SEPTEMBER 1998 EUROPEAN SYNCHROTRON RADIATION … · «Medical imaging with synchrotron...

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 YSEPTEMBER 1998 N° 31

ISS

N 1

011-

9310

Multiple views of the nucleosome core particle crystal structure at 2.0 Åresolution. Over 800 water molecules (cyan) and divalent ions (magenta) havebeen added to the more than 12,000 C,N,O atoms of the proteins and DNA that

make up this fundamental repeating unit found in chromosomes (see article on page 9).

ESRF - SEPTEMBER 1998

2

NEWSLETTER

IN BRIEF

EXPERIMENTS

REPORTS

INDUSTRIAL

FACILITY

MACHINE

DEVELOPMENT

EVENTS

29th Council Meeting, PAGE 3, K. Witte.HERCULES 1998, PAGE 4.Life Sciences lead new requests for beam time, PAGE 5, R. Mason.Next Users’s Meeting, PAGE 5.Workshop for Industry «Chemistry and Surfaces» held on 4 June 1998, PAGE 6, J. Doucet.Soft Matter under Flow as Probed by Small-Angle Scattering - SASFLOW '98, PAGE 7, O. Diat.

Effect of Shear on Diblock Copolymer Gels, PAGE 8, I.W. Hamley, J.A. Pople, J.P.A. Fairclough, A.J. Ryan,N.J. Terrill, C. Booth, G.-E. Yu and O. Diat.The Atomic Structure of Chromatin’s Nucleosome Core Particle at 2 Å Resolution, PAGE 9, K. Luger,A. Mäder, D. F. Sargent and T.J. Richmond.Observation of Speckle Patterns by Coherent X-ray Scattering from Thin Polymer Films, PAGE 12, P. Müller-Buschbaum, T. Thurn-Albrecht, G. Meier, M. Stamm, E.W. Fischer, B. Stark, B. Stühn, D.L. Abernathy, G. Grübel.First Experiments on Troïka II, PAGE 14, D.M. Smilgies, N. Boudet and G. Grübel.Hierarchical Porosity in Quasicrystals Investigated by Coherent X-ray Imaging, PAGE 16, L. Mancini,J. Gastaldi, E. Reinier, P. Cloetens, W. Ludwig, C. Janot, J. Härtwig, M. Schlenker and J. Baruchel.ID26 : X-ray Spectroscopy on Ultra-Dilute Systems, PAGE 18, Ch. Gauthier, V.A. Solé, J. Goulon, E. Moguilineand R. Signorato.Edge Separation Using DAFS, PAGE 20, B. Ravel, C.E. Bouldin, H. Renevier, J.L. Hodeau and J.F. Berar.Coupling Electrochemistry with X-ray Absorption Spectroscopy to Study Polymers, PAGE 22, P.L. Vidal,M. Billon, B. Divisia-Blohorn, G. Bidan, J.M. Kern, J.P. Sauvage and J.L. Hazemann.Modification of the Island Morphology during Metal Epitaxial Growth, due to a Surfactant, PAGE 24,J. Alvarez, E. Lundgren, X. Torrelles and S. Ferrer.Oxidation of NiAl(100) Studied with Surface Sensitive X-ray Diffraction, PAGE 25, A. Stierle, V. Formoso,F. Comin, G. Schmitz and R. Franchy.Synchrotron Radiation Grazing Incidence X-ray Diffraction: A New Tool for Structural Investigations of Ion-Implanted Glasses, PAGE 27, F. D’Acapito and F. Zontone.

ID27: An Industrial TXRF Facility, PAGE 28, F. Comin, M. Navizet, P. Mangiagalli and G. Apostolo.

A Third Acceleration Unit for the ESRF Storage Ring, PAGE 30, C. David, J. Jacob, A. Panzarella, J.P. Perrineand J.L. Revol.

HERCULES 1998.Next Users’s Meeting.

Photography by:

C. Argoud,B. Denis (Art Photos).

PRIZE GIVEN FOR

SR COMPUTED TOMOGRAPHY

PERFORMED ON ANIMALS

At the 6th Conference "Biophysics & Synchrotron Radiation",from 4 to 8 August 1998 at Argonne in the USA, a prize wasgiven for the poster presented by G. Le Duc, A.-M. Charvet,

H. Elleaume and F. Esteve, from the Jeune Equipe RSRM(Univ. Joseph Fourier in Grenoble) and the ESRF Medical

Beamline. This Jeune Equipe were the first to perform braintomography on an anesthetized animal (a rat).

G. Le Duc in frontof the experimental

set-up.

G. Le Duc in frontof the experimental

set-up.

NEWSLETTER IN BRIEF


3

29TH COUNCIL MEETING

At its 29th meeting on 9 and 10June 1998 in Grenoble, the ESRFCouncil took several decisions whichwill have an impact on the medium-term future of the institute.

ISRAEL AS NEW SCIENTIFICASSOCIATE

The Council agreed to, andauthorized its Chairman to sign for theESRF on its behalf, an «Arrangementbetween the Government of the State ofIsrael and the European SynchrotronRadiation Facility concerning the long-term scientific use of synchrotronradiation for non-proprietary research»very similar to the one concluded last yearwith Portugal (cf. ESRF Newslettern° 29, p. 3).

Since the Foundation Phase of theESRF, contacts have been made with theIsrael Academy of Sciences andHumanities concerning an involvementof Israel in the ESRF. In February 1987,the then Israeli Minister of Science andDevelopment, Gideon Patt, considered ina letter to the Chairman of the provisionalESRF Council, Pierre Agrain, anapplication for Israel to become a fullMember of the ESRF. In the outcome,however, Israel was not among thefounding Members of the ESRF.Nevertheless, regular visits by Israeliscientists over the last ten years havemarked the continuous interest of theIsraeli scientific community. However, itwas obvious that its size would not matchthe 4% threshold required for full ESRFMembership.

Discussions with the Israeli partnersintensified after the Council had clarifiedthe conditions for long-term arrange-ments in accordance with Article 8 of theESRF Convention. The arrangementagreed by the Council is due to be signedin November in Jerusalem and will makeIsrael the second Scientific Associate ofthe ESRF. Thus scientists from Israel willhave the same rights of access to theESRF beamlines as those from the twelveContracting Party countries. For the timebeing, the level of the Israeli contributionto the ESRF operating costs has been setat 1% of the contributions of theMembers.

MEMBERSHIP BELOW 4%?During the discussion with potential

Scientific Associates, the question wasraised whether a reduction of thethreshold for full Membership to less than4% could be considered in the long term.Following an exchange of views on thisissue, the Council noted that almost alldelegations wished to maintain the 4%threshold for Membership as provided forby the ESRF Convention and Statutes,and to encourage the participation atlower rates in the form of consortia.

RE-ADJUSTMENT OFCONTRIBUTION RATES

The ESRF Convention determinesthe contribution rates of the differentContracting Parties. In the event of alasting and significant imbalance betweenscientific use and contributions, it alsoprovides for the option of a re-adjustmentof these contribution rates.

With a view to the end of the ESRFconstruction period and some experienceon the use made of the facility since 1994,the Council adopted «Guidelines for thereadjustment of contribution rates» whichin particular • clarify the quantitative assessment ofthe scientific use (taking into account theactivity of Collaborating ResearchGroups and of the GRAAL experimentusing γ-rays produced by Compton back-scattering of laser light),• give a precise meaning to the terms«lasting» and «significant» used in theConvention and• indicate the procedure to be followedfor the implementation of a re-adjustment.

BEAM TIME ALLOCATIONFOR MACROMOLECULAR

CRYSTALLOGRAPHY

In his report to the Council, theDirector General mentioned the stronglyincreasing numbers of proposals formacromolecular crystallographic studies(some 240 for the second half of 1998)which put a heavy burden on thereview committee for life sciences. Withthe extension of the experimentalpossibilities at the Quadriga beamline(ID14A+B) a further increase can beexpected. Since relatively few groups

account for a considerable proportion ofthe projects that are awarded beam time,Management suggested testing anothersystem: a number of groups withinEurope will be identified to whom blocksof beam time will be awarded withoutprior evaluation of the individualexperiment proposals. The a posterioriassessed performance record per groupwill then be essential for whether in thefuture the block allocation of the groupwill be continued, changed or removedcompletely.

The Council noted that such a system(together with several accompanyingmeasures) will be implemented on a trialbasis for the next proposal round(concerning the first half of 1999).

BUILDING CONSTRUCTION

There is a lack of laboratories andoffices at the ESRF due to:• the increase of personnel (staff,externally funded collaborators, visitors),• the closure of buildings provisionallyacquired for the construction period and,• the increasing needs of CollaboratingResearch Groups, etc.

There is still some potential to furtherextend the Experimental Hall, both at theouter and inner circumferences, or to addfurther satellite buildings around the Hall.

The Council took note of the plannedconstruction, starting this year, of furtherlaboratories and offices to satisfy the mosturgent needs.

BUDGET FOR 1999The decision on the budgets is usually

taken at the autumn meeting of theCouncil. With a view to the tight nationalbudget situation in some of theContracting Parties countries, the Councilso far adopted a ceiling figure of 400million French Francs for newcontributions from Members to thebudget of 1999 (which is 1.0% more thanthe amount of Members’ contributions tothe current budget of 1998).

REDUCTION OF WORKING TIME

The Council• noted that the French legislation onworking time will change but that thecorresponding application texts were notyet available, and

NEWSLETTER IN BRIEF


4

• asked Management to prepare, as soonas possible, information about both theregulatory and the financial consequencesof these changes for the ESRF.

This item was also raised by therepresentatives of the ESRF in theirmeeting with the Council.

DIRECTOROF ADMINISTRATION

Prior to its meeting, the Councildecided by written procedure to extendthe appointment of the present Directorof Administration, Dr. W.E.A. Davies, upto 31 January 2002.

GUIDELINES FOR SELECTING ANDAPPOINTING ESRF DIRECTORS

The Council had a second reading ofnew guidelines for selecting andappointing ESRF Directors. Muchconsideration was given to this documentdue to the fact that during the period of 15months between January 2001 andMarch 2002 the contracts of all fiveESRF Directors will end and theselection and appointment proceduresshould begin well beforehand. TheCouncil decided to resume the discussionat its next meeting based on a new draftincorporating the amendments proposed.

SCIENCE ADVISORYCOMMITTEE

The membership of the presentScience Advisory Committee will come toan end on 31 December 1998. Due to theappointment procedure (eleven scientistsdirectly nominated by the Members of theESRF, a further ten appointed by theCouncil) the renewal of membership hasto be prepared in several phases.

The Council took note of the schedulewhich shall lead to the appointment of anew Science Advisory Committee at thenext meeting of the Council.

K. Witte

HERCULES 1998The seventh session of the

HERCULES course (Higher EuropeanResearch Course for Users of LargeExperimental Systems) took place atthe Maison des Magistères, CNRSGrenoble, from 22 February to 3 April1998. The 79 participants (mostlyEuropean or registered in EuropeanUniversities) were divided in twosessions :

• session A: Neutron andsynchrotron radiation for physics andchemistry of condensed matter with 44full-time participants and 11 part-timeparticipants.

• session B: Neutron andsynchrotron radiation for biomolecularstructure and dynamics with 24 full-time participants. A special effort hasbeen made towards the community ofbiologists by opening the course to non-European countries (USA, Brazil...)

The course included lectures,practicals and tutorials as in previousyears. As last year, session A wasparticularly centered on recentdevelopments of neutron and X-rayspectroscopy. New lectures devoted tostructural molecular biology have beenincluded in the program of session B:«Medical imaging with synchrotronradiation», «X-ray microscopy»,«Nuclear spin contrast variationstudies». The publication of the volumeIV of the HERCULES series entitled«Structure and dynamics of

biomolecules» by Oxford UniversityPress is now in progress. For bothsessions, a specific seminar onindustrial applications was introducedand appreciated by the participants.

In Grenoble most of the practicalswere performed on ESRF beamlines(including the French, Italian andSwiss-Norwegian CRG beamlines) andat the ILL. The collaboration of EMBL,IBS as well as CNRS and CEA-Grenoble was also greatly appreciated.

Participants from the two sessionscarried out practicals at LURE (Orsay)and the Laboratoire Léon Brillouin(Saclay).

The poster session at the Maisondes Magistères (55 posters displayed)was one of the highlights of the courseand allowed fruitful exchanges betweenparticipants and Grenoble scientists.

HERCULES 99 will follow nextyear with the same two parallel sessionsfrom 21 February to 1 April 1999.

Session A: «Neutron and synchrotron radiationfor physics and chemistry ofcondensed matter»

Session B: «Neutron and synchrotron radiationfor biomolecular structure anddynamics»

Information:Secrétariat HERCULESCNRS - Maison des MagistèresBP 16638042 Grenoble Cedex 9Tel: +33 (0) 4 76 88 79 86Fax: +33 (0) 4 76 88 79 81e-mail: [email protected]://www.polycnrs-gre.fr/hercules

Deadline for application: 16 October 1998

HERCULES 1999HIGHER EUROPEAN RESEARCH COURSE FOR

USERS OF LARGE EXPERIMENTAL SYSTEMS

Grenoble, 21 February - 1 April 1999

… 29TH COUNCIL MEETING

NEWSLETTER IN BRIEF


5

USERS’ MEETING

11-12 February 1999

The next Users' Meeting will start onThursday 11 February 1999, with theusual «plenary format», plus posters,plus exhibitors at the Atriaconference centre. The change of datefrom pre- to post-Christmas is for atleast two good reasons. Firstly it is inresponse to requests from users tomeet ESRF staff in the «run up» to asubmission date in order to benefitfrom their informal views of projectproposals. Secondly it avoids a period

when the User Office staff, who helptremendously with the administrationof these meetings, are very busyprocessing the recommendations ofthe Review Committees. For manyusers there is also the matter of a fewhours on the slopes which wasprecluded in the pre-Christmas slot!The meeting will continue withworkshops on the Friday, andpossibly Saturday, based on the ESRFsite, plus adjacent venues if we have

more than three viable suggestions.All beamline scientists have beenemailed to ask for their suggestions,and just a few have responded.M. Cooper and other members of theUsers' Committee need to knowNOW if you want to run a workshop.Please contact [email protected] or ring M. Cooper in the UK(+44 1203 523379) to discuss theoptions and put the dates in yourdiary!

LIFE SCIENCES LEAD NEW REQUESTS FOR BEAM TIME

Users requesting beam timebetween August 1998 and January1999 sent in a record number of 799new proposals for the 1 Marchapplication deadline. This was a 37%increase over the previous period, duemainly to the very large increase inLife Sciences applications, up 242from 128, an increase of almost 90%.Requests from this community ofusers reflect a recognition of theexciting results which areincreasingly being achieved with theuse of synchrotron radiation (see inthis Newsletter the article on thenucleosome, on page 9), and theprogressive opening of the Quadrigabeamline, ID14, with its fourindependent experimental stations.This beamline is one of the 28 ESRFand 7 CRG beamlines which areexpected to be scheduling userexperiments during the next period.

Following the meetings of theReview Committees at ESRF on 27and 28 April 1998, 359 proposalswere selected and allocated beamtime totalling 4341 shifts. Of these,seven new long term programswere accepted for a period of twoyears. Details of the requests andallocations, per committee, aresummarized in Table 1.

Numbers of shifts of beam timerequested and allocated perscheduling period since the beginningof user operation in September 1994are shown in Figure 1. It should be

noted that the second schedulingperiod each year to date has beenslightly shorter than the first, so thatin general fewer shifts have been

allocated and scheduled during thesecond half of each year.

R. Mason

Chemistry Hard Hard condensed Life Methods and Soft Surfaces andcondensed matter: sciences Instrumentation condensed Interfaces

matter: Elect. Structures matter& Magn. properties

Review committees

Num

ber

of p

ropo

sals

260

240

220

200

180

160

140

120

100

80

60

40

20

0

49 5866

108

25 32 21

Proposals submitted

Proposals accepted

88

118

193

242

39

71

48

1994 1995 1995 1996 1996 1997 1997 1998 1998Sept Jan Aug Jan July Jan July Jan Aug

to Dec to July to Dec to June to Dec to June to Dec to July to Jan 1999

Scheduling periods

Num

ber

of s

hift

s

12000

10000

8000

6000

4000

2000

0

2252

4236

66347965 7660

8384

99038469

10614

7611965 2034

3065 3046 3569 3759

49144341

Shifts requested: 66 117

Shifts allocated: 27 454

Fig. 1: Total number of shifts of beam time requested andallocated, per scheduling period, 1994 to 1998/II.

Table 1: Number of proposals per review committee.

NEWSLETTER IN BRIEF


6

WORKSHOP FOR INDUSTRY

«CHEMISTRY AND SURFACES»HELD ON 4 JUNE 1998

The ESRF welcomed abouttwenty industrialists on 4 June 1998for the workshop dedicated tochemistry and surfaces.

Participants came from differentfields of activities such as chemicals,pulps, metallurgy, cosmetics, polymers,cements, detergents and electronics(l'Oréal, Philips Analytical, Sagem,Rhône Poulenc, Lafarge, Ocas NV,Dow Corning, Rhodia, Arjo Wiggins,BASF, Unilever, Italcementi).

After an introduction by theDirector General Y. Petroff, A. Fitchgave a talk entitled «Identification andcharacterization of crystallinepowders by x-ray diffraction». Thedifferent uses of powder diffractionwere presented as the advantages ofsynchrotron radiation for thistechnique. Some examples weregiven: characterization of ancientEgyptian cosmetics, research on Li-ion plastic electrolyte battery, thestructure determination from powders(ex: fluorescein diacetate). Thistechnique is of special interest forindustrialists working in polymers andmetallurgy.

S. Pascarelli spoke about theperspectives for industrial researchusing x-ray absorption fine structurespectroscopy at the ESRF. Shedescribed the various EXAFStechniques dedicated to bulk

materials (transmission), diluted systems(fluorescence), surfaces (reflection),and real-time analysis (dispersiveEXAFS). She gave research examplesof industrial applications: Er-doped Sifor opto-electronics, rare earth-dopedglasses for optical fibers, structure ofreaction intermediates in solution(alkene oligomerisation), monitoringof chemical oscillations during COoxidation on Pt supported catalysts.She underlined that the ESRF is muchmore efficient than the secondgeneration synchrotron sources for thestudy of ultra-dilute systems (down toa few ppm) and for the study of fastprocesses (down to the ms resolution).

S. Ferrer gave an overview ofsurface, interface and thin layercharacterization using synchrotronradiation. Three techniques werepresented: surface x-ray diffraction,circular magnetic dichroism and totalreflection fluorescence analysis. Theywere illustrated by the microscopy ofmagnetic domains in ultra-thin films,surface atomic structure of KPDcrystals (KH2 PO4) in aqueoussolution, and trace element analysison Si wafers for which a beamlinededicated to industry is underconstruction at the ESRF (see thisNewsletter, page 28).

J. Baruchel presented x-rayimaging at the micrometer scale and

its place at the ESRF. Differenttechniques were described with theirapplications: microtomography withimages of human vertebra samplesand PVC foam sample, phase contrastimaging of a polymer sphere with twolayers and of a AlSiC compositematerial, as well as the chemicalmapping of fly-ash particles byfluorescence analysis. Severalchemical companies already benefitfrom these new imaging techniques atthe ESRF.

Finally, J. Doucet, ESRF Co-ordinator for Industry, explained thereasons for industrial companies touse synchrotron beamlines andpresented the different types ofcollaboration: isolated short charac-terization, long-term collaboration,pure beam time purchase, full service(data collection + analysis), PhD orpost-doc funding, long-term visitors,and even beamline construction andoperation.

The afternoon was devoted to thevisit of several experimental stations:ID2 (small-angle scattering), ID11(diffraction), ID13 (microdiffractionand microfluorescence), ID16 (high-resolution powder diffraction) andID22 (micro-analysis and phasecontrast imaging).

J. Doucet

VACANCIES AT THE ESRF ON 7 AUGUST 1998

Ref Subject

POST-DOC PDID10B-1 Troïka II beamline (ID10B)PDBM29/ID26 EXAFS (two positions)PDID22 Phase contrast microtomography (ID22)EUPD214 Time-resolved protein crystallographyEUPD215 Proton pumping in membranes

STUDENTS CFR239 X-ray gyrotropy and related optical phenomenaCFR243 High resolution residual strain mappingCFR244 Soluble, electrically conducting polymersCFR245 Micro-structural characterization of cold and hot forming processesCFR246 Phase transitions in colloidal systems

ENGINEER CDD/DP Software Engineer - 6-month contract2209 Mechanical Engineer

If you are interested,please send us a fax(+33 (0) 4 76 88 24 60)or an e-mail([email protected])with your address,and we will provideyou with anapplication form.You can also printout an applicationform on theWorld Wide Webhttp://www.esrf.fr

SOFT MATTER UNDER FLOW AS PROBED

BY SMALL-ANGLE SCATTERING - SASFLOW '98

In the last five years a majorchallenge among the soft-mattercommunity is to elucidate therelationship between rheologicalproperties of complex fluids and theirmesoscopic structure as revealedby small-angle scattering (SAS)experiments. On 28-29 May 1998, acommon ESRF/ILL workshop washeld at the ESRF, organized byO. Diat (ESRF), P. Lindner andJ. Zipfel (ILL).

The main objectives of thiscommon workshop were:

• to advance the new opportunitiesof experiments under shear on largeinstruments and especially at theESRF and ILL.

• to provide a summary of recentexperiments or projects using SAStechniques.

• to explore users' interest in orderto plan long-term projects and tooptimize technical developments andanalysis.

This workshop was divided intosix sub-sessions, on surfactants,polymer solutions, block-copolymers,polymer blends, suspensions undershear and a session on instrumentationplus a visit to the experimentalfacilities on the site (ILL and ESRF).A poster session was also organized inorder to complement the overview ofsome recent experiments in this field.

As D. Roux recalled in the firstcontribution, scattering techniques(light, x-ray and neutron) give themost interesting structural informationto understand the phenomenologicaland mechanical results obtained byrheology. Several examples of sheartransitions i.e. transitions to metastablephases or orientation not observable atequilibrium were presented (shearstripping of charged vesicles -Gradzielsky et al; «onion phases» -Roux et al; shear ordering ofsurfactant system - Penfold et al,Richtering et al; shear ordering orshear melting in block-copolymersystem either in melt or solution -K. Mortensen et al, I. Hamley et al,Porte et al; shear thickeningin polymer-clays mixed system,

S. Cocard et al). On the other hand,phase transitions under shear or shiftof the transition line have also beenobserved as in vesicle to micelletransition in surfactant mixtures(E. Mendes et al), in shear-inducedaggregation in polymer solution(I. Morfin et al, T. Hashimoto), inshear effects in liquid crystal polymers(L. Noirez et al), or in critical polymermixtures under shear flow (C. Hanet al), on weakly flocculateddispersions (J. Vermand et al) and ondiblock copolymer systems with re-entrant phenomenon (H. Leist et al).Although we can always present theseshear effects as enhancement ordecrease of fluctuations along theflow, it is still very difficult to predictthem. All these examples are thesignature of a large and rich variety offlow behavior completely open totheory [1].

All these groups presented someresults using their specific techniques(light or neutron or x-rays). The aimof this common workshop was to findout ways of optimizing the use of thenew third generation synchrotronsources (high spatial resolution andtime-resolved experiments) as well ashigh-flux neutron beams (contrastvariation) for pertinent experiments.Kinetics studies under continuous oroscillatory shear to describe the out-of-equilibrium transitions have beenstressed (Porte et al, O. Pelletier et al,Mortensen et al C. Han et al).Moreover, the development of shear

cells with an access to the vorticity-gradient plane seems very importantdue to the effect of phase separation inthe Couette cell gap (stress effectinstead of shear effect). This alsorequires the on-line set-up of arheometer to correlate directly stressmeasurement and scatteringtechniques.

The last point but not the leastwhich has been emphasized, is thedevelopment of elongational flowsystems (spinning system, stagnationpoint and stretching devices) whichare very relevant in areas likematerials processing, development ofcomposites as well as in investi-gations of polymer dynamics incomplex geometries (M. Cloitre et al).

Some new contacts have alreadybeen established and, in common withthe ILL, we shall be able to judge theeffectiveness of our development onlyin a few years’ time when we look atwhat will be achieved. A commonWEB page will be implemented inorder to inform other large instrumentbeamlines and users about newinstrumental development at theESRF and the ILL.

P. Lindner, J. Zipfel and myselfwould like to thank all the people whohave made this workshop successful.

O. Diat

[1] T. McLeish: Theoretical Challenges in theDynamics of Complex Fluids, edited byT. McLeish, Kluwer Academic Publishers,Netherlands (1997).

NEWSLETTER IN BRIEF


7

EXPERIMENTS REPORTS


8

Of particular interest are blockcopolymers containing hydrophilicpoly(oxyethylene) as well as ahydrophobic block such aspoly(oxybutylene) or poly(oxypropylene)which are manufactured on a large scalebecause they behave as surfactants inaqueous solutions. Despite their industrialrelevance, surprisingly little is knownabout phase formation in these systems,nor on the effect of shear on orderedstructures which is relevant to processing.Small-angle x-ray scattering (SAXS)has been performed on the ID2 SAXSstation to elucidate the effect of shearon the orientation of cubic micellarphases formed by a poly(oxyethylene)-poly(oxybutylene) diblock copolymer inaqueous solution [1,2]. Macroscopicorientation of samples was achieved byshearing in the Couette cell available onID2 using steady shear. By translation ofthe Couette cell it was possible to accesstwo orthogonal planes with either theshear direction (v) or shear gradientdirection (∇ v) horizontal and the vorticitydirection (e = ∇ v x v) vertical.Experiments were conducted on thediblock E40B10 (here E = oxyethylene,B = oxybutylene and the subscriptsdenote the number of repeats) in 0.2 MK2SO4. A body-centred cubic phaseobserved for gels with concentrationsgreater than 30 wt% copolymer wasfound to orient into a polydomainstructure, with the close-packed {110}planes both parallel and perpendicular tothe shear plane [1-3]. This is indicated bythe SAXS patterns in Figure 1. For gelswith 30 wt% copolymer or less, an fccphase was observed, and this was alsoobserved on heating the moreconcentrated gels that formed a bcc phaseat room temperature [1-3]. The hard gelfcc phase could be oriented to form ahighly twinned structure (as shown by the

SAXS patterns in Figure 2), witha significant deviation from theABCABC… stacking sequence of theideal structure due to random stackingsequences resulting from the slip of{111} hexagonal close-packed planes.For the lower concentration solutions, atransition from hard to soft fcc gels atincreasing temperatures was found to becharacterized by a change in thesusceptibility of the sample tomacroscopic shear orientation, asprobed using SAXS [2]. The hard gelcould be oriented by shear into atwinned fcc structure, whereas the softgel comprised an fcc phase with a small

grain size, which could not be shearedto form a macroscopically orienteddomain. Shear only homogenized thesample, producing a powder SAXSpattern [2].

REFERENCES

[1] J.A. Pople, I.W. Hamley, J.P.A. Fairclough,A.J. Ryan, G.-E. Yu and C. Booth, Macro-molecules 30, 5721 (1997). [2] I.W. Hamley, J.A. Pople, J.P.A. Fairclough,N.J. Terrill, A.J. Ryan, C. Booth, G.-E. Yu,O. Diat, K. Almdal, K. Mortensen andM. Vigild, J. Chem. Phys. 108, 6929 (1998).[3] I.W. Hamley, J.A. Pople and O. Diat,Colloid Polym. Sci., 276, 446 (1998).

(a) (qv, qe) plane, during shear at •γ = 8 s-1,(b) q∇ , qe) plane during shear at •γ = 8 s-1. The vorticity direction is vertical.

Fig. 2: SAXSpatterns for a 38wt% solution of

E40B10 in the fccphase at 52 °C.

EFFECT OF SHEAR ON DIBLOCK COPOLYMER GELSI.W. HAMLEY1, J.A. POPLE1, J.P.A. FAIRCLOUGH2, A.J. RYAN2, N.J. TERRILL2, C. BOOTH3, G.-E. YU3 AND O. DIAT4

1 SCHOOL OF CHEMISTRY, UNIVERSITY OF LEEDS (UK)2 DEPARTMENT OF CHEMISTRY, UNIVERSITY OF SHEFFIELD (UK)3 DEPARTMENT OF CHEMISTRY, UNIVERSITY OF MANCHESTER (UK)4 ESRF, EXPERIMENTS DIVISION

Low molecular weight block copolymers can behave as amphiphiles in solution, forming lyotropic

liquid crystalline phases such as cubic micellar phases. The formation of such structures results in stiff

transparent gels which have a number of technological applications for example in drug delivery.

Fig. 1: SAXSpatterns for a 38wt% solution of

E40B10 in the bccphase at 25 °C.

(a) (qv, qe) plane, during shear at •γ = 0.8 s-1,(b) q∇ , qe) plane during shear at •γ = 0.8 s-1. The vorticity direction is vertical.

EXPERIMENTS REPORTS


9

A

THE ATOMIC STRUCTURE OF CHROMATIN’S NUCLEOSOME

CORE PARTICLE AT 2 Å RESOLUTIONK. LUGER, A. MÄDER, D. F. SARGENT AND T.J. RICHMOND

INSTITUT FÜR MOLEKULARBIOLOGIE UND BIOPHYSIK, ETH ZÜRICH (SWITZERLAND)

Protein crystallography is currently turning out novel atomic structures of biological macromolecules at

the rate of 400-500 per year [1]. These structures range in size from proteins comprising under 100

amino acids to virus particles of several million molecular weight. An area of intense interest currently

concerns macromolecular complexes containing multiple components that must periodically assemble

and disassemble in order to carry out their biological function.

classic example in all highercells is the nucleosome, which is thefundamental repeating unit of DNAorganization in chromosomes andaccounts for the two most fundamentallevels of chromatin structure. Not onlydo nucleosomes efficiently packageDNA, they are also intimately involvedin the gene expression mechanisms thatallow only selected regions of the vaststore of genomic information to be readout as a consequence of signalingprocesses. These two functions requireabout 25 million nucleosomes in eachhuman cell nucleus.In 1984, the structure of the nucleosomecore particle (NCP), the larger part ofthe nucleosome, was published at 7 Åresolution [2]. The x-ray data wascollected before synchrotron radiationwas generally available for proteincrystallography by using a singledetector diffractometer and a rotatinganode source. The spatial resolution ofthis first structure was limited by thematerial itself as it was prepared fromwhole nuclear chromatin and wastherefore heterogeneous in compo-sition. Eventually, it became technicallyfeasible to assemble NCP in thetest tube from homogeneouscomponents made individually inbacterial cells [3]. After a long periodof experimentation for developinghomogeneous preparations of NCP,crystals were obtained that diffracted tohigh resolution. Nevertheless, the Braggintensities from these crystals areextremely weak. Fortunately, the ESRFhad opened for business, and we wereinvited to collect full data sets at ID13(C. Riekel, scientist in charge) just afterits construction. As a result, thestructure of the nucleosome core

particle was published in 1997 at 2.8 Åresolution (Figure 1) [4]. At 206 kDa,the NCP is the largest and mostuniversal protein/DNA complex solvedin atomic detail.Our best nucleosome core particlecrystals show Bragg intensities to 1.9 Åand have measurable data to 2.0-2.1 Åspacings (Figure 2). Most recently, 27of these crystals were used to collect adata set to 2.0 Å resolution on ID9(M. Wulff, scientist in charge). The highbrightness of both ID9 and ID13undulator beamlines enabled ourcrystallographic studies of the NCP athigh resolution.

AT 2.8 Å RESOLUTION(ID13)

The diffraction data was measured atID13 (beam size 100 µm x 150 µm)over a period of 2 years (~ 75 shifts) andused to produce the first high resolutionimage of the NCP. In order to solve thephase problem by the multipleisomorphous replacement method(MIR), recombinant DNA technologywas employed to create specificattachment sites for heavy atomderivatives in the protein portion of thecomplex [3]. A large number of heavyatom derivatives were screened to a

Fig. 1: Crystal structure of the nucleosome core particle at 2.8 Å resolution.The DNA double helix (146 base pairs in two chains: turquoise and brown) is

wound around the protein histone octamer (two copies each of H2A:yellow, H2B: red, H3: blue, and H4: green) in 1.65 left-handed superhelical

turns. This is the form of DNA which predominates in higher living cells.The left view is down the superhelix axis. The right view is orthogonal to the

superhelix and overall pseudo-twofold axis.

EXPERIMENTS REPORTS


10

resolution of 4.5 Å in successive beamtimes, solved by difference Pattersonand difference Fourier techniques, andthen ranked by phasing power. Finally,one native and four derivative data setswere collected at – 170 °C to aresolution of 2.8 Å. Crystal cryo-cooling procedures were also tested anddeveloped during these sessions at theESRF. Because each diffraction imagerequired 1-3 min of crystal exposure tothe x-ray beam, each final data setrequired 3-4 crystals to keep the overalleffect of radiation damage on meanintensity below 20%.

OVERVIEW OF THESTRUCTURE

The NCP contains pairs of the four corehistone protein molecules named H2A,H2B, H3, and H4, and a roughly equalmass of DNA in 147 nucleotide pairs(we used 146 bp). Compared to thenucleosome, the NCP is missing onlythe ‘linker histone’ H1 and the shortstretches of DNA that connect thenucleosome cores to each other inchromatin. The core histones arearranged in an octameric unit aroundwhich the DNA is wrapped in 1.65 turnsof a left-handed superhelix (Figure 1)[4]. This arrangement necessitates asubstantial deformation of the DNA,bending the 22 Å diameter double helixto a mean radius of 42 Å in thenucleosomal superhelix. The histone protein chains are dividedinto three types of structures: 1) rigid,folded alpha-helical domains named thehistone-fold, 2) histone-fold extensionswhich interact with each other and thehistone-folds, and 3) flexible ‘histonetails’. The histone-fold domains are

structurally highly conserved betweenthe four types of core histones and havealso been discovered in an increasingnumber of other molecules involved inthe regulation of gene read-out ortranscription. They form crescent-shaped heterodimers which haveextensive interaction interfaces in thepairings H3 with H4 and H2A withH2B. The histone-fold domains areresponsible for organizing 121 basepairs (bp) of DNA in the superhelix, notthe entire 147 bp. It is the responsibilityof the extensions just prior to the H3histone-folds to bind the first and last 13bp of DNA. The flexible tails of thehistones reach out between and aroundthe gyres of the DNA superhelix tocontact neighboring particles. Aboutone-third of these flexible histone tailscan be observed in the electron densitymap, the remainder are too disorderedto be interpreted. The implication fromthe structure is that these flexibleregions are meant to make inter-nucleosomal interactions, perhapsfacilitating the formation of nucleosomehigher-order structures (HOS).

PROTEIN-DNAINTERACTIONS

There are 14 regions of contact betweenthe histone proteins and DNA: three byeach of the four histone-fold dimers andtwo by histone-fold extensions. Thisconstruction allows the DNA moleculein a single nucleosome core to comeloose over one-half of the superhelixwhile the histones maintain their grip onthe other half, permitting the geneticinformation stored in the DNA to beread out without complete dissociationof the DNA from the histone octamer.The nucleosome core was previouslythought to be held together simply byelectrostatic attraction: the negativelycharged DNA molecule wound as yarnaround a positively charged histonespool. Although this type of interactiondoes occur, equally many interactionsof other kinds, such as hydrogen bondsand hydrophobic interactions, are alsoimportant.

DNADEFORMATION

The path of the DNA around the histoneoctamer deviates from that of an idealsuperhelix, displaying strong bends insome regions, while being nearlystraight in others. This path isdetermined predominantly by thehistone/DNA contacts and is probablylargely independent of the DNAsequence of nucleotides. The closespatial proximity of the two turns of theDNA superhelix with a pitch of 24 Å,and the periodic variation of doublehelix parameters with a mean of 10.3 bpper turn, result in an alignment ofmajor and minor grooves from onesuperhelical gyre to the next (Figure 1).

Fig. 2: Diffraction fromnucleosome core particle crystalsto 2.0 Å spacings.The view is nearly orthogonal tothe a* axis of the P212121 unitcell (108 Å x 186 Å x 111 Å). TheX-ray wavelength is 0.842 Å andthe crystal-to-detector distance is320 mm.The image has been medianfiltered to remove the high levelsof diffuse scatter.

Fig. 3: Two consecutive bases ofnucleosomal DNA (white) and

solvent molecules (yellow).The atomic positions are based on a2Fo-Fc electron density map (DNA:

magenta, solvent: red) calculatedbetween 47 and 2 Å resolution

(contoured at 1.1 sigma). Distancesbetween DNA and solvent molecules

greater than 2.2 Å and less than4.4 Å are shown (cyan).

EXPERIMENTS REPORTS


11

The resulting narrow channels formedby the aligned minor grooves serve asthe exit points for four of the eightbasic histone tails, whereas the largepores formed by the aligned majorgrooves are, in principle, free to makebase-specific contacts with otherproteins. The Debye-Waller B-factorsshow that the mobility of the DNAbackbone varies greatly, having lowvalues when it is bound to the histoneoctamer and high values when it isfacing away from it.

AT 2.0 Å RESOLUTION(ID9)

X-ray data to 2.0 Å resolution wascollected at ID9 (beam size150 µm x 300 µm) using a mono-chromatic x-ray beam and a 30 cm MarImage Plate detector (Figure 2).Crystals of an average size of 200 µm x200 µm x 500 µm diffracted veryweakly, requiring exposure times of90 s for each 0.4° rotation picture torecord the highest resolution data. Theprevious observations of radiationdamage incurred by these crystalsdespite cryo-cooling indicated the needto collect a large number of partial datasets. A total of 27 partial data setswere collected, each one covering arotation range of 8.0° and overlappingthe neighboring sets by 4.0° toaccommodate crystal misalignment.Two crystal orientations with respect tothe rotation axis were used to ensurecoverage of the blind zone. Thesubstantial exposure times necessary tocollect the high resolution data requiredcollection of an additional 3.0 Åresolution data set using exposure timesof 10 sec to recover ‘overflows’.Ultimately, the program «denzo» wasused to measure a total of 4,228,118reflections from 570 exposures and tomerge them to obtain our 2.0 Å data set.

The final completeness of the data is98.3% (84.5% in the last shell: 2.04 -1.97 Å) with an overall R-factor of6.8% (reaching 30% in the 2.22 - 2.13 Åshell). The refinement of the structure atthis resolution is currently in progress.

WATER AND IONSThe 2.0 Å diffraction data have allowedus to locate a large number of well-ordered water molecules and ions. Weexpect to gain valuable generalinformation on the role of watermolecules in mediating protein-DNAand protein-protein interactions. Thebinding of divalent metal ions, such asmanganese or magnesium, to the DNAappears to favor the distortion of theDNA seen in the NCP (Figures 3 and 4).The presence of ordered watermolecules at the interface between theprotein subunits may provide a meansto favor their disassembly and thuscould be important to the processes ofDNA transcription and replication.

OUTLOOK

The nucleosome core particle structureexplains in atomic detail how DNA iskept untangled in the cell nucleus andclarifies the unique role of thenucleosome in maintaining andcontrolling the expression of geneticinformation. Nucleosomes do not existas isolated particles in the cell, but arepacked into arrays with an internalrepeat of 157-240 bp. The dynamic

assembly and disassembly of thehigher-order structures made from thesearrays helps determine the functionalstate of DNA. HOS formation isapparently guided by interactions withnon-histone nuclear proteins in additionto the DNA sequence preferencesdisplayed by the histone octamer itself.The NCP structure has already providedinformation that has allowed us toconstruct a nucleofilament zigzagstructure in crystals and add a majorfragment of the H1 linker histone to it.Work on this structure is currently inprogress at the ESRF. Beyond this, wewill endeavor to elucidate the structuresof higher-order arrangements ofnucleosomes in the chromatin fiber andto relate this information to the waythese assemblies participate in generegulation.

REFERENCES

[1] W.A. Hendrickson and K. Wüthrich, eds.,Macromolecular Structures (Current Biology,London) (1997).[2] T. J. Richmond, J. T. Finch,B. Rushton, D. Rhodes and A. Klug, Thestructure of the Nucleosome Core Particle at7 Å Resolution. Nature, 311, 532-537 (1984).[3] K. Luger, T. Rechsteiner, A.J. Flaus,M. M. Y. Waye and T. J. Richmond,Characterization of Nucleosome CoreParticles Containing Histone Proteins Made inBacteria. J. Mol. Biol., 272, 301-311 (1997).[4] K. Luger, A.W. Mäder, R.K. Richmond,D.F. Sargent and T.J. Richmond, CrystalStructure of the Nucleosome Core Particle at2.8 Å Resolution. Nature, 389, 251-260(1997).

Fig. 4: Ordered solvent molecules (white) at theinterface between the protein chains of histone H2A(yellow), histone H3 (blue) and histone H4 (green).

Oxygen atoms are shown in red, nitrogen atoms in darkblue. Distances between protein and solvent greater

than 2.2 Å and less than 3.5 Å are shown in cyan. A2Fo-Fc electron density map, calculated as described

for Figure 3, and contoured at 1.7 sigma, is shown forthe protein (magenta) and for solvent (red).

EXPERIMENTS WORKSHOP


12

I n the wavelength regime ofvisible light this dynamic light scatteringis a well-established technique [1, 2]capable of probing dynamics up to q-vectors of typically 4 x 10-3 Å-1. Thislimitation has been overcome recently bycorrelation spectroscopy with x-rays(XPCS). Using a wide bandpath coherentx-ray beam [3], the translational dynamicsof systems such as colloidal aggregates[4] or block copolymer miscelles [5]could be studied. Motivated by theobservation of static x-ray speckles inblock copolymer systems [6, 7], we wereinterested in exploring the feasibility ofXPCS for the study of surface andinterface dynamics. Two types of polymersamples were chosen: On one sample the

diffusely scattered intensity, containinginformation about surface morphology, iscreated by an additional roughnessintroduced by small droplets on top of thesurface (Figure 1a). The other sample ofmultilayer type scatters due to the densitydifference at internal interfaces which arestrongly correlated (Figure 1b). Thisroughness correlation gives rise toresonant diffuse scattering, whichconcentrates the diffusely scatteredintensity into narrow sheets in reciprocalspace [8]. We show that these samplesmight be good candidates formeasurements concerning the dynamicsof surfaces, like capillary waves, byperforming x-ray photon correlationspectroscopy.

The experiments were performed atbeamline ID10A at the ESRF using awide bandpath set-up [3] for coherentscattering in small-angle reflectivitygeometry.Polymer blends exhibit a complexbehavior upon spin-coating due to thepossibility of phase separation occurringduring the film formation. We usedpolystyrene (PS) and polybromstyreneP(BrxS) with a degree of bromination ofx = 0.72. Figure 2 displays an opticalmicrograph of the sample surface with amagnification factor of 50, displayingdroplets with a mean diameter of 2 µm.They consist of PBrS in a matrix of PS.The laterally disordered domain structureis caused by the partial wetting behaviorof the blend. PBrS entrapped in the PSphase during the rapid evaporation ofthe solvent, segregates to form dropletsat the surface. In the investigatedq-range, capillary waves have onlylittle contributions compared to thecontributions of the droplets, and thedroplet motions on a polymer melt can bestudied. With the increase of roughnessthe diffusely scattered intensity, whichcontained the information about thesurface morphology [9], is increased too,providing an increased intensity forcoherent experiments.Installing the above describedrequirements for coherent x-ray scatteringwe measured the static speckle patterns byperforming common 'detector-scans' (thesample is held fixed at one angle ofincidence αi and the detector angle α isvaried). According to ∆qx ~ ±(2π/λ) α ∆αand ∆qz ~ ±(2π/λ) ∆α, changes in the exitangle ∆α will mainly result in a change ofqz and only very small changes in qx.Typical speckle patterns are presented inFigure 3. The exit angle equals the critical

OBSERVATION OF SPECKLE PATTERNS BY COHERENT

X-RAY SCATTERING FROM THIN POLYMER FILMSP. MÜLLER-BUSCHBAUM1, T. THURN-ALBRECHT1, G. MEIER1, M. STAMM1, E.W. FISCHER1, B. STARK2, B. STÜHN2, D.L. ABERNATHY3, G. GRÜBEL3

1 MAX-PLANCK-INSTITUT FÜR POLYMERFORSCHUNG, MAINZ (GERMANY)2 FAKULTÄT FÜR PHYSIK, UNIVERSITÄT FREIBURG (GERMANY)3 ESRF, EXPERIMENTS DIVISION

Fig. 1: Schematicdrawing of the twosample types used in thecoherent experiment:(a) droplets on a thinpolymer blend film and(b) layered lamella ofdiblock copolymer(about 50 layers, not allshown).

Fig. 2: Optical micrographsof the blend film sample(magnification 50 times)displaying the surfacestructure consisting of smalldroplets which were used forthe surface labelling.

a

b

A speckle pattern is produced whenever a disordered material causes random phase shifts while

scattering coherent incident light. Changes of the disordered structure with time lead to fluctuating

speckle patterns. The observation of these intensity fluctuations at a single point in the speckle pattern

provides a direct measure of the underlying dynamics.



13

angle of PS which corresponds toαi = 0.54°. Therefore the measurementsare performed through a Yoneda peak.The width of the speckles depends on thesize of the front pinhole aperture p. Theangular extent of each spot is comparableto that of the central peak of theFraunhofer diffraction pattern λ/p of thepinhole. An enlarged pinhole aperturecauses a narrower diffraction pattern andtherefore narrower speckles. The solidlines in the two top curves (20 µmdetector pinhole) show the width of themeasured central peaks of the primarybeam with the corresponding frontpinhole for comparison. The goodagreement shows that the observedintensity distribution is due to a staticspeckle pattern from the sample surface.Increasing the detector pinhole d by afactor of 10 averages out the sharp specklestructure corresponding to the observationof the ensemble averaged structure factor.Block copolymers are interestingmaterials due to their particular phaseseparation behavior. Because of thechemical connectivity of the componentsonly microphase separation of mutuallyincompatible blocks takes place.Therefore phase separation occurs onlyon a molecular scale. This order-to-disorder (ODT) transition is controlled bythe degree of polymerisation and theFlory-Huggins segment interactionparameter. Due to selective interactionssurfaces can influence the structureformed during a phase separation process.One component may segregate to thesurface of the film introducing a surface-induced order. In oriented films the

structures extend over large distances andfor a symmetric block copolymer alamellar ordering is induced. Thecharacteristic periodicity d of theselamella is of the order of the radius ofgyration.For the experiments presented here,polystyrene-block-polyisoprene (PS-b-PI) with a symmetric composition wasused. For the bulk material an order-to-disorder transition temperature TODT =178 °C was obtained.For the dynamics ofsuch systems one expectscollective interfacial modesin which adjacent inter-faces move in phase. In thecase of full correlation thein-plane morphology ofthe upper layer would equalthe one of the underlyinglayer through the whole

sample. Whereas for a sample withuncorrelated interfaces all interfacesscatter independently and the diffuseintensities of all individual interfacessuperpose, the case of partially or fullycorrelated roughness gives rise toa coherent superposition of thecontributions from individual interfacesleading to a much higher intensity. Theintensity is concentrated in narrowsheets. These sheets of resonant diffusescattering are oriented perpendicular tothe qz-axis with the center fulfillingthe one dimensional Bragg-condition∆qz = 2π/d [9].Static speckle patterns from this systemwere recorded by performing «detector-scans». The scans were taken around theqz-value of the first order Bragg peak,however slightly off-peak in qx (twotimes the full width half maximum of theBragg peak) to avoid contamination ofthe signal from specular reflection.Again we measured a well-pronouncedspeckle pattern comparable to the onesof the blend sample. However they donot result from an increased intensity dueto surface labeling but are due to anincreased intensity by resonant diffusescattering. Whereas at room temperaturespeckles are nearly independent of time,at T = 125 °C, above Tg of the PS-block,we observe speckle patterns fluctuatingon the time scale of a few minutes.

qz (10-2 Å-1)

αi = 0.54°

5.00

Int

(arb

. uni

ts)

5.05 5.10 5.15

2.0

1.8

1.6

1.4

1.2

1.0

d = 20 µmp = 5 µm

d = 20 µmp = 12 µm

d = 200 µmp = 12 µm

Fig. 3:«Detector-scans»measured at the angle ofincidence αi = 0.54° ofthe blend film sampledemonstrating theinfluence of the frontpinhole p and the detectorpinhole d. In each scanthe qx value is changedbetween 1.69 x 10-4 Å-1

and 1.72 x 10-4 Å-1. Thesolid line shows the widthof the central peak of theFraunhofer diffractionpattern as yielded by themeasurement. For claritythe curves are shiftedagainst each other.

qz (10-2 Å-1)

∆t = 3 mn

4.00

Int

(arb

. uni

ts)

5

4

3

2

1

4.10 4.20

Fig. 4: Time seriesof «detector-scans»

performed at T = 125 °C.The angle of incidence was

αi = 0.224° for theexamination of the diblock

copolymer film sample(p = 5 µm and detector

sided pinhole aperture of20 µm). For clarity the

curves are shifted againsteach other. The dashed

line is a guide to the eye.



14

A

As an example we show a time series ofdetector scans in Figure 4. Each scan wastaken after three minutes delay time witha counting time of one second per point.In summary we presented experimentsusing coherent x-ray scattering in areflection geometry to investigate thinpolymer films. In the blend film systemthe scattering in the demanded q-range isenhanced by surface labelling, whereas inthe diblock copolymer film system theroughness correlation of the multilayerinterfaces provides the high intensity.Therefore both samples are well suited forexperiments using x-ray photoncorrelation spectroscopy, because theydeliver a good signal-to-noise ratio in theobserved speckle patterns. Thus they aregood candidates for further dynamic

measurements. This will enable thedetermination of the time correlationfunction of the surface and interfacedynamics and will open up the door forexaminations of capillary waves, surfacediffusion processes and membranedynamics.

REFERENCES

[1] B. Berne, R. Pecora: Dynamic lightscattering; Wiley-Interscience, New York(1975).[2] D. Langevin: Thin liquid films; in«Light scattering by liquid surfaces andcomplementary techniques»; edited byD. Langevin, Marcel Dekker INC., New York,265 (1992).[3] D.L. Abernathy, G. Grübel, S. Brauer,I. Mc Nulty, G.B. Stephenson, S.G.J. Mochrie,

A.R. Sandy, N.Mulders, M.Sutton; J.Synchrotron Rad. 5, 37 (1998).[4] T. Thurn-Albrecht, W. Steffen,A. Patkowski, G. Meier, E.W. Fischer,G. Grübel, D.L. Abernathy; Phys. Rev. Lett. 77,5437 (1996).[5] S.G. J. Mochrie, A.M. Mayer, A.R. Sandy,M. Sutton, S. Brauer, G.B. Stephenson,D.L. Abernathy, G. Grübel; Phys. Rev. Lett. 78,1275 (1997).[6] G. Vignaud, G. Grübel, A. Gibaud,D. Auserre, J.F. Legrand; ESRF Highlights1994/1995, 8.[7] Z.H. Cai, B. Lai, W.B. Yun, I. Mc Nulty,K.G. Huang, T.P. Russell, Phys. Rev. Lett. 73, 82(1994).[8] V. Holy, T. Baumbach; Phys. Rev. B 49,10668 (1994).[9] S.K. Sinha, E.B. Sirota, S. Garoff,

FIRST EXPERIMENTS ON TROÏKA IID.M. SMILGIES, N. BOUDET AND G. GRÜBEL

ESRF, EXPERIMENTS DIVISION

The Troïka II beamline (ID10B) started user operation on schedule in April 1998.

This second branch of the Troïka beamlines is again an open beamline with a flexible diffractometer

which aims at providing additional beam time for demands in the soft-condensed matter community.

Special emphasis to surface and interface studies using grazing-incidence

diffraction and x-ray reflectivity is given.

lthough Troïka II uses the sameundulator source as Troïka I (ID10A), bothstations can work independently inparallel. A thin diamond (111) crystal actsas monochromator and beam splitter (seeFigure 1), a technique which has been

pioneered at the Troïka beamline.A second (111) diamond crystal deflectsthe monochromatic beam at a constantoffset from the white beam path. Theenergy range of this double-crystalmonochromator is 8 to 12 keV. As a

further optical element, a plane doublemirror for harmonic rejection is inpreparation. Presently, harmonic rejectionis achieved with two small mirrors in theexperimental hutch.The Troïka II eight-circle diffractometerwith an additional beam deflector canoperate both in horizontal and verticalscattering geometry. In horizontalgeometry heavy and bulky sampleenvironments can be handled. Incombination with the deflector, a devicethat employs either a mirror or a Braggcrystal to deflect the beam out of thehorizontal plane, the horizontal geometry

Fig. 1: Troïka II layout: PS primaryslits; M1, M2 diamond double-crystalmonochromator; Mi planned doublemirrors; PhS photon shutter,D deflector, Diff: eight-circlediffractometer.

PS M1 M2 Mi PhS D Diff

Source

Optics Hutch

Experiments Hutch

Sourcewhite beam monochromatic beam

0 27.1 30.5 38.7 40.7 42.4



15

is particularly useful for the investigationof liquid surfaces and self-organization ofamphiphilic molecules at the air-waterinterface. For these studies a large lineardetector can be used in combination withSoller slits or an analyzer stage. Incollaboration with a group at Saclay alarge multi-purpose Langmuir trough is indevelopment. Other possible sampleenvironments could involve cryostats,furnaces, or small vacuum chambers to besupplied by the users.The vertical scattering geometry is usedfor small samples and sampleenvironments to study surfaces or thinfilms in air or under inert gas. In thisgeometry the largest reciprocal spaceaccess is provided. It will be particularlyimportant for the in-house researchactivities which will focus on theinvestigation of thin organic films on solidsubstrates. Various detector options areavailable, such as a scintillation detectorwith motorized slits and a small lineardetector which can be combined with ananalyzer stage.The spectrum of research activities atTroïka II can be best exemplified by thefirst six user experiments. In the very firstexperiment, Daillant et al investigated thediffuse scattering from the surface of waterand various other liquids. The goal of theexperiment was to determine whether

there are higher order corrections to thesurface tension. In the experiment theincident x-ray beam impinges on the liquidsurface below the critical angle and createsan evanescent wave parallel to the liquidsurface with a penetration depth of about50 to 100 Å. The evanescent wave isscattered by capillary waves giving rise todiffuse scattering up to 50° scatteringangle in-plane which is integratedperpendicular to the liquid surface by alarge linear detector. Important for thisexperiment is a very clean, well definedbeam without high-energy contamination.The latter was achieved by a low-passfilter consisting of two small platinumcoated mirrors in deflector and flight pathwhich were set close to the critical angle.In the second experiment, Möhwald et alcharacterized selected phases in the phasediagram of the fatty acids CnH2n+1COOHwith n = 18, 20, 22, and 24. The phasediagram of these amphiphilic molecules iswell understood theoretically and certainpredictions were to be tested in theexperiment. In particular, the variation ofthe lineshapes in the L2 to L’2 phasetransition was to be investigated where thedirection of the tilt of the molecular axischanges from the nearest-neighbordirection to next-nearest neighbor. Thedata are presently being analyzed.The remaining four experiments use the

vertical scattering geometry. Lacaze et altried to determine the 4 x 16 phase of theliquid crystal 8CB absorbed on MoS2which they found previously with STMand first grazing-incidence diffractionexperiments. This experiment is verydifficult due to the large mosaicity of about0.5° of the substrate and the low Z of theliquid crystal film. The experimentaddresses the problem of anchoring aliquid crystal on an atomically well-defined interface. In the followingexperiment by Lucas et al theelectrochemical absorption of CO onPt(111) and Pt(100) surfaces will bestudied. Again the problem will be to see alow-Z monolayer on a high-Z substrate.Finally, the last two experiments byAbstreiter et al and Metzger et al willinvestigate the self-organization ofquantum dot arrays in strain-modulatedsuperlattices. In these last two experimentsa high-resolution set-up with analyzer andlinear detector will be used.The Troïka II beamline addresses theneeds of both the soft and hard-condensedmatter community by making more beamtime available for interface and thin filmsstudies. A variety of scattering geometriesand detectors options is available. Formore detailed information, please contactD. M. Smilgies (tel +33 (0) 4 76 88 27 31– email: [email protected]).

The observation, by hard x-rayphase imaging [2-4] and phasemicrotomography [4,5], of orienteddodecahedral holes, located in the bulk,of high-quality, centimetre-size, singlegrain icosahedral AlPdMn quasi-crystals, is considered by the scientific

community concerned as a crucial result.It was made possible by the high spatialcoherence of the beam at the «long»(145 m) ID19 beamline. Figure 1 showsone of the central results: the sizes of theobserved holes display a discretedistribution, with peaks at 22, 5 and

slightly more than 1 µm. Gas bubbleswould lead to a continuous distribution.The jump from one size to the nextcorresponds to the factor τ3, where τ isthe golden mean (τ = 2 cos 36° = 1.62),a basic ingredient in all theoreticalapproaches of quasicrystals. The three-dimensional reconstruction resultingfrom phase microtomography showsthat the average distance betweenneighboring holes is, again, about τ3

times the hole size (Figure 2).Figure 3 shows that annealing alters theshape and sizes of the holes and reducestheir number. The discrete distributionof sizes flattens out and changesgradually into a continuous one. It wasoften observed that lamella-shapedcrystalline inclusions form in theneighborhood of the holes.Figure 4 shows a diffraction image(«topograph») and a phase imagecorresponding to a hole. The topographshows a white-black contrast whichindicates that a gradient of distortion ispresent in the hole neighborhood. Thiscontrast can be simulated by simplyassuming, as a first approximation,that the deformation field is similar tothe classical one around a sphericalinclusion in a crystal [4].The features observed on as-grown grainscould support the predictions of a modelwhich describes the quasicrystal structurein terms of a hierarchical self-similarpacking of overlapping atomic clusters.An inflation scale factor τ3 preserveslong-range order but generates ahierarchy of holes and a fractal structure.The observations appear to be in fair

EXPERIMENTS REPORTS


16

HIERARCHICAL POROSITY IN QUASICRYSTALS INVESTIGATED

BY COHERENT X-RAY IMAGINGL. MANCINI1,2, J. GASTALDI1, E. REINIER1, P. CLOETENS2,3, W. LUDWIG2, C. JANOT4,6,

J. HÄRTWIG2, M. SCHLENKER5 AND J. BARUCHEL2

1 CRMC2-CNRS, MARSEILLE (FRANCE)2 ESRF, EXPERIMENTS DIVISION

3 EMAT-RUCA, ANTWERP (BELGIUM)4 DIP. SCIENZE DELLA TERRA, UNIV. 'LA SAPIENZA', ROMA (ITALY)5 LABORATOIRE LOUIS NÉEL, CNRS, GRENOBLE (FRANCE)6 LABORATOIRE DE CRISTALLOGRAPHIE, CNRS, GRENOBLE (FRANCE)

An unexpected kind of solids, the quasicrystals, with long range translational order but no periodicity,

was discovered in 1984 [1]. A lot of effort was devoted, from that moment, to understand their

peculiar properties, in particular their growth mechanism and stability.

Num

ber

of h

oles

10

8

6

4

2

00

Distance between neighbouring holes (µm)

50 100 150 200

Fig. 2: Three-dimensionalreconstructionresulting from aphasemicrotomography(sample-to-detector distance50 cm, λ = 0.52 Å)of an AlPdMnquasicrystal,showing the holeslocations.

Fig. 1: Phaseradiography ofan AlPdMnquasicrystal(sample-to-detector distance50 cm,λ = 0.35 Å)showing threefamilies ofdodehahedralholes (arrows).

EXPERIMENTS REPORTS


17

agreement with these predictions [6],albeit with no indication about holes withsize below the experimental resolutionlimit of about 1 µm, and with lowervolume fractions and a flattenedhierarchy. This flattening could resultfrom the quasicrystal single grains havingeffectively, because of their low thermalconductivity, suffered annealing even intheir as-prepared state. But the lowervolume fractions of holes clearly indicatethat the present model is not a final point.Systematic experiments are beingperformed to further clarify this topic.

REFERENCES

[1] D. Schechtman, I. Blech, D. Gratias,J.W. Cahan, Phys. Rev. Lett. 53, 1951 (1984).[2] L. Mancini, E. Reinier, P. Cloetens,J. Gastaldi, J. Härtwig, M. Schlenker,J. Baruchel, Phil Mag.A, in press.[3] E. Reinier, PhD Thesis, Univ. Aix-Marseille(1998).[4] L. Mancini, PhD Thesis, Univ. J. Fourier,Grenoble (1998).[5] W. Ludwig et al., to be published.[6] L. Mancini, C. Janot, L. Loreto, R.Farinato, J. Gastaldi, J. Baruchel, Phil. Mag.Lett., vol 78, n° 2, 159-167 (1998).

Quasicrystals are solids which havea new type of long-range order,such that their diffraction patternsshow Bragg reflections revealingsymmetries which are incompatiblewith periodicity (five-fold or ten-foldaxis, for instance). Large single grainscan be grown for some of thequasicrystalline alloys, like AlPdMn.Their quality is such that dynamicdiffraction effects have been observed. Quasicrystal structures can be describedin different ways. One of them consistsin filling with atoms an appropriatequasiperiodic tiling, produced from twobasic tiles and matching rules(Fibonacci chain in 1D, Penrose tiling in2D,...). Another approach involves onlyone basic unit (a cluster of atoms) and areplication process; it implies someoverlaps and holes.

In a higher dimensional space itis possible to describe the 3D-quasiperiodic structure as a periodicone. The actual quasiperiodic structurein the 3D-physical space is obtainedfrom the hyperspace structure byan appropriate projection/sectiontechnique. The Bragg peaks observedon the AlPdMn quasicrystals, forinstance, can be indexed in a 6Dperiodic hyperspace. Aside from their peculiar structure,quasicrystals also exhibit veryunexpected properties. This includesvery high electrical and thermalresistivities: quasicrystals are there-fore insulator alloys containing about70% of aluminium! Other interestingproperties involving, for instance,adhesion, corrosion, friction, andhardness suggest that quasicrystals are

promising materials for industrialapplications, particularly coatings.

ReferenceSee, for instance, C. Janot, «Quasicrystals: aprimer», Oxford Univ. Press, 2nd edit. (1994)

Num

ber

of h

oles

140

120

100

80

60

40

20

0

Hole size (µm)

0 5 10 15 20 25 30 35 40

140

120

100

80

60

40

20

00 5 10 15 20 25 30 35 40

as-grown

after annealing

Fig. 3: Phase radiography after an annealing treatment showing deformation ofthe hole shapes and the appearance of lamellas of a crystalline phase.

Fig. 4: a) diffraction and b) phaseimage of a hole; the diffraction image

shows that the presence of a hole isassociated with a distortion of thequasicrystal in its neighborhood.

WHAT ARE QUASICRYSTALS ?

ab

EXPERIMENTS REPORTS


18

T

ID26 : X-RAY SPECTROSCOPY ON ULTRA-DILUTE SYSTEMSCH. GAUTHIER, V.A. SOLÉ, J. GOULON, E. MOGUILINE AND R. SIGNORATO


The ESRF beamline ID26 is dedicated to x-ray absorption spectroscopy on ultra-dilute systems (XAUS)

in the spectral range 2.3 - 30 keV. The aim of the beamline is to extract structural and electronic

information on dilute samples for which the concentration of the absorbing

element ranges from a few ppm up to 10 000 ppm.

A wide range of applications is covered in biology, catalysis, chemistry, environmental sciences, solid

state physics... The beamline accepted the first users in November 1997. We give herein a brief

description of the beamline and we report on recent instrumentation developments.

he x-ray source consists of threephasable planar undulators. Each of themhas a magnetic period of 42 mm and alength of 1.65 m. The energy of thefundamental harmonic is 2.35 keV at theminimum gap value (16 mm). The wholeenergy range is obtained by exploitingdifferent harmonics of the undulatorsemission spectra. The design of thebeamline optics is optimized in order toreduce background radiation and toprovide an efficient harmonic rejection.The first component is a flat silicon mirrorthat deflects the beam laterally apart fromthe bremsstrahlung emission cone of thering. This mirror damps the thermal powerof the x-ray source and protects the otheroptical components. The x-ray beam isfocused by two segmented piezoelectricbimorph silicon mirrors in a Kirkpatrick-Baez configuration, which are located inthe monochromatic section of the

beamline. The typical focal spot size at thesample location is 200 µm (horizontal) x15 µm (vertical). The photon flux on thesample is 1013 photons per second.Smaller spot size could be achieved(H : 80 µm x V : 7 µm) with a lower flux(1012 photons/s). Each mirror has threereflective layers: silicon, chromium andplatinum. A two-crystal fixed-exit Kohzu mono-chromator is inserted after the firsthorizontally deflecting mirror. Themonochromator is equipped with twopairs of crystals Si (111) and Si (220)cooled down to - 140 °C. The crystal pairscan easily be exchanged using a lateraltranslation of the monochromator. Two experimental stations are used: asmall fluorescence chamber operated atroom temperature and a fluorescencestation equipped with a continuous flowliquid helium cryostat. The temperature

range covered goes from 4.8 K to 293 K.The cryostat is equipped with a veryaccurate (1 µm) vertical sample translatorand a rotary stage with an angularaccuracy of 0.13 mrad.XAS spectra are recorded in thefluorescence excitation mode. To reachhigh dilution, three main difficulties had tobe overcome : i) The fluorescence signal isusually buried into a large radiationbackground (mainly elastic or inelasticscattering) originating from samplematrix. One has to get rid of thisbackground which degrades thefluorescence signal statistics. ii) Since thenarrow emission peaks of the undulatorspectra (typ 100-200 eV) are not suitableto record a whole EXAFS scan, we had toextend the gap scan techniques recentlydeveloped at the ESRF [1]. iii) Due to thevery high flux delivered at the ID26beamline, sample radiation damage isusually a major source of difficulties,especially in the case of biologicalsamples. In order to overcome thesedifficulties, different data acquisitionstrategies have been developed.

CONTINUOUS OR QUICKEXAFS FOR MODERATE

DILUTIONS

For absorber concentrations above1 mMol/l (50 to 500 ppm depending onsample matrix), the fluorescence detectionis performed using silicon PINphotodiodes operated in the photovoltaicmode. The diodes are associated with alow noise current to voltage amplifier andthen to a high linearity voltage tofrequency converter [2]. The output signalis fed into a gated integrator (ESRF VDL

Fig. 1a: EXAFS spectrum of a 7 mMol/l Carbonmonoxy-Myoglobin samplerecorded at the Fe K-edge measured in one scan of 10 seconds.

Energy (eV)

If /

I0

7000

1.3

1.2

1.1

1.0

0.9

0.8

0.7

0.6

0.5

0.47100 7200 7300 7400 7500 7600 7700 7800 80007900

7 mM Carbonmonoxy-Myoglobin. Single scan (∆t = 10s)

7 mM ∆t = 10 s

Sample in air at room temperatureSi (220)

1000 points, 10 ms per point

EXPERIMENTS REPORTS


19

board). The diode can be operated withabsorption filters in order to minimize theamount of background radiation reachingthe detector. The IO intensity monitor isalso a silicon diode associated with ascattering or fluorescence foil in anoptimized geometry which minimizes thesensitivity to the x-ray beam instabilities.We take full advantage of the highsaturation level, high linearity and lownoise of silicon diodes which allows us toaccept the very high flux delivered by thesource. In the continuous gap scantechnique, the monochromator Braggangle and the undulator gap arescanned simultaneously during themeasurement of the EXAFS spectra.Excellent performances were achieved:Figure 1a reproduces the EXAFSspectrum of a 7 mMol/l myoglobin samplemeasured in 10 seconds. Figure 1b showsthe XANES part of the same EXAFSspectrum but recorded in 40 seconds. Thedata acquisition time of the XANESspectra was less than 3 seconds. Averaging on several spectra indeedimproves the signal-to-noise ratio. Themain advantage is that radiation damagecan be minimized by displacing thesample after each scan. Alternatively, theevolution of fragile samples can bemonitored on a short time scale.Furthermore, there is some possibility toperform time-resolved experiments ondilute samples with a slow time evolution.

SILICON DRIFT DETECTORS

For higher dilution, it is necessary todiscriminate in energy the fluorescencefrom the radiation background. Due to thesevere counting rate limitations of energyresolving detectors, the development of

large arrays of detectors (up to 100channels) is mandatory in order to restrictthe data acquisition time within realisticlimits. Over recent years, we havedeveloped new energy resolving detectors:Silicon Drift Diodes (SDD) [3]. Thesedetectors feature a very small readoutcapacitance which is independent on theactive area. This results in the possibilityof preserving good energy resolutiontogether with high counting rates. Thesedetectors are associated to multichanneldigital shaping amplifiers designed at theESRF. There are clear advantages in favorof this technology: i) The output rate canbe improved in comparison with analogshaping amplifiers. ii) The cost perdetection channel is much lower. iii) Theset-up of the detector array is controlled bysoftware and can be automated. A firstprototype array (12 channels) has beenevaluated at the beamline. Figure 2 shows

an EXAFS spectrum recorded from a 200µmol/l aqueous solution of cobalt acetate(cobalt concentration 12 ppm) during atest experiment. The data acquisition timewas 9 hours, the signal to noise ratio canbe improved with longer data acquisitiontimes. At present the highest dilutionwhich can be achieved is 5-10 ppmdepending on the absorber and the samplematrix composition. A 35 channel SDDarray is under production in collaborationwith Eurisys Mesures Inc. and should beoperational by September 98. We expectto be able to lower the dilution limit downto 1-5 ppm.

REFERENCES

[1] A. Rogalev, V. Gotte, J. Goulon, C. Gauthier,J. Chavanne and P. Elleaume. J. Synchrotron Rad.(1998) 5, 989.[2] C. Gauthier, G. Goujon, S. Feite,E. Moguiline, L. Braicovich, N.B. Brookes,J. Goulon. Physica B208&209 (1995) 232-234.[3] E. Moguiline, Ch. Gauthier, G. Goujon,J. Goulon, M.O. Lampert, P. Dressler, R. Henck,J. Phys. IV France7 (1997) C2 339-340.

Fig. 1b: XANES region of a 40 secondEXAFS scan of 7 mMol/lCarbonmonoxy-Myoglobin sample. The(220) crystal pair of the monochromatorwas used. The data acquisition for theXANES region is less than 3 seconds.

Fig. 2: XAS spectrum of an aqueous solution of cobalt acetate recorded at roomtemperature in air with a 12 channel SDD array. The sample dilution was

200 µmol/l (12 ppm cobalt concentration). Data acquisition time was 9 hours.

Energy (eV)

Nor

m. F

luo

2.0

1.3

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0

–0.281007650 7700 7750 7800 7850 7900 7950 8000 8050 8150

12 ppm Co in Water

Energy (eV)

If /

I00.66

0.64

0.62

0.60

0.58

0.56

0.54

0.52

0.50

0.487100

7 mMol/lMbCO.XANES region

∆t = 2.95 s

7110 7120 7130 7140 7150 7160 7170

ACKNOWLEDGEMENT

The authors wish to thankF. Natali for providing theCarbonmonoxy-Myoglobin sample.

Recent x-ray absorption spectroscopy(XAS) studies [1] of the Ti K-near-edgeand Ba K-extended spectra of BaTiO3demonstrate that the local structure inthis material remains disordered even attemperatures well above the transition toa crystallographically cubic phase. Theanalysis of the Ti K-edge extendedspectrum was hindered by the presenceof the nearby Ba LIII absorption edge at5247 eV, compared to 4966 eV for the TiK-edge. This separation of 281 eV, orabout 8.2 Å-1, places such a severe limiton the information content of the Tiextended spectrum that meaningfulanalysis of those data is not possible.

Use of an energy discriminating detectorin a fluorescence XAS measurement isprecluded by the proximity of thevarious Ti and Ba fluorescence lines. Inthis article we present DiffractionAnomalous Fine Structure (DAFS) as asolution to the problem of overlappingedges. DAFS has been used to separatethe fine structures of single atomicspecies in multiple crystallographic sites,for example the Cu in YBa2Cu3O7 [2]and the Fe in BaZnFe6O11 [3] or at theinterfaces of a multi-layer compound [4].Using similar methodology, we canseparate the signals from the Ti and Basites in BaTiO3, thus isolating the Ti K-

edge fine structure spectrum andextending it beyond the 8.2 Å-1 limit ofthe absorption experiment.In the DAFS experiment, the variation inenergy of the intensity of a diffractionpeak is measured in an energy rangespanning one or more absorptionenergies in the sample. The data weretaken at the ESRF beamline BM2 on a1000 Å thick film of BaTiO3 depositedon sapphire by pulsed laser deposition.The film was found by x-ray diffractionto be oriented with an an axis normal tothe surface. In our experiment, wemeasured the (100) and (200) reflections(Figure 1).The oscillatory fine structure seen in thedata above the absorption energiescontains the same local structuralinformation as an XAFS spectrum. Thisdiffracted intensity can be expressed [2]as eq. (1):

where the sum is over all sites j in theunit cell, Mj is the thermal factor atsite j, fo is the Thomson scattering ofthe atom at site j, and f ’ and f” are theenergy-dependent atomic anomalouscorrections. The final term contains thecomplex fine structure ~χ (of which onlythe imaginary part is measured in a XASexperiment), as a coefficient of ∆f”, theportion of f” being due to the resonantelectron.We extracted the imaginary componentof the scattering factor from the spectraof the (100) and (200) reflections. Wesolved eq. (1) for its real part and

EXPERIMENTS REPORTS


20

EDGE SEPARATION USING DAFSB. RAVEL1, C.E. BOULDIN1, H. RENEVIER2, J.L. HODEAU2 AND J.F. BERAR2

1 NIST CERAMICS DIVISION, GAITHERSBURG MD (USA)2 LABORATOIRE DE CRISTALLOGRAPHIE, CNRS-UJF, GRENOBLE (FRANCE)

We exploit the crystallographic sensitivity of the diffraction anomalous fine structure (DAFS)

measurement to separate the fine structure contributions of different atomic species with closely

spaced resonant energies. In BaTiO3 the Ti K-edge and Ba LIII are separated by 281 eV, or about

8.2 inverse angströms, which severely limits the information content of the Ti K-edge signal. Using the

site selectivity of DAFS we can separate the two fine structure spectra using an iterative Kramers-

Kronig method, thus extending the range of the Ti K-edge spectrum.

Inte

nsit

y

200

150

100

50

0

Energy (eV)

(200)

(100)

4800 5200 5600

Fig. 1: Measured DAFS spectra of the (100) and (200) reflections fromBaTiO3. The (200) spectrum is displaced upwards for clarity.

Kramers-Kronig transform this toobtain a new value for its imaginarypart [5]. The steps of solving eq. (1)and transforming are iterated toconvergence. A weighed sum of the f”spectra are then made to yield theisolated Ti f” spectrum shown inFigure 2. A method developed forisolating χ” from the XAS µ [6] wasapplied directly to the Ti f” spectrum.The extracted χ” function is shownin Figure 3 along with a FEFF [7]simulation using previously-determinedstructural parameters [1]. Theincomplete cancellation of the Ba whiteline results in a problem in the datawhich resembles a monochromatorglitch. Despite this, measurable Ti finestructure signal is seen in χ between 9and 10 Å-1.Using DAFS to separate the finestructures of overlapping absorptionedges is a novel approach to thisproblem which does not require the useof energy discriminating detectors. Wehave shown that it yields data that isanalyzable beyond the secondabsorption edge energy. There are manytransition metal/rare-earth compoundsof technological interest which arecomposed of atoms with overlappingedges. We plan to test our method onone such material, Gd3Fe5O12, amaterial used in microwave waveguides. In that case, the Fe K and the GdLIII are separated by only 130 eV andtheir major fluorescence lines areseparated by about 100 eV.

REFERENCES

[1] B. Ravel, E. A. Stern, R. I. Vedrinskii, andV. Kraizman. Ferroelectrics, 206-207:407-430(1998).[2] L.B. Sorenson, J.O. Cross, M. Newville,B. Ravel, J.J. Rehr, H. Stragier, C.E. Bouldin,J.C. Woicik, in Resonant Anomalous X-rayScattering: Theory and Applications, edited byK. F. G. Materlik, C.J. Sparks (North Holland,Amsterdam, 1994).[3] J. Vacinova, J.L. Hodeau, P. Wolfers,J.P. Lauriat, E. Elkaïm, J. Synchrotron Rad. 2,236 (1995).[4] H. Renevier, J.L. Hodeau, P. Wolfers,S. Andrieu, J. Weigelt, and R. Frahm. Phys.

Rev. Lett., 78(14):2775-2778 (1997).[5] J.O. Cross, M.I. Bell, M. Newville,J.J. Rehr, L.B. Sorenson, C.E. Bouldin,G. Watson, T. Gouder, and G.H. Lander. Phys.Rev. B (1998) to be published.[6] M. Newville, P. Living, Y. YYacobi, J.J.Rehr, and E.A. Stern. Phys. Rev. B,47(21):14126-14131 (1993).[7] S. I. Zabinsky, J. J. Rehr, A. Ankudinov,R. C. Albers, and M. J. Eller. Phys. Rev. B,52(4):2995-3009 (1995).

EXPERIMENTS REPORTS


21

f (Ti)"

3

2

1

0

Energy (eV)

4800 5000 5200 5400 5600

Fig. 2: Extracted f” (E) for the Tiatom. Note the incomplete

cancellation of the Ba LIII white lineat about 5250 eV.

kχ(k

)

0.5

0.25

0

0.25

0.5

Wave number (A1)2 4 6 8 10

dataFEFF

Fig. 3: χ” for the Ti atom compared to a simulation using FEFF [7]. The large featurearound 8.5 Å-1 is due to the incomplete cancellation of the Ba LIII white line.

ACKNOWLEDGEMENTS

We thank A. Carter for providingthe BaTiO3 thin film sample. We alsothank J.O. Cross for inspirationalconversation and M. Newville forkindly supplying source code for thedifferential Kramers-Kronig transform.This work was supported in part by theNational Research Council.

EXPERIMENTS REPORTS


22

B

COUPLING ELECTROCHEMISTRY WITH X-RAY ABSORPTION

SPECTROSCOPY TO STUDY POLYMERSP.L. VIDAL1, M. BILLON1, B. DIVISIA-BLOHORN1, G. BIDAN1, J.M. KERN2, J.P. SAUVAGE2 AND J.L. HAZEMANN3

1 DRFMC / SI3M / LEMSI, CEA GRENOBLE (FRANCE)2 INSTITUT LE BEL, UNIVERSITÉ L. PASTEUR, STRASBOURG (FRANCE)3 LABORATOIRE DE CRISTALLOGRAPHIE, CNRS, GRENOBLE (FRANCE)

The field of chemically modified electrodes has seen an important growth in the two last decades,

especially conjugated polymers including redox-active transition metal centers for their potential

applications in electrocatalysis, sensor or electronic devices. However, there is limited information on

the structure of these layers. On the other hand, x-ray absorption spectroscopy represents a highly

sensitive tool to study atomic and molecular details of non-crystalline materials. Thus, EXAFS

measurements on chemically modified electrodes under different controlled potentials can provide

information about interactions between redox active metal centers and the conjugated structure, i.e.

about the degree of electronic communication between them.

y using thiophene or pyrrolederivates bearing functionalities (a metalion binding site, an anionic or cationic site)as monomers for electropolymerisation,electrodes can be modified with aconductive matrix containing these groupsand thus can be useful in electrocatalysis,sensor or electrochromic devices. A fewyears ago, G. Bidan (CEA-Grenoble), J.-P.Sauvage (Univ. of Strasbourg) and co-workers have used electropolymer-isationof pyrrole-bearing 1,10-phenanthrolinetetrahedral complexes of transition metalsto modify electrodes, leading to sitesrigidly anchored within the polymermatrix and then to stabilisation of unusualoxydation states. More recently, relatedworks, with polythiophene, have beenpublished by us [l] and others, in an effortfor a more direct electronic coupling

between the receptor site and theconjugated polymer backbone and for abetter control of the structure.By combining the phenanthrolinecontaining macrocycle 2 with the keyligand bearing two pendant bithiopheneunits 1 (Figure 1) via the templatingcopper (I) ion, rotaxane complex wasobtained and then electropolymerisedto afford a coordinating polyrotaxanewith a conjugated backbonealternating quaterthiophene moietiesand phenanthroline complexes withthreaded cyclic units. By testing thissystem for transition metal ion sensingability, we discovered an originalstructural effect: as shown in Figure 2, theCu(I) template used to assemble thefragments of the precursor could beremoved but, interestingly, subsequent

remetallation was only possible if lithiumwas present during demetallation. Thefunction of lithium is assumed to be that ofan ionic scaffolding, maintaining thetopography of the coordination site aftercopper removal, though forming a labilecomplex. XAS studies at the Cu K-edgehave been undertaken on this conjugatedpolyrotaxane at the ESRF, on the CRG-IFBM32 beamline, in order to obtain 1)information about interactions betweenredox active copper centers and theconjugated structure at different controlledpotentials and 2) additionally proof of thehigh reversibility of copper(I) bindingwhen lithium is present. The high fluxavailable at the ESRF allows shortacquisition time and due to the highdilution of metallic sites in our samples,data were collected with a highly sensitive

S

S

N N

S

S

N N

1

2

Fig. 1: The monomers.

Threadingstep

Cu(1.2)+

Electropolymerisation

Polyrotaxanes wires

= Cu +

= Li +

S

CollapseNo reversibility Scaffolding effect

Reversibility

=

1 2

Fig. 2: Schematic representation of the synthetic strategyused and of the scaffolding effect.

EXPERIMENTS REPORTS


23

fluorescent detector. We have recordedspectra of the films of polyrotaxaneelectropolymerised onto a carbonelectrode and in contact with adichloromethane / 0.2M tetra-n-butylammonium hexafluorophosphate solutionat different controlled potentials. Asshown in Figure 3, changes in oxidationstate of the copper centers are evidencedby the shift (by about 2.5 eV) in theposition of the edge upon oxidation of thepolymer film from Cu(I) (+ 0.3 V) toCu(II) (+ l.15 V). Furthermore, no increasein the coordination number seems to occur(via a couterion in electrolyte solution)when passing from Cu(1) to Cu(II),contrary to what is observed withmonomeric model compounds. Uponfitting of the data with MC Kale functions,further evidence is obtained for anunchanged coordination with 4 donornitrogen ligands when passing from+ 0.3 V to + l.15 V. Nevertheless, aconsiderable change in geometry occursupon oxydation of the film: at + 0.3 V thegeometry around Cu(I) is best described as

distorted tetrahedral with 2 Cu-N distancesat 2.08 Å and 1.96 Å approximately, whileat + l.15 V, the geometry around Cu(II) isflattened tetrahedral with an average Cu-Ndistance of 2.0l Å Comparison betweenspectra recorded with freshly-preparedand remetallated via lithium/copperexchange films reveals no evident change

in geometry and distances around Cu(I)centers, thus confirming the highreversibility of copper binding.

REFERENCE

[l] P.L. Vidal, M. Billon, B. Divisia-Blohorn,G. Bidan, J.M. Kern and J.P. Sauvage, Chem.Commun., 629 (1998).

Abs

orpt

ion

cœff

icie

nt

1.4

1.2

1

0.8

0.6

0.4

0.2

0

–0.2

Energy (eV)

8960

Film polarised at + 0.3 VFilm polarised at + 0.85 VFilm polarised at + 1.15 V

8970 8980 8990 9000 9010 9020 9030 9040

Fig. 3: A shift inthe position of

the edge showsthe oxidation

state of thecopper centers.

Generally speaking, epitaxial growthof metals or semiconductors aims toproduce thin films as perfect aspossible. This means good crystallinity,low density of defects and surfaceflatness. In many cases, if the substratetemperature is not too high, thedeposited material forms islands at thesurface causing significant surfaceroughness. The dimensions of theislands, their shapes and densities arevery important parameters to determine,in order to control the growth process.The morphology of the islands isthe result of the competitions ofvarious dynamical processes. Thus, itis important to characterize themorphology during uninterruptedgrowth, since interruptions could causefurther morphological changes.The diffracted intensity from a rough

surface contains information on theshort-ranged correlations of the surface.By setting the scattering conditions inan adequate way, one may enhance thediffuse scattering which arises from thesurface roughness relative to the long-range order part of the diffractedintensity. We have used a CCD camerato obtain two-dimensional diffusescattering distributions of the diffractedintensity during uninterrupted epitaxialgrowth at the Surface Diffractionbeamline (ID3).Atoms of silver were deposited onAg(100) at a constant rate of about oneatomic layer per minute. The diffusescattering at an early stage of growthwhen only 0.5 atomic layers weredeposited is shown in Figure 1. Theimage was acquired in 3.7 seconds andit represents a snapshot of a long series

of images acquired during the growthprocess. The diameter of the ring-shaped diffuse intensity providesinformation on the island sizes andseparation. The circular symmetry ofthe intensity distribution indicates thatno azimuthal ordering of the islandstakes place at this stage. After furthergrowth, the ring-shaped intensitychanges to a four-fold symmetricdistribution indicative of relativeordering of the orientation of theislands. Figure 2 shows the result after27.5 atomic layers have been grown. Inthis case the image has been taken witha long exposure time (12 minutes) forbetter clarity of the figure but asnapshot of a few seconds alreadyshows the cross-like intensitydistribution of Figure 2. The arms of thecross are along the closest packed linesof substrate atoms. The diffractionpattern is indicative of a pyramidalisland morphology. A manipulation of the morphology of theislands may be done by depositing asmall amount (~ 0.1 atomic layer) of asurfactant (Indium in our case), prior tothe growth process. The atoms of thesurfactant may act as nucleation centresof the islands or enhance surfacediffusion, modifying their density.Figure 3 is an image taken afterdepositing 10.5 atomic layers of Ag onthe Ag substrate pre-covered with In.Again, as in Figure 2, the image has beentaken with a relatively long exposuretime (5 minutes) for clarity. The scales ofFigures 2 and 3 are the same. As may beseen, the spread of the diffuse scatteringis less in Figure 3 than in Figure 2,indicating that the dimensions of theislands are larger in the case of growthwith surfactant. The diffuse intensityshows also a four-fold symmetryindicative of a regularity of the island

EXPERIMENTS REPORTS


24

MODIFICATION OF THE ISLAND MORPHOLOGY

DURING METAL EPITAXIAL GROWTH, DUE TO A SURFACTANTJ. ALVAREZ, E. LUNDGREN, X. TORRELLES AND S. FERRER


During the epitaxial growth process of metals or semiconductors, the deposited material forms

islands. In order to control this process, it is important to determine parameters such as the

dimensions, shapes and densities of the islands. Two-dimensional images show that the island

morphology and orientation can be modified by the addition of a surfactant before growth.

Fig. 1: Two-dimensional image of the diffuse scattered intensity during thegrowth of Ag on Ag(100) at the early stages of the deposition when only 0.5atomic layers have been deposited. The image was acquired in 3.7 seconds and itrepresents a snapshot of a series of many images recorded during the growthprocess. The ring-shaped pattern indicates azimuthal disorder.

EXPERIMENTS REPORTS


25

G

morphology. However, both the shapeand the orientation differ markedly fromthose obtained without surfactant. Thesymmetry axes of the intensitydistribution in Figure 3 are rotated 45degrees compared to those in Figure 2,indicating that the surfactant has caused

the pyramid-like islands to grow in adifferent orientation. The bases of thepyramids are in this case aligned with the[100] surface directions which consist oflines of atoms separated √2 times thenearest neighbor atomic spacing. Itappears that in some way the surfactant

facilitates surface diffusion along theabove directions.The above examples illustrate how two-dimensional images may reveal veryimportant information that would beinaccessible if only one-dimensionalintensity profiles were recorded.

OXIDATION OF NIAL(100) STUDIED WITH SURFACE

SENSITIVE X-RAY DIFFRACTIONA. STIERLE1, V. FORMOSO1, F. COMIN1, G. SCHMITZ2 AND R. FRANCHY2

1 ESRF, EXPERIMENTS DIVISION

2 IGV-JÜLICH, JÜLICH (GERMANY)

Thin, ordered aluminium oxide layers play a very important role as supports for model catalysts

prepared under UHV conditions. Oxidation of NiAl single crystals provides an elegant way to

prepare ordered Al2O3 layers, which can have different structures depending on the orientation of

the NiAl substrate. In addition NiAl is frequently used as a high temperature resistant material.

The formation of the protective Al2O3 layer is not understood in detail

and a microscopic picture of the oxidation process is still missing.

EED investigations [1] show a(2*1) superstructure after hightemperature oxidation and fromadditional EELS measurements theformation of θ-Al2O3 was deduced. Theaim of our study was the determinationof the Al2O3 layer structure and the

structure of the Al2O3/NiAl(100)interface using surface sensitive x-raydiffraction. The clean sample surfacewas prepared in the in situ UHV surfacediffraction chamber of ID32 by severalcycles of oxidation and flashing to1150 °C, which removes the oxide layer

and other impurities. Figure 1 shows thestructure factor of the clean NiAl(100)surface truncation rods as a function ofthe reciprocal lattice co-ordinate Lnormal to the surface. We tried different models to simulatethe data. In model 1 the ideal Ni bulk

Fig. 2: Diffuse scattering distribution afterdeposition of 27.5 atomic layers.A four-fold symmetry is an evident indicative of ordering.

Fig. 3: Diffuse scattering distribution after deposition of10.5 monolayers of Ag on a Ag(100) surface pre-

covered with a small amount of In. The image showsagain a four-fold symmetry but the axes are rotated 45°

compared to those in Figure 2.

EXPERIMENTS REPORTS


26

L

Fhk

l [ar

b. u

nits

]

101

100

10–1

10–2

10–3

10–4

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Model 1

Model 2

Model 3

Ni

Al

(001)

(100)

11L

21L

31L

Fig. 1: a)Structurefactor F(hkl)as a functionof theperpendicularreciprocallattice co-ordinate L; b) differentstructuralmodels usedto describe theexperimentaldata.

L

Fhk

l [ar

b. u

nits

]

100

10–1

10–2

10–3

0.0 0.5 1.0 1.5 –2.0

(001)

(100)

(1.5,1,L)

(1.5,1,L)

(1.5,1,L)

(-111)

(111)(-311)

(311)(-511)

(511)

(001)

(-201)(201)

(-401) (401)

(-601)

(601)

(010)

(100)

1*2

2*1

x/a (NiAl)

z/a

(NiA

l)

8

6

4

2

0

–2–4 –2 0 2

subs

trat

e

unitcell

oxygenion

Alion

Fig. 2: a)StructurefactorF(hkl) ofhalf-orderrods afteroxidation,b) structuralmodel, c)domainstructure.

termination of NiAl with CsCl structurewas used and only a relaxation of thefirst Ni layer was admitted. Thesimulation could be significantlyimproved by the introduction of Nivacancies on the surface (model 2, bluecurve) and a 2/3 Ni occupation with 4 %inward relaxation gives the bestagreement. An Al terminated surface(model 3), including vacancies andrelaxations, is not able to describe thedata.After oxidation at 950 °C for 600 sat p(O2) = 1 x 10-6 mbar half-order

rods corresponding to the (2*1)reconstruction appear. In Figure 2 thestructure factor as a function of L isplotted for different half-order rods. Onall the rods sharper peaks aresuperimposed on a broad modulation.The red curve is a calculation based ona 2 nm thick monoclinic θ-Al2O3 layer;a projection of the structure is shown inFigure 2b. All the peaks observed canbe indexed in monoclinic co-ordinatesand Figure 2c shows the domainstructure of the oxide layer. Besides the(2*1) and (1*2) in-plane domains also

twin formation in the growth directionis observed, leading to an inversion ofthe Al ion layer stacking.The additional modulation (schematicin green) is attributed to the interfacialstructure. At the moment the formationof an Ni2AlO4 spinel like interfaciallayer is discussed as the origin of themodulation and more detailed modelcalculations are under its way.

REFERENCE

[1] P. Gassmann, R. Franchy, H. Ibach, Surf.Sci. 319, 95 (1994).

EXPERIMENTS REPORTS


27

Composite systems consisting ofmetallic nanoparticles embedded inglass have recently attracted a greatinterest in the field of optoelectronicsbecause of their non-linear opticalbehavior. The implantation processcreates a thin (a few hundreds of Å)cluster-rich layer just about 1000 Åbelow the glass surface. A novel x-raytechnique has been developed to carryout structural investigations of suchsystems. It is based on the signalenhancement from the surface layerwith respect to the substratecontribution when working at thecritical angle for total externalreflection at the implanted layer-substrate interface. By using therefracted beam as a probe, thediffraction profile of the metallicclusters in very diluted samples (asingle implant in the 1016 at/cm2 range)can be extracted by a simple subtractionprocedure. A demonstration of the newmethod on SiO2 glasses implanted

with Ag ions is shown in Figure 1.The measurements were performedat the High Pressure station at theID9 beamline. The samples wereilluminated with a l0(h) x 24(v) µm2

very intense monochromatic beam inthe grazing incidence geometry withgrazing angles ranging from 0.05 to0.2°. The data were collected by usingan imaging plate system in the q range0.1 – 8.5 Å-1 with integration times inthe minute timescale. Curves A and Bare the 2θ scattering patterns afterintegration of the 2D data at the grazingangles for total external reflection at themetal-rich layer-substrate and at the air-substrate interfaces (positions A and Bin the reflectivity curve respectively).The profile A shows small diffractionpeaks from crystalline phases of themetallic clusters emerging from thesmooth amorphous SiO2 backgroundwhich disappear in the profile B. Afterthe subtraction procedure ten mainpeaks are clearly visible in the

difference signal coming from the fccphase of the Ag clusters. The dataquality is high enough to enablecrystalline phases identification, toretrieve lattice parameters and todetermine clusters mean sizes. Theresults are very useful in comple-menting information from x-rayspectroscopic (EXAFS) methods and«single particle» investigations withelectron techniques (TEM).

REFERENCES

[1] «Characterization of Cu Metallic Clustersin Amorphous SiO2 by Synchrotron RadiationGrazing Incidence X-ray Scattering andDiffraction» F. D’Acapito, D. Thiaudiere,F. Zontone, J.R. Regnard, Mat. Sci. Forum,Proceedings of the EPDIC-7 Int. Conf., inpress.[2] «Grazing Incidence X-ray Diffraction inthe Study of Metallic Clusters Buried in Glassobtained by Ion Implantation» F. D’Acapito,F. Zontone, J. Appl. Cryst. (accepted).

SYNCHROTRON RADIATION GRAZING INCIDENCE X-RAY

DIFFRACTION: A NEW TOOL FOR STRUCTURAL INVESTIGATIONS

OF ION-IMPLANTED GLASSESF. D’ACAPITO AND F. ZONTONE


Fig. 1: Extraction of the cluster richlayer signal with the new grazing

incidence x-ray method in the case ofan Ag implanted SiO2 glass:

the diffraction peaks coming from thecluster are maximized at the total

external reflection conditionfor the glass-clusters

layer interface (A case).They are not visible in the profile B.

In the difference profile ten mainpeaks (with indices) are clearly visible

related to the fcc phase of the Agclusters, the other coming from

unknown phases. The insert shows thereflectivity curve.

Inte

nsit

y [a

rb. u

nits

]

7 104

6 104

5 104

4 104

3 104

2 104

1 104

00

<20

0><

111>

<22

0>

<31

1><

222>

<40

0>

<33

1><

420>

<44

2>

<51

1/33

3>

2θ [deg]

A

B

A - BA

B

5 10 15 20 25 30 35

9000

8000

7000

6000

5000

4000

3000

2000

1000

Grazing angle [deg]0

1.2

1

0.8

0.6

0.4

0.2

0

Ref

lect

ivit

y

0.05 0.1 0.15 0.2

Metallic contaminants at thesurface of silicon wafer seriously hinderthe performances of integrated circuits.The present non-destructive method usedin production plants for measuringsurface contamination, Total ReflectionX-ray Fluorescence (TXRF) approachesits limits, while the quest for smaller andsmaller design rule in device technologyimposes lower and lower limits oftolerance. The National TechnologyRoadmap for Semiconductor publishedby Sematech, Inc. [1] has alreadyprogrammed the transition fromlaboratory-based analytical services tosynchrotron-radiation-based installations.In the early 90s the pioneering work ofmany groups opened the way to the useof synchrotron radiation to industrialTXRF applications [2, 3, 4]. At the ESRFa test experiment was performed a littlemore than one year ago [5] in order toverify the possibility of attaining thesensitivities envisaged by the SematechRoadmap while keeping the wafer in aconvenient horizontal attitude. Theoutcome of the test experiment actuallyshowed that the many orders ofmagnitude of increase in brilliance notonly permit the wafer to seat horizontally,but that mapping of the entire wafersurface is also possible in a reasonableamount of time. On the basis of thesefindings, a project financed by the ESRF,semiconductor companies, researchlaboratories and the EuropeanMEDEA** program was launched inJuly 1997 for the set-up of an industrialpermanent instrument.The TXRF facility will offer thepossibility to map the concentration ofcontaminants with an initial detectionlimit of 109 at/cm2 for 1999 to be pushed

down to 108 at/cm2 in the followingyears. The wafer size can be either 200 or300 mm and the pixel size about one cm2.The project envisages also the detectionof low Z elements at least down tosodium, thus requiring excitation beamsbelow the silicon K threshold.

OPTICS

The key issues that have been takeninto account for the definition of theoptical layout of the installation canbe summarized as follows:• Need of a wide energy range from1 keV to 40 keV for optimizing thedetection sensitivity for every element. • Need to reject as much as possible theharmonic content of the x-ray beam tolimit detection problems with standardsolid state energy dispersive detectors.

• Need to preserve the polarization andcollimation of the beam for maximumrejection of the elastically scatteredradiation at the detector.• Need of a monochromator with fixed-exit beam to avoid the re-alignment of thewafer position at each change of energy.

THE PHOTON SOURCE

The best way to reject harmonics is, ofcourse, to start without harmonics. TheInsertion Devices Group analyzed thepossibility of manufacturing an aperiodicundulator that will have the «harmonics»well displaced from where they should beif they were normal harmonics. Fromtheir calculation the residual harmoniccontent of the undulator that is presentlyunder construction is of a few per cent [6].A further reduction of the harmonic

INDUSTRIAL FACILITY


28

ID27: AN INDUSTRIAL TXRF FACILITYF. COMIN, M. NAVIZET, P. MANGIAGALLI AND G. APOSTOLO


The ESRF-TXRF facility will soon provide silicon manufacturing companies with an analytical centralized

TXRF service capable of detection limits for contaminants at the wafer surfaces below 108 at/cm2.

Total Reflection X-Ray Fluorescence(TXRF) analysis is a powerful tool forthe identification and quantification ofchemical elements at the fentogramlevel. TXRF is based on theconventional energy dispersive X-RayFluorescence (XRF) analysis technique.The identification of the differentelements is obtained by exciting theatoms of the sample with an x-ray beam.Each excited atom will emit afluorescence photon with a well-definedcharacteristic energy, which is thesignature of its atomic species. Thequantitive analysis of the fluorescencespectrum with an energy dispersivedetector allows the identification andquantification of the elements present inthe sample. TXRF differs from standardXRF only in the particular excitationgeometry: the x-ray beam impinges onthe sample with a very small glancingangle. Under these conditions the beampenetrates the sample only by a few tensof angströms so that only the shallow

region close to the surface is probed.This excitation geometry simul-taneously reduces the background andenhances the sensitivity of thetechnique, which is now perfectlysuited for the analysis of surfacecontamination. The combination of theTXRF technique with synchrotronradiation x-ray beams allows detectionlimits of the order of 108 at/cm2 fortransition metals, which means that aforeign impurity atom can be identifiedon the wafer surface among ten millionsilicon atoms.

αi < αc

Fluorescencedetector

THE TXRF TECHNIQUE

** Since 1997, the program «Micro-ElectronicsDevelopment for European Applications» is meantto assist the European semiconductor industry indeveloping novel techniques. MEDEA has takenover from JESSI, the «Joint European SubmicronSemiconductor Initiative», which ran out in 1996after 8 years of operation.

INDUSTRIAL FACILITY


29

content will be achieved at themonochromator by detuning the Sicrystals and by an appropriate choiceof multilayers.

THE MONOCHROMATOR

For maximum reliability the x-raymonochromator is based on a singlerotation motion following well-established concepts. The dispersingelements are a silicon (111) channel-cutfor the energy range 2.8 – 40 keV and amultilayer pair for the low energy range[7]. The high brilliance of the ESRFleaves the possibility of using the sharpSi(111) reflection without deteriorationof the TXRF performances. This allowschemical recognition of the impurityatoms by collection of XANES spectraaround the threshold of the atomicspecies of interest.

THE TXRF SET-UPThe TXRF end-station is located in ashielded hutch of about 25 m2 embeddedin a class 100 laminar flow. The stationencompasses: a loading robot thattransfers the wafers from standardcassettes to a load-lock vessel; a highvacuum handler that transfers the waferfrom the interlock loading chamber to themeasuring position in the main TXRFchamber, a set of two 7-element lineararrays of Si(Li) detectors for the parallelacquisition of the fluorescence spectra,and a wafer chuck/manipulator for thecorrect positioning of the wafer onto thex-ray beam and for rotating the waferaround its vertical axis in order to performthe impurity mapping over the wholesurface of the wafer.The 14 channels of the detector are splitinto two independent arrays forservicing/reliability purposes.Due to the grazing incidence, the beamilluminates a strip on the sample, theuse of 14 independent detectorchannels, each of which looks to a10.5 mm wide portion of this stripallows an easy mapping of the entiresurface of the wafer just by rotating thesample and collecting data in parallelfrom the detector array.This geometry greatly simplifies themechanical requirements for the waferhandling, but implies that either thespatial resolution or the sensitivityfor the outer portion of the waferis somewhat lower than for the centralone.

TIME SCHEDULE

The facility will start its commissioningphase early in 1999 with the first testson real 300 mm wafers. This will lead toa field assessment of the specificationsattained and further tests are alreadyenvisaged to ascertain which otheropportunities can be exploited with theinstrument. Two directions seemparticularly interesting, namely thepossibility of performing XANESspectroscopy in selected atomic speciesfor the chemical identification of thecontaminant and the detection ofcarbon species via electron spectroscopytechniques.

REFERENCES

[1] http://notes.sematech.org/97melec.html.[2] A. Iida, Adv. In X-ray Anal., 35 (1992) 795.[3] R. Rieder et al., Nucl. Instr. and Meth. In Phys.Res. A355, 648-653 (1995).[4] P. Pianetta et al. Rev. Sci. Instrum. 66, 1293(1995).[5] L. Ortega et al., Journal of Synch. Radiation,5, 1064-1066 (1998).[6] J. Chavanne et al., EPAC 98 conference,Stockholm.[7] F. Comin et al., SPIE Proc. Intern. Symp. OnOptical Science, Eng., and Instr. San Diego, July(1998).

MACHINE DEVELOPMENT


30

F

A THIRD ACCELERATION UNIT

FOR THE ESRF STORAGE RINGC. DAVID, J. JACOB, A. PANZARELLA, J.P. PERRINE AND J.L. REVOL

ESRF, MACHINE DIVISION

After only two years of design, construction and commissioning, a new acceleration unit is now in

operation on the ESRF storage ring. It allows all klystrons and cavities to be run at moderate power

even for high intensity operation at 200 mA, whilst still gaining in total accelerating voltage,

from 8 to 12 MV. The third pair of cavities can also be used to modulate the radio-frequency voltage

at the revolution frequency, thereby producing additional damping of longitudinal multibunch

oscillations. The planning and commissioning constraints imposed by the installation of this equipment

on a running machine was one of the challenging issues. The project was managed so as

to minimize the inconvenience for ESRF users.

our 352.2 MHz five-cell cavities andtwo 1.0 MW Radio Frequency (RF)transmitters (upgraded to 1.3 MW inJanuary 94) were initially installed onthe ESRF storage ring for an operationat the design current of 100 mA at6 GeV. During the commissioning ofthe storage ring, it was already possibleto store higher currents and 200 mA ofbeam current have been delivered to theusers in multibunch operation sinceNovember 1995. This resulted inpushing the power through the cavityinput couplers to very high values andrequired operating the storage ring withboth RF transmitters, each feeding onepair of cavities with up to 750 kW. In early 1995 it was decided toconstruct a third RF transmitter(SRRF3) feeding a third pair of cavities[1]. Besides the reduced power load forthe cavity couplers and a gain inredundancy, this third RF unit opens upother interesting new features.

DESIGN ANDCONSTRUCTION

Based on experience gained from theexisting RF units, it was decided toconstruct the third RF unit using similarhardware for the large components.This includes the klystrons, the highvoltage power supply, the circulator,parts of the low level RF and the five-cell cavities [1, 2, 3, 4]. The 100 kV - 22 A DC power supply forthe third klystron was ordered inSeptember 95 from the constructor andmounted and cabled under ourresponsibility: it was a very successful

approach both in terms of cost savingsand transfer of expertise to the RF crew.It also eased the implementation of animproved local control and of specialmeasures against interference with theremaining RF equipment. The high-voltage power supply was ready asscheduled at the end of February 97.A new control system was developedfor SRRF3 and commissioned inparallel with the hardware [5]:• The object-oriented programmingapproach based on the separationbetween the transmitter and the cavities,which are themselves split into basicdevices, proved to be very versatile forthe development.• The hardware protection is managedindependently from the control softwareby a PLC for the slow interlocks and ahardwired system for the fastprotections.• The diagnostics tools included in thedesign are very useful for trip analysisand parameter follow up.• The operator controls the RF system viaan easy-to-use graphical user interface. Some sensitive parts of the RF system,such as the arc detection system, havebeen redesigned in order to reduce therate of spurious triggering. Also thecavity vacuum equipment has beenimproved: • The fast pressure interlock system,which is essential for the ceramicsprotection against glow discharges withsubsequent sputter deposition of copper[3], has been doubled (two fast pressuredetectors per cavity). • The NEG pumps of the old designhave been replaced by titaniumsublimation pumps. This proved to be

very efficient during RF conditioningand commissioning with beam, as it waspossible to activate these pumpsregularly, so that the vacuum couldrecuperate quickly after a strongoutgassing.

COMMISSIONINGAND OPERATION

The new SRRF3 transmitter was readyfor commissioning with beam on thescheduled date in August 97. After onlyfour days, 160 mA could be stored andwere then delivered for the first week ofuser service mode. The maximumESRF current of 205 mA was reachedone week later and served routinely inthe 4th week of User Service Mode.The failure rate was higher during thecommissioning period and the beamavailability was reduced (90% insteadof the usual 95%). After only threemonths of operation, the mean timebetween failures was recovered and issteadily increasing.During the three first months ofoperation, a lot of trips were caused bybursts of reflected power from thecavities, which were typically 50 µslong and as high as 300 kW at the peak.These events were mostly associatedwith outgassing and were probably dueto fast detuning provoked bymultipactor in one of the five cells.Continuous operation of the cavitieswas the only cure for these trips whichsteadily became less and less frequent.In the early stage of commissioning, wewere also confronted with highharmonic power generated by the

klystron. The harmonics are fed back tothe klystron by the reflection from thecirculator, which is only matched at thefundamental RF frequency. Carefuladjustment and positioning of matchingelements (iris, post) allowed theharmonic power to be brought to anacceptable level.A part of the early failures wereremoved gradually during the first threemonths simply by debugging thesystem. The high voltage power supplyproved to be very reliable.

NEW POSSIBILITIES

More flexibility and larger safetymarginThe fact that the maximum ESRF beamcurrent of 205 mA can be stored withoutSRRF2 represents a large gain inflexibility and security. Klystronacceptance tests, cavity couplerconditioning and extensive R&D workrequiring RF power can now beperformed without restriction for theESRF users.It is planned to extend the waveguidenetwork such that SRRF2 can replaceeither SRRF1 or even the booster RFsystem. By doing this, we will obtain alevel of redundancy that will allow theoperation of the storage ring to besafeguarded even in case of a majorfailure on any single transmitter.

Increase in cavity voltageThe third pair of cavities has allowedthe cavity voltage to be raised from 8 to12 MV with only a negligible increasein total power and still a reduction of thepower per cavity coupler. However, itturned out that the associated gain inlongitudinal acceptance did not show upon the lifetime which is presentlylimited by the transverse acceptance.Increasing the transverse acceptance ofthe storage ring is therefore one topicof machine physics studies to fullybenefit from the enlarged longitudinalacceptance.Since in single-bunch and 32-bunchoperation the lifetime is also Touscheklimited, the optimum voltage was foundto be close to 8 MV, for which thebunches are longest and the energyacceptance is not yet limited by the RF.

Landau damping** of multibunchinstabilities with cavities 5 & 6For standard high intensity operation,

only 2/3 of the storage ring is filled toproduce strong transient beam loading,which leads to a voltage modulation atthe revolution frequency f0. Thesubsequent spread in synchrotronfrequencies provides Landau damping oflongitudinal multibunch instabilitieswhich are driven by Higher OrderModes (HOM) developing in thecavities, a point of major importance forthe ESRF [6]. An additional feature ofthe third RF system is the possibility tooperate it at fRF + f0, such as tomodulate the total RF voltage activelyeven for symmetrical fillings [7].The detuned fundamental modes ofcavities 5 & 6 can be used intentionallyto launch an n = 1 multibunchinstability. The threshold current isincreased by a factor 4 when applying amodulation voltage of 1.5 MV. Thisscheme is employed for operation with32 equally-spaced bunches in thestorage ring. In this operation mode theoptimum voltage of about 8 MVrequires switching off cavities 5 & 6. Asthese are no longer temperatureregulated, modulation is necessary tooperate untroubled by multibunchinstabilities at the nominal intensity of90 mA.

Basis for an upgrade of the existingtransmittersAfter the successful commissioning ofSRRF3, the new control system is nowfully validated. From the high UNIXlevel down to the hardwired fastinterlock system, it had been designedtaking into account the necessity ofupgrading the existing RF units. Thanksto the existence of SRRF3, this work isnow possible in parallel to operating thestorage ring at full intensity. Based on the clear separation betweencavity and transmitter control, we areable to proceed in two steps. We will firsttransfer the cavities 1 to 4 from the initialto the new control system. This will takeplace in the next October shutdown. Theupgrade of the transmitters will then takeplace one by one.

CONCLUSION

The third RF acceleration unit wassuccessfully built and put intooperation. It was fully designed byESRF staff in order to best match theoperation constraints and specificities,and to fit into the ESRF environment,taking into account the experience

gained on the existing RF units.Performing such a design andconstruction work was very beneficialfor broadening the knowledge andexpertise of our RF group personnel.The project was completed within itsbudget and on schedule, and wasmanaged so as to minimize theinconvenience for ESRF users. Thiswas a challenging issue for a machinewhich delivers 5600 hours of x-raybeam per year.

REFERENCES

[1] J. Jacob et al., «Construction of a Third RFAcceleration Unit for the ESRF Storage Ring»,EPAC’96, Barcelona, June 1996.[2] J. Jacob et al., «Commissioning andOperation of one Booster and two StorageRing RF Acceleration Units at the ESRF»,EPAC’92, Berlin, March 1992.[3] J. Jacob et al., «RF System Developmentfor High Current Accumulation in the ESRFStorage Ring», EPAC’94, London, June 1994.[4] J. Jacob et al., «Commissioning of theThird RF Acceleration Unit for the ESRFStorage Ring», EPAC’98, Stockholm, June1998.[5] J.-L. Revol et al., «An Object OrientedControl System for the Third Storage Ring RFUnit at the ESRF», EPAC’98, Stockholm, June1998.[6] O. Naumann and J Jacob, «FractionalFilling Induced Landau Damping ofLongitudinal Instabilities at the ESRF»,PAC’97, Vancouver, May 1997.[7] O. Naumann and J. Jacob, «LandauDamping of Longitudinal Instabilities for theOperation of the ESRF Storage Ring»,EPAC’98, Stockholm, June 1998.

** The Landau damping decreases the energyspread of the beam induced by coupled bunchoscillations and increases the bunch length (whichalso means the Touschek lifetime).

MACHINE DEVELOPMENT


31

ACKNOWLEDGEMENTS

The authors wish to thank thenumerous colleagues from theTechnical Services, the ComputingServices, the Safety Group and theMachine Division for theircontribution to the success of theSRRF3 project, and particularly:P. Barbier, D. Boilot, B. Boucif,M. De Donno, H. Delamare,M. Dubrulle, G. Gautier, P. Kernel,J.M. Mercier, J. Meyer, N. Michel,O. Naumann, J.L. Pons, E. Rabeuf,D. Vial.

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

Editor: Dominique CORNUÉJOLS Tél (33) 4 76 88 20 25Dépôt légal : 3eme trimestre 1998. Printed in France by Repro-Express. Layout by Pixel Project.

EventsEvents HERCULES 1998(See article on page 4)

NEXT HERCULES COURSE

21 February - 1 April 1999

Deadline for application: 16 October 1998 • e-mail: [email protected]

NEXT

USERS’ MEETING11-12 February 1999

(See article on page 5)

SEPTEMBER 1998 EUROPEAN SYNCHROTRON RADIATION … · «Medical imaging with synchrotron...

Documents

Transcript of SEPTEMBER 1998 EUROPEAN SYNCHROTRON RADIATION … · «Medical imaging with synchrotron...