THE NEWS MAGAZINE OF THE ESRF - ALSO AT http://www.esrf.fr/info/science/newsletter

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 YJUNE 2001 N° 35

ISS

N 1

011-

9310 Microtomographic imaging of snow and ice increases our understanding of the

environment: 3-D reconstructed views of snow taken from the site of an avalanche permitidentification of the granular layers that contribute to their occurrence.

ESRF - JUNE 2001

2

COMMENT:

My feelings on arriving at theESRF in January as the new DirectorGeneral were a mixture ofapprehension and anticipation. I wasaware that there were considerablegaps in my knowledge of the ESRF,but I looked forward with someexcitement to the new colleagues andnew challenges which awaited me.Despite several months ofpreparation – sitting in on meetingsof the Direction and meetings of theScience Advisory Committee andCouncil, building on many years as aUser and as a scientist in charge of aCollaborating Research Group –January was a tough month. As wereFebruary, March, April and May!Many hours of negotiation on theCollective Agreement with ourcolleagues from the unions havegiven me a profound insight into theworkings of the ESRF as a humanenterprise rather than just awonderfully effective researchinstitute. The preparation for ourthree major committees – theAdministrative and FinanceCommittee, the Science AdvisoryCommittee and the Council – haveexpanded my understanding of howthe ESRF is organised with respect toits many member and associatecountries (16 at the last count).My training is far from over, butI now feel that I am beginningto appreciate the major issuesfacing us.

For the ESRF is now in a processof profound change. The hecticactivity of the construction period isgiving way to a phase of scientificexploitation of all 30-plus beamlines.But this does not imply that ourbeamlines and instruments willbecome static, frozen into theircurrent configurations. One of thepillars of the ESRF’s success hasbeen the continuous development of

the machine and beamlines. This willcontinue as we make every effort toremain as international leaders in theuse of synchrotron radiation.

This is also a period of majorstaff changes. We continue to benefitfrom the skill and experience of newyoung staff, predominantly in thescientific and technical areas. By themiddle of 2002, all of the previousDirectorate will have been replaced.Such a rapid turnover of seniormanagement would not be possiblewithout the conscientious support ofthe ESRF staff.

What are our major projects forthe next few years? The Medium-Term Scientific Programme (MTSP),about to be updated, foresees excitingdevelopments spanning the wholerange of the ESRF’s activities. Newextensions of photoemissionspectroscopy to higher energies willopen up new research possibilities,while high magnetic field facilitiescould revolutionise the study ofmagnetic materials. The ESRF’sadvanced techniques in microscopyon the nanometre scale will have amajor impact in the crucial areas ofnanotechnology. In the life sciences,a “New Partnership for StructuralBiology” will unite the ESRF with theEMBL and the ILL in an ambitiousprogramme to create a majorEuropean centre for biologicalstructure determination andgenomics. We have already carriedout extensive consultations with theESRF’s Users in the preparation ofthe MTSP. This partnership will be

even more important during theimplementation of the programme.

We will need to look carefully atour existing infrastructure andorganisation in the light of these andother projects. The ESRF’s buildingsare already bursting at the seams –we must face the problem of ourcritical lack of office and laboratoryspace. Do we need a new building orbuildings? Are our internalorganisational structures appropriateto the current size of the ESRF andour role in the future? In some casesit is evident that we need to thinkabout how to improve ourorganisation to face the newchallenges ahead of us. Is ourcommunication adequate, internallyand towards the world outside? Onceagain, this is a question that requirescareful consideration. Nevertheless,given the experience and expertise ofour excellent staff, I have no doubtthat we will succeed with thetransformations required to preparethe ESRF for a long and successfulfuture as the world’s leadingsynchrotron radiation laboratory.

A final word of thanks to mypredecessor, Yves Petroff: his effectiveand imaginative leadership hasensured that the ESRF and its staffare prepared to meet the challengesof future years. On behalf of all of ushere, thank you Yves, and enjoy yourtime in Berkeley, once again carryingout experiments as a workingphysicist.

Bill Stirling

A PERSONAL VIEW BYTHE DIRECTOR GENERAL

Yves Petroff and Bill Stirling.

ESRF - JUNE 2001

3

2222

4141

• Report from HERCULES 2001,J. L. Hodeau, page 4.• Eleventh ESRF Users' Meeting,K. Hämäläinen and M. Cooper, page 5.• Workshops Associated with the

Users' Meeting: “High-Throughput Structural Biology”,

S. McSweeney, page 6; “Science at High Pressure”,M. Mezouar and M. Hanfland, page 7; “EnvironmentalStudies Using Neutron andSynchrotron Facilities”,Å. Kvick, page 7.• Medical Applications ofSynchrotron RadiationWorkshop,W. Thomlinson, page 8.• Frank Kreith Distinguished Lecturer Award, page 9.• Future Applications of Science with SynchrotronRadiation and Free-Electron Lasers in Europe,P. Lindley, page 10.

• Workshop on Nuclear Inelastic Scattering,R. Rüffer and E.E. Alp, page 10.

• History of the Production of a CD-ROM:Synchrotron Light… to Explore Matter,

D. Cornuéjols, page 11.34th and 35th Meetings of the Council, K. Witte, page 12.• Change of Director in the Machine Division, page 13.• Operation with Users during 2000, R. Mason, page 14.• Diagnostics and Instrumentation for ParticleAccelerators - DIPAC 2001 Workshop, K. Scheidt, page 15.

• Structural Determination ofCatalytically-Active Ag+ Sites in anAg-Y Zeolite: a Combined AnomalousXRPD and EXAFS Study, C. Lamberti, M. Milanesio,C. Prestipino, S. Bordiga, A.N. Fitchand G.L. Marra, page 16.• Glancing-Incidence Diffraction Anomalous FineStructure of InAs/InP Self-assembled Quantum Wires,S. Grenier, M.G. Proietti, H. Renevier, L. Gonzalez,J.M. García and J. García, page 19.

• X-ray and Neutron Studies of the OptimisedSynthesis, the Structure and the

Transformations Involving NovelIon Exchangers, J. B. Parise, page 22.• 3-D Snow and Ice Images by X-rayMicrotomography, C. Coléou andJ. M. Barnola, page 24.• Investigation of Positive ElectrodeMaterials for Lithium Batteries by Means

of X-ray and Neutron Diffraction, C. Masquelier,M. Morcrette and G. Rousse, page 27.

• Structural Genomics: Pitfalls andProspects, A. Bridges and O. Jenkins,

page 35.• Automated Data Collection and Processingfor Macromolecular Crystallography,

A.G.W. Leslie, page 36.• Automation of the MacromolecularCrystallography Beamlines at the ESRF,E.P. Mitchell, page 39.

• Damping Links to Attenuate Vibrations ofMagnet Girder Assemblies, L. Zhang, page 41.

ENVIRONMENTWORKSHOP

MACHINEDEVELOPMENT

• In situ Synchrotron X-ray DiffractionStudies of HP-HT Synthesis of SuperhardPhases in the B–C–N System, V. L. Solozhenko,page 30.• Weird Metals – Modulations within Guests withinHosts, M.I. McMahon and R.J. Nelmes, page 33.

44 NEWSLETTERIN BRIEF

1616 EXPERIMENTSREPORTS

HIGH PRESSUREWORKSHOP3030

3535 BIOLOGYWORKSHOP

Photography by:G. Admans, E.E. Alp,C. Argoud, S. Claisse,B. Denis, K. Fletcher.

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

4

The 11th session of theHERCULES (Higher EuropeanResearch Course for Users of LargeExperimental Systems) course was heldfrom 4 March to 11 April 2001. Thisschool was organised by theUniversities in Grenoble (UJF andINPG), the University of Paris-Sud, andhad the support of several large researchfacilities and laboratories (ESRF, ILL,CNRS, CEA, LURE, LLB, IBS,EMBL). It was dedicated to the use of"Neutron and Synchrotron Radiation"and offered two specialised sessions"Physics and Chemistry of CondensedMatter" and "Biomolecular Structureand Dynamics". This year the 72selected PhD students and post-doctoralscientists came mainly from the EC,with a few from Eastern Europe(Bulgaria, Georgia, Rumania, Russia,Slovakia), America (Brazil, Canada,Mexico, US), and Israel.

Participants were taught throughlectures, tutorials and practicals.Lectures were given by 63 speakers,each an expert in the field concerned.The participants received 50 hours of

personalised training, in groupsof four, from the 100 researchscientists and teachers who tookcharge of the practical andtutorial sessions. The scientistsand teachers came from the localscientific community andrepresented a wide range ofnationalities. Another specificityof the HERCULES course wasits practical training organised atthe neutron and synchrotronbeamlines of the large-scalefacilities, such as the ESRF,ILL, LURE and LLB. Theparticipants gained “hands-on”practical experience of usingneutron and synchrotronradiation for the study ofmaterials, which aimed atdemonstrating the importance ofeach method used. Furthermorethey were able to establishcontact with the scientists incharge in the large-scalefacilities, and so they had theopportunity to discover thepotential existing there.

The programme of the course wasstructured to highlight the impact offundamental and technologicalbreakthroughs in research usingsynchrotron radiation and neutronscentred at the large-scale facilities. Asusual the scientific content included avariety of methods (diffraction, inelasticscattering, absorption imaging, etc.)which were applied to variousmaterials. The programme of theBiomolecular session reflected thecontent of volume IV of theHERCULES series entitled "Structureand Dynamics of Biomolecules" (19contributions) which was published in2000 by Oxford University Press.

Although the schedule was as tightas usual, there was some time forparticipants to exchange scientific andnon-scientific ideas, during the postersession and the skiing outing. They alsojoined in common musical and dancingexperiments with the help of a jazzorchestra during a dinner party in themountains, and, like last year, thecourse ended with a wine and cheeseparty preceded by a multi-disciplinaryconference on Hercules, the Greek hero.

The HERCULES course will, ofcourse, be organised again in 2002.

J-L. Hodeau

HERCULESREPORT FROM HERCULES 2001

Participants and organisers of the HERCULES 2001 Course on Neutron andSynchrotron Radiation for "Physics and Chemistry of Condensed Matter" and"Biomolecular Structure and Dynamics".

HERCULES 2002HIGHER EUROPEAN RESEARCH COURSE FOR

USERS OF LARGE EXPERIMENTAL SYSTEMS

Grenoble, 17 February - 28 March 2002

Session A:"Neutron and synchrotron radiation for physics

and chemistry of condensed matter"

Session B:"Neutron and synchrotron radiation forbiomolecular structure and dynamics"

Information:Secrétariat HERCULES

CNRS - Maison des MagistèresBP 166 - 38042 Grenoble Cedex 9

Tel: 33 (0)4 76 88 79 86Fax: 33 (0)4 76 88 79 81

e-mail: marie-claude.moissenet@polycnrs-gre.frhttp://www.polycnrs-gre.fr/hercules.html

Deadline for application: 16 October 2001

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

5

This year, the Users' Meeting wasfocussed on the ESRF's MediumTerm Scientific Plan for most ofthe plenary session on Monday19 February. This theme attracted justas many participants as the traditionalformat of reports and researchhighlights, with outside usersconstituting almost half of the ~ 400registrations. Once again this healthyparticipation rate is due to theattraction of the accompanyingworkshop programme.

Bill Stirling opened the meetingwith a historical view of developmentsat the ESRF, summarising the majoradvances to date. He stressed theimportance of the Users’ involvementto improve the Medium TermScientific Plan (MTSP), and he invitedtheir suggestions and criticism of theplan drafted beforehand by the ESRF'sscientists and management. After apresentation about machinedevelopments by Pascal Elleaume,there followed a succession ofpresentations by ESRF scientistsengaged in crystal ball-gazing acrossthe spectrum of experimentalsynchrotron radiation activities. Theydrew upon the discussions opened atthe ESRF Local Scientists' Meeting,which took place at the end of May2000 at Lac d'Annecy. Each speakerrepresented a group of beamlines andmade predictions that covered a wide

range of research areas. Thepresentations, by and large, tookexamples of recent research asexemplifying "The Way Forward",with progress coming from evolutionrather than revolution. "Better, Fasterand Smaller" was the recurrent themewith just a few specific suggestions fornew ventures.

The experiment of holdinglunchtime focus group discussionsprovided varied points for theafternoon feedback session and not toomany complaints that talking shouldcome a poor second to eating at thattime of day. Some lack of focus in thefeedback session perhaps reflected theabsence of specific proposals in thecurrent MTSP discussion document,but some recurrent themes did emerge.The Users’ Organisation, through itsnew Chair, Keijo Hämäläinen(University of Helsinki) will besoliciting further feedback fromattendees over the next few months.

The traditional meeting format wasregained mid afternoon with MalcolmCooper introducing the new membersof the Users’ Organisation. He thenhanded over to Keijo before theannouncement of the winner of theYoung Scientist prize. Keijo listed theprevious winners of this award andpointed out that in 1995 the winner wasFrancesco Sette, who will soon

become a research director of theESRF. Keijo concluded by saying thatas much is also expected of this year'swinner. Jens Als-Nielsen, chairman ofthe award committee (the committeealso included John Helliwell, AlanLeadbetter and Christian Vettier), gavea brief introduction to the scientificimportance of the work done by thewinner before inviting GuillaumeFiquet to collect his prize and to makea presentation. Guillaume describedhis work under the title "The Physics ofthe Earth's Interior".

Keijo drew the plenary session to aclose by presenting flowers to theUsers’ Office and CRG Office staffwho had organised the meeting. To adda little humour, he even offered abunch to Malcolm in thanks for hiswork as chairman of the Users’Organisation over the last two years.Later on there was a lively postersession, which ended with a receptionhosted by Bill Stirling. Furtherlubrication was available to aid orhinder scientific discourse over abuffet dinner.

Our thanks go to the ESRF staffmembers who helped organise thismeeting, particularly the CRG andUsers' Office staff.

K. Hämäläinen and M. Cooperon behalf of the Users’ Organisation

Young Scientist Prize Winner G. Fiquet (left),with J. Als-Nielsen (centre) and K. Hämäläinen (right).

ELEVENTH ESRF USERS' MEETING19 February 2001

The Users' Organisation Committee2001: (from left to right)O. Figueiredo, P. Charlier,D. Nicholson, J.-Y. Buffière,M. Cooper, F. Bosherini (retiringmember), F. D'Acapito,P. Fairclough and K. Hämäläinen.(absent: A. Kaprolat and R. Kahn).

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

6

There were three satellite workshops to the Users' Meeting: "High-Throughput StructuralBiology", "Science at High Pressure" and "Environmental Studies using Neutron and

Synchrotron Facilities". The workshops are introduced briefly here, and selectedpresentations from each workshop are included as articles in this edition of the Newsletter.

The recent publication of the draftversion of the Human Genome hashighlighted the growing need forautomation of the process of structuredetermination for biomolecules. Someof the issues around this problemwere addressed in the High-ThroughputStructural Biology Workshop, held at theESRF on the 20 and 21 February 2001.

The scientific programme addressedthe concerns of those involved instructural biology, whether they areaffiliated to proprietary or academicgroups or to synchrotron radiation sources.The meeting was set in the context of theavailability of the human genome withopening lectures by I. Mattaj (EMBLHeidelberg, Germany), and T. Hubbard(Sanger Centre, Cambridge, UK). Thechallenges of high-throughput proteinproduction, purification and crystallisationwere addressed from an academic andcommercial perspective. Plans to adapt

synchrotron radiation beamlinesto allow for a very high degree ofautomation were presented byP. Kuhn (SSRL), B-C. Wang(Univ. Georgia), and E. Mitchell(ESRF). Following on from thisthe techniques for automatic datacollection, analysis, structuresolution and refinement werediscussed.

The use of bioinformatics asboth an aid for target selection,protein structure evaluation, aswell as a tool for genomeannotation and analysis werediscussed in another session.The meeting was brought to aconclusion by M. van der Rest

(IBS, Grenoble, France) who asked thequestion “Where do we go from here ?”This address then led into a livelydiscussion session.

A number of talks made the point thatproteins and enzymes do not work inisolation but are normally components inmore complex processes and thus the"structural genomics approach" may not

always be the answer to the scientificquestion being asked. However, giventhat, in the USA, the National Institute ofMedical Sciences has predicted the needfor around 10,000 structures to be solvedin the next ten years, "high throughput" islikely to be one of the key phrases instructural biology for a number of years tocome.

Our intention in drawing up theprogramme was that the meeting shouldaddress both the practical difficulties andthe scientific opportunities brought by themassive increase in genomic information.

The three presentations reported herefocus on various areas of the programme.In the first article "Structural Genomics:-Pitfalls and Prospects", on page 35,A. Bridges describes how the genomesequence is used for drug design, and theneed for scale-up and automation.The next two articles concernautomation of the process of proteinstructure determination: in "AutomatedData Collection and Processing forMacromolecular Crystallography", onpage 36, A. Leslie discusses some ofthe ideas for complete automation ofdata collection and demonstratesthat this goal may well be feasiblewithin a relatively short time. Finally,E. Mitchell discusses the practicalimplementation of automation processesat synchrotron radiation sources, in thearticle "Automation of the MacromolecularCrystallography Beamlines at the ESRF",on page 39.

S. McSweeney

HIGH-THROUGHPUT STRUCTURAL BIOLOGY

… Eleventh ESRF Users’ Meeting

WORKSHOPS ASSOCIATED WITH THE USERS' MEETING

Posters and aperitif.

A. Leslie,speaking aboutautomated data processing.

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

7

The ESRF and the ILL jointlyorganised a two-day workshop onenvironmental studies on the 20 and 21February 2001. The workshop attractedmore than one hundred scientists in thefield and featured 19 invited speakers.

The first session addressed nuclear andchemical pollution. It included talks aboutnuclear waste transmutation, characterisationof particles, and heavy-metal speciation. Theanalytical techniques discussed were neutronreflectometry, synchrotron-based X-raymicrobeam techniques, and EXAFS studies.The second session was about generalpollution detection and waste disposal,and included talks on neutron activationanalysis, probing natural material withmicrospectroscopy using synchrotronradiation, the environmental relevance of gashydrates, and recent advances in analysis ofheavy metals in plant tissues. The sessionwas concluded with three talks on snow andice. A poster session and a press conferencerounded off the day.

The theme for the second day wasenergy and efficient energy storage. Topicsincluded: the tracking of metals in geofluids;use of SANS and SAXS in petroleumgeology; attempts to produce efficienthydrogen storage in metals; newenvironmentally-friendly battery materials;and a study of hydrocarbons in zeolites.Following lunch, participants were invited totake a tour of the ESRF and ILL facilities.

The workshop was closed by ananimated discussion of the future need ofthe ILL and the ESRF in environmentalresearch. Topics like time-resolved studies,microfocussing, in situ studies, neutronactivation and combinations of methods

were of great interest. The managementwas also encouraged to provide means ofrapid access to the facilities for urgentstudies.

Three talks have been selected forinclusion in this edition of the Newsletter,and demonstrate well the variety of topicscovered. The first, which was also theintroductory talk of the workshop, is entitled"X-ray and Neutron Studies of theOptimised Synthesis, the Structure and theTransformations Involving Novel IonExchangers" by J. Parise (State Universityof New York, USA), on page 22. Thesecond "3-D Snow and Ice Images by X-rayMicrotomography", on page 24, is byC. Coléou and J.-M. Barnola (Météo-France, Centre d’Etudes de la Neige andLaboratoire de Glaciologie et Géophysiquede l’Environnement, CNRS, Grenoble).Finally, "Investigation of Positive ElectrodeMaterials for Lithium Batteries by Meansof X-ray and Neutron Diffraction", page 27,by C. Masquelier (Laboratoire de Réactivitéet de Chimie des Solides, CNRS, Amiens).

Å. Kvick

The workshop entitled "Science atHigh Pressure: Latest Trends from Third-Generation Sources" was held at theESRF on the 16 and 17 February 2001,accompanying the traditional Users’Meeting. The workshop attracted about100 participants. Twenty-one scientists,representing the main research groupsactive in the field across the globe, gavesome very interesting presentations on thelatest contributions of third-generationsynchrotron sources to high-pressurescience. The programme was divided intothree sessions reflecting the most relevantactivities in the field: i.e. earth sciences,chemistry and hard-condensed matter.The presentations focussed on the hottestscientific questions relating to thesefields and were followed by dynamicdiscussions. A variety of techniquessuch as X-ray diffraction, Mössbauerspectroscopy, inelastic X-ray scattering,Compton scattering and EXAFS werepresented. Two of the presentations wereselected for articles: High-pressuresyntheses of superhard materials "In situSynchrotron X-ray Diffraction Studiesof HP-HT Synthesis of Superhard Phasesin the B–C–N System" are presentedby V. Solozhenko, on page 30. Unusualstructures found in metals at high-pressures are described by M. McMahonin the article "Weird Metals –Modulations within Guests within Hosts"on page 33.

M. Mezouar and M. Hanfland

SCIENCE AT

HIGH PRESSURE

ENVIRONMENTAL STUDIES USING NEUTRON AND

SYNCHROTRON FACILITIES

F. Guyot (centre) gave a plenarylecture entitled "New Opportunitiesfor High-pressure Geophysics atThird-Generation Synchrotrons". He is pictured here with G. Fiquet(left) and M. Krisch (right).

Visiting the ILL research facilities: A. Hewat, ILL,showing the ILL D19 diffractometer

to participants of the workshop.

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

8

A scientific meeting of elderstatesmen and young scientists is a goodway to summarise the MedicalApplications of Synchrotron RadiationWorkshop, held at the ESRF between1 and 3 March 2001. The workshopwas jointly funded by the ESRF, theCentre Hospitalier Universitaire deGrenoble (CHU), and the UniversitéJoseph Fourier. The programme wascomprehensive in its breadth ofsynchrotron radiation research in themedical field. Over 100 participantsfrom 12 countries celebrated theextraordinary growth of the field duringthe 15 years since the first humanangiograms were obtained at SSRL.

The workshop opened with historicaloverviews by the organisers W.Thomlinson, F. Estève and JF. Le Bas.E. Rubenstein then initiated the scientificsessions with a lecture describing a newpotential application of our technologyto neurodegenerative disorders of thebrain. This talk by the director of thefirst medical applications programme –human coronary angiography at SSRL –a high point of the workshop as he tiedtogether the older mature programmesand new and exciting prospectiveresearch.

The growth of the field has beenmade possible because of the existenceof dedicated medical research facilities.This was the theme developed in a talkby E. Castelli who highlightedthe facilities at ELETTRA, NSLS,HASYLAB, Photon Factory, andour new one at ESRF. Several newfacilities are being constructed orplanned in Korea, Canada and Japan,giving evidence for the healthy growth ofthe field over the coming years.

The first session on imaging wasdevoted to the bronchography projectthat is a collaboration between the ESRF,the CHU, and the University of HelsinkiHospital and Physics Department. Anoverview of the present status and futureneeds in the field of lung functional andstructural imaging was presented by C-G. Standertskjöld-Nordenstam, followedby an overview of the local project byS. Bayat.

Radiation therapy took centrestage for two sessions. The first of threemajor areas of research presentedwas photon activation therapy.J. Balosso presented results on the use ofplatinum to increase the response toradiation doses by cells in an in vitro cell

model. A second talk on this topic wasgiven by B. Laster who highlighted theuse of incorporated target atoms forovercoming the small cross-sectionfor dose delivery. H. Blattmannpresented the microbeam radiationtherapy project that is being pursued onthe ESRF ID17 beamline by a team fromthe Univ. of Bern, Switzerland. Hepresented the very exciting resultsobtained so far on small and largeanimals. If the programme continues toshow success, it may be possible toadvance it to the stage of human researchfor treatment of some brain tumours ininfants and small children. A. Normanpresented the concept of using mono-energetic X-rays to improve thetherapeutic ratio in X-ray tomographytherapy of brain tumours. At the ESRF, ateam of researchers from the CHU isdeveloping the necessary synchrotron-based protocol for potential extension tohuman research. H. Elleaume presentedthe status of this programme.

The workshop highlighted humancoronary angiography, the onlyprogrammes involving in vivo humanstudies. W R. Dix described the results ofthe extensive programme at HASYLABon 370 patients. His summary described

MEDICAL APPLICATIONSOF SYNCHROTRON RADIATION WORKSHOP

1/3 March 2001

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

9

the strengths and problems associatedwith the synchrotron imaging comparedwith MRI. S. Ohtsuka discussedthe results at the Photon Factory.B. Bertrand described the resultsobtained during the first protocol on 32patients at the ESRF.

F. Berger presented a critical analysisof the importance of angiogenesis withemphasis on therapeutic applications.Recent experiments have been carriedout at the ESRF to develop thetechnology of measuring cerebral bloodvolume. JF Adam and I. Troprèspresented their results and showed a goodcorrelation between MRI andsynchrotron CT measurements.

Medical imaging at synchrotronsis now utilising phase information.P. Cloetens presented an overview of thevarious methods being developed forphase contrast imaging. M. Andodiscussed a novel use of asymmetricreflection optics to simultaneously obtainimages of a sample that containabsorption, phase contrast, and refractioninformation. S. Fiedler described thefundamental concepts of diffractionenhanced imaging being pursued at theESRF.

The two previous workshops ofthis series, held in Japan, featuredoutstanding hospitality in wonderfulmountain settings. Hot springs andkaraoke brought the participants togetherin happy, interactive gatherings. InGrenoble, we have not only mountainsbut also genuine French châteaux! It isfair to say that the evening at the Châteaudu Mollard will never be forgotten.Starting with aperitifs and musicprovided by a joint ESRF-CHU group,the evening rapidly warmed up! Duringthe early course of dinner, the youngItalian participants started singing. Fromthat point on we were rewarded withsongs by our colleagues from around theworld in French, Japanese, English,Finnish, Italian, and Swedish.

The workshop had three sessionsdevoted to research on cell/tissueimaging and the analysis of tissues. Therole of X-ray scattering in medicine wascritically reviewed by J. Doucet. Heranked the contributions of scattering tomedical research as important, but its

importance to diagnosis and screening isfar from being demonstrated. Talks byM. Fernandez and A. Evans discussed theanalysis of cancerous tissues usingscattered radiation. F. Peyrin discussedthe application of synchrotron radiationmicrotomography to the evaluation ofbone quality.

The development of the diffractionenhanced imaging (DEI) techniquehas renewed interest in synchrotronradiation mammography. E. Pisanomotivated the mammography studieswith an overview of the current statusof diagnosis in the clinic and the need forbetter imaging. She and A. Dilmaniansummarised the results obtained by theDEI group in the United States. Recentexperiments at the ESRF were presentedby A. Bravin and F. Arfelli discussed theimages recorded at the SYRMEPbeamline at ELETTRA.

In his conference summary, P. Suorttihighlighted the many new directions ofresearch, built upon 25 years ofinstrumentation and science. A challengewas given to the participants to find asolution to the serious problem of the lackof a compact source technology. Withoutsuch a source, many of the promisingtechnologies will never be applicable inthe clinic. The next workshop of thisseries will be in Trieste, Italy, inSeptember 2003.

W. Thomlinson

Bill Thomlinson has beenselected to receive the FrankKreith Distinguished LecturerAward from the Faculty ofEngineering Sciences of the BenGurion University of the Negev inBeer-Sheva, Israel.

This award recognises andacknowledges Bill's capabilities inconducting experimental researchfor the application of synchrotronradiation to medicine, and for hisability to bring forth to thescientific community a betterunderstanding of the potentialbenefits of this importantinstrument for medical research.

Bill Thomlinson,beamline scientist

at the MedicalBeamline, ID17.

FRANK KREITH

DISTINGUISHED

LECTURER AWARD

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

10

The workshop took place at theESRF on 16 and 17 March 2001. It wasorganised following several suggestionsemerging from the 2000 meeting of theEuropean Round Table for SynchrotronRadiation and Free Electron Lasers(FEL's) held in Karlsruhe in September2000. The main purpose was to initiatean open and broad discussion on thedevelopment of new sources in Europeover the next 10-20 years. The objectiveswere a clarification of the possibledevelopment of lines after the recent

approval of the Soleil and Diamondprojects, and –if possible– a consensuson priorities. The discussionsencompassed large storage rings, linac-driven self-amplified spontaneousemission (SASE) devices and energyrecovery linac (ERL) based systems.The participants included experts onaccelerators and X-ray sources fromEurope and the USA, and a limitednumber of X-ray users. A school on thesame theme at Les Houches precededthe workshop in Grenoble.

Two days of presentations anddiscussions were extremely helpful inclarifying the foreseeable avenues oftechnical development. They alsorevealed several important points ofbroad consensus that could providethe basis for a realistic and effectivelong-term strategy in Europe. A reportto be presented to the EuropeanCommission by the chairman of theRound Table (G. Margaritondo) is inpreparation.

P. Lindley

FUTURE APPLICATIONS OF SCIENCE WITH SYNCHROTRON

RADIATION AND FREE-ELECTRON LASERS IN EUROPE

16/17 March 2001

After the preceding workshop fouryears ago at APS (Argonne, USA) theNuclear Resonance groups organiseda two-day workshop on NuclearInelastic Scattering at the ESRF on 23

and 24 April 2001. The aim was tobring together the developers of themethod, the experimentalists and thetheoreticians in order to discuss thisnew field. About forty participants

from North America, Japan andEurope listened to sixteen talks on thedevelopment, prospects and highlightsfrom the three third generationsynchrotron radiation sources (APS,

ESRF and SPring-8) andon applications specially devotedto high pressure, biology/chemistry and surfaces/interfaces.They were accompanied by talkson the theoretical understandingand treatment of the data withample time for discussions. Theparticipants look forward tocontinuing this successful seriesof workshops and suggested thatthe next one be held at SPring-8(Japan).

R. Rüffer and E.E. Alp

WORKSHOP ON NUCLEAR INELASTIC SCATTERING

23/24 April 2001

Attendees at theNIS Workshop.

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

11

When I arrived at the ESRF at thebeginning of 1993, as InformationOfficer, I was in charge both ofinformation to the scientific community(Newsletter, Annual Report) and ofcommunication with the public,particularly through guided tours, opendays, and events such as the FrenchScience Week. My PhD in physicshelped for the first part of my work, butI felt that I had few resources for thesecond part. The most basic book thatcould be found about the subject was“Introduction to Synchrotron Radiation”by Giorgio Margaritondo, but even thiswas intended for potential users ofsynchrotron radiation. Communicatingto the public in terms that can be easilyunderstood is quite a difficult task and Istill consider it the main challenge of myjob today.

Most of the visitors coming to theESRF – about 3000 every year – arestudents and scientists. However, thereare also secondary school children andretired people, as well as professionals ofall sorts (engineers, doctors of medicine,computer specialists, businessmen,journalists and politicians…). Thefeature common to all these visitors istheir curiosity about science, even whenthey do not have scientific backgrounds.They are generally enthusiastic aboutsynchrotron technology and science andthey sometimes asked so manyquestions that it was difficult for me toend the visit.

A problem became evident: On onehand, due to lack of personnel, we wereunable to accept all the demands forvisits and, on the other hand, not all ofthe school children and universitystudents from European countries wouldbe able to afford to make a trip to theESRF. It was obvious that somethingwas needed to replace the traditionalvisit and that a brochure was notsufficient.

The idea of producing a CD-ROMemerged quite naturally and, as soonas 1994, with the very fruitful

collaboration of the DaresburyLaboratory and some financial helpfrom the European Commissionthrough the European Science Week,we produced a pilot CD-ROM entitled“Materials for the future”. This was afirst step but going further needed moreresources.

In 1996, the launching of theEuropean “Info 2000” programmeoffered us the opportunity to present anambitious CD-ROM project toBruxelles, associating various partners.Our proposal was not retained and,again, we had the problem of how tofinance the project. Finally, in 1998, asolution was found, involving threepartners: the ESRF as the producer ofthe CD-ROM contents, iMediaSoft asthe multimedia developer and Springer-Verlag as the worldwide distributor.

Subsequently followed many yearsof hard work involving many scientificcolleagues, which culminated in therelease of the CD-ROM “SynchrotronLight… to explore matter”, in February2001. I must say that the reception ofthe CD-ROM has been extremelypositive, particularly among teachers.At a time when fewer and fewer young

people are interested in science andscientific careers, the multimediaapproach offers the possibility to mixtexts, images, sound and animation - arevolution in scientific teaching. Virtualrepresentations, interactivity andcomputer-generated films offer greatpossibilities to stimulate the young.

The “Synchrotron Light” CD-ROMis only a beginning. I am alreadythinking about a second edition (maybea DVD) with developments in twodirections: a fun discovery of asynchrotron for secondary schoolchildren (with a quiz?) and more aboutscience for high-level students andscientists. For this new venture, I needthe help of all interested people. Pleasecontact me if you have found any errorsin the CD-ROM and/or you have anyideas about how to improve it.

To know more about the contents ofthe CD-ROM itself, please refer to theone-page review written by MichaelHart in the Journal of SynchrotronRadiation (April 2001).

D. CornuéjolsESRF Information Officer

Editorial director of the CD-ROM

HISTORY OF THE PRODUCTION OF A CD-ROM:SYNCHROTRON LIGHT… TO EXPLORE MATTER

The “SynchrotronLight” CD-ROM

may be orderedfrom Springer’s web

site at http://www.springer.de/synchro

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

12

Scientific andinfrastructure matters

The Council approved theMedium-Term Scientific Programmefor the period 2001 to 2005, aspresented by Management, as a basisfor the financial planning for themedium-term period (see also www.esrf.fr/conferences/usersmeeting01/MTSP/index.htm). The Councilagreed that the ESRF Managementshould continue discussions with theEMBL, the ILL and other relevantagencies with a view to forming a"Partnership for Structural Biology",which would provide the basis fora long-term European strategicinitiative in the post-genomic era.

Legal, procedural andfinancial matters

The Council approved the budgetfor 2001, providing for an expenditureof 68 777.2 kEuro in payments, andrequiring Members’ contributions of62 513.2 kEuro. It confirmed 63 451kEuro as the planning figure for newcontributions from Members to thebudget of 2002. It agreed that thecontributions received from Portugaltowards the initial construction costsof the ESRF shall be incorporateddirectly into the budgets of the ESRFand no longer assigned to theMembers’ accounts, with retroactiveeffect from 2000.

The Council noted Management’sintention to extend the arrangementbetween the Institute of Physics ofthe Academy of Sciences of theCzech Republic (FZU) and theESRF concerning the medium-termscientific use of synchrotron radiationfor non-proprietary research (with theextension by a further two years, thecontribution level has been increasedfrom 0.26 to 0.35% of that of the

Members).The Council also authorised the

Director General to sign 5-yearextensions of the contracts with INFNconcerning the GRAAL experimentand with the Grenoble UniversityHospital on scientific collaborationespecially on the Medical Beamline(ID17).

With regard to correctivemeasures to improve the balancebetween scientific use and financialcontributions, the Council adopted aresolution (cf. side panel), noted that,in accordance with this resolution, theMembers from Nordsync and Francewill make additional contributions tothe 2002 budget of 430 302 Euros and44 416 Euros, respectively, and notedthat for any future allocation of beamtime, the French Contracting Partyaccepts a limitation of its scientificuse – keeping it within the acceptablemargin as set out in the Council’sguidelines.

The Council amended the ESRF’sStatutes in order to permit theScientific Associates to nominatetogether one member of the ScienceAdvisory Committee (in addition tothe previously foreseen 22 members).Another amendment of the Statutesand also of Annex 4 to the FinancialRules was implemented in order toadapt these texts to the new Europeancurrency (Euro).

Appointment of Directors,Committees andChairpersons

The Council appointed• P. Elleaume as Machine Director

for the period from 1 March 2001 to 28 February 2006 (J.M. Filhol having resigned from the post in February 2001);

• A new Science Advisory Committee

(see extra box) with J. Bordas as Chairman and L. Braicovich as Vice-Chairman;

• R. Comès as Council Chairman for the period 1 January 2002 to 31 December 2003,

and re-appointed H. Weijma asChairman of the Administrative andFinance Committee for a further year(1 July 2001 to 30 June 2002).

K. Witte

34TH AND 35TH MEETINGS OF THE COUNCIL

27/28 November 2000 and 5/6 June 2001

1. Having regard to• the principle of allocation of beam

time in accordance with scientific excellence and

• the concern with respect to a possible lasting and significant imbalance between the scientific use of the facility and the contributions to its operation, as expressed in § 6.4 of the ESRF Convention,

the Council decides that any suchimbalance shall, in the first instance, beredressed by pragmatic measures of atemporary nature, i.e. withoutamendments of Convention or Statutes,but rather based on a consensualinterpretation of these texts.

2. To this end, at each summermeeting of the Council, the scientificjuste retour situation shall be assessed

RESOLUTION

ON CORRECTIVE

MEASURES(as adopted by the ESRF Council at its35th meeting on 5 and 6 June 2001)

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

13

Jean-Marc Filhol, MachineDirector since January 1997, resignedfrom his functions on 14 February2001 to take up a leading role in theconstruction of the French synchrotronSOLEIL. In his new position as deputyproject director, he is more particularlyin charge of the construction of theMachine and of the infrastructure.

Pascal Elleaume was appointedas Director of the ESRF MachineDivision and began his term of officeon 1 March 2000, following the earlier-than-expected departure of J.M. Filhol.Pascal graduated in physics at theEcole Normale Superieure in Paris andobtained the "Agrégation de Physique"in 1978. After a one-year stay at the

University of California, Berkeley, hejoined the CEA (Saclay) in1980, where he carried out PhDand postdoctoral studies on thedevelopment of the first storage ringbased free electron laser operating inthe visible range of the spectrum. P.Elleaume joined the ESRF in 1986, inthe early days of the project. He wasgiven the task of defining the expectedperformances from the insertiondevices and has contributedsignificantly to the “Red Book”, i.e.the foundation phase report. During theconstruction phase, he was given theresponsibility of setting up theInsertion Devices group which, underhis leadership, has designed, built andoptimised all the ESRF's insertiondevices (more than 60 segments areinstalled on the ring today). The qualityand variety of the ESRF's insertiondevices are unique in the world andtheir innovative concepts have had astrong impact on insertion devicetechnology. Since 1991, he has alsobeen involved in the development ofvarious machine diagnostics based onX-rays (ID6 beamline, pinholecameras) and scientific software (Xray,Radia, SRW).

CHANGE OF DIRECTOR

IN THE MACHINE DIVISION

based on a three-year gliding period, inaccordance with the “Guidelines for aRe-adjustment of Contribution Rates”adopted by the Council at its 29thmeeting on 9 and 10 June 1998.

3. Those Contracting Parties that,according to these figures, aresignificantly overbalanced (returncoefficient > 1.25 or absolutedifference > 3%) shall be requestedby the Council • to make an additional contribution

to the next year’s budget to such extent that, had it been made for each of the previous three years, it would have kept the return coefficient within the acceptable margin, or

• to accept a limitation of their scientific use during the following year, keeping it within the acceptable margin.

Any additional income created by sucha measure should be used by theManagement to increase the availablebeam time or experimental capacity.

4. More permanent measures shall beconsidered in preparation of therenewal of the Convention or in theevent of the accession of newContracting Parties.

MEMBERS OF THE SCIENCE ADVISORY COMMITTEE

FOR THE YEARS 2001/2002G. Artioli (Università degli Studi di Milano),J. Bordas (Universitat Autonoma de Barcelona), L. Braicovich (Politecnico di Milano), H. Dosch (Max-Planck-Institut fürMetallforschung, Stuttgart), R. Fourme (LURE),J.P. Gaspard (Université de Liège), R. Hilgenfeld (Institut für Molekulare Biotechnologie, Jena), K. Hodgson (SSRL, Stanford), D. Juul-Jensen (Risoe National Laboratory,Roskilde), G.H. Lander (Institute forTransuranium Elements, Karlsruhe),

L.B. McCusker (ETH Zürich), P. Monceau (CNRS, Grenoble),H. Reynaers (Katholieke Universiteit Leuven), G. Rossi (TASC Laboratori, Trieste), G.C. Ruocco (Universita dell'Aquila), D. Shechtmann (Technion, Haifa), G. Schneider (Karolinska Institutet, Stockholm),D. Stuart (Wellcome Trust Centre for Human Genetics, Oxford), S. Suga (University of Osaka), D.P. Woodruff (University of Warwick),G. Wortmann (Universität GH Paderborn),T. Zemb (CEA Saclay).

PascalElleaume, new Director of the Machine.

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

14

OPERATION WITH USERS

DURING 2000

During the year 2000, the fullcomplement of 30 ESRF beamlines,together with 8 Collaborating ResearchGroup beamlines (CRGs), were open foruser experiments, the CRGs making1/3 of their beam time available forgeneral ESRF Users. For the purposes of

scheduling experiments, the year wasdivided into two periods, February toJuly 2000, and August 2000 to mid-February 2001, with run periods of six toseven weeks, broken by shutdownperiods for installation, commissioningand maintenance.

Scientists requesting beamtime submit applications fortwo deadlines – 1 March and 1September – each year. For theproposal round in September 2000,users were able to submit proposalselectronically, and some 67% chose touse this new method. For the March2001 deadline all 837 proposals weresuccessfully submitted electronically.

The increase in requests for beamtime since 1994 is shown in Figure 1.It is to be noted that although the mainbeamline construction effort wascomplete by 1999, requests for beamtime continue to rise, and totalled24 824 shifts in 2001. In the past, suchrequests have outweighed the beamtime available for allocation byroughly a factor of two. This patterncontinued during 2000, which sawbeam time allocated to 53% or 12 309of the 23 216 shifts initially requested.Figure 1 also shows the shifts thathave been scheduled for experimentssince the beginning of user operation.

The total number of proposalssubmitted and experiments carried outover the scheduling periods to date isshown in Figure 2, which also showsthe increase in the number of userssince 1994. The most recent proposalrounds have seen a marked increase inthe number of applications arriving inthe area of Materials Science, both forthe Structures review committee andthe new committee Materials,Engineering and EnvironmentalMatters, which was created inSeptember 1999. In the Life Sciences,the number of applications from BlockAllocation Groups (BAGs) alsocontinues to rise, from 23 groups in1999 to 34 late in 2000.

Also seen in Figure 2 is the sharpincrease - some 50% in the past 12months - in the number of user visits,from 3361 user visits in 1999, to 5049in 2000. This reflects a rise of morethan 9% in projects allocated beam

22988

1994 1995 1996 1997 1998 1999 2000

Years

Num

ber

of p

ropo

sals

/ U

sers

Vis

its

5500

5000

4500

4000

3500

3000

2500

2000

1500

1000

500

0

Proposals submitted

Experiments carried out

Number of Users Visits

792

339

1089

518

1259

656

1380

766

1394

915

271

1149

1777

2376

2726

3361

1470

LS B

AG s

chem

e

1146

5049

1994 1995 1996 1997 1998 1999 2000 2001

Years

Num

ber

of s

hift

s of

bea

m t

ime 30 000

25 000

20 000

15 000

10 000

5 000

0

Shifts requested

Shifts scheduled for user experiments

2 252738

10 870

3 703

15 625

6 563

18 287

8 748

19 083

10 071

21 738

12 390

23 216

13 812

24 824

Fig. 1: Details of shifts of beam time requested and scheduled for userexperiments per year, 1994 to 2000.

Fig. 2: Numbers of applications for beam time submitted, experiments carriedout, and user visits, 1994 to 2000.

NEWSLETTER IN BRIEF

ESRF - JUNE 2001

15

DIAGNOSTICS AND INSTRUMENTATION

FOR PARTICLE ACCELERATORS -DIPAC 2001 WORKSHOP

The workshop serves as a forumfor the exchange of the latestexperiences, results and developmentin the field of accelerator beaminstrumentation worldwide. It takesplace at a biennial rhythm, withDIPAC-2001 being the fifth edition.

A total of 150 people attended theworkshop on 13-15 May 2001. Theyrepresented 42 different institutes ororganisations from 14 countries.Among them, 18 came from the USA,4 from Japan, 1 from Taiwan and 1from Brazil.

The intensive programme consistedof 12 invited talks, 11 contributed talks,42 posters and 6 discussion sessions.A guided visit to the ESRF storagering, an industrial exhibition with8 participating companies and anexhibition of 'BPM blocks' werealso organised. With a large part ofthe workshop dedicated to oralpresentations, the more informal anddirect exchange of contacts and

information between the participantstook place during the discussionsessions and the poster sessions.

The workshop covered a largevariety of accelerators (electron, proton,heavy-particles, free-electron lasers,linacs, specific projects) that areoperated to serve very different typesof objectives. This means that manydifferent types of parameters canbe measured, under very differentconditions, implementing very differenttechnical methods and technology. Thepart represented by synchrotron lightsources was about 40%.

Proceedings of the workshop willbe published.

K. Scheidt

Surfaces and Interfaces 6 %Chemistry 11 %

Hard Condensed Matter:Elect. & Magn. Properties 15 %

Hard Condensed Matter:Structures 21 %

Materials Eng. & Environmental Matters 7 %Life Sciences 21 %

Methods and Instrumentation 7 %Soft Condensed Matter 8 %

Other :Training, feasibility tests,

proprietary research 4 %

time overall, an increase in the numberof Users in each experimental team, andthe success of the Life Sciences BlockAllocation (BAG) scheme; the latter userteams make multiple visits to groups ofbeamlines, and are thus able to makemaximum use of the facilities formacromolecular crystallography.

The breakdown of shifts scheduledin 2000, per scientific area, is shown inFigure 3: some 43% of the shiftsscheduled on beamlines were devotedto hard condensed matter and materialsresearch, whilst 21% were devoted tolife sciences projects. Further, in orderto cater to different types ofexperimental requirements, beamdelivered to the stations was made in avariety of modes: multibunch (2 * 1/3filling) – 65%; “hybrid” mode - 5%;16-bunch fill - 24%, and single-bunchfill - 6%, in support of a very diversescientific programme.

Finally, interested readers arereminded that the next deadline forproposals, for beam time betweenFebruary and July 2002, is 1stSeptember 2001. Further details canbe consulted on the Web at http://www.esrf.fr.

R. Mason

Fig. 3: Shifts of beam time scheduledand experiments carried out in 2000,by scientific area.

EXPERIMENTS REPORTS

ESRF - JUNE 2001

16

Z

STRUCTURAL DETERMINATION OF CATALYTICALLY-ACTIVE

AG+ SITES IN AN AG-Y ZEOLITE: A COMBINED

ANOMALOUS XRPD AND EXAFS STUDYC. LAMBERTI1,2, M. MILANESIO1, C. PRESTIPINO1, S. BORDIGA1, A.N. FITCH3 AND G.L. MARRA4

1 DEPARTMENT OF INORGANIC, PHYSICAL AND MATERIAL CHEMISTRY, UNIVERSITY OF TURIN (ITALY)2 INFM UNITÀ DI TORINO UNIVERSITÀ (ITALY)3 ESRF4 ENICHEM S.P.A., ISTITUTO “G. DONEGANI” (ITALY)

Catalytically-active centres within the complex structure of a silver-exchanged zeolitewere located using a combination of X-ray techniques.

eolites [1] are nanoporous crystallinealuminosilicates formed by a frameworkof corner-sharing [TO4] tetrahedra, whereT represents a silicon or an aluminumatom. The chemical composition can bedescribed by the general formula:Xn+

x/n[(AlO2)x(SiO2)y]x-. Formally, theintroduction of trivalent Al(III) into [TO4]units (substituting tetravalent Si(IV)atoms) induces a net negative chargeon the framework (x-) which must becompensated by the presence of charge-balancing extra-framework cations(Xn+

x/n). Such cations act as Lewis acidcentres, being electron acceptors.

Starting from the basic [TO4]constituent, the framework of anyzeolite is constructed by progressivelyconnecting two adjacent [TO4] units bysharing an oxygen atom, which becomes abridge between the two T atoms (T-O-T).Using the [TO4] unit as the sole buildingblock, the remarkable flexibility of the T-O-T angle (from ≈ 100° up to 180°) allowsthe realisation of an impressive number ofdifferent zeolites, characterised by aregular system of intercrystalline voidsand channels of well-defined size, (in thenanometre and sub-nm range), accessiblethrough apertures of well-defineddimensions. The regularity in channeldimensions controls accessibility andmakes zeolites much more selective in theadsorption of specific molecules ascompared to amorphous carbon or silicagel, which have irregular pore systems.This is the reason for their widespreaduse as molecular sieves. The samecharacteristics explain the ever-increasing

role that zeolites and related zeotypeshave in heterogeneous catalysis (e.g. forthe petrochemical industry, pollutioncontrol and fine chemistry). Moreover,their ability to encapsulate organisedmolecules, crystalline nano-phases andsupramolecular entities inside theirchannels and pores makes zeolitespromising materials in the field of low-dimensional physics, where the quantumeffects due to the spatial confinementbecome observable. Semiconductorquantum wires and quantum dots can thuspotentially be obtained by hostingsemiconductor crystalline nano-phasesinside the channels or cages, so obtaininginteresting applications in the fields ofoptoelectronic, non linear optics,photochemistry, and chemical sensors [2].The same idea applies for metal andbimetallic dots: the incorporation of suchnano-particles inside the pores or channelsof zeotype materials opens a new frontierin the chemistry of metal-supportedcatalysts.

In this article we report on thestructural determination of an Ag-Yzeolite (Si/Al = 2.63) obtained by acombined anomalous X-ray powderdiffraction, (XRPD), and EXAFS studyperformed on beamlines BM16 andBM29, respectively. Silver-exchangedzeolites are used in several catalytic andphotocatalytic processes, which takeadvantage of the presence of bothisolated Ag+ ions and aggregated Agnclusters. Examples include thephotochemical dissociation of H2O intoH2 and O2, the disproportionation of

ethylbenzene, the oxidation of ethanol toacetaldehyde, the aromatisation ofalkanes and alkenes, the selectivereduction of NO by ethylene, and thephotocatalytic decomposition of NO [3].

The first step in the characterisationof a zeolite involves a thermal treatmentto remove all the molecules coming fromthe ambient atmosphere already adsorbedon the catalytically-active centres. Thisactivation process is essential toguarantee the study of a well-definedsystem [4]. Once this step has beenachieved, measurements can beperformed in situ, either on the activatedsample (i.e. the zeolite under vacuumconditions) or after adsorption of a well-defined amount of high-purity gas ontothe sample. The conditions of theactivation process are critical for silver-exchanged zeolites because an increase inthe activation temperature, suitable toremove the most strongly bondedmolecules, has the disadvantage ofpromoting the aggregation of Ag+ ionsinto Agn clusters [5]. Such clusters,needed for some catalytic application(vide supra), are undesired if the aim ofthe study is the location of isolated Ag+

ions. Based on our previous study of thezeolite Ag-ZSM-5 [6,7], an activationtemperature of 120° was adopted.

XRPD experiments were carried outon an Ag-Y zeolite activated in aborosilicate glass capillary at the Ag-Kedge [λ = 0.486093(2) Å], just before[λ = 0.486103(2) Å] and far away[λ = 0.491153(2) Å], in order to single

EXPERIMENTS REPORTS

ESRF - JUNE 2001

17

out, as much as possible, the contributioncoming from the silver cations. Figure 1shows the high resolution XRPD patterncollected at [λ = 0.486103(2) Å]. ARietveld refinement was performedsimultaneously on the three data sets.This made it possible to obtain thezeolite framework and to locate thenear totality of the expected (on the basisof the Si/Al ratio) Ag+ counterions:52.0(4) out of 52.9 per unit cell. Theresult is quite remarkable when comparedwith that obtained in the cases of Cu+-Y[8] and of Rb+-Na+-Y [9] (also BM16data), systems where we were able tolocate only 41.0(5) and 48(1) cations,respectively. This noticeable improvementis ascribed to the following factors: (i) thehigher scattering power of Ag; (ii) thesimultaneous use of three separate datasets; (iii) the extra information in thepatterns arising from the anomalous-scattering effect.

Figure 2 illustrates the structure ofzeolite Y, which is generated byconnecting sodalite units with hexagonalprisms to give a framework containinglarge cavities (supercages) with adiameter of about 13 Å. Figure 2 alsoshows the position of the five differentextraframework Ag+ sites obtained fromthe Rietveld refinement, labelled as sitesI (8.2), I’(17.4), IIa (6.6), IIb (15.2) andI’m(4.6) (the occupancy per unit cell isgiven in parenthesis). Four of the sites arepositions typical of dehydrated cations. Incontrast, I’m is in the middle of thesodalite cage: implying that such ions

must be coordinated to residual watermolecules. In anhydrous conditions, thepositively-charged cations are actually incontact with the walls of the cavities,where they interact with the negatively-charged oxygen atoms of the framework.The fact that so few cations, less than 5out of 52, are coordinated by watermolecules indicates that the temperaturefor the activation procedure wasappropriately selected.

We have used the output of theRietveld refinement to simulate theEXAFS data (collected at BM29) andshown in Figure 3. In theory we haveto simulate five EXAFS signalsrepresenting the contribution to theoverall signal coming from the Agabsorbers located in the 5 different sitesfound by XRPD. The contributioncoming from the hydrated I’m site hasbeen ignored because: (i) less than 10 %of the absorbers occupy this site and (ii) arather high Ag-OH2 Debye-Waller factoris expected. Moreover, since sites I’ andIIb show a similar local environment,(three oxygen atoms at the close distanceof 2.46 or 2.47 Å), the correspondingcontributions were merged. As a result, 3different contributions have been used tosimulate the experimental EXAFS signal.For each contribution, the Ag-O distancewas fixed at the crystallographic value(2.31 Å for IIa, 2.61 Å for I and 2.465 Åfor I’ and IIb), while the coordinationnumber was obtained from the theoreticalnumber of first-shell O neighboursmultiplied by a weighting factor obtainedfrom the relative occupancies obtainedfrom the Rietveld refinement. Ref. [8]describes the procedure for the Cu+-Ycase in great detail. The twosuperimposed spectra at the bottom ofFigure 3 represent the experimentalEXAFS signal, filtered on the Ag-O peak,and the best fit obtained by adding thethree contributions described above,where the refined parameters were onecommon ∆E and three different Debye-Waller factors. The quality of the EXAFSfit, obtained with only four independentvariables on such a complex sample,represents a conclusive proof of thequality of the Rietveld refinement fromthe XRPD data.

We conclude that the use ofanomalous diffraction data collected atthe high resolution powder diffractionbeamline (BM16) has allowed us tolocate nearly all extraframework Ag+

ions hosted in a Y zeolite, i.e. 52.0(4) outof 52.9. The XRPD data have been offundamental help in the understandingof the complex Ag-K edge EXAFSsignal generated by Ag+ ions probingdifferent local environments (BM29data). These studies enhance ourunderstanding of the structure of cation-exchanged zeolites and in particular ofthe extraframework cations which areknown to determine the catalyticproperties of such zeolites.

Fig. 1: Observed, calculated and difference profiles and reflection positions ofthe AgI-Y zeolite.

Fig. 2: Representation of the Y zeoliteframework and cation location. In thefraction of the framework located inthe upper right part of the figure Oatoms are represented in red and theT atoms (Si or Al) in yellow. Forclarity, in the remaining part of theframework, only yellow sticks havebeen adopted. AgI cations arerepresented as spheres and arelabelled with I, I’, IIa, IIb and I’m.The supercage cavity, where guestmolecules can be hosted, can be seenin the centre of the figure.

EXPERIMENTS REPORTS

ESRF - JUNE 2001

18

REFERENCES

[1] C. Lamberti, S. Bordiga, A.Zecchina and G. Vlaic, NotiziarioNeutroni e Luce di Sincrotrone4(2), 3-13 (1999), and refs. therein.[2] G.A. Ozin, Adv. Mater., 4, 612-649 (1992); S. Bordiga, G. TurnesPalomino, A. Zecchina, G.Ranghino, E. Giamello and C.Lamberti, J. Chem. Phys., 112,3859-3867 (2000), and refs.therein.[3] T. Sun and K. Seff, Chem. Rev.94, 857-870 (1994), and refs.therein.[4] J.M. Thomas, Chem. Eur. J. 3,1557-1562 (1997).[5] L.G. Gellens, W.J. Mortier,R.A. Schoonheydt and J.B.Uytterhoeven, J. Phys. Chem., 85,2783-2788 (1981).

[6] S. Bordiga, C. Lamberti, G.Turnes Palomino, F. Geobaldo, D.Arduino and A. Zecchina,Microporous Mesoporous Mater.,30, 129-135 (1999).[7] S. Bordiga, G. TurnesPalomino, D. Arduino, C. Lamberti,A. Zecchina and C. Otero Areán, J.Mol. Catal. A 146, 97-106 (1999).[8] G. Turnes Palomino, S.Bordiga, A. Zecchina, G.L. Marraand C. Lamberti, J. Phys. Chem. B,104 8641-8651 (2000); ESRFHighlights 19-21 (2000).[9] G.L. Marra, A.N. Fitch, A.Zecchina, G. Ricchiardi, M.Salvalaggio, S. Bordiga and C.Lamberti, J. Phys. Chem. B, 101,10653-10660 (1997).

Best fit and simulated EXAFS spectra of the contribution to theoverall signal of silver cations located in, from top to bottom: site IIa

(I’ and IIb merged) and I and sum of the three simulatedcontributions (full line) superimposed with the experimental first

shell filtered kχ(k) function.

VACANCIES AT THE ESRF ON 14 JUNE 2001

Ref Subject Deadline

SCIENTIST 2228 Beamline Operation Managers on ID21 and ID22 15/06/01Previous post-doctoral experience with synchrotron radiation is essential.

POST-DOC PDID01-3 For the Anomalous Scattering beamline ID01 15/06/01PDID03-1 For the Surface Science Diffraction beamline ID03 15/06/01PDID10A-3 For the TROIKA beamline ID10A 22/06/01PDID26 In XAS and related spectroscopies on the X-ray

Absorption Spectroscopy beamline ID26 30/06/01PDID15B-2 For the High Energy beamlines ID15A+B 24/08/01

PhD STUDENT CFR260 Growth, structure and magnetic properties ofultrathin magnetic films 30/07/01

CFR255 Microscopic deformation in rubber 24/08/01CFR278 Micro-imaging investigation of the wetting of

metallic grain boundaries by a liquid metal 24/08/01CFR275 Time-resolved and in situ diffraction on metals

and thin layer compounds 24/08/01CFR271 Diffuse X-ray scattering at grazing incidence

on Si after ionimplantation at ultra-low energy 15/09/01

If you are interested,please send us a fax(+33 (0) 4 76 88 24 60)or an e-mail(recruitm@esrf.fr)with your address,and we will provideyou with anapplication form. Youcan also print out anapplication form onthe World Wide Webhttp://www.esrf.fr

All applications received after the deadline will be considered forthe selected vacancy if not filled or for future similar positions

EXPERIMENTS REPORTS

ESRF - JUNE 2001

19

I nAs quantum dots or quantum wiresprovide attractive quantum properties foroptoelectronic devices like micro-lasersources [1]. They are grown by molecularbeam epitaxy and formed during the firststeps of growth by a "spontaneous" self-ordering process driven by the strain dueto the lattice mismatch between thesubstrate and the epilayer. The resultingobjects are on the nanometre scale. Theyare characterised mainly in terms ofoptical (luminescence, photoreflectance)and vibrational properties (Raman),while AFM (atomic force microscopy)and TEM (transmission electronmicroscopy) give an image of the shapeand an idea of the spatial homogeneity.However, the inner composition is still notwell known, as mixing with the substrate

atoms and segregation mechanisms of theimpinging species could occur during theepitaxial growth.

We have studied InAs/InP(001)quantum wires by Diffraction AnomalousFine Structure (DAFS). This spectroscopicmethod offers the advantage of providingthe local information of the chemicalselective Extended X-ray AbsorptionFine Structure (EXAFS) spectroscopywith the spatial and site selectivity of X-ray diffraction. In such a way, weare able to obtain local informationabout the atoms that belong to thenanostructures.

Our sample consists of an array ofquantum wires (Figure 1) aligned along

the [1-10] direction with a typical lengthabove 5 µm, a height between 0.6 and2 nm, a period of 20 nm with anequivalent InAs coverage of about 2.5monolayers. The DAFS measurementswere carried out at the FrenchCollaborative Research Group beamlineBM2 (D2AM). We measured theintensity at the maximum of twoBragg reflections of the quantum wirearray, near the 420 and 440 InPsubstrate Bragg peaks, as a functionof energy, around the As K-edge(11.867 keV) (Figure 2). Themeasurements were performed inglancing-angle geometry, with anincidence angle kept constant close to thecritical angle of InP (about 0.2°). Thislets us collect the diffracted photonsfrom an extremely thin quasi-surfacelayer enhancing its contribution to thediffracted intensity in comparison withthe substrate contribution.

The data analysis is a two-stepprocess. First, we performed asimultaneous crystallographic fit to thelineshape of the smooth part of the twoDAFS spectra (Figure 2) refining theP concentration, 1-x, for InAsxP1-x,Debye-Waller factors and experimentalfactors. We introduced corrections to thecalculated diffraction intensity in theframework of the Distorted Born WaveApproximation (DBWA) [2] such that, inthe total reflection regime, the thinepilayer is a small perturbation for theevanescent X-ray wave that penetrates afew nanometres into the refractingmedium. Another fit parameter was the

GLANCING-INCIDENCE DIFFRACTION ANOMALOUS FINE

STRUCTURE OF INAS/INP SELF-ASSEMBLED QUANTUM WIRESS. GRENIER1, M.G. PROIETTI1,2, H. RENEVIER1, L. GONZALEZ3, J.M. GARCÍA3 AND J. GARCÍA2

1 LABORATOIRE DE CRISTALLOGRAPHIE, CNRS, GRENOBLE (FRANCE)2 INSTITUTO DE CIENCIA DE MATERIALES DE ARAGÓN, C.S.I.C.-UNIVERSIDAD DE ZARAGOZA (SPAIN)3 INSTITUTO DE MICROELECTRÓNICA DE MADRID, C.S.I.C., MADRID (SPAIN)

Anomalous X-ray diffraction at glancing angles is shown to be an appropriatetechnique to study the inner part of nanostructures self-assembled in a very

thin epilayer. The site and chemical specificity of anomalous diffraction has permittedus to probe atoms of the quantum wires only and not those belonging to the

oxide layer or to the substrate.

Fig. 1: AFMtridimensional viewof InAs quantumwires on InP buffer[7].

EXPERIMENTS REPORTS

ESRF - JUNE 2001

20

thickness of an oxided dead layer thatwas found 15 ± 5 Å thick. To reproducethe shape of the anomaly at the edge, wehad to include P atoms, 1-x = 0.5 ± 0.1:a considerable amount of P atoms iscontributing to the satellite diffractionpeaks, showing the same periodicity ofthe wires. However, at this stage of theanalysis, we are not yet able to discernwhether the P atoms belong to the wiresor to the InP substrate which could showa periodic modulation as well as periodicstrain spots due to the strain with the InAswires.

In a second step, we extracted theExtended DAFS oscillations thatappear after the edge (Figure 3) and weanalysed them according to an EXAFSdata processing scheme to get localparameters such as distances andpopulations [3]. Theoretical multiple-scattering EXAFS signals were calculatedby the FEFF code [4] simulating the finestructure signal from an As atom insidea cluster containing In atoms, As (x) andP (1-x) atoms. The polarisation of theincoming photons was perpendicularto the surface, so the As and P nextnearest neighbour atoms contribution toEDAFS is due only to the out-of-planeatoms. Refinement of the data wascompleted by the least-square fitprocedure of the FEFFIT program [5](Figure 3a and 4a). The relevant resultsare the P concentration, (1-x) = 0.4, andthe As-P distance, found at 4.17 Å. Whilethe As-As distance was fixed at 4.29 Å,the P content is, within the error, equal tothe value found by the crystallographic fitof the anomalous diffraction lineshape.

The two independent analysesdetected a consistent amount of P atoms:(1-x) = 0.4-0.5. In a previous study ofInAs0.5P0.5/InP superlattices samples [6],the out-of-plane As-As and As-Pdistances were found much closer to eachother, about 4.28 Å and 4.25 Årespectively. In our analysis, the As-Pdistance is 4.17 Å, close to the P-Pdistance in bulk InP (4.15 Å), therefore,we could exclude the hypothesis of afully relaxed InAsP epilayer. The P atomscontributing to EDAFS have to belong tothe interface region, 0.5-2 monolayers,and the core of the quantum wire isessentially strained InAs. Two types ofinterface or a combination of both canexplain the results: a) an abrupt InAs/InPinterface with periodical strain stripsgenerated in the InP buffer layer due tothe interface mismatch strain; b) a

corrugated InAs/InP interface with thesame wire's periodicity.

We have also measured a glancing-angle EXAFS spectrum at the As K-edge(Figure 3b) at beamline BM8 (GILDA).The spectrum shows a clear As oxideshape with a strong low-frequencycomponent that corresponds to a hugepeak at 1.2 Å in its Fourier Transform(FT) (Figure 4b). The oxide layer causes

a significant loss of information inparticular for shells beyond the first one,whereas, for a DAFS spectrum, it lowersthe overall diffracted intensity and thejump at the edge but it does not perturbthe fine structure signal of the interestingatoms.

In conclusion, we have shown for thefirst time that DAFS can been applied inthe glancing-angle regime to the study of

Fig. 2: Glancing-angle DAFS

spectra ofquantum wires,

for (440) and(420) Bragg

peaks, at the AsK-edge (dotted

curves) andcrystalloghraphic

fits (continuouscurves).

Fig. 3: (a) InAsquantum wire

Glancing-angleEDAFS oscillations,

after backgroundsubtraction, with bestfit (continuous line),

(b) EXAFS of thequantum wires. The

curves have beenrescaled for clarity.

Fig. 4: (a) FT ofquantum wire

EDAFS, with best fit(continuous curve),(b) FT of quantumwire EXAFS. Thecurves have been

rescaled for clarity.

EXPERIMENTS REPORTS

ESRF - JUNE 2001

21

auto-organised nanostructured materialswith an extremely low equivalentcoverage. As the interface effects on theEDAFS spectrum are remarkable dueto the low epilayer thickness and to theX-ray beam polarisation, we were able toshow, as a preliminary result, that theinner composition of the quantum wiresis not a relaxed InAsP structure.Furthermore, we have detected a P-richregion at the interface with the sameperiodicity as the quantum wires.

REFERENCES

[1] Q. Xie, A. Madhukar, P. Chen andN.P. Kobayashi, Phys. Rev. Lett. 75, 2542-2545(1995).

[2] H. Dosch, Critical Phenomena at Surfacesand Interfaces, Springer Tracts in ModernPhysics, 126, Springer Verlag (1992).[3] M.G. Proietti, H. Renevier, J.L. Hodeau,J. García, J.F. Bérar, and P. Wolfers, Phys. Rev.B, 59, 5479-5492 (1999).[4] J.J. Rehr, J. Mustre de Leon, S.I. Zabinsky,and R.C. Albers, J. Am. Chem. Soc. 113, 5135-5145 (1991).

[5] M. Newville, B. Ravel, D. Haskel,J.J. Rehr, E.A. Stern and Y. Yacoby, Physica B208 & 209, 154-155 (1995).[6] S. Pascarelli, F. Boscherini, C. Lamberti,S. Mobilio, Phys Rev. B, 56, 1936-1947(1997).[7] L. Gonzalez, J.M. Garcia, R. Garcia,J. Martinez-Pastor, C. Ballesteros, Appl. Phys.Lett., 1104-1106 (2000).

ACKNOWLEDGEMENTS

We acknowledge the BM2 (D2AM) and the BM8 (GILDA) beamlines for grantingbeam time and technical assistance. This work was supported by the LEA-MANESEuropean Agreement and by CICYT project No. MAT99-0847.

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

22

T

X-RAY AND NEUTRON STUDIES OF THE OPTIMISED

SYNTHESIS, THE STRUCTURE AND THE TRANSFORMATIONS

INVOLVING NOVEL ION EXCHANGERSJ. B. PARISE

DEPARTMENTS OF GEOSCIENCES AND CHEMISTRY, STATE UNIVERSITY OF NEW YORK (USA)

Presentation given at the ESRF-ILL Workshop "Environmental StudiesUsing Neutron and Synchrotron Facilities", 20-21 February 2001.

A combination of X-ray and neutron techniques can be used to elucidate the oftenrather complex structures of ion-exchangers - materials designed to trap and remove

certain chemicals from an aqueous environment.

he shear number of sites worldwidecontaminated by industrial and militarywaste represents a major environmentaland public health concern. For example,approximately 40,000 uncontrolled wastesites have been reported to U.S. federalagencies. About 1,300 of these havemade it to the National Priorities List of

sites for remediation. The US experienceis not unique.

In those cases where the toxicmaterials are transported in anionic orcationic forms, inorganic ion exchangerscan be used to replace benign ions in thecrystalline lattice for contaminant ions,

thereby limiting availability of thesetoxins to the biosphere [1]. In favourableinstances the inorganic ion exchangeroccurs naturally, can be mined andapplied directly. For example the zeolites,naturally occurring aluminosilicateminerals possessing pores and channelsof molecular dimensions, are well knownion exchangers. One of the mostabundant, and therefore cheap, naturallyoccurring zeolites is the mineralclinoptilolite, which is commonly usedfor ammonia control and sorption inanimal feed, fish tanks and kitty litter. Italso happens to have a strong affinity for90Sr and 137Cs, and is used routinely totreat radioactive effluent [2]. Unlike ionexchange resins, which are susceptible toradiation damage, the zeolites sequesterthe strontium and caesium over longperiods. Clays, layered aluminosilicates,are also commonly used to treatcontaminated soils and to sequestercontaminants [1].

Fig. 1: Structure ofNa16Nb12.8Ti3.2O44.8(OH)3.2•8H2O(SOMS-1) determined from a 5x5x8 µm3

twined crystal. Large green balls areoctahedral framework sodium cations,purple spheres are Nb/Ti sites and redspheres are oxygen.

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

23

The naturally occurring materialssuffer from some disadvantages,including low capacity and a tendency torelease contaminants if environmentalcircumstances such as pH are altered. Thezeolites, and indeed the other mineral ionexchangers, are inspirational howeverwhen we seek fertile synthetic territory toexplore for materials with these desirableion exchange properties. The generalformula Ax+z[Si1-yTy

w+O2]•mH2O withz = (4-w)y/x, summarises the compositionfor most silicate zeolites. The frameworkcation sites substituting for silicon in theformula for quartz (SiO2) are designatedT; these cations, with typical valencew = 3 (Al3+), are tetrahedrally coordinatedby oxygen. The symbol A is used forpossible charge balancing cationsoccluded in the regularly spaced poresand channels. Conventionally the ion-exchange capacity, determined by theSi/T ratio, has been maximised inaluminosilicate zeolites by producingmaterials with low Si/Al (xSiO2 + 1-xAl2O3) + 1-x(A2O) → (Si2xAl2-2xO4)(2-x)-

.(2-x)A. Unfortunately this limit is1.0 and so the capacity for thealuminosilicates is limited.

Strategies designed to increase theion exchange capacity include loweringthe valence of the substituting “T-atom”in the formula above, increasing thenumber of anions in the framework byintroducing octahedrally coordinatedcations, and combinations of these two.Both strategies lead to new materials withincreased capacity and selectivity [3,4].In the first instance the substitution of Li+for Si4+ increases capacity and leadsto other unusual properties includingion conductivity [3,5]. In the secondinstance, our collaborators T. Nenoff andM. Nyman at Sandia National Laboratoryfound inclusion of Ti4+, although it doesnot decrease cation charge in theframework, does boost the anioncontent because of its octahedralcoordination, and this requires extracharge-balancing cations in the pores andchannels. Crystal structures, determinedfrom the combined use of synchrotronX-ray diffraction from microcrystals(2 – 10 µm on edge) and neutrondiffraction from powders, reveal that bothstrategies come together in the SOMScrystalline inorganic molecular sieves(Figure 1) developed at Sandia. In thisinstance a framework of octahedrallycoordinated Nb5+/Ti4+ and Na+ is formed,which is selective for removal ofstrontium from acidified solutions and in

the presence of such benign cations asNa+, Mg2+ and Ca2+. Part of the sodiumin the structure is exchangeable, and italso serves to decrease the total cationcharge in this anion-rich framework.This results in increased cation exchangecapacity, in a manner analogous tothe principles observed in the Li-silicates[3].

It is important to determine accuratecrystal structures for the materialsresulting from the syntheses inspired bythe strategies outlined above. Thiscompletes a feedback loop involvingsynthesis, ion exchange properties andstructure and provides a basis forrationalising the functionality of theexchangers and for the development ofnew ones. Many of the structures ofinterest involve atoms of widely varyingX-ray scattering powers, Li and Cs in thecase of the Li-silicates for example [3].They also crystallise from gels at lowtemperatures and so are often kineticallystabilised phases. This leads to samplesthat are mixtures of phases crystallisingin fine hair-like crystal bundles.Structural studies have therefore reliedheavily on synchrotron radiation, to bothscreen and collect useful diffraction dataon microcrystals for determination offramework connectivity, and on neutrondiffraction to provide the positions oflight atoms either in the channels(Figure 1) or in the framework [3]. Time-resolved synchrotron X-ray powderdiffraction is invaluable also since itallows us to observe directly the phasescrystallising during a hydrothermalsynthesis. Information from these studiescan immediately be implemented inlarger batch processes to prepare thesingle phases samples required formeaningful neutron diffraction, propertyand spectroscopic measurements.

The pressure to make the search forselective ion-exchangers more rationaland efficient requires further developmentof time-resolved techniques. Along withwide-angle diffraction, this includesinterrogating the crystallising gels withboth SAXS and SANS to determine theincipient behaviour that leads to theformation of open materials rather thandense phases. Subtle changes in thestarting materials chosen for synthesisoften lead to quite different products.Some experiments in this area have begun(see for example [6]) and the joint use ofX-ray and neutron diffraction on the samesample is clearly a priority in the future.

REFERENCES

[1] M.J. Mench, A. Manceau, J. Vangronsveld,H. Clijsters, B. Mocquot, Agronomie, 20, 383-397 (2000).[2] D.V. Marinin, G.N. Brown, Waste Manage.20, 545-553 (2000).[3] S.H. Park, J.B. Parise, H. Gies, H.M. Liu,G.P. Grey, B.H. Toby, J. Am. Chem. Soc., 122,11023-11024 (2000).[4] M. Nyman, A. Tripathi, J.B. Parise, R.S.Maxwell, W.T.A. Harrison, T.M. Nenoff, J. Am.Chem. Soc., 123, 1529-1530 (2001).[5] S.-H. Park, P. Daniels, H. Gies,Microporous Mesoporous Mat., 37, 129-143(2000).[6] P. de Moor, T.P.M. Beelen, R.A. van Santen,K. Tsuji, M.E. Davis, Chem. Mat., 11, 36-43(1999).

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

24

S

3-D SNOW AND ICE IMAGES BY X-RAY

MICROTOMOGRAPHYC. COLÉOU1 AND J.-M. BARNOLA2

1 MÉTÉO-FRANCE, CENTRE D’ETUDES DE LA NEIGE, GRENOBLE (FRANCE)2 LABORATOIRE DE GLACIOLOGIE ET GÉOPHYSIQUE DE L’ENVIRONNEMENT, CNRS, GRENOBLE (FRANCE)

Presentation given at the ESRF-ILL Workshop "Environmental StudiesUsing Neutron and Synchrotron Facilities", 20-21 February 2001.

Evolution of snow in temperate as well as in polar regions is strongly dependent onits structural properties. We present 3-D imaging results obtained by X-ray absorption

tomography, a technique that will allow the extraction of quantitative informationabout the structural parameters of snow and ice.

now is a porous medium. At negativetemperatures it consists of ice and airwith water vapour. Snow is constantlymoving and its porosity decreases withtime. It is slowly compacted into firn andthen ice. A very wide range of porositycan be found: from more than 95% forfresh snow to around 40% for firn andthen to less than 10% for ice. Changes indry snow are caused by vapour diffusion

among the grains - they are drivenby temperature gradients and graincurvatures [1]. On polar ice caps, below adepth of 10 metres, the medium becomesisothermal although its porosity is stillopen. Then the porosity decreases moreslowly, firn evolves into ice by a processsimilar to sintering under load. Thetransition from an open to a closedporosity is called close off and it occurs

when firn becomes ice. During thetransition, air is trapped in bubblesbetween the ice crystals.

Three-dimensional images of snowand ice with a spatial resolution of 10 µmwere obtained using X-ray absorptiontomography, an established techniqueused on the ID19 beamline [2]. 900 two-dimensional X-ray absorption images

Fig. 1: View ofthe experimental

setup used forsnow and icesamples. Thecryostat was

mounted on ahigh-accuracy

rotation andtranslation stage.

Images wererecorded with a

Frelon CCDcamera.

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

25

were recorded at angular positions of theobject around an axis spanning 180°. Anappropriate algorithm was then used toreconstruct a 3-D image from the data.

SAMPLE PREPARATIONSpecificities of snow and ice need

special requirements for samplepreparation. Since firn and ice is ratherhard, it was possible to carve out a cubicsample (edge about 15 mm). Ice and airoffer a good level of contrast and so theimages could be obtained by localabsorption tomography at an energyof 18 keV. Snow needed strengtheningprior to being machined into the shape ofa cylinder (9 mm high, 9 mm diameter).Dense snow was cohesive enough toallow removal of the filling mediumbefore imaging. X-ray absorptiontomography at 10 keV was used. Snowwith low density was too fragile to standup to the rinsing stage. A selection of purechemicals have been tested, examplesbeing trans-1,2-dibromocyclohexane andcycloheptane. They fulfilled the requiredconditions regarding their melting point,reaction with ice, hardness and theirattenuation coefficient which has to bedifferent from that of ice. In all cases, aspecial device (a liquid-nitrogen cooledcell at about –50°C) was used to preventboth melting and metamorphism ofthe samples during the experiments(Figure 1).

RESULTS ON SNOWFrom the grey level data files

produced at the ESRF, a 3-D binaryimage was obtained by morphologicalprocessing on each image plane. The first3-D images [3] have shown the feasibilityof this technique which provides us withgood quality images of the snow structureor grain assembly.

What is the interest of such images?Most physical properties and physicalprocesses occurring inside snow arestrictly linked to its microstructure. Theaim was then to extract relevantinformation from these images. The twofirst geometrical parameters studied wereporosity, which is the simplest descriptiveparameter but often the most importantfor many of snows physical properties,and local curvature, which is a governingparameter of snow metamorphism. Wehave shown that the size of the sample

studied is statistically representative forthese parameters [4]. Another importantgeometrical parameter is surface area,which denotes the ability of a given snowto evolve. 3-D images will be used asa reference for the development of asnow metamorphism model at uniformtemperature. Under such conditions,snow tends to minimise its specific areaand should lead to a symmetric histogramof curvature peaked at zero.

A large area of interest where snowimages could improve our knowledge isthe study of avalanches and mechanicalproperties of snow. Snow stability on aslope is linked to the superposition of thedifferent snow types. For instance, acohesive layer above a weak layer istypical of most of the slab avalanchestriggered by skiers. Figure 2 shows viewsfrom a snow sample collected at thefailure of a slab avalanche. Classicalobservations describe the layering of thesnow pack. X-ray microtomography is aninteresting tool for visualising thebonding between different snow layers.

RESULTS ON FIRNAND ICE

The transformation of the snow intoice can be very slow in polar regions (upto 3000 years) and it is very important toknow precisely when and how the gaswas trapped in order to interpret the airarchive (atmospheric information)compared to the ice archive (climaticinformation) [5]. Thus the aim of the firnand ice part of the project was to studythe evolution of both open and closedporosity of the material and in a first stepto test the feasibility of the method.Twelve samples have been analysed.They were taken at depths between 62and 120 m on a Vostok (Antarctic) core,thus covering the entire range of the poreclosure. The results were excellent: inFigure 3, two examples are shown andcompared with the 2-D technique alsoused at the LGGE to study the firnstructure [6]. This last method presentsthe advantage of revealing both pores andgrain boundaries. However, it is obvious

Small roundedgrains (slab)

Thin crust

Faceted crystals(weak layer)

Fig. 2: Plane and 3-D reconstructed views from a snow sample collected at thefailure of a slab avalanche.

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

26

Coaxial episcopy X-ray tomography

2D slice 3D reconstruction90 m: all the pores are open

120 m: all the pores are closed

from Figure 3 that the 2-D informationalone is unable to give preciseinformation on the open to closedporosity ratio. The detailed images of theshape of the pore phase near their closureshow some particularities: at 90 m all themain pores are open, however very smallbubbles, which are certainly formed wellbefore the “close off”, are visible. At120 m, even if all the pores seem to bewell isolated, some very small channelsstill exist and the gases can be fractionatedwhen they diffuse through them.

Quantitative studies are in progressto look in particular at the influenceof the sample size on the quality ofthe information given by suchreconstructions.

CONCLUSIONX-ray microtomography appears as a

very powerful tool for the study of snow

microstructure and firn-ice porosities.Furthermore, 3-D imaging is gainingmomentum in several domains dealingwith the structure of materials. New waysto provide relevant information fromthe images and model developmentwill certainly benefit from exchangesbetween the different research areas.

REFERENCES

[1] S.C. Colbeck, Journal of GeophysicalResearch, 88, 5475-5482 (1983).[2] J. Baruchel, J-Y. Buffière, E. Maire,

P. Merle and G. Peix, X-Ray Tomography inMaterial Science, Hermes (2000).[3] J.B. Brzoska, C. Coléou, B. Lesaffre,S. Borel, O. Brissaud, W. Ludwig, E. Boller andJ. Baruchel, ESRF Newsletter, 22-23 (April1999). [4] C. Coléou, B. Lesaffre, J-B. Brzoska,W. Ludwig and E. Boller, Ann.Glaciol. 32,(2000), in press.[5] L. Arnaud, M. Gay, J.-M. Barnola and P.Duval. Journal of Glaciology, 44, 326-332,(1998).[6] J.-R. Petit et al., Nature, 399, 429- 436(1999).

ACKNOWLEDGEMENTS

We thank J.-B. Brzoska, B. Lesaffre, P. Duval, C. Goujon and F. Flin for theirparticipation in the research and E. Boller, P. Cloetens and O. Brissaud for their helpfor the experimentation.

Fig. 3: 2- and 3-dimensional images of ice around the firn-ice transition: porosities appear in black in the 2-D images. Theepiscopy technique reveals that the pores are located at crystal boundaries. 3-D reconstructions from the 2-D slices giveaccess to the real shape of the porosities.

25 mm 10 mm 3 mm

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

27

T

INVESTIGATION OF POSITIVE ELECTRODE MATERIALS

FOR LITHIUM BATTERIES BY MEANS OF X-RAY AND

NEUTRON DIFFRACTIONC. MASQUELIER1, M. MORCRETTE1 AND G. ROUSSE2

1 LABORATOIRE DE RÉACTIVITÉ ET DE CHIMIE DES SOLIDES, UMR CNRS 6007, AMIENS (FRANCE)2 ILL, GRENOBLE (FRANCE)

Presentation given at the ESRF-ILL Workshop "Environmental StudiesUsing Neutron and Synchrotron Facilities", 20-21 February 2001.

In situ X-ray diffraction studies allow the following of the structural changes associatedwith lithium extraction from positive electrode materials for Li batteries.

Two recent examples are presented here.

he “layered” rock salts compositionsLiCoO2 and LiNiO2 as well as the spinelLi[Mn2]O4 have been extensively studiedfor use as positive electrodes forrechargeable lithium batteries. Theirinterest lies in the delivery of voltagesclose to 4V vs Li+/Li when lithium isextracted reversibly. Other families ofcompounds, i.e. polyanionic structuresbuilt up from MO6 octahedra (M = Fe, V,Ti) and PO4

3- or P2O74- polyanions have

also been identified as alternativeelectrodes, among which are LiFePO4,LiMP2O7 and the NASICON familyLixM1M2(PO4)3 [1]. A review ofelectrode materials for lithium batteries isgiven in ref [2].

Besides the need to identify new

materials for electrochemical devices,it is also a challenge to understand thetopotactic lithium insertion/de-insertionmechanisms (associated with reductionor oxidation of the transition elementat the positive electrode) that occurwithin these materials. Lithium batterieshave ever increasing industrial andtechnological importance. Additionally,they do indeed offer a rather nice tool forsolid state electrochemists because theredox processes involved often lead tosubtle phase transitions and to metastablenew forms of materials with “exotic”oxidation states or compositions [3].These processes may also lead tostructural instabilities that are penalisingfor the integrity of the host material overextensive charge/discharge cycles. In this

regard, X-ray and neutron diffractionare powerful – and complementary –techniques for the investigation ofpositive electrode materials for lithiumbatteries as they benefit from thecrystalline nature of the material used.Powder neutron diffraction is used forhighly reliable crystal structuredeterminations of pristine materialscontaining lithium into which localdisorders, local distortions, phase purityetc… play a major role for their effectiveuse in a battery system. On the otherhand, in situ X-ray diffraction that useeither standard equipment in thelaboratory or high-resolution synchrotronradiation, has been used by severalgroups for the investigation of structuralchanges during cell operation. For thispurpose, specially-designed cells fordiffraction in either reflection ortransmission geometry, were developedfollowing J. Dahn’s pioneer concept [4].

The intent of this communication isto demonstrate the use of neutron and/orX-ray diffraction for the investigation ofpositive electrode materials in lithiumbatteries. For this purpose we havechosen to present two typical examplesthat have recently been addressed:LixMn2O4 and LixCoO2.

Fig. 1: Ratio between cubic and orthorhombic forms ofLiMn2O4 as a function of temperature, determined bysynchrotron diffraction (LURE, WD4C). From [6].

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

28

LiXMn2O4

Palacin et al. [5] reported an in situsynchrotron study carried out at theESRF (Beamline BM16, transmissiongeometry through a Bellcore Plion“plastic” cell) on the structuralphenomena that occur during lithiumextraction out of a fluorine-substitutedspinel LiMn2(O3.74F0.26). Besides the“classical” features previously observedin the 4.1 V – 4.2 V region for LixMn2O4,two extra phenomena of interest wereobserved at the beginning and then at theend of the charge process:

i) The pristine material, with averageoxidation state for manganese equal to+3.4, is indeed a mixture of two phases: aminor cubic spinel phase and a majororthorhombic form that had just beendiscovered from combined electronic,X-ray (at LURE, Orsay) and neutron(at LLB, Saclay) diffraction. Theorthorhombic distortion of the spinelresults from a charge-ordering transitionon Mn3+ and Mn4+ sites that occurs veryclose to room temperature when theconcentration of Mn3+ present in thespinel phase is above the critical level of50% [6] (Figure 1). The in situ datapresented in Figure 2 (beginning ofcharge) nicely revealed that thisdistortion disappears exactly for 0.18lithium extracted from LiMn2(O3.74F0.26),i.e. when the average valence ofmanganese reaches 3.5+.

ii) On charging above 4.4 V, Palacin etal. also observed an apparent irreversiblecapacity of ∆x = 0.08, associated withextra electrochemical phenomena at 4.5 V(on oxidation) and at 3.3 V (on reduction).A complementary study [7] wasundertaken that allowed, with the help ofhigh-resolution transmission electronmicroscopy, to identify the formation ofa new Li1-xMn2O4 double hexagonal(DH) structure isotypic with LiFeSnO4(Figure 3). The transition betweenspinel LixMn2O4 (ABCABC… stacking)and DH-LixMn2O4 (ABACABAC…stacking) was explained in detail [6]through a translation mechanism ofKagome (OC3) and Te2OC type blocksthat are common to both structures(Figure 3).

LiXCoO2Another interesting in situ experiment

was carried out at the ESRF, to understandthe structural behaviour of Li1-xCoO2when charged (lithium extracted) to highvalues of x, i.e. up to a voltage close to 5Vvs Li+/Li. This was the subject of a recentcommunication [8] whose main outlinesare recalled here. The originality of theexperiment was the use of themicrodiffraction beam line ID11, wherethe 0.2 x 0.2 mm beam was small enoughto pass through the grids of the Al and Cucurrent collectors of a PLion plastic cell.The diffracted beam was collected on a 2-D image plate detector that permitted veryshort acquisition times (~ 10 seconds perpattern !).

Several processes of de-intercalationof lithium can be distinguished in

Figure 4: the incremental capacity peaksat 3.92 V, 4.05 V, 4.18 V and 4.55 V arethe signature of transitions betweenphases at various stages of charge. A verylarge number of diffraction patterns werecollected between 8° and 20° for awavelength of 0.3757 Å (Figure 5).Besides the interesting evolution of latticeparameters vs x in LixCoO2 (Figure 6),successful Rietveld Refinements allowedthe determination of the atomic positionsin each of the phases formed [8]. Ofparticular importance for the cyclingbehaviour of LixCoO2 is the region closeto x = 0.5 / V = 4.18 V where a monoclinicphase, M1, extends from x = 0.55 to 0.47.This region corresponds to the maximumof the c equivalent lattice parameter, i.e. tothe maximum distance between CoO2close packed layers. At deeper extentsof charge, i.e. at potentials higherthan 4.28 V vs Li+/Li, irreversible

Fig. 3: TEM image of stacking faults along [111]spinel due to the intergrowth ofthe double hexagonal structure. The structural relationships between Spinel

and DH arrangements are drawn. From [6] and [7].

Fig. 2: Parts of in situ synchrotrondiffraction patterns (ESRF, BM16)

as a function of x duringelectrochemical extraction of

lithium from LixMn2(O3.74F0.26).From [5].

ENVIRONMENT WORKSHOP

ESRF - JUNE 2001

29

transformations occur and hence, thepractical composition range for cyclingLixCoO2 in a real battery system when xis between 0.5 and 1.

CONCLUSIONThe two examples developed in this

paper show the effective use of in situsynchrotron X-ray diffraction to helpunderstand and solve problems associatedwith structural changes that occur uponlithium extraction from positive electrodematerials. Despite its great potential usefor Li-containing materials, neutrondiffraction is still limited for in situ studiesdue to the small quantity of activematerial and to the large amount ofhydrogen-containing components withinthe battery. Both techniques arecomplementary and should be used moresystematically alongside the essential insitu and/or ex situ experiments on “in-house” X-ray diffractometers.

REFERENCES

[1] J.B. Goodenough, A.K. Padhi, K.S.Nanjundaswamy and C. Masquelier, US Patent# 5,910,382 (1999).[2] D. Guyomard, in “New Trends inElectrochemical Technology : Energy StorageSystems for Electronics", T. Osaka et M. Datta(Editors), Gordon & Breach SciencePublishers, Chapter 9, p. 253-350 (2000).[3] C. Delmas, M. Ménétrier, L. Croguennec,S. Levasseur, J.P. Pérès, C. Pouillerie,G. Prado, L. Fournès and F. Weill, Int. J. Inorg.Mater. 1, 11-19 (1999).[4] J.R. Dahn, M.A. Py and R.R. Haering,Can. J. Phys. 60, 307 (1982).[5] M.R. Palacin, F. Le Cras, L. Seguin,M. Anne, Y. Chabre, J.M. Tarascon,G. Amatucci, G. Vaughan and P. Strobel, J.Solid State Chem. 144, 361-371 (1999).[6] G. Rousse, C. Masquelier, J. Rodriguez-Carvajal, E. Elkaim, and J.P. Lauriat,J.L. Martinez, Chemistry of Materials 11(12),3629-3635 (1999), Thèse de Doctorat,Université Paris-XI Orsay, Sept 2000.[7] L. Dupont, M. Hervieu, G. Rousse,C. Masquelier, M.R. Palacín, Y. Chabre andJ.M. Tarascon J. Solid State Chem. 155, 394-408 (2000).[8] M. Morcrette, G. Vaughan and Y. Chabre,Colloque Gaston Planté, Paris, October 2000.

Fig. 4:Potentiostaticintermittentextraction oflithium fromLiCoO2 at anequivalent rateof C/20. From[8].

Fig. 5: Synchrotron (λ= 0.3757 Å)diffraction patternsas a function of x inLixCoO2. From [8].

ACKNOWLEDGEMENTS

G. Vaughan (ESRF, Grenoble), J. Rodriguez Carvajal (LLB, Saclay), E. Elkaim (LURE, Orsay), J.P. Lauriat (LURE, Orsay), Y.Chabre (UJF, Grenoble), L. Dupont (LRCS, Amiens) and J. M. Tarascon (LRCS, Amiens) are gratefully acknowledged for theircontributions.

Fig. 6: Evolution of cell parameters (rhombohedral description) as a function ofx in LixCoO2. From [8].

HIGH PRESSURE WORKSHOP

ESRF - JUNE 2001

30

M

In situ SYNCHROTRON X-RAY DIFFRACTION STUDIES

OF HP-HT SYNTHESIS OF SUPERHARD PHASES

IN THE B–C–N SYSTEMV. L. SOLOZHENKO1,2

1 PHYSIQUE DES MILIEUX CONDENSÉS, UNIVERSITÉ PIERRE ET MARIE CURIE, PARIS (FRANCE)2 ON LEAVE FROM INSTITUTE FOR SUPERHARD MATERIALS, KIEV (UKRAINE)

Presentation given at the ESRF Workshop "Science at High Pressure:Latest Trends from 3rd Generation Sources", 16-17 February 2001.

The present paper briefly reviews results of our very recent studies of high-pressurehigh-temperature synthesis of superhard phases in the B-C-N system using powder

X-ray diffraction with synchrotron radiation.

aterials Science under extremeconditions is one of the most importantlines of research using the third-generation synchrotron radiation sources.The high resolution and improvedsensitivity resulting from the use of high-intensity synchrotron-derived X-rayradiation from these sources areindispensable for in situ studies of phaseformation and reaction kinetics ofcompounds of low-Z elements at highpressures and temperatures.

Here we report the results of in situstudies of synthesis of superhard phasesin the B-C-N system that have recentlybeen performed at the high-pressurebeamline ID30.

SYNTHESIS OF NEWSUPERHARD PHASE,

CUBIC BC2NPhase transitions of graphite-like

BN-C solid solutions (g-BCxN) werestudied up to 32 GPa and 3000 K using alaser heated diamond anvil cell andangle-dispersive X-ray diffraction [1]. Atroom temperature an increase in pressureis accompanied by a pronounceddecrease in the line intensities of theturbostratic g-BC2N (Figure 1). Uponcompression to 19.9 GPa, the intensity ofthe strongest 002 line decreases by a

factor of 6, and at 25.8 GPa this linealmost disappears. Also, with increasingpressure, a variation in the 10 asymmetricline of the turbostratic structure isobserved. The intensity of scattering inthis region increases, the profile of theline becomes increasingly symmetric andits peak shifts towards a value of 2.07 Åwhich is close to those observed for the111 reflections of diamond-like phases.These effects point to the reconstructionof the graphite-like sp2-structure into thediamond-like sp3-structure, which startsat about 5 GPa and ends at about 25 GPa.

At 25.8 GPa, the heating of g-BC2Nup to 1600 K is not accompanied by anychange in the diffraction patterns whichexhibit only a broad line in the region of111 reflections of diamond-like phases.At higher temperatures, the profile of thisbroad line changes to a rathercomplicated fine structure, and two newweak lines with dhkl = 1.26 and 1.09 Å (atambient temperature) also appear.Finally, above 2200 K a drastic change inthe spectrum is observed (Figure 1, toppattern) which clearly points to the

formation of a new phase. The diffractionpattern of the quenched sample exhibitsonly 111, 220, and 311 lines of the cubiclattice, which indicates that the sample issingle-phase.

Laser heating experiments at differentpressures have shown that the formationof c-BC2N is observed only at pressures

Fig. 1: Laser-heating sequence ofdiffraction patterns taken at several

pressures and temperatures.Bottom and top patterns correspondto g-BC2N and c-BC2N, respectively.

HIGH PRESSURE WORKSHOP

ESRF - JUNE 2001

31

above 18 GPa. At 14.5 GPa andtemperatures above 2000 K g-BC2Ndecomposes to form a mixture ofcubic boron nitride (cBN) and diamond.On further decrease in pressure downto 11.0 GPa, thermal decomposition ofg-BC2N proceeds to form cBN anddisordered graphite, as reportedpreviously [2].

The lattice parameter of c-BC2N atambient conditions is a = 3.642(2) Å,which is larger than those of bothdiamond and cBN. The large deviation ofthe lattice parameter of cubic BC2N fromthe value expected from ideal mixingbetween diamond and cBN testifies thatthe synthesised phase differs from the so-called “diamond-cBN solid solutions”reported earlier [3-7] (Figure 2).

The bulk modulus of c-BC2N is282(15) GPa which is exceeded only bythe bulk moduli of diamond and cBN[8,9]. The Vickers hardness of c-BC2N is76 GPa, which makes it the hardestknown solid after diamond.

CBN CRYSTALLISATIONFROM BN SOLUTIONS

IN SUPERCRITICALFLUIDS

In situ studies of the BN-N2H4 systemat pressures to 5 GPa and temperatures to1500 K using a Paris-Edinburgh pressand X-ray diffraction (energy- and angle-dispersive) provided us the ability toinvestigate the local structure of boronnitride solutions in supercritical fluid, andBN crystallisation from these solutionson cooling.

At 4.1 GPa, cooling the solutioncontaining 81 mol.% BN is accompaniedby spontaneous crystallisation of cubicboron nitride. The emergence of cBNlines is accompanied by a change in theshape of the solution spectrum, inparticular by the appearance of a broadhalo with a maximum at dhkl = 1.1 Å. Asimilar change in the shape of thesolution spectrum was also observed at4 GPa and 1350 K by cooling thesolution containing 33 mol.% BN. Inthis case, however, no formation of anycrystalline phase was recorded even oncooling the solution down to roomtemperature. This fact should beattributable to the formation of anunknown amorphous phase of the BN-N-H system (phase X). A subsequentheating up to 1600 K in both cases gaverise to a recovery of the characteristicshape of the solution spectrum due todissolution of solid phase(s) which

is unambiguous evidence for thereversibility of the observed precipitation/dissolution processes.

Our findings show that spontaneouscrystallisation of cBN is observed downto 2.1 GPa (Figure 3). This is the lowestpressure of cBN crystallisation everreported before, though the processoccurs without any catalyst. At 1.7 GPa,cooling of the solution containing73 mol.% BN is accompaniedby crystallisation of graphite-likehexagonal boron nitride (hBN) with asimultaneous formation of the phase X.This fact points to the predominantnucleation of metastable hBN in theregion of cBN thermodynamicalstability. The broad halo of phase Xdisappears upon quenching down toambient conditions. This fact allows thesuggestion that phase X is metastable atlow pressures.

Fig. 3: Sequence of ADX diffractionpatterns taken at 2.1 GPa in thecourse of cooling the supercriticalsolution containing 74.5 mol.% BNfrom 1460 to 1390 K.

Fig. 2: Lattice parameters of c-BC2N (solid circle) and “cubic BN-C solidsolutions” reported by Kakudate et al. [3] (open down triangle), Knittle etal. [4] (open circles), Nakano [5] (open up triangle), Kagi et al. [6] (opensquare) and Komatsu et al. [7] (open diamond). The dashed line representsideal mixing between cBN and diamond, while solid curve shows thedeviation from ideality for the data reported by Knittle et al. [4].

Fig. 4: The radialdistribution functions forthe supercritical solution

containing 74.5 mol.%BN at 2.1 GPa. Dash

line is the 4 πr2ρ0 curve.The average atomic

density ρ0 was estimatedfrom the linear fit to the

reduced RDF G(r) in therange of r = 0-1 Å.

HIGH PRESSURE WORKSHOP

ESRF - JUNE 2001

32

The radial distribution function4 πr2ρ(r) (RDF) was used to characteriseBN solutions in supercritical fluid atdifferent pressures and temperatures. Ourfindings have shown that above 1450 Kbetween 2.1 and 4.8 GPa in theconcentration range from 74.5 to82.5 mol.% BN, interference function i(s)for BN solutions exhibits a sharp firstmaximum (s = 2.84 Å-1) with a smallshoulder on its high-s side, a secondmaximum at 5.1 Å-1 and a thirdmaximum at about 7 Å-1.

RDFs for the solution containing 74.5mol.% BN at 2.1 GPa and differenttemperatures are presented in Figure 4.In cooling the solution, the intensitiesof the maxima in the RDFs decrease,and at 1390 K the maxima nearlydisappear while the most prominent 111line of cBN becomes visible in thecorresponding X-ray pattern (Figure 3),i.e. this temperature can be considered asthe temperature of cBN liquidus at theabove pressure and concentration of thesolution. Thus, the cBN crystallisation isaccompanied (or even preceded) bydisappearance of short-range order in thesolution.

Similar changes of RDFs of solutionsclose to BN liquidus are observed overthe whole ranges of concentrations andpressures being studied. Therefore, it canbe suggested that crystallisation of boronnitride is preceded by decomposition ofthe BN associated solutions in thesupercritical fluid of the N-H system.

KINETICS OF DIAMONDCRYSTALLISATION FROMMETAL-CARBON MELTS

In situ studies of diamondcrystallisation from the Fe-Ni-C melt atpressures up to 6 GPa and temperaturesup to 1700 K were performed using aParis-Edinburgh press and energy-dispersive X-ray diffraction. At 5.2 GPaand the heating rate of 25 K/min,spontaneous crystallisation of diamondstarts at 1510 K immediately after thehalo of the liquid phase appears, and isfully completed at 1605 K (Figure 5). Wenote that despite the presence of a liquidin the system over the whole temperaturerange of diamond crystallisation, alldiffraction patterns exhibit lines of thesolid phase that can be ascribed to thefcc Fe-Ni-C solid solution (γ-phase).

This fact indicates that in accordancewith the p,T-phase diagram of the Fe-Ni-C system [10], under experimentalconditions the L = C + γ monovarianteutectic reaction takes place. In this case,the melt is in equilibrium with bothdiamond and γ-phase.

Diamond crystallisation in the systemunder study is very fast and proceeds in anarrow temperature range that plaguesessentially isothermal kinetics studies.Because of this, the present workdescribes a non-isothermal approach. Asall the diffraction patterns exhibiteddiffraction lines of graphite, we state thatdiamond crystallisation from the Fe-Ni-Cmelt occurs at the constant carbonsupersaturation with respect to diamond,which is ensured by dissolution of theinitial graphite. Based on this, the degreesof the graphite-to-diamond conversionproceeding via melt has been calculatedby normalising integral intensities of the(111) reflection of diamond at varioustemperatures to the appropriate value at1605 K (α = 1).

From the non-isothermal kinetic datait follows that at 5.2 GPa diamondcrystallisation is controlled by carbondiffusion in the melt to the surface of agrowing crystal. Kinetic data might bebest fit by the model that assumes aconstant nucleation rate and three-dimensional growth of the resultingnuclei. From the temperature dependenceof the rate constant in the 1505-1605 Krange, the activation energy of diamondcrystallisation from the Fe-Ni-C melt at5.2 GPa was calculated to be 148(64)kJ/mole.

These studies firmly establishsynchrotron radiation experiments at thethird-generation sources as a powerfultool for studies of materials synthesis in

the B-C-N system at high pressures andtemperatures on a real timescale.

REFERENCES

[1] V.L. Solozhenko, D. Andrault, G. Fiquet, M.Mezouar and D.C. Rubie, Appl. Phys. Lett. 78,1385-1387 (2001).[2] V.L. Solozhenko, Eur. J. Solid State Inorg.Chem. 34, 797-807 (1997).[3] Y. Kakudate, M. Yoshida, S. Usuba, H. Yokoiet al. Trans. Mat. Res. Soc. Jpn. 14B, 1447-1450(1994).[4] E. Knittle, R.B. Kaner, R. Jeanloz and M.L.Cohen, Phys. Rev. B 51, 12149-12156 (1995).[5] S. Nakano, in Proc. 3rd NIRIM Int. Symp.on Advanced Materials. NIRIM, Tsukuba, Japan,287-292 (1996).[6] H. Kagi, I. Tsuchida, Y. Masuda, M. Okudaet al. in Proc. XV AIRAPT Int. Conf. W.A.Trzeciakowski (Ed.) World Scientific, Singapore,258-260 (1996).[7] T. Komatsu, M. Nomura, Y. Kakudate and S.Fujiwara, J. Mater. Chem. 6, 1799 1803 (1996).[8] Ph. Gillet, G. Fiquet, I. Daniel, B. Reynard,and M. Hanfland, Phys. Rev B 60, 14660-14644(1999).[9] V.L. Solozhenko, D. Häusermann,M. Mezouar, and M. Kunz, Appl. Phys. Lett. 72,1691-1693 (1998).[10] Yu.A. Kocherzhinsky, O.G. Kulik andV.Z. Turkevich, High Temp. - High Pres. 25,113-116 (1993).

ACKNOWLEDGEMENTS

The author thanks M. Mezouar, D. Andrault, Y. Le Godec, G. Fiquet,W.A. Crichton, V.Z. Turkevich, A.A. Kurakevich and the late J.-M. Besson for theirinvaluable assistance and contributions.

Fig. 5: Diffraction patterns of the Fe-Ni-C system taken at 5.2 GPa in the

course of a linear heating at a rate of25 K/min.

HIGH PRESSURE WORKSHOP

ESRF - JUNE 2001

33

I t is now more than 40 years since thefirst high-pressure structures weredetermined using diamond anvil pressurecells. In that time, great advances havebeen made in the pressure rangeaccessible with such devices, and in thequality of diffraction data obtainablefrom them. It is then somewhat surprisingthat so many significant uncertaintieshave remained, even in the structures ofelements – at quite modest pressures insome cases.

Among the most persistent of theunsolved problems have been those ingroups II and V of the periodic table,where the high-pressure phases Ba-IV,Sr-V, Bi-III, Sb-II and As-III have beenshown to be complex, but have resistedall previous attempts at full solutions.Recently, using a combination of single-crystal and powder diffraction datacollected at the SRS (Daresbury, UK), in-house, and at the ESRF, we have foundthat in all these cases, the high-pressurestructure is very similar and of an entirelynew type – described as the ‘weirdestknown atomic structure of … any pureelement’ [1].

Figure 1 shows a diffraction patterncollected from a single-crystal of Ba-IVat 12 GPa at the SRS. The diffractionpattern comprises layers of diffusescattering (seen edge-on in this image),Bragg reflections that lie on the planes ofdiffuse scattering, “satellite” reflectionsaround these reflections (enlarged in theinset), and Bragg reflections not on thediffuse planes are marked by arrows.

The structure explaining all thesefeatures is shown in Figure 2. It is acomposite arrangement comprising atetragonal ‘host’ structure (which givesrise to the Bragg reflections not on thediffuse planes), and chains of ‘guest’atoms lying in channels that run alongthe c-axis of the host. These chains formC-centred tetragonal and C-centredmonoclinic guest structures [2] whichgive rise, respectively, to the Braggreflections on the diffuse planes and tothe satellite reflections. The diffuse

scattering itself can be attributed to a lackof inter-chain ordering in some fractionof the chains. The most surprising thingabout this composite structure is thatthe host and guest structures areincommensurate with each other alongthe c-axis: the ratio of their c-axis latticeparameters, cH/cG, is 1.388 at 12 GPa,and varies continuously with pressure.

Further studies of Ba-IV have revealedthat at 12.5 GPa, the monoclinic gueststructure undergoes a structural phasetransition, without any accompanyingchange in the host [2]. We have termedthis an intra-phase transition. Anotherintriguing aspect of the structure is that thenumber of atoms in the host unit cell –which is non-integer and equal to8 + 2x(cH/cG) = 10.776 at 12 GPa – ispressure dependent, through the pressuredependence of cH/cG. Some of the chainatoms must then be ‘squeezed out’ of thechannels with increasing pressure.

WEIRD METALS – MODULATIONS WITHIN

GUESTS WITHIN HOSTSM.I. MCMAHON AND R.J. NELMES

THE UNIVERSITY OF EDINBURGH (UK)

Presentation given at the ESRF Workshop "Science at High Pressure:Latest Trends from 3rd Generation Sources", 16-17 February 2001.

High-pressure studies on certain elemental metals have unravelled unusual structuressuch as host-guest complexes. The characteristics of these structures cannot easily be

explained and pose quite a challenge for theoretical studies.

Fig. 1: 2-D diffraction pattern from a single-crystalof Ba-IV at 12 GPa. The inset enlarges the marked

area. Vertical arrows mark host reflections, andarrows in the inset mark satellite reflections from the

monoclinic guest adjacent to stronger reflectionsfrom the tetragonal guest.

HIGH PRESSURE WORKSHOP

ESRF - JUNE 2001

34

Sr-V is stable above 46 GPa andknown to have a similar diffractionpattern to Ba-IV. Studies at the SRS andon ID9 at the ESRF have revealedthat it too has a composite host-guestincommensurate structure of the Ba-IVtype, with cH/cG = 1.404 at 56 GPa [3].We have also found that Sr-V undergoesan intra-phase transition at 71(1) GPa [3].However, the structure of the guest phaseabove this pressure remains unknown.

The structure shown in Figure 2bears a striking resemblance to the(commensurate) structure proposedpreviously for Bi-III and Sb-II [4], butwhich cannot be correct on densityconsiderations. Powder and single-crystalstudies of Bi-III and Sb-II have shownthat in both cases, the true structure is acomposite incommensurate structure withthe same tetragonal host as Ba-IV andSr-V, but with a different, body-centredtetragonal, guest [5]. The calculateddensities of these incommensurate phasesagree extremely well with thosedetermined directly over 40 years ago byBridgman and others. Recent studies ofAs-III at the ESRF and the SRS haveshown that while it also has the sametetragonal host as Ba, Sr, Bi and Sb, theguest is monoclinic.

While the composite host-gueststructures fit all the main features in theobserved Bi-III and Sb-II diffractionprofiles, there are a number of extremelyweak reflections in profiles collected atthe ESRF (see Figure 3) that are notaccounted for. These peaks are from a

furthur modulation of the structure [7]. Itis clear that there are further levels ofcomplexity yet to be uncovered in thesestrange phases!

It is a challenge for theoretical studyto understand why such complexstructures are stable over pressure rangesas large as 5 - 30 GPa and more. Thischallenge is made more difficult by theincommensurate nature of the structures.Heine has suggested possible criticalfactors for incommensuration such ascharge density waves and the strength ofthe host-guest interaction [1]. Firstinsight into the stability of Ba-IV hasbeen obtained from calculations of acommensurate approximation [6]. Someother insight might come from anintriguing similarity with commensurateanalogues found in binary alloys such asAl2Cu and In5Bi3, which raises theinteresting possibility that the Ba-IV-type

phase might also be considered as an‘alloy’ – comprising atoms with, perhaps,two different electronic arrangements.There is much yet to do, including furtherdiffraction work to search for more ofthese ‘weird metals’ in other elements.

REFERENCES

[1] V. Heine, Nature 403, 836-837 (2000).[2] R.J. Nelmes, D.R. Allan, M.I. McMahonand S.A. Belmonte, Phys. Rev. Lett. 83, 491-494 (1999).[3] M.I. McMahon, T. Bovornratanaraks,D.R. Allan, S.A. Belmonte and R.J. Nelmes,Phys. Rev. B. 61, 3135-3138 (2000).[4] J.H. Chen, H. Iwasaki and T. Kikegawa,High Press. Res. 15, 143-158 (1996).[5] M.I. McMahon, O. Degtyareva and R.J.Nelmes, Phys. Rev. Lett. 85, 4896-4899 (2000).[6] S.K. Reed and G.J. Ackland, Phys. Rev.Lett. 84, 5580 (2000).[7] O. Degtyareva, M.I. McMahon andR.J. Nelmes, to be submitted (2001).

ACKNOWLEDGEMENTS

We would like to acknowledge the part played in the experimental work and analysis by our colleagues D.R. Allan,S.A. Belmonte, T. Bovornratanaraks, O. Degtyareva and C. Vanpeteghem. We thank V. Degtyareva for first drawing ourattention to the similarities between the published structures of Bi-III and Ba-IV and for stimulating discussions. This workis supported by grants from the EPSRC, funding from the CCLRC, and facilities provided by Daresbury Laboratory. M.I.M.acknowledges support from the Royal Society.

Fig. 2: The composite structure of Ba-IV. The host structure(dark atoms), with guest-atom chains (light atoms), is shownin a c-axis projection. The inset shows the tetragonal andmonoclinic guest structures.

Fig. 3: Integrated profile from Sb-II at 10.3 GPa [7]. The tick marks show thepositions of reflections from the host-guest composite structure. The strongest of

the additional reflections are marked by asterisks.

BIOLOGY WORKSHOP

ESRF - JUNE 2001

35

The solving of the human genomesequence, and the genomes of many otherspecies, has resulted in a huge increase inthe number of clinically-relevant targets(proteins) of interest to pharmaceuticalcompanies. The production of targets forstructural analysis has typically been avery slow process, mainly due to the veryhigh purity and large quantities of proteinrequired. To give the highest impactin a drug design process, structuralinformation should ideally be availablewhen lead compounds are identified fromhigh-throughput screens. However, thesolving of structures of proteins ofinterest can frequently occur at a very latestage within drug development projectspermitting only limited input to structure-based drug synthesis and optimisation.

Large pharmaceutical companies areoften interested in targets from diverseprotein classes, ranging from microbialenzymes to kinases, membrane boundproteins such as 7-transmembranereceptors and ion channels. Expression ofthese proteins increases in difficulty -microbial proteins are relatively easyto express and crystallise, solublemammalian proteins can be much harderto obtain, and integral membrane proteinsare extremely difficult to generate in anactive, folded form in sufficientquantities for structural determination.

There are many different systemsavailable for expression of genes ofinterest. E. coli is the quickest and mosteconomical, being easy to scale to largevolumes and suitable for generation oflabelled protein. However, E. coli does notperform post-translational modifications

that may be required for functionalactivity. Over-expression in yeast is alsofast and straightforward to scale, thoughthe first choice for mammalian proteinsis often baculovirus expression. This ismuch more expensive than yeast andE. coli, especially at large scale, but thissystem does perform post-translationalmodifications such as phosphorylation.Expression of eukaryotic proteins inmammalian cells will usually producethe native protein, but is a very slow,expensive procedure and only infrequentlyproduces proteins in high enough yield forstructural determination.

Generation and expression ofproteins for structural analysis hasusually been performed as an iterativeprocess. One to three constructs -alternative forms of the gene, eithertruncated or mutated - are generated fromthe full-length gene and expressed in the‘best bet’ system, usually E. coli. The firstconstruct will often be that generated forhigh-throughput screening, which is notnecessarily ideal for crystallisation. If theconstructs express well, and can besupplied to the crystallographers in therequired purity and concentration, asecond round of construct generation isnot necessary unless crystallisation failsor the crystals are of poor quality.However, if the proteins fail at any stepduring the supply (purification orexpression), further rounds of constructgeneration or use of an alternativeexpression system is required. Thisiterative process clearly lengthens thetime taken to produce structures,decreasing its impact on the drug designprocess.

It is clear that the time from geneto structure must be improved totake advantage of the many new targetsin the completed human genome.Homology modelling can produceputative structures for genes withinfamilies, but the structures currentlydeposited in the PDB (BrookhavenProtein Data Bank) represent only afraction of the populated folds,characteristic tertiary structure elements,such as helix-loop-helix, ß-barrel,thought to exist in natural proteins [1].This limits the usefulness of homologymodelling as a predictive tool. There aremany initiatives worldwide to determinestructures of unknown proteins from thevarious genomes, and it is estimated thatsolving a minimum of 10,000 structureswithin 10 years should produce at leastone example of the folds from eachfamily of proteins [2].

To determine such a large number ofstructures within a relatively short timeframe requires a re-think of the currentmethods of protein generation. Instead ofworking with one to three constructs,and performing several rounds ofgeneration and purification to generateall of the constructs of interest to thecrystallographers, multiple constructscould be produced in one go (includingconstructs designed for high-throughputscreening). These could be cloned intopowerful E. coli expression systems, andexpressed in multiple E. coli strains. Thereare many new strains designed to improveexpression and folding of proteins. Oneor two affinity tags could be included(both N- and C-terminal) to assist inpurification, and protease cleavage sites

STRUCTURAL GENOMICS: PITFALLS AND PROSPECTSA. BRIDGES AND O. JENKINS

GENE EXPRESSION SCIENCES, GLAXOSMITHKLINE, HARLOW (UK)

Presentation given at the ESRF Workshop "High Throughput Structural Biology", 20-21 February 2001.

A discussion of the issues and feasibility of truly high-throughput expression,purification and structural determination using current technologies, highlighting

potential bottlenecks in the existing process of gene to structure.

BIOLOGY WORKSHOP

ESRF - JUNE 2001

36

S tructural genomics initiatives such asthat recently funded by the NationalInstitute of General Medical Sciences in

the USA will lead to a dramatic increasein the rate at which macromolecularstructures are determined. This increase

in throughput will be achieved byapplying automation to almost all steps inthe structure determination pathway,

(such as TEV or PreScission) could alsobe encoded to remove large flexibleregions from the N-termini of the proteinswhich would potentially interfere withcrystallisation. In addition, site-directedmutagenesis could be employed to removeundesirable surface residues or loops.Using this approach, it could beanticipated that many constructs of aparticular protein in a form suitable forcrystallisation may be generated.

What would be the implications of thisincrease in the number of proteins forcrystallisation? It is estimated that theproduction of 1000 crystal structures peryear would require approximately 10,000proteins to be screened for crystallisation(about 800 crystallisation screens permonth). If each crystallisation screencomprised 200-500 conditions, then eachscreen would require 2-5mg of protein at aconcentration of 10mg/ml and 95% purity.In order to collect sufficient X-ray data tosolve this quantity of crystal structures adedicated MAD (Multiple-wavelengthAnomalous Diffraction) beamline wouldbe essential. Using current structuresolution methods, automated electrondensity map interpretation and refinementpackages each structure would takeapproximately 2-4 weeks to complete. A

team of 50-100 protein crystallographerswould be required to undertake a task ofthis magnitude.

The key to completion of this numberof structures is miniaturisation andautomation at every step of the process.Cloning, expression and purification isrelatively straightforward to automate;robots exist to perform liquid handling,but expression analysis would be timeconsuming by electrophoresis (SDS-PAGE), so an alternative must bedeveloped. Crystallisation can also beautomated (possibly at the nanolitrescale), but crystal mounting and freezingis currently a time-consuming andmanual procedure, and therefore aserious bottleneck in this process. Finally,new programs need to be developed toautomate structure solving and performdocking in a high-throughput mode.

It is not impossible to do high-throughput structural determination.Indeed, there are several biotechcompanies (e.g. Syrrx, StructuralGenomics) with business plans based onstructural genomics. Large pharmaceuticalcompanies have a slightly different focus,with needs closer to 'functional' genomics.Although crystal structures of pure

proteins can be used for compounddocking and SAR (Structure-ActivityRelationship studies), more informationfor drug design and optimisation can beobtained from structures of protein/ligandor protein/ inhibitor complexes. This isparticularly true if inhibitors of manystructural classes are available for co-crystallisation. However, this is just adifference in emphasis in the needs ofstructural genomics projects versus drugdiscovery, and the methods discussed inthis article can be applied and adapted toachieve both sets of aims.

REFERENCES

[1] S. K. Burley, Nature Struc. Biol. 7,932-934 (2000).[2] J. C. Norvell and A. Z. Machalek,Nature Struc. Biol. 7, 931 (2000).

ACKNOWLEDGEMENTS

We would like to thank all of themembers of the Structural Biology,Gene Expression Sciences, ProteinBiochemistry and ComputationalChemistry departments at GSK inHarlow that took part in thediscussions described in this article.

AUTOMATED DATA COLLECTION AND PROCESSING FOR

MACROMOLECULAR CRYSTALLOGRAPHYA.G.W. LESLIE

MRC LABORATORY OF MOLECULAR BIOLOGY, CAMBRIDGE (UK)

Presentation given at the ESRF Workshop "High Throughput Structural Biology", 20-21 February 2001.

Fully automated data collection and processing is an achievable and highly desirableobjective for modern synchrotron protein crystallography beamlines. A possible scheme for

reaching a high level of automation with modest programming resources is outlined.

BIOLOGY WORKSHOP

ESRF - JUNE 2001

37

from protein expression and purificationthrough to model building andrefinement. Inevitably, the stepsinvolving collection and processing ofthe diffraction data will also need tobecome far more automated than atpresent, if these steps are not to becomerate limiting. Based on conservativeestimates of exposure time and detectorreadout time (10 s each per 1˚ image), atypical third-generation beamlineequipped with automated sample loadingis capable of producing 24 complete(180˚ rotation) datasets per day. Thiscould represent 12 ab inito structures(based on 2 wavelength MAD phasing)which translates into over 3000 structuresper annum. In practice, the requirementto screen multiple crystals prior tocommencing data collection and the factthat not all MAD datasets will lead to astructure, will significantly reduce thisnumber. Nevertheless, this example helpsto illustrate that data collection andprocessing must become more automatedif the potential throughput of such abeamline is to be achieved.

At present, the scientists working onthe beamline are responsible for mountingsamples, for operating the beamlinecontrol software and for running the dataprocessing software. They decide whetheror not to collect data from a particularcrystal on the basis of visual inspection ofimages and information obtained from thedata processing programs. They set theoptimal data collection parameters(resolution, exposure time, oscillationangle, phi range) and ideally try to processat least some of the images as they arecollected. In an automated system,crystal mounting and centring would be

performed robotically, and "intelligent"software (an expert system) would takethe decisions currently made by thescientists, based on information providedby the processing software and "projectparameter information" stored in adatabase. One possible scheme is outlinedin Figure 1. Its modular nature isdeliberate because a more monolithic orintegrated structure would take longerto develop and could not easily beinstalled on different beamlines, whichtypically have their own version ofthe beamline control software.Communication between the moduleswould be at a high level such that theexpert system could instruct theprocessing module to characterise acrystal using images from a certaindirectory. In response, the processingmodule would return information such ascrystal cell parameters, possible spacegroup, mosaicity, and resolution limit. Onthe basis of this information, the expertsystem could then choose the acquisitionparameters and instruct the beamlinecontrol module to proceed with the dataacquisition.

To assess the feasibility of this type ofautomation, it is necessary to examine thesteps involved in manually characterisinga crystal for data collection. The first stepis to collect two images, separated by 90˚in rotation angle, and to verify whetherthe crystal diffracts to a useful resolution,whether it is a single lattice, and whetherthe mosaicity and spot shapes areacceptable. If the crystal passes this"quality check", the images areautoindexed to derive a unit cell andpossible spacegroup(s). The autoindexingis also checked by predicting the spots for

the two images. A more quantitativeestimate of the mosaicity is determined,and a data collection strategy is workedout (total phi range(s), oscillation anglerequired to avoid spatial overlaps). Theimages are integrated to obtain anestimate of the true resolution limit, andthe exposure time and resolution for datacollection are chosen. Finally, the data arecollected and, ideally, processedsimultaneously.

The most challenging step toautomate in this procedure is anassessment of the "quality" of the twoinitial diffraction images. This issomething that is quite straightforwardfor an experienced crystallographer, but itwould require rather sophisticated imageprocessing techniques to extract the sameinformation automatically. A possiblesolution to this problem is to extract thisinformation indirectly, based on thesuccess or failure of the autoindexing ofthe two images. Algorithms forautoindexing are now generally veryrobust and further improvements shouldhelp to reduce the occurrence ofalgorithm failure. An examination of thesituations in which the autoindexing failsreveals the following causes: incorrectdirect beam co-ordinates, crystal-to-detector distance or wavelength;insufficient spots found (implying weakdiffraction); multiple lattices or splitspots; excessive mosaic spread or toolarge a rotation angle. On an automatedbeamline, the physical parameters wouldbe passed from the beamline controlsoftware and therefore should not beerroneous. Many of the other reasons forthe failure of autoindexing would suggestthat the sample is actually unsuitable for

Fig. 1: A schematic representation ofthe control software for an automatedprotein crystallography beamline.

BIOLOGY WORKSHOP

ESRF - JUNE 2001

38

data collection. The two images shouldbe autoindexed separately (to detectlattice imperfections which are notvisible in some regions of reciprocalspace) and also together, to give moreaccurate cell parameters. Rejectioncriteria based on the rms positionalresidual of indexed spots and the numberof spots rejected from both the indexingand the cell parameter refinement will beapplied to select successful solutions.Samples which fail this test are flaggedfor visual inspection (possible at a latertime) by the beamline operator who may(e.g. by manual spot editing) be able tofind a satisfactory solution. Autoindexingwill produce a list of possible solutionswith different symmetries, each with itsown penalty score which evaluates howwell the cell parameters comply with therestrictions appropriate for thatspacegroup. Generally (unless the truesymmetry is triclinic) there will be a clearseparation between a number of solutionswith low penalties and other solutionswith much higher penalties. The solutionwith the highest lattice symmetry fromthe group with low penalties willnormally be selected as an initialhypothesis for the true symmetry.

The next step is to obtain an estimateof the mosaic spread. The algorithmimplemented in MOSFLM [1] relies onintegrating an image with different valuesof mosaic spread and evaluating the totalintensity of all predicted reflections. Thistotal intensity will reach a plateau whenthe true mosaic spread is reached.

A data collection strategy can now bedetermined based on the assumedspacegroup, the known spot size and themosaic spread. Initially, a minimum totalrotation range, typically divided into twoor three wedges of data that should resultin a high (e.g. 90-95%) completeness,will be calculated. For this rotationrange(s) the maximum oscillation anglethat will avoid spatial overlap will also becalculated as a function of the phi angle.This data will be collected first, in orderto minimise radiation damage. A seconddata collection range, designed to bringthe completeness to 100% and increasethe multiplicity of the observations,will also be calculated. For appropriatecases (MAD data collection), thecompleteness of the anomalous data willbe maximised. All the functionality forthese calculations already exists inMOSFLM. Finally, the exposure timeand resolution of the dataset need to be

defined. To do this, the two initial imagesare integrated and the mean I/σ(I) iscalculated as a function of resolution. Theeffective resolution is defined as that atwhich the mean I/σ(I) drops below acutoff value (e.g. 2). Using Poissonstatistics, it is possible to estimate theexposure time required to achieve thedesired I/σ(I) value at any particularresolution. This would be compared withthe "maximum allowable" exposure timeto arrive at a final choice of resolutionand actual exposure time. Data collectioncan then be started. For lower symmetryspacegroups (orthorhombic or below), afew images should be collected 90˚ awayin Φ from the first data wedge. Theseimages, together with images from thefirst wedge, are used to determineaccurate cell parameters by post-refinement. Using these cell parametersthe images are processed as they arecollected. During the data collection thefirst two images are re-collected atregular intervals to monitor radiationdamage. The images should also bescaled and merged at the earliestopportunity, to check the initialassumption of the spacegroup symmetry.If it turns out to be incorrect, the datacollection strategy will need to berecalculated, taking into account the datathat has already been collected.

The modularity of the systemshould allow maximum flexibility inimplementation while minimising theprogramming effort. The expert systemshould be beamline and synchrotronindependent and could be a focal pointfor collaboration between differentsynchrotron sites. Only a command"translator" module would be required tointerface a standard expert system withdifferent beamline control software.Information specific to a given sample orproject will need to be supplied to theexpert system in order to assist indecision making. Some of the possibleproject parameters are listed below:

(a) Minimum acceptable resolution (defined as I/σ(I)>X)

(b) Highest resolution required(c) Maximum exposure time per dataset(d) Maximum acceptable mosaic spread(e) Maximum acceptable anisotropy(f) Maximum acceptable radiation

damage (expressed as a B factor or ∆I/σ(I) at the highest resolution)

(g) Number of wavelengths required(h) Anomalous scatterer

(i) Minimum exposure time per image(j) Maximum number of overloaded

reflections(k) Minimum acceptable completeness

(anomalous or overall)

The number of project parametersand their default values would have to bedetermined on the basis of experience.

Beamline performance could bemonitored by evaluating a standardsample at regular intervals. Anydeterioration in quality would result ina message being passed to the beamlineoperator, possibly via a mobile phone.

The different processes outlined inFigure 1 would probably run on differentcomputers. In particular, a highperformance machine with rapid diskaccess would be required for the dataprocessing if this is to keep pace withdata collection. Given the continualimprovement in performance, it is notunrealistic to expect data collectionand processing to proceed in step.A collaboration is currently beingestablished with personnel at the ESRFand SRS (UK) to work towards apractical implementation of the schemeoutlined above.

REFERENCES

[1] A.G.W. Leslie, Joint CCP4 +ESF EAMCB Newsletter on ProteinCrystallography, 26, (1992); http://www.mrc-lmb.cam.ac.uk/harry/mosflm/mosflm_user_guide.html.

ACKNOWLEDGEMENTS

I would like to thank P. Evansand H. Powell for many usefuldiscussions.

BIOLOGY WORKSHOP

ESRF - JUNE 2001

39

Since the first ESRF beamline formacromolecular crystallography, ID2B,came on-line in September 1994,macromolecular crystallography hasdeveloped into a standard tool forscientists in academia and industry alike.It now has a routine role in the drugdevelopment cycle and is an intrinsiccomponent of the burgeoning structuralgenomics field. In order to maintain pacewith this explosive growth and augmentthroughput of generic experiments, suchas ligand soaking and <100 kDa structuresolution, the beamlines must becomeautomated, reliable and easy to use.Automation will clear the path forthe challenging long-term projectsrequiring many shifts of beam time thatwill follow structural genomics: themembrane proteins, the functionalstudies and, perhaps most crucially, thestudy of complexes of proteins andmacromolecules that is central tothe understanding of their real-lifeinteractions.

THE CURRENTBEAMLINES

Five JSBG (Joint Structural BiologyGroup) beamlines are available formacromolecular crystallography: theID14 complex with four end-stations andID29 which, since the end of 2000, hassuperseded BM14 (now an Anglo-Spanish CRG line). Every six monthsthese beamlines handle over five hundredprojects and demand is on the increase,

requiring improved efficiency andthroughput.

GOAL: GENERIC HIGHTHROUGHPUT

Crystallographic data collectionrepresents only a small time slice on thestructural genomics pathway, butautomation at the beamline is notonly concerned with this facet ofthe process. Beamlines increasinglyresemble factory assembly lines, with thegoal of automation to streamline theentire structure solution process from

sample mounting, alignment andcollection strategy through to dataprocessing, reduction, phasing andeven initial 3-D model construction(Figure 1). To accomplish this willrequire a close collaboration betweeninstrumentationalists and softwareengineers.

Automation also goes to the heart ofthe beamline, in the optics hutch. It is oflittle use automating the experiment if theunderlying optical elements are not asoptimised and reliable as possible. Thesuite of JSBG beamlines will be madeconsistent in terms of software and

AUTOMATION OF THE MACROMOLECULAR

CRYSTALLOGRAPHY BEAMLINES AT THE ESRFE.P. MITCHELL

THE JOINT ESRF AND EMBL STRUCTURAL BIOLOGY GROUP (JSBG)

Presentation given at the ESRF Workshop "High Throughput Structural Biology", 20-21 February 2001.

Macromolecular crystallography is now a routine tool in academia and industry alike.Automation is required to maximise throughput on the available beamlines and

to cope sensibly with the increasing influx of measurements.

Beamline Detector Wavelength range (Å)

ID14-1 165mm MAR CCD Fixed: 0.93ID14-2 ADSC Q4 Matrix CCD Fixed: 0.93ID14-3 165mm MAR CCD Fixed: 0.93ID14-4 ADSC Q4R Matrix CCD MAD: routinely* 0.94 - 1.15ID-29 ADSC210 Matrix CCD MAD: routinely* 0.80 - 2.1

*other wavelengths possible by prior arrangement.

Fig. 1: Targetprotein to initial 3-D model forgeneric structuresolution.

BIOLOGY WORKSHOP

ESRF - JUNE 2001

40

hardware, with built-in diagnostic tools toallow rapid alignment or intervention inthe event of a breakdown.

Optics hutch and experimentautomation together will reduce humanerror: tired users in the small hours canmake unwise decisions, deterioratingtheir data quality and eventual structuralinformation or indeed the beamline itself.

IN THE PIPELINE:HARDWARE

Plans are already underway to useID14-3 as a test-bed for beamlineautomation. New equipment will beinstalled on the line over the comingmonths and then rolled out over the otherstations once thoroughly tested andproven. The Automation Task Force havedesigns in hand and the first steps will betaken shortly with particular regard tothe beamline optics where automaticalignment tools and instrumentation willbe added.

For the experimental hutch, the JointInstrumentation Group (JIG) of the ESRFand EMBL has developed an automatedsample changer. Currently, user groupscan take many minutes to retrieve, mountand align a pre-frozen sample, with thisprocess becoming harder as averagesample sizes drop from 200 µm severalyears ago to around 80 µm now. Thesample changer is under final benchchecks and will soon move to a beamlinefor on-line testing. The current device,the design of which is extendable, canhold up to twenty samples in HamptonResearch cryo-vials (Figure 2) and issimple in conception as well as robust,compact and rapid. With the footprint ofa dry-transport dewar and standing lower,it can place or remove a sample in undertwo seconds on the experiment

goniometer. Once mounted, automaticalignment routines can precisely centrethe sample for data collection. Work onthis aspect is underway, though patternrecognition of the cryo-cooled crystal insupporting loop will not be simple.

The changer is compatible with amicrodiffractometer, also developed bythe JIG, and already used for scheduledexperiments on the ID13 and ID14 linesmuch to the delight of users. Themicrodiffractometer can work withsamples down to 5 µm in size and allowsremote accurate centring in small beams.A commercialised version is underdevelopment.

IN THE PIPELINE:SOFTWARE

With the revolution in data collectionspeeds, the ergonomics and ease of use ofa beamline have become paramount. Tothis end, the ProDC interface, the user’sbeamline communication and controlcentre, is now installed on all beamlinesproviding a unified interface no matterwhich detector or equipment is installed.The software takes care of routine datacollection tasks, such as safety shutteropening, beam loss checks, auto beamrealign and one button edge scans foranomalous diffraction data.

The beamline software also interactswith databases in order to storeexperiment and beamline parameters.This can allow subsequent automateddata processing using stored parameterseither on-site or remotely through web-based interfaces, and also the recovery ofknown beamline configurarations in theevent of failures. These web interfacesare going through the second generationof development using the ZOPEenvironment and eventually users can

expect to use the interfaces to querydatabases with questions such as:• What did I collect around 28 Feb

2001?• Can I take a look at the images?• I’d like to process the data using

software X and script Y.• What are the data processing statistics?• Can I see the Pattersons?• Can I trace a 3-D structure?

These questions highlight theadvantages and flexibility of adatabase environment, in comparison tocomplementary information storage inthe headers of image files, whereparameters and results can be changed oradded at a later date.

THE FUTUREThe beamlines are continuously

evolving and projects, such as theautomated sample mounting andalignment, are underway to furtheradvance beamline efficiency. The JSBGis aiming for a fully automated beamlineto be operational in two to three years'time. In the future, user presence may notbe necessary for all measurements withsamples routinely sent to beamlines,handled by robots and the processed data(or initial coordinates) sent electronicallyback to the research groups at home.

WHAT COULD BEDONE?

Assuming one useful structure data set(MAD, SAD, ligand soak, etc.) per hour,an optimised suite of JSBG beamlinescould produce 125 structures per day or25,000 per year. Multiply this bythe European synchrotrons and theirmacromolecular crystallography resourcesand over 100,000 structures could beproduced per year. Even if only onequarter of this total were to be realised,the input of structural information toworld databases would be extraordinary.Imagine what could be achieved withthat amount of knowledge!

Further InformationThe JSBG is a joint ESRF and EMBL team oftechnicians, engineers and scientists operatingand maintaining the JSBG beamlines, togetherwith technical developments and in-houseresearch. (http://www.esrf.fr/exp_facilities/jsbg/jsbg_beamlines.html)

Fig. 2: (Left)schematic of

sample changercassette and coldbox. (Right) the

completechanger unit

ready to move toa beamline.

M A C H I N E DEVELOPMENT

ESRF - JUNE 2001

41

In the ESRF storage ring, there arethree types of magnet girder assemblies(MGA) involving quadrupole magnets:G10, G20, G30. The G10 and G30MGAs each contain five magnets,whereas the G20 MGA is longer andconsists of seven magnets. Thefundamental resonant frequency of thesemagnet girder assemblies is in thefrequency range of 6.5 to 9 Hz, with alateral rocking motion. This resonantrocking motion of the MGAs induces theelectron beam motion with a dominantfrequency of 6.8 Hz mainly in the lateraldirection. The rms displacement in thefrequency range of 4 to 12 Hz is typically0.2 µm for quadrupoles, and 12 µm forthe electron beam at high-ß section. Theelectron beam motion influences theposition stability of the X-ray beam, theintensity stability of the X-ray beam aftera monochromator, especially for ahorizontally diffracting monochromator.Intensity variation of the X-ray beamwith a peak frequency around 7 Hz hasbeen observed in some beamlines, forinstance, ID14, ID24 and ID26. In orderto improve electron beam and X-raybeam stability, it is necessary to attenuatethe vibrations of the magnet girderassembly in the storage ring.

A damping device, the so-called‘damping link’, has been developed toattenuate the vibrations of the magnetgirder assembly. It consists of three parts(Figure 1):

• a sandwich structure withaluminum plates and ViscoElasticMaterial (VEM): Al + VEM + Al• a girder mounting fixture (GMF)which allows the sandwich structureto be linked to the girder

• a floor mounting fixture (FMF)which allows the sandwich structureto be linked to the floor

The idea is to use the sandwichstructure with VEM to absorb thedynamic strain energy of the MGA. Thedamping links are installed on the twoextremities of the girder and floor asshown in Figure 1. The mountingfixtures (GMF, FMF) should bothaccommodate the environment in thetunnel and be stiff enough to transmit

maximal dynamic strain energy of theMGA to the VEM layer which thendissipates this energy. Mechanicalproperties of the VEM are keyparameters for the successful design ofdamping link.

Significant efforts were necessary forthe installation. The available space wasvery limited such that cooling pipes andsome electrical tracks had to be moved.The installation of damping links at alllocations in the ESRF storage ring was

DAMPING LINKS TO ATTENUATE VIBRATIONS OF MAGNET

GIRDER ASSEMBLIESL. ZHANG

ESRF, TECHNICAL SERVICES DIVISION

Installation of damping links to reduce quadrupole magnet vibrations was completedat the ESRF during the March 2001 shutdown. The results from measurement

clearly show attenuation of vibration of the quadrupole magnets, and significantimprovement of electron beam and X-ray beam stability.

GMF:Girder Mounting Fixture

VEM:ViscoElastic Material

FMF:Floor Mounting Fixture

Fig. 1: Damping link and location on a storage ring girder.

M A C H I N E DEVELOPMENT

ESRF - JUNE 2001

42

completed during the March 2001shutdown.

Vibration tests have been performedon quadrupoles before and after dampinglink installation. Results in terms of Q-value (peak value in the transfer functionat fundamental resonant frequency) ofquadrupoles are shown in Figure 2. Notethat there are 32 cells in the storage ring,and 3 MGAs with quadrupoles per cell.Results are given here for 2 quadrupolesper MGA: QF2 and QD3 for G10, QD4and QF5 for G20, QF7 and QD8 forG30. The Q-value of the quadrupoleswas about 50 before installation of thedamping links, and was reduced to about10 after installation of the damping links.

In cell 26, damping plates (anotherdamping device) were installed betweenthe jacks and the floor in 1997. The jackswere bolted to the floor. The dampingplates were partially shunted by thebolts, but there are still some dampingeffects. This explains why the Q-valuesbefore installation of the damping linksin cell 26 were significantly smaller thanin other cells.

Electron beam motion has beenmeasured before and after theinstallation of the damping links in thestorage ring. As the installation work wasdone during four machine shutdowns(Summer 2000, October 2000, Winter2000/2001, and March 2001),

measurements were also made in thecase where only part of storage ring wasequipped with damping links. PowerSpectral Density (PSD) of the horizontaldisplacement of the electron beam isshown in Figure 3 for four cases. Beforethe installation of the damping links,there was a huge peak at 6.8 Hz in thehorizontal displacement PSD. When halfof the storage ring was equipped withdamping links, limited damping effectson the electron beam could be observed.When the storage ring was totallyequipped with damping links, the peak at6.8 Hz in the PSD was shifted to about 9Hz and dramatically reduced by a factorgreater than 40. The RMS displacementin the frequency range of 4 to 12 Hz wasreduced from a typical value of 12 µm to3 µm for the electron beam, and from0.2 µm to 0.05 µm for the quadrupoles.In the results shown in Figure 3, there isa wide peak around 30 Hz. The dampinglinks have no effects on this wide peak.This is because the wide peak around 30Hz in the PSD of the electron beam isdue to the lateral rocking motion of the

Fig. 2: Q-valueof quadrupoles

before and afterthe installation

of the dampinglinks in the

storage ring.

Fig. 3: Horizontal displacement PSD of theelectron beam before, during and after theinstallation of the damping links in thestorage ring.

M A C H I N E DEVELOPMENT

ESRF - JUNE 2001

43

quadrupoles QF2 and QF7 relative to thegirder. As the girder does not move forthis vibration mode at 30 Hz, thedamping links are therefore not effectivefor the vibration of the quadrupolesaround 30 Hz, as well as for the motionof the electron beam around 30 Hz.Some countermeasures to reduce thevibrations of quadrupoles QF2 and QF7have been studied by finite elementsimulation, and could be very effective.

The significant enhancement of theelectron beam stability was alsoobserved on the X-ray beam. Figure 4shows the spectra of the X-ray beamintensity variation measured at the ID14-EH1 beamline in January 2000 and inApril 2001. Damping links for themachine girders were installed betweenthese two dates. The spectra areexpressed in percentage of the DC value.The variation of intensity should be assmall as possible, so that the spectralvalue should be significantly smallerthan unit 1. The peak at 6.8 Hz in the X-ray bean intensity spectra is removedin totality after damping links have beeninstalled in the storage ring. Note that thelocal feedback on the electron beamsignificantly reduced the intensityvariation around the peak frequency6.8 Hz, but the peak is still visible.

In conclusion, the damping linkshave been successfully developed andimplemented in the ESRF storage ring.Vibrations of the magnet girderassemblies were effectively attenuated.Electron beam and X-ray beam stabilitywere significantly improved.

REFERENCES

[1] L. Zhang, T.M. Lewis, C. Michael, 1998International Conference on Noise andVibration Engineering, Proceedings ofISMA23, Vol. 3, pp1481-1488 (1998).[2] L. Zhang, Proceedings of EPAC 2000,pp2489-2491, 7th EPAC, Vienna, Austria (2000).

ACKNOWLEDGEMENTS

The author gratefully acknowledges M. Lesourd’s contribution for variousmeasurements, P. Duru and J.-C. Erard for the installation, and the Drafting Office forthe manufacture of the mounting fixtures. The contribution from Buildings &Infrastructure, Survey & Alignment Groups and many ESRF colleagues is alsoacknowledged. Many thanks to L. Farvacque and E. Plouviez for fruitful discussionsand collaboration on the electron beam stability measurements.

Fig. 4: Spectra of the X-ray beam intensityvariation measured with the ID14-EH1 beamline.

VACANCIES AT THE ESRF ON 14 JUNE 2001

Ref Subject Deadline

TECHNICIAN 2554 Beamline Technician 30/06/01CDD/MR Mechanical Technician, 12-month contract,

with possible extension to 18 months 11/07/012551 Detector Technician 24/08/01

ENGINEER 2228 Beamline Operation Managers on ID21 and ID22 15/06/016129 Mechanical Engineer 24/08/01CDD/CH Engineer in Electronics, a 12 month-contract 15/06/01

OFFICE EMPLOY. 7538 Accounts/Payroll Assistant 29/06/01

If you are interested,please send us a fax(+33 (0) 4 76 88 24 60)or an e-mail(recruitm@esrf.fr)with your address,and we will provideyou with anapplication form. Youcan also print out anapplication form onthe World Wide Webhttp://www.esrf.fr

All applications received after the deadline will be considered forthe selected vacancy if not filled or for future similar positions

The ESRF Newsletter is published by the European Synchrotron Radiation FacilityBP 220, F38043 Grenoble cedex

Editor: Gary Admans Tél (33) 4 76 88 27 57Dépôt légal : 2ème trimestre 2001. Printed in France by Imprimerie du Pont de Claix. Layout by Pixel Project.