/plugins/se-mtf/index.html materials that will be developed by LPE. describing the optics to the MTF of the converter screen obtained by Monte We combined a Monte Carlo approach Carlo simulations [2]. using Geant4 Monte Carlo toolkit [1] for the energy deposition leading We validated our model experimentally to light emission with analytical at beamline BM05, using the slanted calculations for the optical collection. edge method [3] to investigate the The energy spread in the converter spatial resolution of the detector screen is described by Monte Carlo in a variety of configurations. In simulations in which a monochromatic Figure 154a we show the excellent X-ray pencil beam hits the scintillator, agreement between experimental MTF generating secondary X-rays and and the simulated MTF obtained at electrons. To quantify the spatial 15 keV. The microscope optic has a resolution, we estimated from the numerical aperture of 0.4 and it was spatial distribution of the deposited combined with various scintillators: energy the linear spread function (LSF). LSO grown by LPE on YbSO substrate We also calculated the modulation (ytterbium orthosilicate), yttrium and transfer function (MTF) corresponding lutetium aluminium garnet (LuAG and to the LSF in the frequency domain: this YAG) bulk single crystals grown by the function gives the contrast as function Czochralski technique. The thickness of of the spatial frequency. the scintillators ranges from 11.8 µm to 500 µm. In Figure 154b, the same However the intrinsic MTF of the model is used to simulate the MTF converter screen does not describe at 55 keV: gadolinium aluminum the ultimate resolution of the detector perovskite (GAP), a candidate material completely: the visible light still needs for the development by LPE, is compared to be projected on a CCD by means of to LSO and GGG, both currently used LSO 11.8 µm on YbSO experiment microscope optics. The optics magnify at ESRF. Simulations show that GAP LuAG 25 µm experiment the image, reducing the real pixel size could improve the spatial resolution LuAG 50 µm experiment of the CCD, with the consequence that compared to the available screens YAG 500µm experiment LSO 11.8 µm on YbSO simulation it can degrade the resolution because in the energy range higher than the LuAG 25 µm simulation of the combination between the depth gadolinium k-edge. LuAG 50 µm simulation of field of the optics and the thickness YAG 500 µm simulation A model for indirect detection has 1 been developed and validated with (a) 0.8 Incident X−rays 15 keV experimental data. The agreement with Numerical aperture 0.4 experiment is very good and we are 0.6 confident that our model can be used

MTF 0.4 for optimisation of detectors, helping to estimate the spatial resolution for 0.2 different configurations of the detector 0 0 100 200 300 400 500 600 700 800 900 1000 for various experimental conditions. Additionally, it gives important information about the improvements GAP 5 µm on YAP simulation GGG 5 µm on GGG simulation that could be obtained regarding LSO 5 µm on YbSO simulation resolution and efficiency using a 1 new scintillator or a new detector Incident X−rays 55 keV (b) configuration as well as the crucial 0.8 Numerical aperture 0.4 role of the substrate when the ultimate 0.6 resolution is desired.

MTF 0.4

0.2 Fig. 154: Contrast as a function of the spatial 0 frequency (MTF): a) X-rays 15 keV, experimental 0 100 200 300 400 500 600 700 800 900 1000 data and simulations; b) X-rays 55 keV, Spatial frequency [lp/mm] simulations for different materials.

132 ESRF ENaBLING TECHNOLOGIES

Authors THE ESRF HELIUM NETWORK F. Favier, G. Guernet and T. Marchial. ESRF

Helium is the second-lightest element In 1996, the US passed a law which in the Universe (seven times lighter diminished the return on investment than air), it has the lowest boiling-point, of recycling helium; the law stated the lowest liquefaction temperature that the US National Helium Reserve, (-269°C) and, unlike any other element, which is kept in a disused underground helium will remain liquid down to gas fi eld in Texas – by far the biggest absolute zero at ambient pressures. Its store of helium in the world – must thermal conductivity and transmission all be sold off by 2015, irrespective of of speed of sound are superior to that the market price. The ESRF applies a of all other gasses (save hydrogen). standard practise for distribution and It is also the least water-soluble of all recovery of helium. ESRF benefi ts from gasses. Colourless, odourless and non- compressed gas storage facilities at toxic, helium is particularly chemically the ILL and the liquefaction services inert and ranks as the second least available at the CNRS and the CEA. reactive noble gas, aft er neon. The Liquid helium is supplied by the CEA or remarkable properties of this gas have CNRS in 100 dewars. Users connect the led to its widespread use in both liquid dewars to the equipment to be cooled and gas forms. (cryostats) then connect the gas outlet to the helium recovery network (50 mm In industry, helium is used to infl ate stainless steel tube), running for 1.5 km airships and weather balloons. It around the storage ring and to the ILL. provides a protective atmosphere in the manufacture of integrated An individual meter on each beamline circuits, screens, and in arc welding makes it possible to control the volume procedures. Helium can be used as a of gas transferred over the network. leak detector in vacuum equipment or This low pressure gas (about 30 mbar) in high pressure tanks, or to determine is fed back into the ILL system where water tightness in food packaging. In it is stored in a buff er tank (volume science, helium is an integral element 20 000 l) then compressed into bottles in cryogenic procedures for cooling or at about 200 bars. The compressed for experiments that require extremely gas is then transported to one of the low temperatures (supra conductors). It liquefaction centres in either the CEA is also used in optics for its refraction or CNRS, where it goes through a index and in the aeronautics industry, drying and purifi cation process at low laser systems and leisure activities temperature (77 K, -196°C) to separate (infl ating balloons). oxygen, nitrogen and water vapour. It is then re-liquefi ed (passage from gas Helium is an irreplaceable but non renewable gas. It is the most abundant element in the universe, second only to hydrogen, but it is much rarer on earth. There is no synthesising method capable of producing helium except for nuclear fusion. Helium is a sub-product of natural gas that contains it from some parts per million to several percent. Helium- producing countries are USA (78%), Algeria, Qatar, Russia and Poland.

Fig. 155: The helium recovery process at the ESRF.

The European Synchrotron 133 HIGHLIGHTS 2014 ENaBLING TECHNOLOGIES

phase to liquid phase) to fi ll the dewars 10%) and accidental loss (leaks on which come back again to supply the experimental pipes, poor connections). ESRF experiments (Figure 155). Nonetheless, in case of an important contamination of the recovered gas The ESRF’s consumption of liquid (e.g. introduction of air), the global helium is about 40 000 per year volume of recovered helium is aff ected with a recovery rate of over 70% (see because the correct functioning of the Figure 156). An analysis of the recovery liquefi er is linked to the purity of the rate shows that it can vary between 50 gas (at very low temperatures solidifi ed and 80% depending on the month and gas particles can cause damage). the year. The major factors causing helium loss are evaporation (about Adhering to simple procedures reduces loss and improves the recovery rate. Dewars pending use are systematically connected to the recovery network; dewars must be returned rapidly aft er use at a cold temperature, i.e. not completely empty. The tightness of the connections between the dewars and the experiment and the recovery network must be checked. Indications on the recovery meter should be checked against consumption. A Fig. 156: signifi cant diff erence could indicate Consumption a leak or contamination by ambient and recovery of air and a local contamination could helium at the shutdown the whole of the ESRF ESRF since 1999. recovery network.

Authors IPv6 - PREPARING THE INTERNET F. Calvelo Vazquez and B. Lebayle. ESRF CONNECTIVITY OF THE FUTURE

The Internet Protocol (IP) is the main The top-level exhaustion occurred on communication protocol in the 31 January 2011. Just three months Internet for relaying packets across later, on 15 April 2011, the APNIC networks and version 6 (IPv6) is the (Asia-Pacifi c region) was the fi rst RIR to latest version [1]. IPv6 was developed run out of their free IPv4 address pool by the Internet Engineering Task Force (Figure 157). (IETF) to deal with the long-anticipated problem of IPv4 address exhaustion. By 1998, the Internet Engineering Task Force had formalised the The depletion of IPv4 addresses was successor protocol IPv6. IPv6 uses foreseen as long ago as the late 1980s. a 128-bit address, allowing 2128, or With the rapid growth of the Internet approximately 3.4 × 1038 addresses, aft er commercialisation in the 1990s, or about 1029 times as many as IPv4, the democratisation of Internet access which uses 32-bit addresses and in the developed world, the increasing provides approximately 4.3 billion use of mobile devices and buildings addresses. IPv6 off ers a virtually or home automation, the initial over- inexhaustible number of addresses allocation managed by the Internet and paves the way for connecting Assigned Numbers Authority (IANA) ubiquitous devices to the Internet, a and by fi ve Regional Internet Registries trend lately branded as the “Internet (RIR), and the ineffi cient address use of Things” where uniquely identifi able by companies, are some of the reasons embedded computing devices are that have contributed to accelerate connected within the existing Internet address consumption. infrastructure.

134 ESRF ENaBLING TECHNOLOGIES

IPv6 is intended to replace IPv4, which still represents 95% of the Internet traffi c worldwide. At the beginning of 2013, the percentage of users reaching Google services with IPv6 was just a little bit above 1% [2]. This shows the slow adoption of IPv6 by major Internet Service Providers (ISPs), home networking equipment manufacturers and companies around the world.

IPv4 and IPv6 protocols are not designed to be interoperable, complicating the transition to IPv6. However, several IPv6 transition mechanisms have been Fig. 157: Exhaustion of IPv4 addresses. devised to permit communication between IPv4 and IPv6 hosts.

The “IPv6 Task Force France” (TTF) was created by the French Senate in 2002 to regroup main service providers and this subject between the diff erent EPN industrial actors to promote the use of participants started only in April 2013. IPv6 in France. It is part of the initiative IPv6 has a signifi cant impact for the launched by the European Commission System Administrators of the campus to coordinate all IPv6 Task Forces in laboratories because of some new Europe. features and default behaviours. In fact System Administrators have to learn Eff orts made by the government and a completely new internet protocol. some other institutions that were Once administrative formalities to involved have proved fruitful as France obtain the new IPv6 address space had is currently among the top 20 countries been settled, there was still a long road using IPv6 (rated as fourteenth in the ahead for EPN institutes to implement global classifi cation). the new prefi x assigned by RENATER: give external visibility to our new prefi x RENATER [3], our National through our metropolitan network telecommunications network for (TIGRE), fi nd new partners supporting Technology, Education and Research, is the Domain Name Service (DNS), currently connecting more than 1000 adapt several hardware and soft ware sites (institutions, universities and components to support this protocol, French organisations), but only 174 of and make an address plan to share this them have already applied for an IPv6 IPv6 prefi x between the four institutes prefi x. Furthermore, only 118 have taking part: ESRF, ILL, EMBL, IBS. already broadcast their prefi x to the rest of the world. Rhône-Alpes is one of The last component to be dealt with the most active regions just behind Île- was the upgrade of the common EPN- de-France. site fi rewall. Discussions with diff erent suppliers of security systems started At the ESRF, IPv6 will not replace as early as May 2013. But nonetheless, IPv4 in the short term. Instead both fi nding companies with suffi cient protocols must co-exist on the Internet experience of this kind of deployment for several years. Being IPv6 ready over IPv6 was challenging. Finally, on is important for sites like ESRF to 20 October 2014 everything was in guarantee that data transfers are place and the migration was performed not slowed down by heavy protocol with only 43 seconds of service translations. However, progressing interruption. Today, the fi rewall is able towards the implementation of this to support new security features over References new Internet protocol standard was a the dual IPv4 and IPv6 stack. [1] Wikipedia entry on IPv6: http:// en.wikipedia.org/wiki/IPv6 long and complicated procedure for the [2] Google’s statistics: http://www. EPN-campus. There were signifi cant The ESRF is now ready for the next google.com/intl/en/ipv6/statistics. hurdles to overcome; organisational generation of Internet services based html as well as technical. Discussions on on IPv6! [3] RENATER: https://www.renater.fr

The European Synchrotron 135 ACCELERATOR AND X-RAY SOURCE Highlights 2014

The Accelerator and Source Division • Extension of the straight sections In addition to our work on Phase I (ASD) is in charge of the production from 5 m to 6 m (completed) and Phase II, we also hosted the 22nd of synchrotron light from the ESRF’s 6 • Installation of canted straight annual European Synchrotron Light GeV storage ring. We have continued sections (completed) Source (ESLS) workshop in November. our efforts to ensure reliable operation • Installation of one 7 m straight Thirty-five representatives from during 2014. Indeed, during the section (completed) fifteen European light sources spent year, 5399.20 hours of beam was • Implementation of top-up mode two days at the ESRF discussing the delivered, representing an accelerator capability (in progress) status of their institutes operational availability of 99.11% and beating performance as well as future upgrade the all-time ESRF record of 99.04% in This year was also a crucial year for projects. We were pleased to note 2009. the future of the Phase II Accelerator that the accelerator lattice concept Upgrade, and a lot of time was developed at the ESRF is being As well as providing beam to users, dedicated to presenting our ideas, adopted by many other world-leading with 2014 marking the end of studies and designs to the Accelerator synchrotron facilities. Phase I of the Upgrade Programme, Project Advisory Committee (APAC), the Division has completed or is in a which was held in January, May, After a busy but fruitful year, we very advanced stage of completion of September and October. Following entered the execution phase of the major upgrades of several subsystems positive reports from the APAC in Phase II Upgrade Programme on 1st of the accelerator complex, including: May, the project was approved and January 2015 and work on the project • Upgrade of the RF system with solid launched by the ESRF Council in is now in full swing. state amplifiers (completed) June 2014. The upgrade will see the • Development and installation implementation of a new lattice for P. Raimondi of new higher order mode (HOM)- the storage ring to reduce the natural damped cavities (completed) horizontal emittance from 4 nm to • Development and installation of about 134 pm, thus greatly increasing improved electron beam position the brilliance and coherence of the monitoring system and a fast orbit beam. feedback system (completed)

B eAM PARAMETERS OF THE storage ring

Table 1 presents a summary of the These points differ by the observation characteristics of the storage ring angles, of respectively 3 and 9 mrad electron beam. from the entrance of the magnet.

Table 2 gives the main optic functions, Electron beam profiles are Gaussian electron beam size and divergence at and the size and divergence are various source points. For insertion presented in terms of rms values. device source points, the beta functions, The associated full width at half dispersion, sizes and divergences are maximum sizes and divergences are calculated in the middle of the straight 2.35 times higher. Horizontal electron section. For bending magnets, two beam sizes and divergences are given representative source points have been for the multibunch filling modes and selected for each apply to almost all filling patterns, Energy [GeV] 6.04 type (even or odd cell except when the current per bunch Maximum current [mA] 200 number) of magnet, is larger than 4.5 mA, for which a corresponding to slightly larger size and divergence Horizontal emittance [nm] 4 magnetic fields of are attained because of the increased Vertical emittance [pm] 4 0.4 T and 0.85 T. energy spread of the electron beam. Revolution frequency [kHz] 355 Number of bunches 1 to 992 Table 1: Principal Vertical electron beam sizes and characteristics of the divergences are given for a vertical Time between bunches [ns] 2.82 to 2816 electron beam.

136 ESRF aCCELERaTOR aND X-RaY SOURCE

Even ID Odd ID Even BM Even BM Odd BM Odd BM (ID2, ID6…) (ID1, ID3…) (ID2, ID6…) (ID2, ID6…) (ID1, ID3…) (ID1, ID3…) 3 mrad 9 mrad 3 mrad 9 mrad Magnetic fi eld [T] Variable Variable 0.4 0.85 0.4 0.85 Horiz. Beta function [m] 37.6 0.35 1.33 1.06 2.12 1.61 Horiz. Dispersion [m] 0.134 0.031 0.062 0.051 0.089 0.075 Horiz. rms e- beam size [µm] 413 50 99 85 132 113 Horiz. rms e- divergence [µrad] 10 107 116 114 104 99 Vert. Beta function [m] 2.95 2.97 41.7 42.0 32.1 32.2 Vert. rms e- beam size [µm] 3.4 3.4 12.9 13.0 11.3 11.4 Vert. rms e- divergence [µrad] 1.17 1.16 0.50 0.50 0.36 0.36

Table 2: Beta functions, dispersion, rms beam size and divergence at the various source points. emittance of 4 pm, which is now Filling pattern Uniform 7/8 + 1 Hybrid 16-bunch 4-bunch the standard for 2 x 1/3 and 7/8+1 Number of bunches 992 870+1 24x8+1 16 4 fi lling modes. The vertical sizes and Maximum current [mA] 200 200 200 90 40 divergences are about 1.4 times larger in uniform fi lling mode (due to ion eff ects, Lifetime [h] 50 45 30 16 9 which are partially corrected by the use Rms energy spread [%] 0.11 0.11 0.11 0.12 0.16 of a vertical bunch-by-bunch feedback). Rms bunch length [ps] 20 20 25 48 55 To increase the lifetime of the stored beam, the vertical beam sizes and Table 3: Current, lifetime, bunch length and energy spread for a selection of fi lling modes. divergences are deliberately increased by about a factor of 4 in the 16-bunch, 4-bunch and hybrid fi lling patterns. patterns, the energy spread and bunch length decay with the current (the value The lifetime, bunch length and energy indicated in the table corresponds to the spread mainly depend on the fi lling maximum current). The bunch lengths pattern. These are given in Table 3 for are given for the usual radiofrequency a few representative patterns. Note accelerating voltage of 9 MV (8 MV for that in 16-bunch and 4-bunch fi lling 16-bunch and 4-bunch).

SUMMARY OF ACCELERATOR OPERATION

Last year was a record-breaking year for accelerator operation, seeing 5399.20 out of 5487.5 planned hours of beam delivered to Users in 2014 (excluding shift s for radiation and personal safety system tests). This excellent fi gure represents an accelerator availability of 99.11%, which beats the all-time ESRF record of 99.04% in 2009. This was achieved through a rapid resolution of failures, of which only a few were signifi cant long failures. The longest of these was in May and lasted just over four hours.

Similarly, the mean time between failures (MTBF) of 105.5 hours is the Fig. 158: Uninterrupted beam delivery second best MTBF ever obtained at for 567 hours in 7/8+1 and 16-bunch the ESRF (the record is 107.8 hours fi lling modes during weeks 5, 6, 7 and in 2011). In April, during the second 8 (RUN 2014-01).

The European Synchrotron 137 HIGHLIGHTS 2014 aCCELERaTOR aND X-RaY SOURCE

ToTAL rUn nUMBEr 2014-01 2014-02 2014-03 2014-04 2014-05 2014 Start 17/01/2014 28/03/2014 06/06/2014 22/08/2014 17/10/2014 End 19/03/2014 27/05/2014 30/07/20147 08/10/2014 17/12/2014 Total number of shift s 183 180 162 141 183 849 Number of USM shift s 146.44 147 131.13 113.25 148.13 685.94 Beam available for users (h) 1158.70 1153.30 1031.50 892.40 1163.30 5399.20 Availability 99.50% 98.95% 98.87% 99.38% 98.90% 99.11% Dead time for failures 0.5% 1.1% 1.1% 0.6% 1.1% 0.89% Dead time for refi lls 0.6% 0.9% 0.5% 0.9% 0.7% 0.72% Average intensity (mA) 153 133 179 99 151 95.4 Number of failures 10 14 10 9 9 52 Mean time between failures (h) 117.2 84 104.9 100.7 131.7 105.5 Mean duration of a failure (h) 0.6 0.9 1.2 0.6 1.4 0.94

Table 4: Overview of storage ring operation in 2014.

run of the year, six beam trips during including one period of four weeks the same week reduced the MTBF. (see Figure 158) and two periods of This was due to an unfortunate series 288 hours (i.e. two consecutive weeks) of events. Once these problems had without a single beam interruption. A been solved, a noticeable absence of summary of operation is presented in repetitive events (especially during Table 4. Runs 1, 3 and 5 when the MTBF greatly exceeded 100 hours) resulted in 14 out of 37 weeks where beam was delivered without any interruptions,

FILLING PATTERNS

The distribution of fi lling modes During machine-dedicated time, several (Figure 159) has not drastically evolved shift s were carried with participation over the last four years, since the 7/8+1 of the beamlines to assess top-up in fi lling mode was introduced. This mode 16-bunch mode, testing refi lls every remains the best compromise for 10, 20 and 30 minutes. Many scientists users who need intensity and/or time- expressed their enthusiasm thanks to structure. The uniform mode remains the excellent results obtained. More mainly for the specifi c experiments on details are presented in a later section. one beamline.

Fig. 159: Distribution of fi lling modes in 2014.

138 ESRF aCCELERaTOR aND X-RaY SOURCE

COMPLETION OF THE TECHNICAL DESIGN STUDY FOR THE NEW STORAGE RING

In 2014, the ASD completed the technical design study (TDS) of the Phase II Accelerator Upgrade. The upgrade is focused on greatly improving the brightness of the source by decreasing the natural horizontal equilibrium emittance of the storage ring from the current 4 nm to about 134 pm. This will be possible by replacing the present 32 double-bend achromat arcs with a much more compact lattice Fig. 160: The ESRF based on a hybrid 7-bend arc achromat hybrid multi-bend (Figure 160). The smaller emittance, achromat (HMBA) together with a shorter-period, smaller- lattice design. gap undulators and new beamline optics, optimised to take full advantage Fig. 161: The prototype of the new source, will allow a general octupole for the new lattice, improvement in performance by a delivered in October 2014. factor of 100 or more.

The sources for the bending magnet beamlines will be replaced by fi rst magnetic designs three-pole wigglers, resulting in a for the correctors comparable increase in brightness have been completed and fl exibility of the X-ray spectra and the engineering for these beamlines too. The lattice designs will be done in and the engineering designs are well 2015. advanced and the magnetic designs are already at the prototyping stage. In May 2014, the Accelerator Project A prototype octupole was delivered in Advisory Committee (APAC), appointed October 2014 (see Figure 161). Two by the Council, reviewed and produced prototypes of the permanent magnet a very positive report on the overall dipoles with longitudinal gradients feasibility of the project. The TDS, have been constructed and a third is together with their report was planned for 2015. A high-gradient submitted to the June meeting of the quadrupole prototype is currently Council, which subsequently gave its under construction and prototypes for approval to the project. The execution a combined dipole-quadrupole and a phase was offi cially launched on st 1 sextupole are foreseen for 2015. The January 2015.

PERMANENT MAGNET DIPOLES FOR THE ACCELERATOR UPGRADE

The new magnet lattice for the beam. The use of permanent magnet accelerator upgrade includes 128 technology was found to be suitable for dipoles with longitudinal fi eld gradients building these dipoles for compactness contributing to the reduction of the and economical reasons. Each dipole horizontal emittance of the electron magnet consists of a series of fi ve short

The European Synchrotron 139 HIGHLIGHTS 2014 aCCELERaTOR aND X-RaY SOURCE

modules with diff erent The Sm2C017 permanent magnet fi eld amplitudes, ranging material was chosen to ensure from 0.65 T to 0.17 T adequate temperature stability and (see Figure 162). long-term stability with respect to radiation-induced demagnetisation. The development of permanent magnet structures was initiated in 2013-2014 Fig. 162: Typical magnetic with the construction of several types fi eld along a dipole with a of modules. Each module has a length longitudinal fi eld gradient. of 350 mm and a transverse size of 190 mm (horizontal) and 240 mm (vertical) with a typical weight of Fig. 163: Prototype 85 kg. The magnetic structure is based dipoles with longitudinal on a C-shape design with solid iron fi eld gradient under study poles and a yoke driving the magnetic (two modules). fi eld from permanent magnets in the 25 mm gap. Figure 163 shows an assembly of two modules under study. The magnetic measurement results are in very good agreement with the simulation.

PROGRESS IN THE TOP-UP OPERATION PROJECT

To reduce the heat load variation on the several hours, with no disturbance; beamline optics, the implementation of all parameters were even more stable top-up operation is under preparation. in this confi guration. Radioprotection This mode of operation will be benefi cial aspects are satisfactory (injection is for the 16-bunch mode by reducing the routinely performed with the front end current variation and also by allowing open) but a formal agreement from the delivery at low vertical emittance. A new authorities for such operation in USM cleaning technique has been developed is still needed. Large horizontal orbit for this project and is now routinely perturbations are still being induced used during user service mode (USM) during injection, mainly by the injection for the 16bunch and hybrid modes. kickers and injection septa. Solutions This reduces the strong vertical blow- using feed-forward corrections have up just aft er injection from 2 nm to only been successfully developed and should a few pm. The development of bunch be implemented for routine operation cleaning in the booster is progressing (see dedicated paragraph). An injection and will benefi t from the upgrade of the sequencer has been developed in injector, in particular, the new booster order to automatically manage the power supply system operating at 4 Hz, equipment and to optimise electrical which is under development for the consumption. Using this tool, injection dipole and quadrupole magnets. is now done in top-up in less than 30 seconds. Top-up in 16-bunch mode has been tested successfully (Figure 164). The top-up frequency is a crucial The reliability of the storage ring parameter for this project. The fi rst top- was validated with 90 mA average up tests with users at the beamlines current in 16bunch mode, refi lled were carried out in September 2013 every 10 minutes over a period of and May 2014 in 16-bunch mode, with a refi ll every 5, 10, 20 and 30 minutes. The main feedback from the beamlines Fig. 164: Synopsis of eight hours of top-up was that it had not been possible to tests with beamlines in September 2014. move the mechanics during the top-

140 ESRF Accelerator and X-ray Source

up sequences, preventing them from general improvement in the stability taking maximum advantage from of all the X-ray beam parameters. The top-up operation. This was due to a beamlines could be operated without restriction on an ID device server, which interruption during injections and were was then removed in time for the next unable to detect the exact moment tests. when refills had been carried out.

In September, a third top-up shift – Nevertheless, detailed analysis and also in 16-bunch mode – was carried tests still have to be completed for the out with 23 beamlines. This time it most sensitive beamlines. Further user had been possible for the beamlines to tests are planned for early 2015. There move their mechanics and the lowest remains much to do in order to ensure gap of 6 mm was obtained at ID09 reliability of equipment before top- and ID11 in-vacuum devices, while up operation can be put into routine ID15 obtained 6.5 mm. The beamline service but the benefits for most users feedback was positive, showing a are clear. new storage ring BPM prototypes tested with beam

The development of 6 mm-button stored beam and to carry out detailed UHV feed-throughs is making very verification of the beam position good progress; a total of 27 buttons measurements obtained. A second were delivered in the spring and chamber with different buttons will tested successfully. During the August shortly be installed and tested. 2014 shutdown, a test beam position monitor (BPM) chamber was installed in the ID25 straight section (Figure 165). Three BPMs were connected to the Libera BPM electronics and the test BPM chamber itself placed on a motorised translation table allowing +/-3 mm displacement in both planes, Fig. 165: The test BPM to enable RF signals to be checked with chamber installed at ID25.

New booster bpm electronics

Following the goal to modernise the started and 10 units were successfully 25-year-old electronics for the RF commissioned in December. The full signal acquisition and processing conversion to the new BPM system of the booster BPM system, two is expected to be finished by the end prototypes, called Libera Spark, have of February 2015. An agreement been tested under various beam has been reached on specifications filling modes in the booster and and a price for an advanced version they have shown their potential for of this device that opens up various precise beam position monitoring. applications in the injector and the After satisfactory testing, a contract storage ring. for the delivery of a further 81 units was signed and these were delivered in November 2014. The installation of these Sparks in the booster has now

The European Synchrotron 141 HIGHLIGHTS 2014 aCCELERaTOR aND X-RaY SOURCE

NEW VERTICAL BEAM HALO MONITOR

The halo monitor measures the of the year, the halo monitor was electron population along the vertical further optimised with more rigorous beam profi le at 0.5 mm to 6 mm lead shielding for its detector and the from the central core of the electron choice of CMOS camera technology, beam, using X-rays emitted from the which has greater resistance to fi rst dipole in cell 7. Following initial radiation. The tungsten blade has also commissioning at the beginning been modifi ed to give more calibration possibilities. A device server records halo measurements at a frequency of 1 Hz and stores them in a database (Figure 166). Such recordings show the very strong sensitivity of this diagnostic device to vacuum conditions and to any variation of Touschek scattering conditions. It is a new and novel addition to our range of diagnostic instruments dedicated to electron beam loss phenomena.

Fig. 166: Images from the halo monitor for four vertical scraper positions. In the lower part of each image, one can observe that the beam core is shielded while the halo signal disappears progressively.

NEW PICOSECOND TIME-RESOLUTION STREAK CAMERA

The new visible light streak camera was several MDT studies, such as HOM delivered to the ASD optical laboratory instability and also for beam dynamics in January 2014. The results obtained studies relating the RF voltage to beam from various tests were satisfactory phase, bunch length and lifetime. Aft er and in line with the theoretically initial setup for each experiment, predicted shortest bunch length at zero the streak camera can be controlled bunch current. Since its installation, and read from the control room. the Streak camera has been used in

NEW BUNCH CLEANING METHOD FOR THE BOOSTER

A new method of time-gated cleaning it won’t be compatible with the future was implemented in 2013, using the top-up mode of operation, as the bunch-by-bunch feedback system. time interval between injections will However, since cleaning is performed become much smaller. Therefore, in aft er injection into the storage ring, 2014, a new system was implemented

142 ESRF aCCELERaTOR aND X-RaY SOURCE

to perform bunch cleaning 5 to 10 cleaning effi ciency should be stable at milliseconds aft er injection in the each injection. Measurements taken booster. This has been upgraded by the in the autumn of the effi ciency of the addition of a system generating the booster cleaning showed that adequate excitation frequency of the parasitic cleaning is generally achieved for most bunches more accurately. Now that injection shots. Future work will aim the transverse resonant frequency to increase the reproducibility of the of the beam is tracked precisely, the system.

DAMPING OF THE HORIZONTAL OSCILLATION OF THE STORAGE RING STORED BEAM CAUSED BY INJECTION

The fi ring of the pulsed magnets of forward scheme. The fi rst test of the the injection system (kickers and system demonstrated a very eff ective septum) is the cause of horizontal cancellation of this orbit distortion and vertical transient orbit distortion (Figure 168). A similar set of vertical or beam position oscillation. Various correctors was added during the factors contribute to this disturbance December 2014 shutdown and the including non-closure of the injection system will be operated in USM in bump during kicker rise and fall, a 2015. slight tilt of the kickers, septum magnet fi eld leakage and asymmetry. The We aim to suppress the vertical horizontal disturbance is the largest, betatron oscillation due to the kickers’ but is expressed in the relative increase tilt by mechanically trimming the tilt of the beam emittance, whereas the of the kickers according to the result of vertical disturbance is more signifi cant turn-by-turn position measurements since the vertical nominal emittance is done following the fi ring of the kickers. A much smaller. Solutions are underway fi rst set of tilt corrections will be applied to tackle each source of disturbance. during the shutdown in January 2015.

We have implemented a system that damps the horizontal betatron oscillation by applying a correction signal over nine turns using a 0.6 MHz bandwidth magnet driven by a 400 W amplifi er. The system was tested at the end of 2013 and proved to be eff ective, dividing the amplitude of the horizontal Fig. 167: Non-damped betatron oscillation induced during the (purple) versus damped rise and fall time of the kickers by three (blue) horizontal oscillations over turn 7 to turn 10 aft er (Figure 167). injection, for all bunches, showing the eff ect of the feed-forward damping. At the end of 2014, an orbit correction system cancelling the horizontal orbit distortion caused by the septum Fig. 168: Eff ect of the septum fi eld leak leakage was also implemented, using on the horizontal a set of fast corrector magnets and beam position amplifi ers that were originally part of without (left ) and the old fast orbit correction system. with (right) the orbit A signal with the same shape as the correction system activated. Scale: beam position signal recorded when Vertical: 0.5 mm/div, the correction was not active was used Horizontal: 10 ms/div. to drive the amplifi er input in a feed-

The European Synchrotron 143 HIGHLIGHTS 2014 aCCELERaTOR aND X-RaY SOURCE

UPDATE ON THE NEW HOM-DAMPED CAVITIES IN THE STORAGE RING POWERED BY 150 kW RF SOLID STATE AMPLIFIERS

In 2013, three new RF HOM-damped retrofi tted the four booster SSAs cavities were installed in cell 23, the last with non-fl ame-propagating cables, being put in place during the December bringing them in line with the three shutdown and powered in January storage ring SSas. 2014. The fi rst two cavities performed well throughout 2014. They were The implementation of the in-house operated in USM at the nominal voltage 75 kW prototype RF SSA, which uses a of 0.5 MV with 80 kW input power at cavity combiner to combine the power a stored beam current of 200 mA. from 132 ESRF-developed 700 W RF During machine studies, these cavities power modules, is in progress. This were tested with up to 150 kW incident came within the remit of work package power by increasing their voltage to WP7 of the EU/CRISP project, which 0.65 MV and adjusting their phase to ended in September 2014. Over the increase their beam loading. Following course of the project, it was decided the discovery of an air leak in the to develop an integrated system third cavity, it was removed from the comprising fully planar RF circuits ring for repair in December 2014. The that are coupled directly to the cavity manufacture of 12 additional HOM- combiner instead of connecting existing damped cavities by RI is progressing RF modules available from industry by well and the delivery is foreseen during means of coaxial cables. This resulted the course of 2015. in an extremely compact solution that will allow cost eff ective industrial The four 150 kW solid state amplifi ers production. A 12 kW prototype with 18 (SSAs) installed in the booster in 2012, connected modules was successfully and the three SSAs installed in the tested, and power tests of the 75 kW storage ring in 2013, are all performing prototype are scheduled for the second well. In October 2014, the supplier half of 2015.

ID STRAIGHT SECTIONS

The required undulator segments straight section and a new insertion for the upgrade of ID1 have been device layout will be completed before installed and are in operation. The the end of 2015. new 2.5 m-long helical undulator will be installed in March 2015 in the ID32 straight section. Most of the work is now focused on ID31 and ID15. A high- performance cryogenic permanent magnet undulator is under construction at ID31. The undulator will be installed in the middle of the 5 m-long straight section to reach a minimum gap of 4.5 mm. Its installation will take place before mid-2015. At ID15, the straight section will be modifi ed with a canting scheme in 2015. A required short, high- fi eld wiggler has been designed and the associated components are under procurement. The modifi cation of the

144 ESRF HIGHLIGHTS 2014 FACTS AND FIGURES

MEMBERS AND ASSOCIATE COUNTRIES (AS OF JANUARY 2015)

Member countries: Associate countries: 27.5% France 1.5% Israel 24% Germany 1.3% Austria 13.2% Italy 1% Poland 10.5% United Kingdom 1% Portugal 6% Russia 1.05% Centralsync 5.8% Benesync (Belgium, (Czech Republic, The Netherlands) Hungary, Slovakia) 5% Nordsync (Denmark, 0.3% South Africa Finland, Norway, Sweden) 4% Spain 4% Switzerland

Experimental Hall 2 Central building Experimental BM26 Hall 1 BM25 ID26A+B THE BEAMLINES A+B ID27ID28 BM28 ID24 ID29 BM23 Laboratory and 29 28 30 BM29 ID30A+B office building ID23 27 3 Details of the public ESRF beamlines as 1

6 2 3 BM30 A+B ID22 2 well as those operated by Collaborating 5 2 1

Research Groups (CRG) are given in ID31 4 ID19 ID21 2

2

Tables 5 and 6. Figure 169 shows

BM20 BM32

3 2

the location of the beamlines in the 3 ID32

ID20 Control room

Linear

experimental halls. 2 BM01 A+B 2 Accelerator

4

(Linac)

ID01

1

ID17 2

5

Booster

Fig. 169: Experimental hall showing

BM02

Synchrotron

0

ID18

2

location of the beamlines (public and

ID03 ID02

6

ID16 A+B

CRG beamlines).

9

1

7

8

1

8

ESRF beamlines

7

1

BM05

9

CRG beamlines

6

1

1

0 ID15A+B

ID06

5

1

1

1

Instrumentation 4

BM14

1

1

2

3

Storage1 ring and machine test ID14 beamlines ID13 BM08 ID12 ID09 Refurbished or ID10 upgraded ID11 beamlines

The European Synchrotron 145 HIGHLIGHTS 2014 Facts and Figures

SOURCE number of BEAMLINE STATUS POSITION independEnt name end-stations

ID01 1 Microdiffraction imaging Operational since 12/14 ID02 1 Time-resolved ultra-small-angle X-ray scattering Operational since 07/14 ID03 1 Surface diffraction Operational since 09/94 ID06 0.3 Large volume press Operational since 10/13 ID08 1 Dragon Closed since 10/13 ID09 1 White beam Operational since 09/94 ID10 1 Soft interfaces and coherent scattering Operational since 06/12 ID11 1 Materials science Operational since 09/94 ID12 1 Circular polarisation Operational since 01/95 ID13 1 Microfocus Operational since 09/94 ID15A 1 High energy diffraction Closed since 01/15 ID15B 1 High energy inelastic scattering Closed since 01/15 ID16A 1 Nano-imaging Operational from 05/14 ID16B 1 Nano-analysis Operational from 04/14 ID17 1 Medical Operational since 05/97 ID18 1 Nuclear scattering Operational since 01/96 ID19 1 Topography and tomography Operational since 06/96 ID20 1 Inelastic X-ray Scattering Operational since 06/13 ID21 1 X-ray microscopy / IR Spectroscopy Operational since 12/97 ID22 1 High resolution powder diffraction Operational since 05/14 ID23 2 Macromolecular crystallography MAD Operational since 06/04 Macromolecular crystallography microfocus Operational since 09/05 ID24 1 Dispersive EXAFS Operational since 02/96 ID26 1 X-ray absorption and emission Operational since 11/97 ID27 1 High pressure Operational since 02/05 ID28 1 Inelastic scattering II Operational since 12/98 ID29 1 Multiwavelength anomalous diffraction Operational since 01/00 ID30A 1 Macromolecular crystallography Operational since 12/4 ID31 1 Powder diffraction Closed since 01/14 ID32 1 Soft X-ray spectroscopy Operational since 11/14 BM14 1 Macromolecular crystallography (MAD) Operational since 01/10 BM23 1 X-ray absorption spectroscopy Operational since 03/11 BM29 1 Bio SAXS Operational since 06/12 Operational in 2015: ID30B 1 Macromolecular crystallography Operational from 01/15 ID31 1 High energy Operational from 08/15

Table 5: List of the ESRF public beamlines.

SOURCE number oF BEAMLIne FIELD STATUS POSITIon IndependEnT nAME OF RESEARCH end-stations

BM01 2 Swiss-Norwegian BL X-ray absorption & diffraction Operational since 01/95 BM02 1 D2AM (French) Materials science Operational since 09/94 BM08 1 Gilda (Italian) X-ray absorption & diffraction Operational since 09/94 BM20 1 ROBL (German) Radiochemistry & ion beam physics Operational since 09/98 BM25 2 SPLINE (Spanish) X-ray absorption & diffraction Operational since 04/05 BM26 2 DUBBLE (Dutch/Belgian) Small-angle scattering Operational since 12/98 EXAFS Operational since 06/01 BM28 1 XMAS (British) Magnetic scattering Operational since 04/98 BM30 2 FIP (French) Protein crystallography Operational since 02/99 FAME (French) EXAFS Operational since 08/02 BM32 1 IF (French) Interfaces Operational since 09/94

Table 6: List of the Collaborating Research Group beamlines.

146 ESRF Facts and Figures

7000 6927 User operation 6395 6000 6318 5472 5735

Following 20 years of successful 5000 Proposals submitted operation for scientific users (September Experimental sessions 4798 1994 to September 2014), we look back 4000 Number of user visits on user operation at the facility over the past year 2014. The ESRF has been back 3000 User visits User in full user operation for the majority 2047 2035 2035 2113 2013 1905 1967 of users since 2013, but Phase I of the 2000 1731 1559 1559 1392 Upgrade Programme continues and 1344 1186 several beamlines were either fully or 1000 partially closed during 2014 for upgrade. ID01 (UPBL1) and ID32 (UPBL7) were Number of proposals / ExperimentalNumber of proposals sessions / 0 closed for almost the whole year, 2008 2009 2010 2011 2012 2013 2014 taking their first user experiments in Years December 2014; ID31 has been closed awarded beamtime for each beamline. Fig. 170: Numbers of applications for for upgrade for the whole year and the Proposals for experiments in 2014 were beamtime, experimental sessions and user visits, 2008 to 2014. N.B. Final new upgrade beamline ID22 opened reviewed by 10 review committees numbers of experimental sessions and in May 2014 (UPBL2); ID02 (UPBL9a) grouping the following beamlines of user visits for 2014 were not available at was re-opened more than half way similar techniques or activities: the time of going to press. through the year while new beamlines • C01 (ID01, ID03, BM25B, BM32) ID16A-NI and ID16B-NA were also only • C02 (ID11, ID15A, ID15B, ID22) partially open (UPBL4); ID14 closed • C03 (ID12, ID20, ID26, ID32, BM28) and one of the new MX beamlines, • C04 (ID24, BM01B, BM08, BM20, ID30A-1, opened half way through the BM23, BM25A, BM26A, BM30B) year (UPBL10). In total, the equivalent • C05 (ID06-LVP, ID09A, ID18, ID27, of around 23.3 ESRF publicly funded ID28, BM01A) beamlines (out of an optimum of 29.5 in • C06 (ID17, ID19) the portfolio) and 9 CRG beamlines were • C07 (ID16A&B, ID21) available for users in 2014. Figure 170 • C08 (ID02, ID13, BM26B) shows the number of applications for • C09 (ID09B, ID10, BM02) beamtime received since 2008. Despite • C10 (Structural Biology Beamlines). the heavy construction programme and the unavailability of a non-negligible The scientific areas of the ESRF research number of beamlines, 2014 saw the activities reviewed by these ten highest ever number of proposals committees are shown in Table 7. submitted to the ESRF in its history. This shows that the user interest in the ESRF In September 2014, an 11th review has not diminished in any way and is, at committee C11 was created and the the moment, even stronger than in the beamlines reviewed by committees past. C03 and C04 were redistributed so that some of those beamlines are now Proposals for experiments are selected reviewed by C11. This modification and beamtime allocations are made concerns beamtime allocation for 2015 through peer review. As in previous and so will be described in more detail years, beamtime for 2014 was allocated in next year’s report concerning 2015 based on recommendations made by operation. review committees of specialists, for Total shifts Total shifts the most part from European countries, Table 7: Number Scientific field requested allocated of shifts of Israel and South Africa. A new structure Chemistry 4 338 1 775 beamtime Earth Sciences 1 835 819 has been in place since 2013 for the requested and Environment 832 249 proposal review process, with beamline- allocated for user Hard Condensed Matter Science 9 380 3 165 experiments, based committees allowing all proposals Cultural Heritage 332 144 year 2014. received for a particular beamline Life Sciences 1 423 549 to be assessed by the same review Applied Materials Science 5 055 1 652 committee. This gives the committees Medicine 1 058 465 Engineering 237 117 a better overall view of all the projects Methods & Instrumentation 492 201 proposed for a single beamline, and Structural Biology 3 061 2 077 allows them greater flexibility to Soft Condensed Matter 3 157 1 077 optimise the selection of proposals Totals 31 200 12 290

The European Synchrotron 147 HIGHLIGHTS 2014 FaCTS aND FIGURES

The review committees met twice for non-MX experiments. The annual during the year, around six weeks number of experimental sessions and aft er the deadlines for submission of user visits since 2008 is shown in proposals (1 September 2013 and 1 Figure 170. The eff ect of the reduction March 2014). They reviewed 2113 in beamtime available in 2011 and applications for beamtime, and selected 2012 due to the long shutdown is clearly 864 (40.9%), which were then scheduled visible, and even though the fi gures for for experiments. 2013 show the return to something close to pre-shutdown user operation, Requests for beamtime in 2014, which the unavailability of the remaining is scheduled in shift s of 8 hours, totalled upgrade beamlines due to the ongoing 31 200 shift s or 249 600 hours, of which Phase I programme means that, for the 12 290 shift s or 98 320 hours (39.4%) time being, the number of experimental were allocated. The distribution of shift s sessions and user visits is still slowly requested and allocated, by scientifi c growing back towards its optimum level. area, is shown in Table 7. This will also be the case for 2014 due to the upgrade beamline work listed at The breakdown of shift s delivered the beginning of this report, but once all for experiments by scientifi c area beamlines are running towards the end in the fi rst half of 2014 is shown in of 2015 the trends suggest that we may Figure 171. This same period saw return to a higher level of usage than 2667 visits by scientists to the ESRF before, despite the permanent closure of under the user programme, to carry two beamlines since 2010. out 643 experiments. Overall, the number of users in each experimental One of the new upgrade beamlines that team averaged just over 4 persons as came online in 2014 is MX beamline in all previous years, and the average ID30A-1, also known as MASSIF-1, which duration of an experimental session is the fi rst beamline in Europe to off er was 9 shift s. This can be further broken fully automatic hands-off screening and down to show an average duration of 3 data collection for structural biology. shift s for MX experiments and 14 shift s Users do not come to the beamline but instead send their samples, which Fig. 171: Shift s are automatically analysed. This delivered for analysis can include the collection of full experiments, March diff raction datasets, and is carried out to July 2014, by based on information supplied by the scientifi c area. user via the ISpyB LIMS for structural biology beamlines.

One of the principle measurable output parameters of the ESRF is the number and quality of publications accepted in peer-reviewed journals. The number of publications rose continuously over the past years to reach a plateau of just under 1900 publications per year, corresponding to the maximum operation level of the ESRF before the Upgrade Programme, as shown in Figure 172. Despite the 6 2000 1848 1822 1884 1870 1832 month shutdown between 2011 and 1754 1800 1699 2012 for the construction of the new 1600 EX2 experimental hall, the number 1400 of publications has not fallen in any 1200 signifi cant way. These are excellent 1000 fi gures, with a publication output 800 600

Refered publications Refered 400 200 Fig. 172: Numbers of publications appearing in 0 refereed journals reporting on data collected either 2008 2009 2010 2011 2012 2013 2014 Years partially or totally at the ESRF, 2008 to 2014.

148 ESRF Facts and Figures

systematically on a level of well over User responses to questionnaires one publication per experimental show once again that the ESRF session. The year 2014 promises to be continues to maintain its excellent highly fruitful, with 1754 publications reputation concerning the assistance registered so far in the ILL/ESRF Library and service given by scientists and database, which is higher than the 2013 support staff on beamlines, and travel figure at this same time last year. Since and administrative arrangements, the ESRF began user operation back in addition to the quality both of the in 1994, a total of 25148 publications beam and of the experimental stations. have been accepted in peer-reviewed Facilities on site, such as preparation journals. Of these, nearly 300 every laboratories, the Guesthouse and a year are published in high impact canteen open 7 days a week, also make factor journals. an important contribution to the quality of user support.

Administration and finance

Expenditure and income 2013 Expenditure kEuro Income kEuro Accelerator and Source 2013 Members’ contributions 84 981.0 Personnel 6 181.0 Funds carried forward from 2012 843.0 Recurrent 1 801.1 Building costs provision 1000.0 Operating costs 1 723.4 Other recurrent costs 77.7 Other income Capital 5 366.0 Scientific Associates 5 408.0 Accelerator and Source developments 5 366.0 Sale of beamtime 1 588.0 Compensatory funds 2 005.0 Beamlines, experiments and in-house research Scientific collaboration and Special projects 8 232.0 Personnel 18 232.3 Recurrent 5 716.1 Pre-financing 4 744.0 Operating costs 1 462.5 Pre-financing carried forward to 2014 5 560.0 Other Recurrent costs 4 253.6 Capital 13 698.9 Beamline developments 13 695.6 Beamline refurbishment 3.3 Technical and administrative supports Personnel 28 562.5 Recurrent 12 220.7 Capital 16 498.1 Industrial and commercial activity Personnel 428.9 Recurrent 68.7 Unexpended committed funds Funds carried forward to 2014 5 586.0 Total 114 360.0 Total 114 360.0

R evised expenditure and income budget for 2014 Expenditure kEuro Income kEuro Accelerator and Source 2014 Members’ contributions 86 863 Personnel 6 193 Funds carried forward from 2013 5 586 Recurrent 1 546 Building costs provision 315 Operating costs 1 466 Other recurrent costs 80 Other income Capital 6 026 Scientific Associates 6 103 Accelerator and Source developments 6 026 Sale of beamtime 1 310 Compensatory funds 3 912 Beamlines, experiments and in-house research Scientific collaboration and Special projects 4 149 Personnel 18 983 Pre-financing –5 995 Recurrent 4 861 Operating costs 918 Other Recurrent costs 3 943 Capital 13 992 Beamline developments 13 992 Technical and administrative supports Personnel 28 726 Recurrent 13 156 Capital 8 235 Industrial and commercial activity Personnel 422 Recurrent 103

Total 102 243 Total 102 243

The European Synchrotron 149 HIGHLIGHTS 2014 Facts and Figures

Expenditure 2013 Revised budget for 2014 The budget for 2014 includes by nature of expenditure by nature of expenditure additional contributions kEuro kEuro from Members and Scientific Associates of 13 971 kEUR PERSONNEL PERSONNEL ESRF staff 51 450.9 ESRF staff 52 249 dedicated to the Upgrade External temporary staff 51.1 External temporary staff 60 Programme. The Upgrade Other personnel costs 1 902.8 Other personnel costs 2 015 expenditure budget amounts RECURRENT RECURRENT Consumables 7 637.6 Consumables 7 355 to a total of 27 427 kEUR Services 9 790.3 Services 9 863 including 13 891 kEUR of ESRF Other recurrent costs 2 378.7 Other recurrent costs 2 448 operating budget. CAPITAL CAPITAL Buildings, infrastructure 12 328.0 Buildings, infrastructure 4 181 Lab. and Workshops 1 979.3 Lab. and Workshops 2 932 Accelator and Source incl. ID’s and FEs 5 366.0 Accelator and Source incl. ID’s and FEs 6 026 Beamlines, Experiments 13 698.9 Beamlines, Experiments 13 992 Computing Infrastructure 2 129.8 Computing Infrastructure 1 036 Other Capital costs 61.0 Other Capital costs 86

Unexpended committed funds Funds carried forward to 2014 5 586.0

Total 114 360.0 Total 102 243

2014 manpower (posts filled on 31/12/2014) Scientists, Engineers, Technicians and PhD Students Total Senior Administrators Administrative Staff

Staff on regular positions Accelerator and Source* 36 42 1 79 Beamlines, instruments and experiments* 233 88 21 342 * Including scientific General technical services 29 52 81 staff on time limited Directorate, administration and central services 37 53 90 contracts. Sub-total 335 235 22 592

Other positions Short term contracts 35 35 Staff under “contrats de professionnalisation” 33 33 European Union grants 0 Temporary workers 0

Total 370 268 22 660

Scientific collaborators and consultants 6 6

150 ESRF Facts and Figures

Organisation chart of the ESRF (as of January 2015)

Administrative and Finance Council Science Advisory Committee Committee Chairperson: Chairperson: D. McMorrow Chairperson: A-M. Hundere B. Girard up to 3 delegates per Machine Advisory Committee Contracting Party Chairperson: R. Walker

Beam Time Allocation Panels Director DG’s Services: C01 (ID01, ID03, BM25B, BM32) Chairperson: A. Stierle General • Head of DG’s Office/Secretary of Council: I. Echeverria C02 (ID11, ID15A, ID15B, ID22) Chairperson: A. Beale F. Sette • Communication Group C03 (ID12, ID20, ID26, ID32, BM28) Chairperson: H. Tjeng • Business Development Office C04 (ID24, BM23, BM01B, BM08, BM20, BM25A, • Internal Auditor BM26A, BM30B) Chairperson: C. Lamberti • Safety Group C05 (ID06-LVP, ID09A, ID18, ID27, ID28, BM01A) Chairperson: U. Schwarz C06 (ID17, ID19) Chairperson: A. Olivo C07 (ID16A, ID16B, ID21) Chairperson: I. Fairchild C08 (ID02, ID13, BM26B) Chairperson: P. Stepanek C09 (ID09B, ID10, BM02) Chairperson: A. Plech C10 (Structural Biology Beamlines) Chairperson: M. Czjzek

ACCELERATOR AND SOURCE ADMINISTRATION EXPERIMENTS Director: P. Raimondi Director: L-J. Sánchez Ortiz

Beam dynamics and applications Personnel Director: J. Susini Director: H. Reichert Power supplies Finance Insertion devices Purchasing Beamline groups: Beamline groups: Front-ends • Structural biology • Dynamics and extreme Radio frequency and linac • Structure of soft matter conditions Diagnostics Joint ILL/ESRF • X-ray imaging • Electronic structure and Operation Medical service magnetism Secretariat Joint ILL/ESRF • Structure of materials Library Scientific infrastructure Theory User office CRG liaison Secretariat

INSTRUMENTATION SERVICES AND TECHNICAL INFRASTRUCTURE DEVELOPMENT DIVISION DIVISION Acting Head: A. Barrett Head: R. Dimper

Advanced analysis and modeling Management information systems and web Detector and electronics Systems and communication Software Buildings and infrastructure X-ray optics EX2/CPER project team Mechanical engineering Survey and alignment Vacuum

The European Synchrotron 151 HIGHLIGHTS 2014 Facts and Figures

Cr ove Cover design by M. Scandella featuring the image

“The MoFe7S9C cluster in a nitrogenase enzyme active site” by R. Bjornsson from the article HERFD XAS reveals a Mo(III) in the nitrogenase active site, R. Bjornsson et al., p22. Back cover, thumbnail images: X.B. Zeng et al., p36; S. Cusack et al., p81; M. Moretti-Sala et al., p12.

We gratefully acknowledge the help of:

C. Argoud, T. Baudoin, B. Boulanger, K. Clugnet, M. Cotte, E. Dancer, R. Dimper, I. Echeverría, S. Gerlier, L. Graham, V. Honkimaki, A. Joly, A. Kaprolat, M. Krisch, G. Leonard, C. Mueller-Dieckmann, C. Mary, J. McCarthy, T. Narayanan, S. Pascarelli, P. Raimondi, H. Reichert, S. Rio, M. Scandella, F. Sette, L. Stone, J. Susini and all the users and staff who have contributed to this edition of the Highlights.

Editor G. Admans Layout Pixel Project Printing Imprimerie du Pont de Claix

© ESRF • February 2015

Communication Group ESRF CS40220 38043 Grenoble Cedex 9 • France Tel. +33 (0)4 76 88 20 56 Fax. +33 (0)4 76 88 25 42 http://www.esrf.eu

152 ESRF ESRF - HIGHLIGHTS 2014 The European Synchrotron The European Cedex 9, France - F38043 Grenoble CS 40220 www.esrf.eu HL2014-19-01-4-impression.pdf 1 26/01/15 16:35 1 HL2014-19-01-4-impression.pdf K M Y C CM MY CY CMY N J M C CJ MJ CM CMJ