ESS/SAC/Report/1/01

ESS – SAC/ENSA Workshop

on

“Scientific Trends in Condensed Matter Researchand Instrumentation Opportunities at ESS”

edited by

D. Richter(Chairman of SAC)

jointly organized by

the Scientific Advisory Committee (SAC) of the European Spallation Source(ESS) and the European Neutron Scattering Association (ENSA)

sponsored by

European Science FoundationSwiss National Science FoundationEuropean Neutron Round TablePaul Scherrer Institute VilligenForschungszentrum Jülich

Address:Institut für Festkörperforschung (IFF)Forschungszentrum JülichD-52425 JülichGermany

Tel.: +49-(0)2461-61-2499Fax: +49-(0)2461-61-2610IFF Homepage: http://www.kfa-juelich.de/iff/Institute/ins/ins-d.shtmlESS Homepage: http://www.ess-europe.de

ESS – SAC/ENSA Workshop

“Scientific Trends in Condensed Matter Research andInstrumentation Opportunities at ESS”

Engelberg/Switzerland, 03. – 05.05.2001

Contents

Preface (D. Richter, SAC Chairman) I

Preface (R. Cywinski, ENSA Chairman) II

Executive Summary IV

1. Introduction 1

2. Science and Instrument Groups 3

3. The ESS - SAC/ENSA Workshop on“Scientific Trends in Condensed Matter Research and Instrumentation Opportunities at ESS”

4

4. The Long Pulse Target Station 8

5. Recommendations of the Scientific Advisory Committee 10

6. Instrumentation issues, established techniques, need for innovation, technical risks(F. Mezei, HMI Berlin)

13

7. Science Group Reports

7.1 Solid State Physics 34

7.2 Material Science and Engineering 45

7.3 Biology and Biotechnology 56

7.4 Soft Condensed Matter 69

7.5 Chemical Structure, Kinetics and Dynamics 80

7.6 Earth Science, Environmental Science and Cultural Heritage 91

7.7 Liquids and Glasses 107

7.8 Fundamental Physics 119

8. Appendix

A1 Members of the Science Groups 126

A2 Members of the Instrument Groups 127

A3 Participants at the SAC/ENSA Workshop in Engelberg 128

9. Acknowledgments 130

I

Preface

This document reports on a decisive step towards the final definition of the European

Spallation Source (ESS). At the beginning of May 2001 in Engelberg (Switzerland), about 70

scientists from all fields of neutron science encompassing solid state physics, material

science and engineering, biology and biotechnology, soft condensed matter science,

chemistry, earth and related sciences, the science of liquids and glasses and the physics of

the neutron itself gathered in order to deliberate upon the optimum choice for the neutron

parameters of ESS. The course of consideration went through a two stage process. Initially,

the scientists identified high profile research areas in their scientific fields at the limit or

beyond of what may be accessed today. From these so called flagship areas scientific

demands on the performance of ESS were derived. Then, these demands were discussed in

the frame of the instrumental opportunities offered by different configurations of ESS. This

part of the discussions was guided by instrument specialists representing the different fields

of instrumentation. From the synthesis of science considerations and instrumental

opportunities a recommendation for the neutron parameters of ESS was reached.

The report documents the scientific and instrumentation background which led to the

recommendations for the choice of target stations and power levels at the ESS accelerator

complex. The document emphasises strongly the science case considerations and in

particular the high profile flagship areas which are behind the conclusions on the neutron

parameters. With respect to the science case of ESS, this document is an important but

intermediate step which will need to be completed adding information on the breath and the

width of neutron research on science in general. Furthermore, the opportunities offered by

innovative and novel instrumentation that will become possible at ESS, need to be outlayed

in more detail. This next and final step will be completed within the next year. The resulting

report will underpin the ESS project proposal and is aimed to be published together with

European Science Foundation.

I like to take the opportunity to thank all participating scientists for their extraordinary

engagement and enthusiasm that became evident during the three intense working days in

Engelberg. I also like to thank the conveners of the science groups for their efforts to

complete the science group reports within a very short time after the workshop, allowing us

to go into print less than one month after the event.

Prof. D. Richter(Chairman of SAC)

Jülich, 05.06.2001

II

A Preface from the Chairman of ENSA

Neutron scattering is a uniquely sensitive and extensivelyused experimental technique for studying the structure anddynamics of matter. However it is a technique that is strictlylimited by the intensity of available neutron sources. Indeedsince the pioneering neutron scattering studies of the NobelPrize winners Shull and Brockhouse in the early 1950’s thissource intensity has increased only by a factor of 4, anincrease achieved almost 30 years ago with thecommissioning of the world’s premiere neutron researchreactor, the Institut Laue Langevin in Grenoble.

Over the intervening decades the scientific and technologicalproblems associated with the fine analysis of matter thatdepend crucially upon neutron scattering for their solutionhave grown enormously in complexity, subtlety and range. Asa response to this burgeoning demand for ever more sensitiveneutron scattering methods the European Spallation Source(ESS) Project was established in the early 1990s as the mostdirect and feasible approach to providing a new thirdgeneration neutron source. It is intended that the ESS willoffer an effective increase in performance of between 10 and100 over all existing neutron sources, depending upon thespecific application. This achievement will represent by far thegreatest single stepwise increase in source performance sincethe early 1950’s. The implications for neuron scatteringscience will be profound.

By 1997 the scientific case for, and the technical design of, theESS was published [1]. Indeed the ESS design had reachedsuch an advanced state that the United States adopted theprincipal design features of the ESS as the basis for their ownSpallation Neutron Source (SNS) project. The construction ofthe SNS commenced in 1999, but unfortunately there hasbeen no parallel commitment to progress the ESS withinEurope.

The overriding need for a third generation neutron source inEurope has been emphasised within the global neutronstrategy compiled by the OECD Megascience Forum [2] in1998 and endorsed by the OECD Ministerial Conference in1999. The OECD recommended that an advanced Europeanneutron source should stand alongside complementary thirdgeneration neutron sources in America and Asia. Both the USand Japan have adopted these recommendations and arenow, in 2001, well advanced with the design and constructionof their own advanced regional sources (the SNS and JSNSrespectively).

It is the goal of the ESS project to provide the scientific casefor and design of a neutron spallation source which will satisfythe OECD global neutron strategy and provide Europe with atop ranking neutron facility which will remain at the forefront ofneutron scattering technology for at least the first half of the21st century.

Neutron scattering,although a uniquelysensitive probe ofcondensed matter, isstrictly limited by neutronsource intensity

The ESS represents morethan an order ofmagnitude improvement insource performance, anachievement unequalled insince the 1950s

The OECD has endorsed aglobal neutron strategywhich places a thirdgeneration neutron sourcein Europe, America andAsia. The US and Japanhave already commncedconstruction

The ESS will be the worldleading neutron facility forthe first half of the 21st

century

III

The ESS project has been strongly and vociferously supportedby the European neutron scattering community, through itsrepresentive body, the European Neutron ScatteringAssociation, ENSA. ENSA represents, via the electeddelegates from 17 European national user communities,approximately 4500 physicists, chemists, material scientistsand members of the life, earth and engineering sciencecommunities, all of whom view neutron scattering as anessential tool in their exploration and exploitation of thestructure and dynamics of condensed matter.

Although a close and effective collaboration between ENSAand the ESS Project has been in place for some time, the ESSScience Advisory Committee/ENSA workshop held inEngelberg between 3 and 5 May 2001 proved remarkablysuccessful not only in defining an exciting and far reachingscientific case for the ESS, but also in building upon thisscientific case in order to establish the optimal neutronicperformance parameters for the ESS target stationsnecessary to enable the ESS to address the ambitiousscientific challenges outlined in the scientific case.

This report, prepared through an active collaboration betweenthe ESS and ENSA presents a detailed discussion of therefined scientific case and its implications for the ESS targetstations as established at the Engelberg Workshop.

Prof. R. Cywinski(ENSA Chairman)University of Leeds, UK

ESS has been stronglyendorsed by themultidisciplinary neutronuser communityrepresented by ENSA

Collaboration between theuser base and the ESS hasnow defined the scientificcase for ESS andestablished the optimalneutronic parameters

References

[1] ESS – A Next Generation Neutron Source for Europe, Vol. 1-3, ISBN 090 237 6500, ISBN 090 237 6551(1997)

[2] Final Report OECD Megascience Forum Working Group on Neutron Sources (1997),www.OECD.ORG//dsti/sti/s_t/ms/prod/online.htm

IV

Executive Summary of the ESS-SAC/ENSA Workshop on “ScientificTrends in Condensed Matter Research and InstrumentationOpportunities at ESS”

This document is addressed primarily to the Council of the European Spallation Source(ESS) project and contains science based advice for selection of the design parameters ofthe ESS. It also reaches out to a wider scientific public by presenting the scientific case forthe ESS, exemplified by high profile flagship research areas which are coupled withinstrumentation opportunities. The document reports the results of a combined effort bytopical science groups covering all science fields that will benefit from the ESS, and byinstrumentation experts who evaluated the performance of generic instrumentation at thisthird generation neutron source. This culminated in a joint workshop on “Scientific Trends inCondensed Matter Research and Instrumentation Opportunities at ESS”, which wasorganized by the Scientific Advisory Committee (SAC) of the ESS and the European NeutronScattering Association (ENSA). The eight science groups discussed the key areas ofscientific research where ESS will be essential for future progress. The richness and depth ofthe scientific opportunities that will be opened up by ESS cannot possibly be represented inan executive summary. Instead we will give one brief example from each science groupreport.

Solid State Physics. Advances in solid state physics are at the root of most of thetechnologies that shape today’s world. Neutrons are a key to our understanding of solids, sothe ESS will have a large impact on cutting-edge research in solid state physics. One exampleis molecular and organic magnets i.e. solids build from structurally well defined clusters ofmagnetic ions in a complex environment. Such systems are of fundamental importance andcould also serve as atomic scale information storage systems. Neutron scattering is a uniqueprobe for their magnetic characterization and for studying the excitation spectra whichdetermine their properties.

Material Science and Engineering. The next generation of high power neutron spallationsources, like ESS, will allow us for the first time to investigate materials in real time with realisticdimensions and under real conditions. One example is the deformation of materials and theunderstanding of the mechanisms involved, which is a vital part of engineering science. ESSwill enable the assessment of real scale components on realistic time scales. For example, newsolid state joining techniques require more accurate information about the generation ofresidual stresses that will add to in-service stresses and shorten component life. Furthermore,finite element modelling has become the main method for the design and assessment ofengineering structures. Such models cannot be developed reliably without accurate informationto validate them. Neutron diffraction is the only technique that can do this, providingmeasurements deep inside most engineering materials.

Chemical Structure, Kinetics and Dynamics. Chemists respond to the present demandsfor higher performance materials, cleaner environments and improved efficiency in the use ofchemicals, in a wide variety of ways. This includes use of smart materials that respond to theenvironment, use of thin films to build devices, and exploitation of pharmaceuticals and otheragents such as catalysts that are active in much smaller quantities. One example is thedevelopment of molecular materials with useful and tuneable physical properties such asmagnetism, superconductivity, nonlinear optical activity, polymorphism, etc. Much of this workis focused on understanding the intermolecular interactions that hold 3-D arrays of moleculestogether, often weak hydrogen bonding interactions. In supramolecular chemistry, accuratequantification of weakly bonded motifs will allow for more rational crystal engineering, enablingchemists to tailor properties by designing structures, e.g. of pharmaceuticals. ESS will advancethe expanding area of molecular science by allowing characterization of all atoms in suchstructures, including the often crucial hydrogen atoms.

Soft Condensed Matter. Space time resolution at proper scales, variation of contrast andhigh penetrability, qualify neutrons as a unique tool for studying the structural and dynamicalproperties of soft matter at a molecular level. One of the great challenges of basic soft

V

condensed matter science is the development of a molecular rheology, i.e. an understanding ofmechanical and rheological properties on the basis of molecular motion. Such a developmentwould also facilitate the molecular design of new materials. This task requires space timeresolution on widely differing length and time scales, together with the selective observation ofkey components. Neutrons are uniquely suited to achieve these goals. Advances ininstrumental techniques and flux are required, far beyond the present state of the art.

Liquids and Glasses. Neutrons are a key probe for the study of the atomic structure anddynamics of liquids and glasses. A future scientific highlight will be the detailed investigationand understanding of the influence of molecular entities on the solvent structure in solutions. Anamazing fact is that some ion combinations or molecular species in solution can e.g. induceprotein folding while others cause protein denaturation. There is little doubt that an importantpart of this behavior arises from the effect of the ions or other molecular species on thestructure of the water surrounding the macromolecule. Neutron diffraction already provides asensitivity which is unparalleled by any other technique. The ESS will enhance this capacityeven further, allowing the aqueous environment of larger molecular entities, of interest tochemistry and biology, to be examined for the first time.

Biology and Biotechnology. Neutrons have a unique role to play, when information onhydrogen atoms - their positions, hydrogen bonds, the role of water, hydrogen motions - as wellas contrast variation on larger scales and dynamical features in general, are a focus of interest.Unfortunately, present neutron intensities are too low to facilitate progress in a broad sense.ESS will change this picture in many respects. One example is membrane biophysical studiesvia native state reflectometry, with impact on biosensors and nanobiostructures. Such biochipswill become a crucial technology for many applications and diagnostics, and also proteomics.The enabling technology that will emerge from this research will provide the tools for findingmolecular markers for early stage detection of illnesses, and for the unrevealing of the humanproteon. Neutron data on such complex systems can become a prerequisite for the design ofeven more advanced combinations of biological matter with solid surfaces for biochips,including biosensors.

Earth and Environmental Sciences and Cultural Heritage. Neutron scattering has onlyrecently been added to the portfolio of methods used in earth sciences, thanks mainly to thelatest generation of diffractometers and spectrometers at the most modern neutron sources.However, many areas of earth science research remain out of the reach of present day neutroninstrumentation. Among them is one of the most significant issues in the earth sciences ,related to the prediction of earth quakes and volcanic eruptions. The reliability of the relevantmodels crucially depends on knowledge of the physical and chemical properties of thematerials involved (oceanic crust, upper mantl, continental crust). Foremost among theseproperties are the role of water in such materials and the behaviour of related magmas. Frontierapplications of an ESS class neutron source will be in the field of in-situ studies where mineralstructures and material behaviours are investigated under extreme temperature and pressureconditions simulating the real conditions deep in earth.

Fundamental Neutron Physics. Neutrons appear both as composite particles and asquantum waves. Both features have been investigated with thermal, cold and ultra coldneutrons at many neutron sources. The significantly higher intensity and the pulse structure ofthe ESS will provide new possibilities for fundamental neutron physics experiments. Onequestion concerns the left handedness of the universe. At present most grand unified theoriesstart with the left-right symmetric universe and explain the evident left handedness of naturethrough a spontaneous symmetry breaking caused by a phase transition of the vacuum, ascenario which would mean that the neutrinos today should carry a small right handedcomponent. Looking at the decay of a neutron into a hydrogen atom could provide a yes-noexperiment, since one of the four hydrogen hyperfine states cannot be populated at all if theneutrinos are completely left handed.

The future scientific endeavours identified by the science groups place high demands on thecapabilities of instrumentation at the ESS. The instrument task groups provided performancecalculations for generic instrumentation for the target stations considered for ESS: a 50Hz,

VI

5MW short pulse target station, a 162/3Hz, 5MW long pulse target station and a 10Hz 1MWshort pulse target station. These conservative estimates predicted gain factors up to 3 ordersof magnitude in count rate compared to the best-in-class instruments of today. With theperformance data at hand, the science groups mapped their science demands on theinstrumental opportunities provided by the ESS. The target station evaluation led to clearpriorities. Most demands were for the 50Hz short pulse target station, while a significantnumber of requests - particularly strong from the broad field of soft condensed matterresearch - were for a long pulse target station. The priority for the short pulse 10Hz targetstation was low. In addition the fundamental physics group requested an ultracold neutron(UCN) factory, implying a third dedicated target station for UCN research.

The Scientific Advisory Committee suggests following this line and arrived at the followingrecommendations:

• ESS should be built as a 10MW facility serving two target stations, a short pulse 50Hzstation and a 162/3Hz long pulse station.

• Both target stations should be considered with equal priority. This recommendation isbased on the large potential for innovation at the long pulse target station.

• The SAC asks for an optimization of the long pulse station and the development ofinnovative instrumentation for this new tool of neutron research.

• An instrument suite should be defined for both target stations, in order to allow for anoptimization of the moderators.

• As a long term project, a methane or methane-like moderator should be developed forthe short pulse target station.

• The target group should be asked to estimate the cost of a dedicated UCN target station.

• Finally, an acceleration of the project time scale is requested which should culminate in apresentation of the ESS to the scientific and political public in May 2002.

1

1. IntroductionAt its meeting in Berlin on May 24, 2000, the EuropeanSpallation Source Council proclaimed that

“The objective is to design and construct a European nextgeneration spallation source, that upon completion will be thebest neutron source worldwide for all classes of instruments.”

and made a commitment to deliver a fully costed ESS projectproposal to the European Governments early in 2003.

Since the 1997 ESS proposal, the ESS project has developedin several new directions encompassing the old ESS withthree mayor new features.

• In addition to the short pulse target stations, the ESS isnow investigating the scientific prospects for a long pulsetarget station.

• Progress in superconducting accelerator technology hashad an impact on the LINAC design and is being seriouslyconsidered.

• Advances in many areas of the multidisciplinary sciencefield of ESS have led to an extension of the science case.

This report deals with the first and third points which areintimately related. The Science Case and the demands ofscience should determine the ESS design. Therefore, it wasfelt necessary that an urgent but thorough re-examination ofthe original 1997 ESS scientific design should be instigated,particularly in light of the significant advances in many areasof the scientific and technological base to which the ESS isdesigned to contribute.

The original scientific case for the ESS design has therefore.been reviewed in close consultation and collaboration with themultidisciplinary neutron user community and the original ESSdesign has undergone further refinement.

New features of ESSproject: long pulse stationsuperconducting LINACextended Science Case

These developments, which have taken place over severalmonths up to and including the Engelberg SAC/ENSAworkshop constitute the first of several milestones defined bythe ESS council in pursuit of its primary objectives, i.e. that ofproviding a complete specification of the neutron parametersof ESS, i.e. the power level, the repetition frequencies and theproton pulse lengths at the target stations to be fully definedand fixed by July 2001. It should be noted that theseparameters have a major impact on the whole of the ESSproject design and must be based upon scientific demands ofthe user community.

In 1993 the philosophy for fixing the neutron parameters forthe ESS reference design was based on two considerations.(i) The overall power level was chosen such as to equalize

the average flux emitted by a coupled moderator with thatof the most advanced continuous neutron source, the ILL:

Milestone I: Specificationof neutron parameters byJuly 2001

2

This consideration specified the overall proton beampower to 5MW.

(ii) The proton beam power should be delivered in shortpulses at the lowest repetition rate commensurate with5MW power delivery. This consideration led to the 50Hztarget station with µs-proton pulses drawn fromaccumulator rings.

The impacting proton beam power within one short pulse islimited by the mechanical strength of the mercury vessel of thetarget station which has to withstand shockwaves caused bythe microsecond pulses. From an engineering point of view100 KJ/pulse is close to or even beyond the maximum thatcan safely be taken by the target materials. The proton pulseenergy then defines the neutrons intensity in the pulse. Forcomparison, the SNS and the JNS project operate with 30-40 KJ/pulse.

The frequency at which a target is operated defines theneutron bandwidth available for an instrument. From that pointof view low repetition rate targets have an advantage.Therefore, in parallel to the 50Hz target, the ESS referencedesign considered an additional 10Hz short pulse target,where special emphasis is placed on cold neutroninstrumentation. However, due to the shockwave limit such atarget cannot increase the power i.e. the number of neutronsper pulse compared to the 50Hz target station. Thus, such a10Hz station dilutes the average power compared to the 50Hzstation. It offers space for more instruments but at the ESSpower level does not provide novel opportunities which couldnot be accomplished at the 50Hz target station.

100KJ/pulse is themaximum power at a shortpulse target

In order to overcome the short pulse limit of 100 KJ/pulse, theshort pulse paradigm has to be abandoned and long pulsetarget stations have to be considered. In the case of the ESSsuch a long pulse target station could deliver 5MW protonpower at 162/3Hz, if 2.5ms pulses from the LINAC are drawndirectly into the target station. Compared to the short pulses,in this case the number of neutrons per pulse increases by afactor of 3. It is conceivable that following an optimisationprocedure the neutron flux available from the moderatorscould increase by further factors of 2-3 compared to the shortpulse target station.

Long pulse target are ableto provide more neutronsper pulse

The final definition of the ESS layout must be based on thescience demands placed on ESS by the users. In thefollowing, this report describes the approach taken by theScientific Advisory Committee (SAC) of the ESS, in order toarrive at such a science based definition of the ESS neutronparameters. For that purpose an analysis of future trends indifferent science fields of ESS are necessary, i.e. sciencecase deliberations for ESS needed to be performed. Thereby,special emphasis was placed on important and decisive areasof science, so called flag ship areas, where ESS is expectedto impact strongly in solving problems which cannot beaccessed presently. This exploration of future trends in thescientific fields of ESS led to general science case

Science driven definitionof ESS neutronparameters

3

considerations.

In the following chapters this report will describe the proceduretaken in order to arrive at a recommendation for the ESStarget station, it will describe the outcome of the SAC/ENSAworkshop in Engelberg, it will in particular emphasize aspectsof the long pulse target station it will present the science caseconsiderations and not the least it reports the SACrecommendations following from these considerations.

2. Science and Instrument GroupsIn order to investigate the scientific opportunities which areconnected with the different target options, the SAC hascreated 8 science groups in the disciplines important for ESS,which were convened by SAC members. In parallel, theinstrument task under F. Mezei (HMI Berlin) has assemblednine instrumentation groups in order to study the performanceof generic instrumentation at the different target stations.

Science groups were established, in order to cover thefollowing fields (Table 1):

Table A1 in the appendix presents the members of all sciencegroups. In a similar way, instrument groups were establishedcovering the following instrumentation fields (Table 2).

Table 1:Science Groups and their Conveners

Science Group Convener/s

Solid State Physics A. Furrer PSI VilligenC. Vettier ILL

Material Science and Engineering H. Zabel Univ. of BochumT. Lorentzen Danish Stir Welding Techn.

Biology and Biotechnology J. Helliwell Univ. of Manchester

Soft Condensed Matter J. Colmenero Univ. of the Basque Country& DIPC

D. Richter FZ Jülich

Chemistry, Structure, Kinetics and Dynamics H. Jobic CNRS/Univ. Lyon 1W. David ISIS

Earth Science, Environmental Science and CulturalHeritage

R. Rinaldi Univ. of Perugia

Liquids and Glasses R. McGreevy Univ. of Uppsala

Fundamental Physics H. Rauch Atomic Inst. of the Austrian Univ.

4

Table 2:Instrument Groups and their Conveners

Instrument Group Convener

Powder Diffraction P. Radaelli ISIS/Rutherford Appleton Lab.

Direct Geometry Spectrometers R. Eccleston ISIS/Rutherford Appleton Lab.

Indirect Geometry Spectrometers K. Andersen ISIS/Rutherford Appleton Lab.

Neutron Spin Echo Spectrometer M. Monkenbusch FZ Jülich

SANS R. Heenan ISIS/Rutherford Appleton Lab.

Reflectometry H. Fritzsche HMI Berlin

Single Crystal Diffractometerand Protein Crystallography

C. Wilson ISIS/Rutherford Appleton Lab.

Structure Factor Determination A. Soper ISIS/Rutherford Appleton Lab.

Engineering P.J. Withers Manchester Mat. Science Cent.

Table A2 in the appendix presents the members of theinstrumentation groups.

The science groups were asked to explore the likely lines ofdevelopment in their scientific field with emphasis on highprofile flagship areas, where the power of a third generationneutron source could make the most impact. This analysis onfuture trends in science was aided by the AUTRANSworkshop report on “Scientific Prospects for NeutronScattering with Present and Future Sources” [3] which waspublished in 1997 and the scientific case evaluation in the1997 ESS report [1,4].

In order to provide a starting point for the instrumentationgroups, the neutronics and moderator specialists performedcalculations on the performance of different moderator typesat the different target stations. These results served as aninput for the considerations of the instrument groups. With thatinformation at hand, they looked into the performance ofgeneric key instruments at the different target stations. Theirwork culminated with an instrumentation workshop which tookplace on February 16, 2001 in Heathrow [5].

At a Scientific Advisory Committee meeting on March 7./8.,2001 in San Sebastian the conveners of the science and theinstrument groups came together in order to exchangeinformation and to define the work which needed to be doneuntil the SAC/ENSA workshop in Engelberg.

3. The SAC/ENSA Workshop on “Scientific Trends inCondensed Matter Research and InstrumentationOpportunities at ESS”

On May 3-5, 2001 the European Neutron ScatteringAssociation (ENSA) jointly with the SAC organized a

5

workshop on “Scientific Trends in Condensed Matter Research and Instrumentation Opportunities at ESS” in Engelberg/Switzerland. This workshop brought together the members of the SAC science groups, the leaders of the instrument groups and various observers. The list of participants is displayed in Table A3 in the appendix. The prime goal of this workshop was to arrive at recommendations for the final neutron parameters of ESS. The workshop took place in two stages accomplishing two tasks. Task I based on the preparatory work of the science groups and earlier science case considerations with the goal to assess trends in condensed matter science related to neutron research, to predict new opportunities, in particular flagship areas and to define the science demands on ESS. In a similarly way, scientists focusing on the neutron itself, in the area of fundamental physics, discussed the main issues in their field and derived their demands on the facility.

Task I: Science Case for ESS

Task II dealt with the consequences for the ESS layout, in particular with the different target stations and the associated instrumentation. In order to facilitate Task II, the instrument groups provided a performance evaluation of generic instruments at the different target station which is published as a separate ESS report [5]. Furthermore, instrument performance sheets were handed out, which summarized the performance of the different instrument classes at the different target stations. Table 3 gives a simplified overview on these results, in displaying the expected performance of generic instruments at the various target stations at ESS. First choices from the point of view of the instrument groups are given as black dots, second choices by open dots. Furthermore, gain factors are given which distinguish source gains and further gains due to increased instrument performance. The numbers correspond to instruments implemented on the best (or one of the best) target options. Blue numbers compare to ILL beams and best existing instruments at ILL. Black numbers compare to ISIS beams and best existing instruments at ISIS. The average gain factors at the bottom are geometrical averages overall instruments. A summary of the instrument group considerations including thoughts on future innovative instrumentation is given in Chapter 6 by F. Mezei.

Task II: The mapping of science demands on instrumental opportunities at the different target stations Instrument gain factors distinguish

6

Table 3:

Expected performance of generic instruments on the various target station options for ESS. � : first choice or one of essentially equivalent first choice options. � : second choice: about a factor of 2 inferior in data collection rate to the first choice. The numbers correspond to instruments implemented on the best (one of the bests) target options. Blue numbers: compared to ILL beams and best existing instruments at ILL. Black numbers: compared to ISIS beams and best existing ISIS instruments (see report F. Mezei, Chapter 6).

Instrument 50 Hz 5 MW

10 Hz 1 MW

162/3 Hz 5 MW

Source gain

Total gain

High energy chopper �� � � 30 30 Thermal chopper �� � � 30 240 Cold chopper �� �� �� 50 1600 Variable, cold chopper �� � �� 20 800

Backscattering 0.8 ��eV �� � � 25 50

Backscattering 17 �eV �� � � 150 600 Molecular sp. (TOSCA) �� � � 50 100 Electron Volt Spectr. �� � � 30 300 High resolution NSE � � �� 10 100 Wide angle NSE � � �� 9 300 Triple-Axis �� � �� 0.5-1 1-4 High Resolution Single X �� � � >>10 >>10 Chemical Single X �� � � >>10 >>10 High Resolution Protein �� � � >20 >20 Low Resolution Protein �� � �� 3-5 3-5

Single Peak incl. Cryopad �� � �� 0.3-3 0.3-3

High Resolution Powder �� �� �� 50 150 High Q Powder �� �� � 60 120 Magnetic Powder �� �� �� 60 60 High Res. Reflectometer �� �� �� 20 40 High Intensity Reflectom. �� � �� 15 40 Liquids Diffractometer �� � � 20 20 High Intensity SANS � � �� 8 100

High � Resolution SANS �� �� � 150 300 Engineering Diffractometer �� � � 30 90 Fundamental Physics �� � �� 1 NA Diffuse scattering (D7) �� � �� 15 300 Backscatt. (Musical) � � �� 40 40 Average (geometrical)

�

��

� �

��

� �

��

�

>19 >47 With the instrument performance data at hand and the analysis of the science trends and subsequent science demands available, the science groups made up their priorities for different instruments at different target stations. For instruments important for their field of science, the groups had the option to make first and second choices of target stations. First choices are indicated by (A), second by (B). Table 4 presents the outcome of this target station evaluation. Obviously, the 50Hz 5MW short pulse target station drew the

Target station evaluation leads to priorities for 50Hz short pulse and 162/3Hz long pulse stations

7

most first choices with 70 (A), while the 162/3Hz, 5MW longpulse target station came very clearly as second with 31 firstchoices. Compared to that, the 10Hz, 1MW short pulse targetstation received only 3 first choices and clearly was qualifiedas less preferable.

Table 4:Target Station Evaluation

A B A + B

50 HZ 5 MW SP 70 12 82

10 HZ 1 MW SP 3 8 11

162/3 HZ 5 MW LP 31 2 33

Thus, the ensemble of scientists covering the differentresearch areas in neutron science came to a very clearconclusion. They recommend to build ESS as a 10MW sourcewith two target stations, a 162/3Hz long pulse target stationwhich is directly fed by the LINAC and a 50Hz, 5MW shortpulse target station which is placed after the accumulator rings(see Fig.1).

Figure 1: ESS target station outlay as suggested by the members of thescience groups and recommended by SAC. The long pulse targetstation is directly fed by the LINAC, while the short pulse stationis placed after the accumulator rings.

8

From the fundamental physics group a suggestion for a thirdtarget station, an ultra cold neutron (UCN) factory, emerged.Such a dedicated UCN station would accept the whole beampower for about 1% of the beam time. It would be dedicated toexperiments on the neutron decay and the measurement ofthe neutron electric dipole moment. Such experiments couldbe important, in order to investigate grand unified theories(GUT) which could be behind the baryon asymmetry and theleft handedness of the universe. The possible arrangement ofsuch an UCN station is shown in Fig.2. Further subjects ofresearch for such an UCN factory could be elastic andinelastic surface reflections and the investigation of quantumgravitational states.

Figure 2: Possible arrangement of an UCN factory at the ESS acceleratorcomplex.

4. The Long Pulse Target StationThe calculations of the neutronic performance of the longpulse target station have up to now been based onconfigurations typical for short pulse stations. For a coldcoupled moderator as an example, Table 5 compares theavailable peak neutron current densities as a function ofwavelength for the long and short pulse stations [6]. Inaddition the best ILL average fluxes are given [7] – theabbrivations in front of the figures relate to the respectivebeam holes. Furthermore, the ratio of the short and long pulsepeak fluxes as well as the ratio between the long pulse peakflux and the ILL flux is given. We note again, that the longpulse target station is not at all optimized.

9

Table 5:Peak Current Neutron Density in [1/(cm2 s sr Å)] for Different Neutron Wavelengths

from a cold coupled moderator

Wavelength 2Å 4Å 6Å 10Å

long 2.05 ⋅ 1014 9.12 ⋅ 1013 3.12 ⋅ 1013 2.95 ⋅ 1012

short 1.486 ⋅ 1015 3.708 ⋅ 1014 1.035 ⋅ 1014 7.532 ⋅ 1012

ratio short/long 7 4 3 2.5

ILL H12*: 3.5 ⋅ 1013 H1: 4 ⋅ 1012 H15: 6 ⋅ 1011 H17: 1011

ratio long/ILL 6 23 52 30*thermal beam hole

Fig.3 presents the pulse shapes for the two stations. In thecase of the short pulse target for long wave lengths pulseswith half widths in the order of 200µs evolve, while for the longpulse saturation is achieved and neutron pulses with a broadplateau length in the order of 2ms are visible.

(a)

(b)

Figure 3: Time spectra of the neutron current density for coupled cold H-moderators at the short (a) and long pulse (b) target station.

In their peak flux for cold neutrons the two target stations differby factors between 2.5 and 4. An optimisation of the longpulse station with respect to the target configuration (flux trap)and moderator reflector setups has not yet been undertakenand a significant increase in the available neutron flux is

Vast innovative potentialfor long pulse targetstation

10

anticipated. This will reduce or maybe even annihilate thepeak flux differences in the cold neutron regime between theshort and long pulse station. Following F. Mezei’sconsiderations in Chapter 6 pulse shaping choppers, repetitionrate and neutron frame multiplication as well as ballisticneutron guides will open a whole set of new opportunities toutilize the high flux long pulses from such a target station andit may be expected that a number of instruments classified forthe 50Hz station in Table 3 may at the end be better placed atthe long pulse station.

5. Recommendations of the Scientific AdvisoryCommittee

On May 6, directly after the SAC/ENSA workshop, theScientific Advisory Committee met in order to assess andevaluate the outcome of the workshop. It was realized that the50Hz short pulse target station had the highest requests fromtoday’s perspective. On the other hand, it was also recognizedthat the 162/3Hz long pulse target station has the largerpotential for innovation, while the 10Hz short pulse targetstation was of little interest. Therefore, the SAC advises theCouncil:

• The ESS project should incorporate a 50Hz shortpulse target station as well as a 162/3Hz long pulsetarget station both at a level of 5MW proton beamenergy with equal priority. This recommendationimplies a 10MW proton LINAC serving the twostations.

The recommendation for a 162/3Hz long pulse target stationhas a number of implications for the ESS project planning.

1. The efforts of the target group need to be redirectedtowards an optimisation of the long pulse target station.Target geometries, moderator and reflector environmentsneed to be reconsidered - a mere transfer of short pulsetarget station outlays towards the long pulse station doesnot necessarily lead to the optimum design.

2. Likewise, the instrument task has to move from the studyof generic to innovative instrumentation. Again particularemphasis needs to be placed on the long pulse station,where methodical ideas as laid out in Chapter 6 will haveto be transformed into instrumentation.

3. The study of instrument suites for both target stations willallow to find the optimum moderator configurations for bothtarget stations. Thereby, it would be preferable to reducethe number of different moderators at each target stationsto a minimum.

4. The discussions of instrument needs at the workshopclearly showed the importance of a methane like coldmoderator which delivers short neutron pulses over a largewavelength range. The ESS project should devotesignificant efforts into the development of such amoderator. This, however, will be not a short time project.

11

• Concerning the proposed ultra cold neutron factory, whichcould make an important impact in the field of particlephysics, the SAC advises the Council to have a costestimate done., in order to have a baseline against whichfuture discussions on the scientific importance may carryon.

• Furthermore, the SAC recommends, to accelerate theintermediate time frame of the ESS project. In particular,the SAC suggests to present the ESS to the scientific andpolitical public at a European User Meeting in May 2002.At that point an in depth science case for the ESS needsto be ready. The technical outlay of the ESS has to be in astate, where a reliable costing of the facility is available.

In order to prepare the final science case, the SAC intends toorganize topical workshops to strengthen the science casewith special emphasis on emerging fields. In life sciences atopical workshop on “Structure-Dynamic-Function” is foreseenin the late Fall of 2001. Another workshop covering the earthand environmental sciences would be highly desirable. In allresearch areas the science case which in this document ispresented with flagship areas needs to be updated withrespect to the breath and width of neutron sciences.

Topical workshops inBiology and EarthSciences

These science case considerations should again be related toinstrumental opportunities, with the instrument task groupsemphasizing innovative instrumentation. Both lines ofconsiderations will come together at a second SAC workshopwhich is foreseen for March 2002. This second SAC workshopwill come up with the final science case for ESS. It will discussan instrument suite for ESS and suggest priorities for day oneinstrumentation.

2nd SAC workshop inMarch 2002

Also beyond neutron scattering ESS will provide importantscientific opportunities. These are not considered as part ofthe project but should not be designed away in the ESStechnical design. The ESS accelerator complex presentsopportunities in the field of muons, radioactive beams, isotopeproduction, neutrino physics and elementary particle physics,where the UCN factory could play an important role. Thesescientific opportunities need to be assessed during the nextyear.

Assessment ofopportunities beyondneutron science

References

[1] ESS – A Next Generation Neutron Source for Europe, Vol. 1-3, ISBN 090 237 6500, ISBN 090 237 6551(1997)

[2] Memorandum of Extension Berlin (2000)

[3] Scientific Prospects for Neutron Scattering with Present and Future Sources, ESF Framework Studies intoLarge Research Facilities, ISBN 2-903148-90-2

[4] The Scientific Strategic Case for a Next Generation European Spallation Source for Science and Research,ESF Studies on Large Research Facilities in Europe (2001), ISBN 2-912049-20-2

12

[5] ESS Internal Report: “Performance evaluation for a set of generic instruments on ESS” ed. by F. Mezei andR. Eccleston (2001)

[6] Particle Transport Simulations of the Neutronic Performance of Moderators of the ESS Mercury Target-Moderator-Reflector System, D. Filges et al., ESS Internal Report ESS-SAC-MOD-No. 3 (2001)

[7] ILL Yellow Book, www.ill.fr

13

6. Instrumentation issues, established techniques,need for innovation, technical risks

F. Mezei

ESS Instrumentation Task Leader, Hahn-Meitner-Institut, BENSC, Glienicker Str. 100, D-14109 Berlin, Germany

AbstractIn the first phase of the instrumentation development task at ESS the aim was to assess the performance of a setgeneric instruments, assumed to be of well established, conservative design for the various target station optionsconsidered. This evaluation serves the dual purpose of supporting the scientific case and serving as input data forthe selection a target station configurations to be pursued in preparing the specific ESS project proposal. Ofcourse, it is not to be expected and it should be avoided to limit the ESS instrument design effort to wellestablished, by now traditional solutions. Innovation and what it is most often called "speculative ideas" will helpus to actually design and build in nearly 10 years down the road instruments that, at least to a considerablefraction, are more efficient than those known today. Thus the analysis of the expected performance of ESSinstruments presented in the reports of the various Instrument Groups and in the instrument performance sheetsshould rather be considered conservative lower limits, susceptible to be superseded by innovation. The presentreport tries to review some of the not yet implemented / established new ideas on more efficient instrument andtarget-moderator concepts. A common feature of these ideas is to make the instruments benefit from effectivepulse lengths, pulse repetition rates, wavelength bands, beam divergences, etc quite different from those thesource would offer in a straightforward manner. Pulse shaping, beam delivery optics and multiplexing are typicalnotions we will be concerned with. Some of these not yet realised and tested approaches have already been usedas design assumption in the evaluation of ESS instrument capabilities (such as shortening the pulse lengths byfast disc choppers) on the basis of ample experience with similar devices at continuous reactor sources. Othersare described in what follows as opportunities to be explored in the future. It will be seen that beyond opening upthe way to eventually build instruments which are more efficient than assumed today, these innovative optionsalso offer alternative solutions to back up the ESS design, to provide a safety net against the technical risksunavoidably involved in the rightly ambitious specifications. In particular it can be expected that even if the ESSdesign goal of delivering 100 kJ proton beam energy per short pulse proves to be not feasible, equivalentinstrument performances can be achieved by the combination of a 100 Hz 5M W short pulse and a low frequencylong pulse target station, assuming that the not yet performed optimisation of the long pulse target station designand/or proton beam power (5 MW or more) provides for an about two fold neutron flux gain compared to thecurrent conservative estimates, which are based on assuming the use of the short pulse target station designwithout modification also for the long pulses.

I. IntroductionSince the groundbreaking work of Bert Brockhouse at ChalkRiver in the 1950's the thermal flux performance of neutronsources only progressed by now by a mere factor of 4.Actually all this progress was accomplished a long time ago,by 1972 with the commissioning of ILL. Spallation sources didnot go any further by now: the peak flux of the leading facilityISIS is just about the same as the steady state flux of ILL, andits time average flux is only about 0.4 to 0.8 % of that of ILL.The obvious and huge progress since Bert's time is thus notprimarily due to advances in source performance in terms ofthe number of neutrons produced, but to developing ways ofusing these neutrons more efficiently. This has been achievedby the evolution of neutron scattering instrumentationtechniques and the development of specific neutronmoderators, the hot and cold sources, in order to enhance theneutron flux at energies above and below the thermal energyrange. As the next decisive step in the evolution of neutronsources, ESS will offer a quantum leap in neutron science: itwill provide for an enhancement in source performance for thedifferent applications by factors between 10 and 100, i.e. a lotmore than it has been achieved by now since the pioneeringdays of Brockhouse. This will essentially also beaccomplished by improving the efficiency of use, and not

Neutron production powergrew little in the last 4decades

Spectacular progress wasachieved by advances ininstrumentation

ESS goal: quantum leapin performance compared

14

primarily by producing more neutrons. Indeed the timeaverage flux of ESS will be comparable to that of ILL, and itpulsed character will allow for using these neutron in a 10 to100 times more efficient way.

This goal of the ESS project, to provide at least about an orderof magnitude enhanced neutron beam performance comparedto any existing source, however, cannot be achieved withexisting and established techniques. Indeed, on some reactorinstruments up to 15 % of the neutron spectrum hits thesample, so a 30 fold increase of the power compared to ISISalone would be just enough at best to break even with ILL forthese kinds of instruments. For ESS we also need to developnew, more efficient approaches both in neutron productionand moderation and in instrumentation. Actually these twoaspects are closely related, innovation in source performancewill call for new instrumentation concepts and new instrumentdesign approaches will allow us to better use the potentials ofthe source, also by relaxing some design requirements for thetarget / moderator system, which are hard to meet or undulydetrimental to neutron intensity.

As in the past, a key feature of enhancing the efficiency ofneutron sources is to improve the neutron moderators. Currentspallation neutron instruments face neutron moderatorsoptimised for producing short neutron pulses by limiting thetime allowed for neutron thermalization with the help ofneutron absorbers placed around and/or inside themoderators. The enhanced efficiency of neutron use at ESSwill also have to include implementing more efficient, so calledcoupled moderators. The first two of this kind have recentlybeen installed at Lujan center in Los Alamos, now making thissource about 2 times as bright as ISIS at half the acceleratorbeam power. The higher moderation efficiency of thesemoderators is accompanied by longer moderation times, withsignificant neutron intensities emitted for up to 3 – 4 ms afterthe beginning of the pulse. As of today, there is no experienceavailable with the use of such long pulses.

Another potential option to enhance neutron productionefficiency, in particular in view of the long moderation times incoupled moderators, is to avoid the compression of the mslong linear accelerator pulses to µs length short pulses byproton storage rings. The ESS design goal of 100 kJ totalproton energy per short pulse is certainly at the technologicallimit both in ring accelerator design and material strength. Bymaking economy of the pulse compression rings, up to some500 kJ beam energy per pulse becomes feasible in 2-3 mslong linac pulses. Again, as of today, we have no practicalexperience with the production and use of such long pulses.

The great achievements of spallation sources by now, inparticular ISIS, offers a solid base for planning ESS. On theother hand, to achieve ESS goals we also need tocomplement the established techniques by novel approachesfor more efficient neutron moderation and production, such ascoupled moderators and long pulses. It turns out, that the

to all existing neutronsources…

…it can only be achievedby combining enhancedpower with newapproaches

New type of moderatorproduces morethermalized neutrons inlonger pulses

Long proton pulses canprovide several timesmore neutrons per pulse

15

powerful instrumentation techniques developed at steady statereactor source provide a number of technical opportunities forthe use of these options.

By now the goal of the of the Instrumentation Task Group wasto assess expected instrument performance on ESS. Theresults are contained in the reports of the 9 instrument groups,including the instruments performance sheets. In order to beon the safe side, this task was deliberately based on prudentextrapolation from established approaches and conservativeestimates of the performance of the novel kind of moderators,namely the coupled ones and those on the long pulse targetstation. Nevertheless, it is fair to think that in nearly 10 yearsfrom now innovations and new ideas will play a major role forthe instruments that will really be built. The aim of the presentreport is to complement the Instrument Group reports byexploring some essential parts of the uncharted territory ofnew, not yet implemented instrumentation approaches. Thisterritory will certainly keep growing in the months and years tocome, as new ideas will continue to emerge.

First ESS instrumentperformance evaluationwas based on establishedtechnology. Enhancedapproaches keepemerging.

Moderator performancesThe instrument assessment effort of the Instrumentation Taskwas based on a set of assumed moderator parameters, whichhave been established in collaboration with the Target andModerator task group in December 2000 [1]. In what followswe will refer to this moderator data base as "Dec. 2000"compilation. At that time no detailed neutronics calculations ofthe ESS target-moderator system were available, and thecompilation was based on calculations made at SNS, LosAlamos and in Japan, and also on performance estimates ofexisting sources. The most complete and most recent of thesecalculations, those made at SNS, were given the most weight.As the ESS moderator calculations progress, cf. the report ofthe Moderator Task Group at this meeting, the referencemoderator data will be regularly up-dated. We expect the nextup-date to be completed and adopted within the next 2-3month. A first comparison suggest, that this up-date will notbring dramatic changes, i.e. both assumed pulse lengths andpulse intensities will have to be revised by less than a factor of2.

For short proton pulses three different types of moderators areenvisaged (poisoned de-coupled, de-coupled and coupled),which provide different neutron pulse lengths. For ms longproton pulses the neutron and proton pulse lengths areeffectively the same (not the pulse shape though) andtherefore only the brightest (coupled) moderator needs to beconsidered. At this stage the short and long pulse targetstations are assumed identical.

One important difference between the current ESS referencedesign and SNS is that ESS assumes the use of Pb reflectoraround the target and moderators and SNS a compositereflector, consisting of Be in the inside and Pb outside. PurePb reflector provides for higher time average flux from all

Current ESS referencemoderator performancedata are based onprevious projectevaluations

Novel target-moderatordesign options for ESS arebeing evaluated

16

moderators, however the pulse lengths, primarily those of thethermal moderators also become larger. The final decision onthe ESS choice of reflector therefore will have to be madeafter careful consideration of the advantages anddisadvantages for the reference suits instruments on twotarget stations, to be completed by the end of 2001. The Dec.2000 compilation basically assumes Be-Pb reflector, by itspreferential reference to SNS data.

0 500 1000 1500 2000 2500 30000

1x1013

2x1013

3x1013

4x1013

λλλλ = 6 Åpoisoned moderatordecoupled moderatorcoupled moderatorlong pulse

Instantaneousflux[n/cm

2 /s/str/Å]

Time [µs]

Figure 1: Example of ESS reference neutron pulses in the beam lines forthe various cold (liquid H2) moderators considered. [1]. For thethree short pulses the proton pulse length is negligible (< 2 µs).The steady state flux of ILL is 6.1011 in the same units and at thesame wavelength. The proton beam energy is assumed to 100 kJper pulse short pulse and 300 kJ per long pulse.

Comparison to existing sources, namely to ISIS and ILL was abasic goal of the performance evaluation by the InstrumentTask Group. Concerning ISIS, this is based on the assumptionthat the ESS de-coupled moderators will provide 30 times theflux at similar line shape to the corresponding ISIS decoupledmoderators. ISIS does not have coupled moderators (whichare now foreseen for the second target station at ISIS), so thetime average flux of the brightest (coupled) ESS moderators isexpected to amount to about 120 times that of the ISIS un-poisoned moderators today, at 2-3 times longer pulse lengths.A comparison to ILL is part of the Dec. 2000 compilation. It isbased on the flux data published in the ILL "yellow book" [2],which refer to neutron fluxes measured in the various ILLbeam tubes and guides at considerable distance form thereactor core. Due to the Liouville theorem the flux isindependent of the position it is measured at, as long as thereis no beam attenuation due to interaction with materials, suchas beam windows or neutron mirrors. The time averageneutron flux at 5 MW ESS power was estimated in Dec. 2000to be equal to that of ILL for coupled cold (liquid H2)moderators and about 50 % that of ILL for coupled thermal(ambient H20) moderators. In contrast, the peak flux of the

17

ESS poisoned moderators will achieve 30 – 60 times thesteady state flux of ILL, and that of the coupled ones some 60– 120 times, depending on neutron wavelength. For thevarious experimental applications, the effective useful flux ofESS compared to a continuos source will thus range betweenthe ratios of average and the peak fluxes, i.e. from 0.5 to 120in the case of comparing to ILL.

Figure 2: Peak (top) and time average (bottom) flux of reference ESSthermal neutron (ambient water) moderators as a function ofwavelength [1] compared to the steady state flux of the variousILL moderators [2]. For the short pulses 5 MW time averageproton beam power is assumed at 50 Hz pulse frequency and forthe long pulse the same 5 MW power at 16.66 Hz.

The hot (epithermal) neutron flux is a special case. While inthis, under-moderated regime the peak flux of ESS is morethan 100 times superior to the steady state flux at ILL, the time

The peak instantaneoussource brightness of ESSwill be about two orders ofmagnitude higher thanthat of ILL and ISIS

The time averagebrightness of the coupledmoderators at ESS will becomparable to that of ILL

New design options will beconsidered to enhance the

0 1 2 3 4

1010

1011

1012

1013

1014

1015

1016

1017

ILL hot sourceILL thermal sourceILL cold source

peak fluxpoisoned m.decoupled m.coupled m.long pulse

Flux[n/cm

2 /s/str/Å]

Wavelength [Å]

0 1 2 3 4

1010

1011

1012

1013

1014

1015

1016

1017

ILL hot sourceILL thermal sourceILL cold source

average fluxpoisoned m.decoupled m.coupled m.and long pulse

Flux[n/cm

2 /s/str/Å]

Wavelength [Å]

18

average flux is some 10 times less. The former assures greatperformance for pulsed sources when good hot neutronwavelength resolution is required, e.g. in powder diffraction.In contrast, the latter only offers modest capabilities in singlepeak analysis on single crystals such trademark neutronpolarisation analysis work (e.g. CRYOPAD). Implementing ahot moderator on ESS would alleviate this deficiency,however, the alternative suggestion of installing a beamlineviewing directly the Pb reflector, rather than a moderator,appears to be a much superior (and by now untried) solution[3], which needs to be explored by neutronics calculations inthe near future.

The moderator performances of Dec. 2000 assume that eachmoderator is located at the most favourable position next tothe mercury target. This position is well defined in space, inparticular within some 5 cm in the direction of the proton beamhitting the target. Therefore only two moderators on a targetstation can occupy the brightest position, one below and oneabove the target. Placing 4 moderators on a target stationnecessarily reduces the flux of all 4 moderators by amountsbetween approximately 10 – 40 %. It is a design choice howthe reduction is distributed between two moderators behindeach other, e.g. trying to maximise the flux of one of them(usually the upstream one with respect to the proton beamdirection, at the expense of the downstream one.)Furthermore, the performance of moderators on the long pulsetarget station was assessed in Dec. 2000 by assuming thesame target-moderator design deemed optimal for shortpulses. Thus in comparison to the short pulse moderatorperformances the long pulse flux is certainly underestimated,possibly by as much as a factor of 2 or more. In particular ifone also takes into account that 4 moderators are a likelynecessary minimum for a short pulse target station, while onthe long pulse station there is no need to have more than 2moderators (both coupled, of course) with 2 viewed faceseach.

Another conservative assumption made in Dec. 2000 in theabsence of methane or similar moderators. Liquid H2 has arather low proton density per cm3, and it is therefore ratherinefficient as de-coupled or poisoned de-coupled moderator,where the thickness of the moderator is limited to less than 5cm in order to keep small the contribution of neutron flight pathuncertainty to the pulse length. Potential flux gains fordecoupled and poisoned moderators by methane or similarcould reach a factor of 2-3 (while no essential gain is expectedfor coupled moderators). However, radiation at ESS levelsrapidly destroys methane, and this is a tremendous, unsolvedtechnical obstacle. Thus as of today methane or similarmoderators cannot be envisaged as feasible options. Thismight, however, change in the future, e.g. by the developmentof techniques to make solid methane pellets circulated by acryogenic fluid.

ESS time average hotneutron flux

The number of moderatorsat a target station is animportant designparameter

Planned optimisation ofthe long pulse targetstation is expected tobring substantial furthergains

If difficult feasibility issuescan be solved, methanemoderators will become agreat option for decoupledmoderators

19

Pulsed sources: big advantages with tough stringsattachedThe overwhelming reason to prefer pulsed sources tocontinuos ones for highest performance is the vastlyenhanced efficiency of making use of the produced amount ofneutrons in the vast majority of neutron scattering experiments(while the pulsed structure is virtually of no advantage forirradiation work). On a continuos source beammonochromatization is synonymous to throwing away all thoseneutrons in the Maxwellian source spectrum which do notpossess the required velocity with the required precision,which in the majority of the cases amounts to only making useof 1-3 % of the spectrum. On pulsed sources the neutrons withdifferent velocities arrive at different times to the instrument,so they can be distinguished from each other andconsequently used at the same time. This allows us to utilise alarge fraction, up to 60 – 90 % of the spectrum. Thistremendous gain in efficiency, as most good things, does notcome free. The pulsed nature of the beam imposes a wholeset of boundary conditions, which lead to substantialcompromises in instrument design options and can partially(or in rare cases even completely) erode the very gain inefficiency one seeks to achieve. As we plan to make ESS tobenefit all areas of neutron scattering science, we will facenew challenges to minimise this erosion of performance in anumber of new situations. Some key approaches to achievethis will be considered in the next chapter, after havingreviewed here the detrimental boundary conditions in need ofremedy.

a) Beam delivery to sampleOne of the great achievements over the past decades was theperfection of curved, focussing monochromator crystals, whichcan deliver typically 2-10° beam divergence, both horizontallyand vertically. While crystal analysers are veryadvantageously used in inverse geometry time-of-flight (TOF)spectrometers, crystal monochromator are in general notadequate for pulsed source instruments. The direct view of themoderators from the closest reasonable moderator to sampledistance on ESS, some 12 m, will only provide 0.6° deliveredbeam divergence capability both horizontally and vertically,which can prove to be as much as 100 times less than theincoming beam solid angle obtained by Bragg focussingoptics. This practically eliminates most of the potential gain bythe higher ESS peak flux.

The divergence and hence intensity of the beam impinging onthe sample can only be enhanced on TOF instruments by theuse of broad wavelength band mirror optics. The current upperlimit is practically set by the performance of the best(commercially) available supermirrors to

δα ≅ 0.6°×λ [Å] (1)

i.e. to 2 times the supermirror cut-off angle in both directions.There are two significant novel aspects to observe here. On

Pulsed sources allow formuch more efficient use ofthe neutrons produced ...

… but they also imposetough accelerator, targetand instrument designcompromises to beaddressed

Focussing crystalmonochromators are agreat strength ofcontinuous sourceinstrument

Advanced supermirroroptics can now deliverenhanced divergencepulsed beams over largedistances

20

the one hand side, supermirror guides can deliver higher fluxon the sample than the direct view of the moderators for allwavelengths λ>1 Å. On the other hand side, advanced guidedesign, in particular the principle of ballistic guides allows usto transport beams with supermirror divergences overdistances up to several 100 m with moderate losses in therange of 30 % or less [4]. These two facts combined meanthat for > 1 Å wavelengths the moderator to sample distancenow became a free parameter in instrument design andoptimisation. In contrast, many current instruments had to beoptimised for the direct view of the moderator, i.e. made asshort as possible. Consequently, to keep resolutionreasonable, the source needed to be designed for very shortpulses, to the detriment of brightness. This explains why therewas little interest in coupled moderators until recently.

b) Repetition rateAt a given time average power the repetition rate of the pulsedsource is a parameter largely neutral to the ultimate sourceperformance – unless it is too high. Indeed, if there is notenough time T left between pulses to allow the full desiredwavelength band ∆λ to be accepted at the desired moderatorto sample (or detector) distance d, i.e. the

T[ms] > d[m]×∆λ[Å] / 3.96 (2)

relation is not fulfilled, the source brightness is not fully takenadvantage of. Thus a low repetition rate, practically somethinglike 20 Hz or less is the guarantee for efficient use of a pulsedsource in most applications.

It has to be emphasised that this consideration is only valid ifthe total proton beam power of the source can be keptconstant, independently of the choice of the repetition rate. Inreality, however, the beam energy per pulse is the parameteractually determinant for the technical complexity, feasibilityand price tag of the accelerator system. At constant energyper pulse condition, however, the optimal repetition ratebecomes very strongly dependent on the type of theinstrument and it can be as high as 1000 Hz. Such a highvalue is, on the other hand, wasteful for other instruments,which cannot efficiently deal with more than 20-30 pulses persecond, i.e. their performance does not increase if additionalpower comes in the form of higher repetition rate. In this sense50 – 100 Hz seems to be a rather reasonable compromise,between instrumental needs and accelerator designimperatives to minimise the beam energy delivered per pulse,in particular for short pulses. In contrast, for long proton pulsesin the ms range the energy per pulse could reasonably reachvalues well in access of 500 kJ/ pulse and thus low (< 20 Hz)repetition rate is a good option up to some 10 MW timeaverage power.

c) Wavelength resolutionThe source neutron pulse length δt and the moderator to

For many instrumentspulse repetition ratesbelow ~20 Hz only offerfull efficiency

...but high poweraccelerator imperativesgenerally make higherrepetition rates preferred

With long acceleratorpulses <20 Hz repetitionrate is feasible even for>5 MW power

21

sample (or detector) distance determine the incomingwavelength resolution on a usual spallation source instrument:

δλ[Å] = 3.96×δt[ms] / d[m] (3)

If δλ is less than necessary, the pulsed source instrument willlose in efficiency relative to a continuous source machine. Ifone does not consider background type issues (e.g. thenecessity on continuos sources to filter higher order reflectionswhen using crystal monochromators, etc) and beam lossesdue to absorption, finite reflectivities, etc. a pulsed sourceinstrument can deliver a full neutron intensity gaincorresponding to its peak flux compared to a similar type ofcontinuous source machine if and only if all of 3 conditions aremet:

i) The required incoming beam divergence is less in bothdirections than the supermirror guide limit, eqn. (1).

Ii) The source repetition rate is low enough to satisfy thecondition (2) and high enough to make effective datacollection time fill most of the elapsed time.

iii) δλ as given by eqn. (3) matches the one freely chosen onthe continuous source instrument.

These three points are actually quite some strings attachedand can prove to be contradictory, for example d chosen tofulfil eqn. (3) happens to be much too large for (2) to besatisfied (e.g. this is usual in high resolution powderdiffraction) or the minimum value of d as determined byshielding and other instrument design imperatives leads toboth too narrow wavelength band ∆λ and too good wavelengthresolution δλ.

The canonical example for the latter is small angle scattering,SANS. Here the required good angular resolution can only beachieved without counterproductive limitation of the samplediameter to much less than the usual 0.5 – 1 cm if d ~ 40 m(cf. D22 at ILL). At 50 Hz repetition rate and for short pulsesthis implies δλ ~ 0.6% (at λ = 4 Å), i.e. some 20 times toosmall compared to D22 and ∆λ ~ 2 Å, i.e. some 3 – 4 timesless than the useful potentially useful part of the Maxwellianspectrum of a cold coupled moderator. In this crucial examplethese two factors substantially lower the intensity gain of ESScompared to ILL, namely detailed Monte Carlo simulationsshow that instead of the peak flux ratio of 60 it becomes 3-4 atbest. For the 16.7 Hz long pulse target option the situation ismore favourable: with δλ ~ 5 % and ∆λ ~ 6 Å an intensity gainof about a factor 10 remains from the 25 fold ESS peak fluxadvantage.

An even worse example than SANS is thermal neutronspectroscopy, as represented by the TOF instrument IN4 atILL. Around 1 Å wavelength the divergence loss at direct viewof the moderator compared to a focussing crystalmonochromator is about 20 fold (even with substantial crystalreflectivity losses assumed), and it is compounded with a

Pulsed source instrumentdesign challenges include

- avoiding too goodresolution

- avoiding too low pulserepetition rates

- avoiding too high pulserepetition rates

... at the same time for avariety of very differentinstruments

22

typically 8 fold loss by too low pulse repetition rate (50 Hz vs.some 400 Hz). This is not compensated by the 100 fold peakflux advantage of ESS.

These kind of efficiency losses are ultimately due to the factthat the pulsed source repetition rate and the neutron pulselengths just cannot be simultaneously optimal for all kinds ofinstruments of conventional design. In what follows we willdiscuss, how by now not implemented instrument designtechniques can mitigate the problem. The need for innovativesource and instrument design is a centrally essential feature ofthe ESS project: ESS is not just a next generation pulsedspallation source. Spallation sources as we know them todayare efficient complements to the current mainstay tools forneutron scattering, continuos reactor sources. By now weknow, that there is no way to enhance the neutronperformance of reactor sources at reasonable costs beyondthat of ILL. Thus the real challenge of ESS is that it has totransform the spallation technique from complementary tosuperior across the board. ESS will be the next generationneutron source compared to both existing reactors andexisting spallation sources. It is probably a fair to estimate thatat most 30 % of moderator and instrument design approachesneeded to achieve this goal are established by now.

New multiplexing and pulse shaping approachesIn order to alleviate the contradictions between demands onsource parameters set by accelerator physics, target andmoderator design needs and the variety of instrumentalrequirements, we need to develop techniques which make theinstrument benefit from effective source parameters, namelyrepetition rate and neutron pulse lengths, different from theactual ones. Suppressing source pulses, up to actually 4 outof 5, is a well established method in order to reduce theeffective pulse repetition rate. We will consider here thepotentials offered by some other, not yet experimentally testedtechniques.

The use of fast pulse shaping choppers close to the source(which can practically mean not less than some 6 m distance)in order to shorten the moderator pulse length for highresolution applications has already been considered in thegeneric instrument evaluations (cf. the report on indirectgeometry spectrometers [8]), although it has not beenexperimentally realized yet. It will allow us to achieve shorterpulse lengths than provided by poisoned moderators.Additional, not less important advantages are that the shortpulses cut out of coupled moderator pulses benefit from thehigher peak flux of these moderators compared to thepoisoned ones (cf. Fig. 1) and this way one could eventuallyreduce the number of required moderators on one targetstation. The great inconvenience of the technique is that itnormally drastically reduces the available wavelength band(e.g. to ~0.2 Å when used with cold coupled moderators). Newtechniques to alleviate this drawback will be discussed below.

Multiplexing and pulseshaping allows us tobetter satisfy a largervariety of instrumentrequirements

Pulse shaping by fastchoppers makeunprecedented short andsharp pulses and highresolutions available

23

Multiplexing approaches can be roughly defined by trying tomake use of multiple beam parameter domains, which wouldconventionally exclude each other. A most beautiful exampleis the pioneering proposal of the MUSICAL spectrometer byAlefeld [5], in which the beams of several different crystalmonochromators are distinguished from each other by timedelays induced by judiciously choosing the positions of thevarious crystals. We will now consider a number of relatedschemes, aimed at enhancing the efficiency of use of pulsedsources by getting around restrictions implied by relations (1)– (3) above. Paradoxically, we will not really need to considerAlefeld's original idea here: the combination of pulse shapingchoppers and very long flight paths bridged by advanced, lowloss neutron guides now offers a simpler approach to veryhigh resolution backscattering spectroscopy, as shown in therelevant instrument group report.

a) Repetition rate multiplicationThe example of a IN4 type instrument above illustrates thateven 50 Hz pulse repetition rate is a substantial disadvantagein direct geometry TOF spectroscopy, where 100 Hz to 500 Hzpulse frequency is the rule at continuous sources. This higherfrequency is determined by the time needed to analyse thescattered beam, and at lower repetition rates no data arecollected for most of the time. This drawback can be removedby extracting a number of pulses with different monochromaticneutron velocities from the same source pulse [6]. Forexample in the IN500 project at Los Alamos up to 240pulses/s can be delivered to the sample from 20 source pulsesper second In practice this means, that instead of oneexperiment made, as usual, with a single wavelength anumber of experiments with different incoming neutronwavelengths are accomplished simultaneously, and theinformation needs to be combined in the data evaluationprocess. This can be the more efficiently accomplished thesmaller is the difference between the neutron wavelengths insubsequent pulses (e.g. 0.265 Å at 240 Hz multiplied rate onIN500). This is best achieved by long source to sampledistances, in which case the most intense and longest coupledmoderator pulses are the best choice for matching therequired resolution. Numerical evaluation of a few specificexamples confirmed [6] that multiple pulses with differentwavelengths can deliver comparably useful information, thuse.g. in case of an IN4 type instrument much of the count rateloss due to the low repetition rate can be recovered. Thus incontrast to the conventional situation discussed above,repetition rate multiplication will allow ESS to deliver superiorneutron flux even in this case (and at better angular resolution,without the focussing crystals).

Repetition rate multiplication can be readily realised by theuse of disc choppers of the type well established at continuoussource cold neutron spectrometers, such as IN5 at ILL orNEAT at HMI. At the same time it also implies source pulseshaping, as explained in the caption of Fig. 3.

Repetition ratemultiplication can enhancethe effective flux on thesample by several times

24

Figure 3: Principle of Repetition Rate Multiplication as realized by a discchopper system. For an example, the source to sample distancein the IN500 project is 63 m, the time between source pulses50 ms, and the maximal speed of the disc choppers used is14400 RPM. Choppers #3 and #6 consist of two counter-rotatingdiscs each. The role of chopper #5 (not shown), mounted close to#6, is to reduce the basic 240 Hz pulse repetition rate on thesample, if necessary, by optionally suppressing every 2nd, 3rd, 4th

or 6th pulse. Note that the effective source pulse length for allpulses on the sample is determined by the timing between thesharp rising edge of the source pulse and the closing time ofchopper #3. By properly choosing the chopper parameters noneutrons with λ < 80 Å can make their way through the choppersystem, other than the 12 pulses shown [6].

b) Wavelength frame multiplicationThe use of a broad wavelength band allows to achieve a widedynamic range in wavenumber q. Data can be taken withvarious delays after the pulse emitting the neutrons, i.e. thewavelength band ∆λ allowed by (2) can be rather freelypositioned around various average wavelengths. Thispositioning is accomplished by frame definition choppers(choppers #1 and #2 in Fig. 3 serve this function) and in aseries of data collection periods one can cover any reasonablewavelength band. It remains, however, desirable to also beable to collect data in a broad wavelength band quasisimultaneously, for example with a sample whose statechanges rapidly with time. An example of how to reach this isillustrated in Figure. 4 [7] The set of frame definition choppersshown makes sure, that in subsequent periods T between twosource pulses different wavelength bands alternate on arrivalat the sample or detector. For example with three alternatingwavelength bands (three fold frame multiplication) a SANSmachine on the 50 Hz source can cover the same or broaderband than a conventionally operating instrument on a 10 Hzsource (Figure 5).

Wavelength framemultiplication can extendthe wavelength bandwithout the need ofchanging the framedefinition chopper timing...

25

40 60 80 100 120 140 160 1800

5

10

15

20

25

30

35

Framedefinition

and

fram

eoverlapchoppers

8.8-11 Å4.4-6.6 ÅDetector

Source

Sample

Distance[m

]

Time [ms]

Figure 4: An example of wavelength frame multiplication for a 36 m longinstrument at a 50 Hz source. At the junction of subsequentframes at the detector the source pulse length and the chopperopening/closing time uncertainties produce overlap of thepenumbra. More than 17 ms from the 20 ms frame durationbetween source pulses are free from overlap or higher orderleakage neutrons with wavelength < 80 Å [7].

Figure 5: Example of extending the q range of a SANS instrument on the50 Hz target station by wavelength frame multiplication (redtriangles) compared to single frame 50 Hz data collection (blacksquares) and single frame operation on a 10 Hz station with equalpulses (blue dots).

Another, more important application of wavelength framemultiplication is related to the use of pulse definition choppers,which can cut the length of coupled moderator pulses or longpulses to actually shorter duration (10 – 20 µs) than that of thepoisoned moderator pulses. For example, this approach hasbeen found to allow us to reach sub µeV resolution in

…and to make broadwavelength rangeaccessible to choppershaped, very short andsharp pulses

1E-3 0,01 0,10,1

1

10

100

1000

10000

100000

Extending q range by frame multiplication: (flat scatterer)ESS 50 Hz, 15 m/ 15 m, 2.2<λ<4.4ESS 50 Hz, 15 m/ 15 m, 2.2-4.4, 6.6-8.8 and 11-12 Åframes measured simultaneouslyESS 10 Hz, 15 m/ 15 m, 2<λ<10 Å

Counst/

0.001Å-1channel[a.u.]

q [Å-1]

26

backscattering spectroscopy at some 200 m source to sampledistances (as mentioned above) which resolution previouslywas deemed to only be achievable at good intensity by theMUSICAL approach. This kind of pulse shaping areparticularly easily performed with long pulses, but the sourcepulse length limits the wavelength band, according to relation(2). With the pulse shaping chopper at 6 m from the moderatorand 2 ms pulse length, the band width is 1.333 Å. The framemultiplication scheme shown in Figure 6 allows us to appendin a gap free manner a number of 1.3 Å wavelength bandsone after the other (2 are shown in the figure) in order to fillthe whole time frame between two subsequent source pulses.For example the 5.33 Å band obtained by 4 fold framemultiplication corresponds to the full bandwidth of the 16.7 Hzsource at 45 m source to detector distance.

Figure 6: Principle of wavelength frame multiplication used to extend thestrongly reduced wavelength band conventionally allowed for by asingle opening of a pulse shaping chopper outside the bulkshielding. For 2 ms long pulses this technique enables us tosimultaneously and continuously cover typically requiredwavelength bands of 3 – 8 Å with short pulses of adjustablelength on a long pulse source [7].

This "frame multiplication for pulse shaping chopper" allows usto cut out very short and sharp, symmetrically shaped pulsesfrom long pulses, as an alternative choice to the asymmetric,exponentially decaying short source pulses in virtually anyapplication. To achieve the same resolution these symmetricpulses can be about 50 % larger in FWHM than theexponentially shaped source pulses. With the long pulse peakflux equal to the poisoned moderator peak flux (cf. Fig. 1), thelong pulses actually offer equal or better beam intensity orshorter pulses for the highest wavelength resolutionapplications than the shortest moderator pulses. In addition, ifone assumes that the optimisation of the long pulse targetstation brings a factor of two in intensity at least, for coldneutrons the chopper shaped long pulses will be equal or

i i k fl t th h t l f th 50 H t t

Detector

Frame overlap choppers

Pulse shaping chopper at 6 mSource

. . . . .. . . .

Distance

Time

27

superior in peak flux to the short pulses from the 50 Hz targetstation for all types of cold moderators, and equal or less bynot more than a factor of 2 to all thermal neutron moderatorpulses. This will open up the option to use the long pulsetarget station for all high wavelength resolution applications(e.g. very high resolution powder diffraction) for wavelengths> 1 Å. eventually alleviating the demand for a high number ofdifferent types of moderators on the 50 Hz short pulse station.

c) Multiple beams and detectorsSimilarly to making several pulses impinging on the sampleone after the other, one can conceive to make severalconverging incoming beams placed next to each other. Suchconverging beams can be produced at little technical risk byusing a series of collimating beam frames or slits, starting at arather large cross section virtual source created outside thesource bulk shielding. Such a virtual source can beconveniently achieved by using supermirror optics, forexample in the manner the central section of a ballistic guidecan be fully illuminated by the diverging "reflector" section,even if its dimensions are bigger than that of the moderator[4]. It is harder to implement such optics on existing sourcebeam tubes, not designed for this purpose. The following twoexamples show the importance of this kind of multiplexing inspace.

Most reflectometers on continuos sources use amonochromatic beam with scanning the grazing impact angle.This offers the advantage of optimising sequential datacollection by distributing the available time in a most efficientmanner between the points in the scan. With the TOFtechnique on a spallation source the data collection efficiencyis necessarily less optimal, since in a more or less broadsimultaneously measured band every point gets the sameshare of time. The advantage on the other hand is, that asimultaneous scan can eventually be completed in a shortertime for time dependent studies, since one saves the timeneeded to physically change the angle between beam andsample. (Which approach is the faster really depends on theexperiment, case by case.) To achieve a broad q range in afixed angle TOF experiment one needs the broadest possiblewavelength band, i.e. the lowest possible source frequency,which for the short pulse option ultimately means to reducethe source flux by eliminating most of the 50 Hz sourcepulses. The alternative way to broad band rapid datacollection is to implement a number of beams impinging on thesample at different angles, e.g. at 0.5, 1 and 2°, as tested orplanned at several places (LANSCE, HMI, SNS, ...). Since fora typical reflectivity curve only a few % of the total measuringtime is required to obtain more than enough statistics at thesmaller angles, this method allows us to retain the full powerof the 50 Hz source and cover the q range of a 10 Hz sourcemore than 4 times faster (cf. the Reflectometry InstrumentGroup report in this volume for simulation results). Switchingfrom one beam to another by opening and closing beam slitscan be done with little time lost, it could even be accomplished

Fast switching between aset of incoming beams canspeed up data collectionin a broad range of q

28

between two source pulses.

Coverage of a broad q range is also an issue in SANS. Oncontinuous sources this is most often achieved in a veryefficient way by moving the detector more or less close to thesample. Again, this allows us to optimally divide the measuringtime between the different detector positions, and if the timeneeded to move the detector is much shorter than thenecessary data collection times, this is the most efficientapproach on all sources. An alternative approach is to placeseveral detectors at different distances from the sample, asalready implemented at several facilities (Dubna, ISIS,...). Themethod can work particularly well for pulsed sources, sincethe gaps in scattering angle between the detectors is bridgedin the q space by the use of a more or less broad wavelengthband. Figure 7 illustrates the far superior data collection rateand q range one can obtain by this technique on the 16.7 Hz,5 MW long pulse source compared to both ILL and the 10 Hz,1 MW option.

Figure 7: Monte Carlo simulated small angle scattering spectra from adilute monodisperse colloid sample for 15 m-15 m long SANSinstruments, with an additional detector at 3 m from the samplefor the ESS 16.7 Hz instrument (red squares). The detectors areof 1 m x 1 m dimension, with the additional detector at 3 m havinga 23 cm x 23 cm hole in the middle. The two red curves have aconvenient overlap, in spite of the gap in angular coveragebetween the two detectors. The data also illustrates the order ofmagnitude intensity gain compared to ILL (black triangles) atgood resolution.

The final example of multiplexing is to enhance the data ratesin SANS by making a large number of identically collimatedbeams converge on the same spot on the area detector. It hasbeen proposed to realise this by a set of diaphragms withmultiple holes (e.g. at Los Alamos), as indicated in Figure 8.The real significance of this approach is that it will allow us toenhance the data collection rate at ESS by another order of

Multiple detectors vastlyextend the simultaneouslyexplored q range in highspeed SANS datacollection

0,01 0,11E-4

1E-3

0,01

0,1

1

10

100

1000

10000

Extending q range by multiple detectors:(Monodisperse colloid sample)

ESS 16 Hz, 15 m/ 3 & 15 m, 4<λ<9 ÅESS 10 Hz, 15 m/ 15 m, 2<λ<10 ÅILL, λ= 6 Å

Counst/

0.0025

Å-1channe

l[a.u.]

q [Å-1]

29

magnitude compared to the most powerful SANS machinestoday – at least for flat samples larger than 20 – 30 cm2. Thiswill make possible to take meaningful SANS spectra in lessthan a second time and thus open up fully new opportunities intime dependent studies.

The same techniques of converging beams can also help usto achieve better angular resolution and reduce the lower limitof the accessible q range to some 2-3×10-4 Å-1. Here 100 –200 finely collimated (typically 1 – 3 mm cross section) beamsshould be used. Gravity will become a problem for so narrowbeam paths and focussing mirrors should offer a betterapproach if the problem of high background due to non-specular reflections can be solved.

1E-3 0,01 0,1

1E-3

0,01

0,1

1

10

100

1000

10000

100000

8m /8 m converging beam SANSfor high intensity (Guinier model)

ESS 16 Hz, 50 cm2 sample, 2.2< λ<9 Å

ILL, 1cm2 sample, λ= 4.2 Å

Counst/

0.0014

Å-1channel/10

s

q [Å-1]

Figure 8: Neutron absorbing frames to build a converging collimator, whichessentially just amounts to repeating the same incoming beamconfiguration several times next to each other. The simulatedSANS data (bottom) confirm the some 100 fold enhanced datacollection rate at ESS compared to a single beam, identicalangular resolution experiment at ILL.

Multiple convergingincoming beam combinedwith the superior neutronpower of ESS can increasethe speed of SANS datacollection by two orders ofmagnitude

30

Summary of ESS instrument performancesThe table below, concluding this report, summarizes theinstrument performance gain factors presented in more detailsin the topical Instrument Group reports [8]. A few similarinstruments are represented just by one of them and 4, notincluded in the reports, have been added (TAS, D7, MUSICALand fundamental physics beam). The table has been updatedin two second choice cases, to take into account the impact ofrepetition rate multiplication. The neutron intensity gains dueto the enhanced flux of ESS have been evaluated in thereports separately from additional gains in data collection ratedue to foreseeable improvements in instrumentation methodsand techniques. (In this latter respect the only deviation in thetable compared to the reports is including the convergingcollimator for the high intensity SANS, cf. Fig. 8). The sourcegain factors tell us how much of the increased performancewill only be available by realizing ESS. The total gain factors,i.e. source plus progress in instrumentation, on the otherhand, are those which will determine what kind of newscientific opportunities ESS will offer. The highest number inthe table, well above 1000 in the case of cold neutronspectroscopy, is one of the best studied and establishedfigure, and it illustrates the point in an instructive manner.Compared to the extremely popular and powerful TOFspectrometer IN5 at ILL (which was roughly equivalent to themore recently designed NEAT at HMI), the source gain factorwas estimated to be around 50, which is somewhat less thanthe peak flux ratio in order to take into account the need to userepetition rate multiplication to recover losses due to the ratherlow source repetition rate of 50 Hz. Supermirror optical beamdelivery has been shown to provide for an additional gainfactor of 6 – 7 [4] compared with Ni coated guides, a factor of4 can be gained at least in detector solid angle, and improvedchopper design offers another factor of 1.5 – 2 at equalresolution (optimized chopper positions, counter-rotatingchoppers, trapezoidal pulses). This safely adds up to morethan the 1600 fold improvement quoted in the table, of whichabout a factor of 9-10 is being realized by the currentreconstruction of IN5. Knowing from experience all the newscience that could be achieved by IN5 in the past, one canexpect an explosion of qualitatively new opportunities thistremendous increase of power will make happen.

Not only the highest, but the lowest numbers in the table alsomerit some attention. Single peak single crystal work with hotneutrons suffers from the already mentioned low time averageflux around 0.5 Å wavelength. The pulsed time structure of thesource offers some fringe benefits even if one uses crystalmonochromators, by making easier the identification ofspurious effects, such as taking care of the problem of higherorder reflections without the use of filters. But the real solutionhere is to work out an equivalent of the hot source, asdiscussed above. For the very similar case of triple axisspectroscopy the often invoked speculative possibility of verylow background between source pulses, if realized, mightbecome an additional and very important fringe benefit. TOFspectrometers with large detector solid angle coverage, and

ESS instruments willbenefit from both highestneutron power andenhanced instrumentdesign

50 – 1000 fold increase ofsensitivity in mostexperiments will open upnew fields of scientificopportunities

These gains are enabledby the on average 20 foldincrease of the effectiveneutron flux compared tothe best today

Single crystal methodsand studies can also havegreat potentials at ESS

31

consequently wide open detector geometry, are less likely to achieve very low backgrounds needed to bring extremely week signals to evidence. For this reason, even with their very limited data collection rates, TAS instruments might well have a role play at pulsed spallation sources with time average fluxes approaching that of ILL.

Table 1: Expected performance of generic instruments on the various target station options for ESS. � : first choice or one of essentially equivalent first choice options. � : second choice: about a factor of 2 inferior in data collection rate to the first choice. The numbers correspond to instruments implemented on the best (one of the bests) target options. Blue numbers: compared to ILL beams and best existing instruments at ILL. Black numbers: compared to ISIS beams and best existing ISIS instruments.

Instrument 50 Hz 5 MW

10 Hz 1 MW

162/3 Hz 5 MW

Source gain

Total gain

High energy chopper �� � � 30 30 Thermal chopper �� � � 30 240 Cold chopper �� �� �� 50 1600 Variable, cold chopper �� � �� 20 800 Backscattering 0.8 ��eV �� � � 25 50

Backscattering 17 �eV �� � � 150 600 Molecular sp. (TOSCA) �� � � 50 100 Electron Volt Spectr. �� � � 30 300 High resolution NSE � � �� 10 100 Wide angle NSE � � �� 9 300 Triple-Axis �� � �� 0.5-1 1-4 High Resolution Single X �� � � >>10 >>10 Chemical Single X �� � � >>10 >>10 High Resolution Protein �� � � >20 >20 Low Resolution Protein �� � �� 3-5 3-5 Single Peak incl. Cryopad �� � �� 0.3-3 0.3-3 High Resolution Powder �� �� �� 50 150 High Q Powder �� �� � 60 120 Magnetic Powder �� �� �� 60 60 High Res. Reflectometer �� �� �� 20 40 High Intensity Reflectom. �� � �� 15 40 Liquids Diffractometer �� � � 20 20 High Intensity SANS � � �� 8 100 High � Resolution SANS �� �� � 150 300 Engineering Diffractometer �� � � 30 90 Fundamental Physics �� � �� 1 NA Diffuse scattering (D7) �� � �� 15 300 Backscatt. (Musical) � � �� 40 40 Average (geometrical) �

��

� �

��

� �

��

� >19 >47

32

Technical risks and challenges The reference instrumentation techniques assumed in the evaluation of the expected performance of a set of generic instruments at ESS, as summarized in the topical reports by the various Instrument Groups, were chosen to be fairly conservative, and with a few exceptions well established at current spallation sources. The most notable exception is the use of fast disc choppers to cut out very short pulses from the longish (>0.2 ms) coupled cold moderator pulses in order to achieve sub �eV resolution over 200 m flight path in backscattering spectroscopy. The more unusual approaches discussed in the present paper have not yet been implemented, but they are largely based on using components well established on reactor sources, e.g. fast disc chopper systems consisting of 6 – 8 phased rotors. Background issues, in particular the direct fast neutron bursts from the proton pulses will certainly represent some technical challenge, as does the need to increase the maximum acceptable count rate of detectors. Nevertheless, the technical risks involved by now in the instrument performance assessment is rather small. There are higher risks in the accelerator and target-moderator system design. Delivering 100 kJ beam energy in every 1-2 �s short proton pulse is at the limit both of accelerator technology at < 2 GeV proton energy and of target material resistance. It is therefore not excluded that the ESS goal of 5 MW total power in short pulses might only be achieved in 50 kJ pulses at 100 Hz frequency. This choice would be of no consequence for epithermal and hot neutron experiments and also for the majority of neutron scattering work in the thermal wavelength regime of � < 2 Å. On the other hand, the source performance would be reduced by a factor of 2 for longer neutron wavelengths. This would actually imply that the long pulses at 16.7 Hz and 5 MW, eventually shaped by choppers and using the frame multiplication scheme described above, provide superior performance compared to any type of the liquid H2 moderator for all cold neutron applications. (Only the poisoned methane would do marginally better than the long pulse). On the other hand, in ms long pulses even 500 kJ or more per pulse appears to be well on the safe side (i.e. up to 8 –10 MW at 16.7 Hz). It is expected that by using a target station design specifically optimized for long pulses, the neutron efficiency could be increased substantially, maybe by a factor of 2. This shift of relative neutron flux would offer, as discussed in the previous paragraph in connection with the 100 Hz short pulse option, an advantageous long pulse alternative to the short pulse cold moderator beams even at 5 MW 50 Hz. What this all comes down to is that there is a low risk technical alternative to achieve at least the planned ESS neutronic performance even if only 50 kJ short pulses prove to be feasible: namely 100 Hz 5 MW short pulse operation for most thermal neutron and higher energy application and 16.7 Hz 5 MW (or more) for the rest.

Established spallation and reactor instrument techniques combined offer a solid basis for instrument design and innovation. The ESS accelerator – target design goals are at the technical limits. New instrument design concepts and further optimisation of target-moderator performance offer alternatives for reducing technical risks.

33

Another advantageous implication of a successfuloptimization of the long pulse target station couldparadoxically be enhanced hot and thermal flux on the shortpulse station. Namely, if the long pulse station could take careof all the long wavelength applications, it could be sufficientto install two moderators (instead of 4 ) on the 50 Hz targetstation, which would enhance their flux by an estimated 20 %or more.

ConclusionThe evaluation of expected performance of ESS by theInstrument Task Group was deliberately based onconservative assumptions both in instrumentation techniquesand neutron beam parameters. The results confirm the highexpectation of decisive benefits for all kinds of applications inneutron scattering. The few new, not yet establishedinstrumentation concepts discussed here will certainly befollowed by many others. They show potentials to furtherenhance the ESS performance compared to the currentcautious estimates and to reduce the technical risks byproviding more technical alternatives.

References

[1] see www.hmi.de/bereiche/SF/ess/ , select: Instrumentation task group

[2] see www.ill.fr, select: yellow book

[3] G. Bauer, private communication

[4] F. Mezei and M. Russina, Physica B 283, 318(2000)

[5] B. Alefeld, Proc. of Workshop on Neutr. Sc. Instrumentation for SNQ (KFA, Jülich, 1984)

[6] F. Mezei, J. of Neutron Res., 6 (1997) 3; M. Russina, F. Mezei and F. Trouw, Proc. ICANS XV, Nov. 2000,Tsukuba, Japan, (in press)

[7] F. Mezei and M. Russina (to be published)

[8] Instrument Group Report: “Performance evaluation for a set of generic instruments on ESS" ed. byF. Mezei and R. Eccleston

34

7.1 Solid State Physics

R. Currat1, R. Cywinski2, C. Fermon3, A. Furrer4, B. Keimer5, G.H. Lander6,D.F. McMorrow7, H.R. Ott8, C. Vettier1

1Institut Laue Langevin, 6 rue de Holowitz, F-38042 Grenoble 9, France2 Department of Physics and Astronomy, University of Leeds, Leeds LS2 9JT , UK3 Laboratoire Léon Brillouin , CEA, Saclay, F-91191 Gif-sur-Yvette Cedex, France4 Laboratorium für Neutronenstreuung, ETH Zürich and PSI Villigen, CH-5232 Villigen PSI, Switzerland5 MPI für Festkörperforschung, Heisenbergstraße 1, D-70569 Stuttgart, Germany6 Institut für Transurane, Postfach 23 40, D-76125 Karlsruhe, Germany7 Condensed Matter Physics and Chem. Dep., Risø National Laboratory, P.O. Box 49, DK-4000 Roskilde, Denmark8 Laboratorium für Festkörperphysik, ETH Zürich, CH-8093Zürich, Switzerland

AbstractSolid state physics encompasses fundamental research that has underpinned much of the technologicalprogress in the last 50 years. Recent trends include the emphasis on complexity, including organic materials,and reduced dimensionality down to the scale of quantum dots. In the large variety of instruments used in solid-state physics, scattering has a special place as it gives information on spatial correlations. Within scatteringtechniques, neutrons are unique as they are able to provide, simultaneously, information on both the static anddynamical correlations. We discuss these advantages for neutrons, stress the materials-driven nature of thisapproach, and present a selection (by no means inclusive) of flagship experiments that will be possible only atthe ESS. We conclude with a discussion of the best instruments and pulse structure for frontier experiments atthe ESS.

I. IntroductionThis report identifies future research frontiers in solid statephysics. It starts by reviewing briefly general capabilities ofneutron scattering methods for the study of phenomena incondensed matter physics, with particular emphasis on thecapabilities provided by powerful modern spallation neutronsources, such as the ESS. Although complementary methodssuch as the other scattering probe, synchrotron X-rays, aswell as local probes like NMR, EPR and Mössbauer, provideimportant information, our deliberations have confirmed theunique opportunities afforded by neutrons in general and byESS in particular. The report highlights examples of flagshipexperiments, and addresses the impact that the ESS canhave in these frontier areas. Recommendations forinstruments and target options are presented.

Research in solid state is continuing to have a high impact,both in basic physics as well as with respect to technologicalapplications. Recent examples include high-Tc and otherunconventional superconductors, low dimensionalsemiconductor structures, magnetic thin films, and materialswith giant magneto-resistance. A great deal of basic researchis now technology driven, and much is oriented towardsnanometer-scale systems. The properties of nano-patternsand self-assembled quantum dots are of great interest fromboth a theoretical and experimental perspective.

In solid state physics, the degrees of freedom and interactionsnecessitate the use of a large variety of experimentalmethods. Bulk measurements of thermal and transportproperties are invariably the first step, but several techniquesare needed which are sensitive to the atomic environment andelectronic structure, and which provide information on the spinstate of the electron. Electron correlations in the solid state

Advances in solid statephysics are at the root ofmost technologiesshaping today’s world.Neutrons are key to ourunderstanding of solids.The ESS will have a largeimpact on cutting-edgeresearch in solid statephysics

Frontier topics includenovel superconductors,low dimensionalstructures, andphenomena at thenanometerscale

Because of a uniquecombination of properties,neutrons are a powerfuland indispensable probeof solids

35

exist over vast ranges of spatial and temporal (energy) lengthscales. Local probes such as STM, and atomic spectroscopicstudies using electromagnetic radiation provide crucialinformation. In addition, scattering methods (neutrons, x-raysand electrons) provide inter-atomic information on spatialcorrelations. However, the neutron has a unique combinationof properties that make it indispensable for many problems insolid state physics. The de Broglie wavelength of thermalneutrons is on the same scale as inter-atomic spacings,allowing diffraction experiments to be conducted to locate thepositions of atoms. Because of their mass, neutron have arather low kinetic energy; they can be moderated and neutronbeams are produced with energy in the range 0.1 meV to10 eV, well matched to solid state excitations. (This is to becontrasted with x-rays, which at comparable wavelengths aremuch more energetic, in the keV range.) Because of theirmagnetic moment (spin), neutrons are sensitive to magneticmoments arising from electronic and nuclear magneticmoments. Because they are uncharged, neutrons penetratedeep into materials. Because they are weakly interacting (incontrast to electrons), measured scattering cross-sections canbe compared directly to theory.

Neutrons have played a pivotal role in the investigations ofphase transitions and co-operative phenomena, magnetism,structure (static and dynamics), as well as in many otherfields. Particularly intriguing is the connection between phasetransitions and theoretical concepts, such as symmetrybreaking, order parameter, universality class, scaling andcritical behaviour. In the past, the interplay betweentheoretical concepts and experimental observationsconcerning phase transitions has been extremely successfuland many significant contributions have been made usingneutron scattering.

A vast panoply of neutron techniques have contributed to thiswork. Neutron diffraction (from powders and single crystals) isa basic, but essential, technique, providing information onchemical and magnetic structures. Neutron reflectometryusing polarised neutrons has given us a clearer picture of thegrowth and the physics of magnetic thin films andsuperlattices. Inelastic neutron scattering is the only probethat provides a complete picture of both structural andmagnetic dynamics in solids. Emerging techniques includeanalysis of three dimensional polarisation, and the directmapping of the full dynamical susceptibility over the entireBrillouin zone. All of these techniques suffer from the intrinsiclow brilliance of neutron sources; as a result, studied materialsmust have sizeable volume and/or sizeable scatteringdensities. The (lateral and vertical) spatial resolution isrestricted to a few 100 µm while the temporal resolution is inthe 0.1 sec range. The advent of the ESS will offer entirelynew capabilities to explore spatial and temporal properties ofcondensed matter with µm and ms resolution, respectively.

The interplay betweenneutron experiments andtheory has driven thedevelopment of many newconcepts in solid statephysics

Low intensity continues tobe the major limitation ofneutron research. Due toits high flux, the ESS willopen up entirely newopportunities

36

II. Frontiers in Solid State PhysicsThe future challenges in basic solid state physics are theexploration and the understanding of collective behavior oflarge numbers of interacting particles. Although future trendsare notoriously difficult to predict, two important directionsemerge. Firstly the tendency to higher complexity, specificallymaterials which have physical properties determined bycompeting interactions, and secondly the trend to reduceddimensionality, both by synthesizing materials with lowdimensional structural elements and by reducing the physicalsize of objects to surfaces and interfaces, single atom wiresand dots. Of basic interest in solid state physics is to establishthe ground state of relevant systems. This may be done byexploring possible excitations out of the ground state, andneutrons are a versatile, and often unique probe with which toaccomplish this goal.

The table summarizes some of the research areas that areexpected to be of major interest in ten years time.

Complexity and reduceddimensionality are over-riding themes of futureresearch on solids

Table 1:Frontier Research Areas in Solid State Physics

Dimensionality ComplexityStructures andlattice effects

Non-equilibriumand time-dependentphenomena

New Materials

Quantum dot arrays

Transport andmagnetic properties

in 1-d systems

Domains walls,domains

correlations, grainboundaries

Surfaces and thinfilms

Interplay of spin,orbital and chargedegree of freedom

Coupled excitations

Strongly interactingelectron systems

Flux line lattices

Phase transitions,quantum critical

points

Frustration

Disorder, interfacialroughness

Proximity effects

Lattice modes

Confinement

Fast response toexternal probes and

fields

Magnetic fluctuationsand relaxations

Tunneling

Molecular magnets

Interfaces/hybridstructures

Self-organisingmolecular systems

Novel magnets andsuperconductors

Organic materials

Here we emphasize a few topics where neutron scatteringtechniques are expected to play a major role. In magnetism,significant advances are expected in synthesizing molecularand organic magnets, i.e. solids built from structurally well-defined clusters containing magnetic ions in a complexenvironment. These are of both fundamental importance, andwith respect to potential application in magnetic storagedevices. New developments are also expected in exploringnovel magnetic phases and their dynamics in low-dimensionalsystems. The study of phase transitions will continue to bea major field of research with neutron based techniques.Systems of high complexity, exhibiting extreme many bodyeffects (e.g. unconventional superconductivity) and lowdimensional features are known and expected to undergo alarge variety of phase transitions. Their exploration usingneutron techniques will provide crucial insights into the

Research at the ESS couldlead to breakthroughs inquantum magnetism,many body physics, andother frontier fields

37

microscopic mechanisms causing these phenomena. Of highcurrent and most likely future interest is the relationshipbetween spin polarization and transport of conductionelectrons in specially tailored materials, spintronics. Highintensity neutron beams will play a central role in elucidatingthe spin polarization and dynamics of these electrons.

III. Future Opportunities at the ESSThe ESS will lead to breakthroughs in three distinct ways:

a) to allow scientists to address new problems, and to asknew questions.

b) to provide new tools to tackle problems at the researchfrontiers.

c) to offer high quality experimental data for unambiguousdiscrimination between theoretical models.

In the following, we present selected flagship areas which arerepresentative of the topics listed above. Dynamics of Superlattices, Thin Films, Wires and DotsFollowing the discovery of giant magneto-resistance in 1988,the physics of micro- and nano-structured magnetic materialshas become a field of intense activity. Thin films andsuperlattices, as well as wires and dots are now extensivelystudied for their fundamental properties and their potentialapplications in systems like sensors and magnetic randomaccess memory devices (MRAMS). Understanding thedynamics of these systems will continue to be a keychallenge. In contrast to Brillouin light scattering andferromagnetic resonance techniques, neutron scatteringgives access to the whole Brillouin zone.

Brillouin light scattering has provided the first dispersioncurves in magnetic dots, but is limited to 30 GHz [1]. Untilnow, there is no theory able to explain the excitationsmeasured in dots, even in simple NiFe square dots. Anexperimental input at higher frequencies appears to beessential in understanding the dynamics of these systems. Areflectometer at the ESS will offer the capability of measuringspin wave spectra in very thin films, wires and dots, and willcertainly have an important impact on the field of nano-magnetism.

The observation of magnetic inelastic scattering fromsuperlattices is presently at the limits of neutron technology.An interesting experiment has been performed recently at theILL on a Dy/Y superlattice, where the effects of folding on theinelastic response function due to the superlattice periodicityhave been observed [2]. Such neutron experiments allow theexchange coupling parameters, both within a single layer andbetween the layers, to be deduced. However, this will requirea considerable increase in intensity, coupled with betterresolution, especially if technologically important films such astransition-metal superlattices are to be examined and

Neutron beams at the ESSwill provide maps of themagnetic polarizationand dynamics of nano-structured materials anddevices

38

understood.

Molecular MagnetsA typical example of a molecular magnet is Mn12 acetate withtotal spin quantum number S=10, giving rise to thousands ofexcited spin states which can only be disentangled by highresolution neutron spectroscopy. For instance, the lowest lyinggroup of spin states comprises (2S+1) = 21 levels of energiesηω ≤ 1.2 meV as shown in figure 1 below. Mn12 acetateexhibits quantum tunneling between these spin states whichcan be tuned in a controlled manner by an applied magneticfield [3]. This opens the way to a novel class of informationstorage systems on the molecular level. Unfortunately, thequantum tunneling in Mn12 acetate is restricted totemperatures of a few Kelvin. The search is on for materialsthat would preserve the virtues of Mn12 acetate at liquidnitrogen temperature.

Figure 1: Energy spectra observed at three different temperatures inMn12 acetate.

Spin Density Waves in Organic MaterialsAmong the low-dimensional electronic systems, the chargetransfer Bechgaard salts (TMTTF)2X and (TMTSF)2X (X=PF6,AsF6, SbF6, SCN) show the richest phase diagrams withalmost all known electronic phases: a metal, a paramagneticinsulator, spin and charge density wave states, a spin-Peierlsstate and finally an unconventional superconducting state.

Molecular magnets couldserve in atomic-scaleinformation storagesystems. Neutronscattering is a uniqueprobe of their excitationspectra whose accuracywill be tremendouslyenhanced by the ESS

39

Other salts in the same family have been shown to exhibit thequantum Hall effect. The phase diagrams have been mappedout mostly based on transport and specific heat data as wellas NMR results [4].

For the spin density wave (SDW) phases detailed NMRpredictions exist concerning the wavevector of the modulationand the amplitude of the ordered moment. Yet so far directneutron evidence for a SDW superlattice peak is missing, dueto a combination of the weak magnetic moment (~ 0.08µB),the unfavorable magnetic form factor and small sample sizes.With the intensity available at the ESS, experiments of thistype will become possible, allowing a direct determination ofthe SDW amplitudes and periodicities, thereby opening anentirely new territory of research.

The ESS may enable thefirst direct observation oftiny magnetic momentscentral to current theoriesof organic conductors

Figure 2: Phase diagram of the Bechgaard salt (TMTTF)2PF6 deducedfrom resistivity [4].

Revealing Exotic InteractionsThe properties of magnetic materials are usually described interms of bilinear spin interactions. Neutron spectroscopy withits dipole selection rule ∆M=1 has been the technique ofchoice to measure the magnetic excitation spectrum andthereby, to allow the direct determination of the magneticexchange coupling constants. However, there are numerousexamples such as molecular magnets, high Tc cuprates and f-electrons compounds where higher-order interactions (e.g.quadrupolar, octupolar, three- and four-body exchange) arerelevant, but their sizes could so far not be determineddirectly. In principle, neutron scattering allows the directobservation of higher-order term transitions, however, theassociated transition matrix elements are typically two ordersof magnitude smaller than for dipolar scattering [5]. Suchnovel experiments would be made possible by the ESS.

Theoretically predicted(but hitherto unobserved)interactions in solids willbe discovered andexploited at the ESS

40

Coupled ExcitationsExcitation phenomena in solids can be classified into singleparticle continua and collective modes. Because of intensityconstraints, neutron scattering experiments have been almostexclusively limited to collective modes. Recent investigationsof two-spinon continua in insulating one- and two-dimensionalquantum magnets are pushing the limits of current sources. Ata chopper spectrometer at the ESS with wide reciprocal spacecoverage, detailed maps of single particle Stoner continua inmetals and superconductors will be obtained up to energies ofthe order of the Fermi energy. In systems where correlationeffects are strong (which are currently at the forefront of thecondensed matter science) it will be possible to extract awealth of information on the band dispersions, Fermi liquidparameters, superconducting coherence effect, etc., that iscurrently inaccessible. The high neutron flux at the ESS willalso enable high resolution measurements of the intrinsiclifetimes of collective modes over the entire Brillouin zone.Although predictions for the lifetimes of magnetic and latticevibrational excitations (for instance, due to electron-phononscattering) have been available for many years and arebecoming ever more accurate, they could thus far be testedonly in a few special cases.

Physics of defects at the dilute limitThe increasing sophistication of “first principles” theoreticalcalculations of the fundamental electronic structure, totalenergies and atomic short range order places stringentdemands upon the accuracy of experimental measurementsof the extended atomic and magnetic defects around impurityatoms in metals, alloys and compounds. Experimentally,information on this problem can be obtained only from diffuseneutron scattering experiments, with polarisation analysis. Theassociated cross sections are extremely small and countingtimes are often prohibitively long. Moreover, such experimentsshould ideally be carried out at extreme dilution to circumventthe often intractable problem of non-linear superposition ofoverlapping defects. These experiments are crucial for a fullsolution of the defect problem and experimental corroborationof the most sophisticated of our “first principles” bandtheoretical calculations, but is not feasible at present neutronsources.

Spin glass dynamicsIt has been said that the deepest and most interestingunsolved problem in solid state physics is probably the natureof glass and the glass transition. Indeed the status of theglass transition as a true thermodynamic transition is stillquestioned. Spin glasses provide a simple analogue of thestructural glasses yet the accurate measurement of relaxationprocesses in spin glass systems is at the very limits of what isfeasible at present using spin echo facilities. A wider Fouriertime, coupled with significantly improved counting statisticsare required to determine the precise functional form of the

Inelastic neutronscattering at the ESS willopen a new window on theelectronic structure ofmetals andsuperconductors

Defects are ubiquitous insolids, and high intensityneutron beams at the ESSwill provide incisiveinformation about theirmicroscopic structure

The ESS will allowmeasurements of thecomplex dynamics nearthe glass transition withunprecedented precision

41

relaxational dynamics, both for comparison with structuralglasses and to discriminate between the proposed theoreticalmodels. In addition the measurements should also beperformed over a wide range of magnetic dilution, the lowerranges of which are entirely inaccessible at present. Theimplications of such studies are profound, as many of the spinglass relaxational models are finding applications in areas asdiverse as virus mutation, protein folding and the travellingsalesman problem.

Figure 3: Temperature dependence of the magnetic response of UPd2Al3near the superconducting phase transition. Above the phasetransition, there is coexistence of quasi-elastic scattering anddispersive modes. In the superconducting state the quasielasticscattering is replaced by a low-lying inelastic mode (data takenfrom [6]).

Quantum Phase TransitionsOf special interest are situations where either a competition ofdifferent interactions prevents the system from readily

42

adopting a well-defined ground state or, where a restriction inspatial dimensionality does not allow for long-range orderingphenomena. Interactions of similar, medium or largemagnitude may lead to very complicated phase diagrams.Very challenging are cases where all the interactions are ofsimilar strength and weak. Here the situation may occur thatthe phase transition only sets in at T=0K, in the quantumcritical regime. A related example would be the important andtechnically challenging experiment on a material such as therecently discovered ferromagnetic superconductor UGe2.Inelastic scattering experiments need to be performed at lowtemperature (< 0.5 K) and at pressures of up to 3 GPa. Thekey requirement is to map out the inelastic response functionover a wide range of momentum space and energy as afunction of temperature, external pressure and appliedmagnetic field. While neutron scattering can make unique andessential contributions to our understanding of themechanisms underlying these phenomena, the samplevolumes that can be subjected to these extreme conditionsare necessarily small.

At current neutron sources, inelastic scattering studies arehence restricted to a few special cases. Detailedinvestigations of the dynamic aspects of magnetic field and/orpressure induced quantum phase transitions require the ESS.

It should be stressed that in such experiments the positions inmomentum space where the maxima will occur are unknown,so that a wide “mapping” technique is required. Many newmaterials call for this approach; unavailable at present, but aplanned development at the ESS.

Experiments at the ESSwill set new benchmarksfor extreme conditions ofexternal pressure andmagnetic field, thusproviding importantinsights into zero-temperature phasetransitions driven byquantum fluctuations

IV. Instrument Requirements at ESSThe requirements for instrumentation in solid state physics arebased upon two principle demands: that of probing (Q,ω)space for excitations, and that of providing a detailedstructural (atomic and magnetic) characterisation of thesample. Whilst the former demand can only be met by a suiteof inelastic scattering spectrometers, the latter must be by arange of rather disparate total- and elastic scatteringinstruments, namely powder and single crystaldiffractometers, diffuse scattering instruments andreflectometers. In all cases polarised incident neutrons andpolarisation analysis are either essential or a distinctadvantage, and the instruments should be capable ofaccepting extreme sample environments (e. g. pressure,magnetic field, temperature etc.).

The key scientific topics we have highlighted demand acoverage of Q space from 0.1≤ Q≤ 12Å-1 and an energytransfer range of six orders of magnitude, from µeVs to eVs.This can be achieved through an instrument suite consistingof the 17 µeV backscattering spectrometer, the variableresolution cold chopper spectrometer and the thermal andhigh energy chopper spectrometers. In addition we include aconstant Q spectrometer of the PRISMA type, principally for

A suite of high-performance instrumentsis required to fully realizethe potential of the ESSfor solid state physics

A wide coverage of energyand momentum space isessential

43

the discrimination against inelastic processes as a procedurefor reducing the intrinsic sample generated backgrounds incrystallographic studies of small moment systems. Theresulting coverage of Q-ω space is illustrated in the diagrambelow.

10-6

10-5

10-3

10-4

10-2

10-1

10-0Energytransfer(eV)

Momentumtransfer ( A-1 )

2 4 6 8 1210

Backscatteringspectrometer(17 µeV)

Variable resolution coldchopper spectrometer

Thermalchopperspectrometer

High energy chopperspectrometer

ConstantQ spectrometer(for elastic discrimination)

Figure 4: Required coverage in the Q-ω space.

The "chemical single crystal" diffractometer is an essentialcomponent of the suite for determination for the structural andmagnetic order and for crystallographic studies of multilayersystems. The specifications for this instrument are appropriatefor such studies, but in addition polarised incident neutronsand high magnetic fields at the sample position will enablemagnetic spin density determination of the type currentlyperformed, uniquely, on D3 at ILL.

Solid state physics places several conflicting demands onpowder diffractometry. Firstly the growing complexity ofmagnetic structures (e. g. spirals and spin density waves)studied by neutron diffraction requires high resolution (∆ d/d>0.1%) across a Q-range from a lower limit of at least 0.3 Å-1

to 12 Å-1. While this range is adequately covered by the highresolution powder diffractometer, it is not at all clear that theresolute constraints can be met over most of this range. The"magnetic diffractometer" has too low a resolution for studiesof extremely complex structures. The efficient mapping of theevolution of magnetic and structural phases in parameterspace defined by pressure, temperature magnetic field andconcentrate is best met by a "Super Gem" diffractometerwhich does not yet appear on the instrument suite. There ismuch new physics to be explored under extremes of pressure+ (pulsed) magnetic fields. In many respects the "Engineeringdiffractometer" with its well defined gauge volume is wellsuited to this purpose.

The exploration of structural and magnetic defects at thedilute limit in alloys, metals and compounds, requires a high

The instruments neededfor solid state physics arebest placed at the50 Hz 5 MW and16.7 Hz 5 MWtarget stations

44

countrate diffuse scattering instrument of the D7 type with fullX-, Y-, Z-polarisation analysis. Such an instrument must becapable of measuring diffuse scattering at the mb/st/atomlevel, and should be optimised to minimize multiple Braggprocesses.

The study of the structures of artificial films, multilayers andmesoscopic structures requires a spectrometer optimised forintensity rather than resolution. Polarisation analysis andsurface capabilities are essential. The quoted dynamic rangeof 108 up to a Q of 0.512 Å-1 represents a real advance inreflectometry and should be considered to be a design goal.

V. Target Requirements at ESSMost of the above instruments are best placed at the 50 Hz 5MW station, one could be equally well suited to the 50 Hz or16 Hz stations, whereas one can only be placed at the 16Hztarget. We have no demand for the 10 Hz 1 MW target. Ourconclusions are summarised below:

Inelastic instruments: 50 Hz 5 MW targetHigh Energy ChopperThermal ChopperVariable resolution cold chopperBackscattering 17 µeVConstant Q spectrometer

Diffractometers: 50 Hz 5MW targetChemical single crystal (with PN)High resolution powder (modified)Engineering diffractometer(Super Gem)

Diffuse Scattering diffractometer: 16.6 Hz 5MW targetD7-type

Reflectometers: 16.6 Hz 5 MW / 50 Hz 5 MWHigh intensity version

References

[1] J. Jorzick et al. J. Appl. Phys. 87, 5082 (2000)

[2] A. Schreyer et al. J. Appl. Phys. 87, 5443 (2000)

[3] I. Mirebeau et al. Phys. rev. Lett. 83, 628 (1999)

[4] D. Jaccard et al. J. Phys. Condens. Matter 13, L89 (2001)

[5] P.M. Levy and G.T. Trammell, J.Phys. C: Solid State Phys. 10, 1303 (1977)

[6] N. Bernhoeft et al., Phys. Rev. Lett. 81, 4244 (1998)

45

7.2 Materials Science and Engineering at the ESS

M. Ceretti1, R. Cowley2, M.R. Daymond3, T. Lorentzen4, A. Magerl5, F.M. Mulder6,P.J. Withers7, H. Zabel8

1Laboratoire Léon Brillouin (CEA-CNRS), CEA Saclay, F-91191 Gif sur Yvette, France2Oxford Physics, Clarendon Laboratory, Park Road, Oxford OX1 3P4, UK3ISIS Facility, Rutherford-Appleton Laboratory, Chilton OX11 0QX, UK4DanStir ApS, Danish Stir Welding Technology, CAT Science Park, Frederiksborgvej 399, Box 30, DK-4000 Roskilde, Denmark5 Lehrstuhl für Kristallographie und Strukturphysik, University of Erlangen-Nürnberg, D-91054 Erlangen, Germany6IRI, Delft University of Technology, Mekelweg 15, 2629 JB Delft, The Netherlands7Manchester Materials Science Center, Grosvenor St. Manchester, M1 7HS, UK8Ruhr-Unversität Bochum, Fakultät für Physik und Astronomie, D-44780 Bochum, Germany

AbstractMaterials science and engineering provide the keys to future technologies, economic wealth and sustainablegrowth. They are also the keys to mastering many of the challenges for the next generation, such as thedevelopment of new energy sources and the reduction of pollution. Because of their many different facets,materials scientists and engineers use a large number of experimental techniques. Neutron scattering has alwaysbeen an important tool for the provision of structural information on the atomic scale and for the understanding ofdynamical properties of solids and liquids. However, in the early days of neutron scattering this technique was anexquisite tool in the hands of a few specialists. Because of its sustained success in providing unique answers tomaterials science problems, neutron scattering has become increasingly popular among materials scientists andlikewise engineers. Today the advance of numerous materials science topics relies heavily on the availability ofstrong neutron sources. The strongest neutron sources available at the present time are often at the limit of theircapabilities. Presently the data acquisition time is often too slow for in-situ and real-time studies of kineticprocesses and process monitoring. Modern technologies demand information from smaller sampling regions,sometimes buried in larger component volumes or environmental chambers, samples in complex environments,samples in real time and from samples in extreme fields. The next generation of high power pulsed neutronsources will dramatically alter the nature of the experiments which are possible, allowing for the first timeinvestigations of materials in real time, with realistic dimensions and in real conditions. Improved data acquisitionwill enable pump-probe type experiments, to study the spin structure in magnetic nanostructures for informationtechnology devices, and to provide high resolution three dimensional maps of stresses and textures ofengineering components during fatigue cycling. In addition, the time structure unique to pulsed sources will beexploited, for example in chemically sensitive radio- and tomographic imaging. Therefore, the ESS will be anextremely powerful tool for the analysis of advanced materials, which are of scientific, commercial and practicalinterest. In this respect the ESS will be highly significant for maintaining the competitive edge of materials sciencewithin the European Union as compared with that in other parts of the world.

In order to capitalize on the opportunities the ESS offers for neutron research in the area of Materials Science andEngineering, several instruments are required, including an engineering diffractometer for stress-strain analysis, ahigh resolution and focussing small angle scattering for in-situ kinetic measurements of microstructure and poregrowth, a polarised neutron reflectometer for the analysis of magnetic nanostructures, a tomography andradiography instrument for imaging large industrial components, and a variable resolution cold chopperspectrometer for the analysis of the internal dynamics of atoms and molecules.

Most of the demands for the listed instruments can be satisfied with a 50 Hz short pulse and a 16.7 Hz long pulsetarget option. The short pulse option is required for instruments requiring high resolution such as for highresolution powder diffraction and for high resolution backscattering spectrometer, while the long pulse catersinstruments in demand of high intensity, such as high intensity neutron reflectometers, focussing small anglescattering, and neutron diffractometers for the analysis of diffuse scattering.

I. IntroductionThe road to our present day use of a vast variety of novelmaterials and engineering processes is marked by anoutstanding array of discoveries, inventions, and theoreticalinsights. The start of modern materials research was datedback to about 50 years ago. Since then the transistor hasbeen invented, superconductivity explained, high temperaturesuperconductors and the quantum hall effect have beendiscovered, semiconductor lasers and magnetoresistivereading heads for hard disks have been developed, and finiteelement modelling has revolutionised structural and process

Materials sciencedevelopment during thepast 50 years

46

design. New structural materials have become available suchas high temperature and corrosive resistant steels, light weightfoam metals, carbon fibre reinforced materials, shape memoryalloys, superalloys, photo-voltaic materials, energy conversionand storage materials. Material processing techniques haveseen similar advances through new friction based solid statewelding techniques, surface hardening treatments, andcorrosive resistant coatings. At the same time, theunderstanding of more traditional building materials such asconcrete has dramatically improved. These are but a fewexamples to illustrate the remarkable progress materialsscientists have achieved in recent years.

II. The role of neutron scatteringNeutron scattering has established itself as one of the mostimportant tools for the analysis of materials. Metals, ceramics,and their composites, semiconductors, superconductors,nanophase materials, liquids, polymers, paints, lubricants,concrete, coal, wood, bones and biomaterials as well as manyother materials have been analysed with great success usingneutron scattering. The results are accessible throughpublications and data banks. Engineering developments andapplications draw heavily from this body of work.

Contributions of neutronscattering to materialsscience and engineering

In contrast to other experimental techniques, neutronscattering provides access not only to the positional order ofthe atoms but also to their dynamics. Depending on thematerial chosen, the motion of atoms in thermal equilibriumranges from local vibration of atoms in solids, diffusion ofatoms in liquids and solids, creeping motion in polymers,rotation of molecular sidegroups, and tunnelling of atomsthrough potential barriers. As a result of these investigationsthe elastic properties of materials are very well understood,the danger from hydrogen embrittlement of steels can bepredicted, and the dependence of the lifetime for metal –semiconductor junctions on the operation temperature can bedetermined.

Strengths of neutronscattering for materialsscience and engineeringquestions

In many situations neutron scattering is an indispensable tool,since no other method provides information within the samespace – time window as neutron scattering does. Withoutneutron scattering our knowledge of the vibrational propertiesof atoms, the diffusional dynamics of liquids and polymers, thejump vectors of interstitials in host lattices, and the rotationaldynamics of molecules would be much more limited. Withoutneutron scattering we would not have information on theoxygen ordering in high temperature superconductors and onthe hydrogen diffusion motion in metals. Furthermore,neutrons have a magnetic moment which interacts with themicroscopic moments in solids and liquids and with magneticfield distributions in superconductors. Without neutronscattering we would not know about the ordering of magneticmoments in ferro- and antiferromagnetic materials, fluxordering and melting in superconductors, about the vortexstates in low dimensional magnetic systems, and spin disorderin frustrated magnetic alloys.

Examples forcontributions of neutronscattering to materialsscience and engineeringproblems

Dynamical properties ofliquids solids

Oxygen and hydrogenordering

Ferro- andantiferromagneticstructures

47

Neutrons can also penetrate many centimetres throughengineering materials, allowing non-destructive studies oflarge components and samples in complex environmental orprocessing chambers. Neutrons are thus particularly wellsuited to non-destructive studies of components in their as-fabricated and in-service condition such as engine parts. Atthe same time, neutrons take a volume average such thatresults are relevant for the properties of real materials.

Non-destructive analysisof engineeringcomponents

Volume average

Whenever new materials become available, neutron scatteringplays a key role in providing a microscopic understanding oftheir structural, dynamic and magnetic properties. Thisunderstanding is vital for the development of these materialsin technological applications. This is shown by the improvedinsight that has been gained into high-Tc compounds, GMRand CMR materials, fullerenes, and other materials of presentinterest. This will undoubtedly also be the case for any newclass of materials that are discovered. The requirements forthe studies of these materials and their microscopic propertiesare detailed elsewhere in this document.

Neutron diffraction also provides important insights into newand established processing methods, from the stressescaused by new joining techniques to the molecular alignmentof polymer molecules during plastic manufacture.

Although neutron source fluxes are lower than those of thenew generation of high brilliance synchrotron sources,neutrons nevertheless remain the preferred tool in many areasof materials science. For example, neutron scattering ispreferable in many cases over x-ray scattering because of alower background and a cleaner signal, particular for magneticscattering. Furthermore, neutron scattering provides absolutestructure factors and total magnetic moments for directcomparison with theory. Only neutron scattering provides astrong contrast by isotope substitution, which is best known forthe case of hydrogen and deuterium, but is also extremelyuseful in other fields such as the study of binary metallicalloys. Furthermore neutrons provide a probe of bulkproperties deep within samples, for instance in engineeringstrain measurement.

Novel materials

Processing andmonitoring

Intensity limitation

Neutron advantage

Contrast variation withisotopes

III. New opportunities with neutrons at the ESSWith the ESS gain factors exceeding two orders of magnitudeare projected. The increased flux and intrinsic time structurewill allow the design of new materials science investigationswith innovative instrumentation inconceivable on existingneutron sources. With the provision of these improvements, ahuge impact on materials science is expected:

• Time resolved experiments, with second to milli-secondresolution;

• Higher spatial resolution to the important sub-millimetreregion for residual stress, high pressure and high

More intensity meansfaster data rate and newopportunities

48

temperature experiments and for monitoring of interfacialdiffusion and reactions at interfaces of metal and organicmultilayers;

• In-situ real time experiments, for example nucleation andrecrystallisation of undercooled liquid alloys, the ageingand fatigue of alloys under cycling conditions;

• Analysis of the spin structure and spin fluctuations oflaterally patterned arrays or magnetic nanostructures;

• Routine determination of 3D maps of stress and texturewithin engineering components;

• Neutron tomography for the production of 3D images ofmachine parts under working conditions in real time, andwith structural sensitivity;

• Pulsed radiography to study fast time dependentphenomena with isotope sensitivity.

The sustainable growth of society can only be achieved if newmaterials and material combinations are explored on all lengthscales, time frames, and under real conditions. The followingexamples are identified as important areas which can beforeseen to play a vital role within the next decade andbeyond. These areas will benefit tremendously from neutronscattering experiments at the ESS, but represent only a smallnumber of the potential applications of neutrons to materialsscience and engineering.

Chief areas of materialsscience and engineeringand new opportunities atthe ESS

LubricationIneffective lubrication leads to permanent wear and failure ofmechanical parts, which causes an estimated damage in theUSA of about 6% of the gross national product [1]. It has beendemonstrated that neutrons can unravel the structure anddynamics of lubricants in moving engineering parts [2,3].Understanding the dynamics on all length scales embracingmacroscopic flow to molecular diffusion under real loads, willlead to the development of new lubricants for extremeconditions. Presently the lubricant layer in thesemeasurements has a minimum thickness of about 0.3 mm, farmore than realistic scales of industrial interest. Studying ofreduced film thicknesses relevant to real applications (~10µm)will only become feasible at the ESS with the improvedintensity available.

Neutrons provide insightin the functioning oflubricants via the analysisof the macroscopic flowand the dynamics on amolecular level

49

Figure 1: Schematic velocity profiles for liquids under shear load. Themiddle panel shows the classical Newtonian profile with a linearvelocity gradient between the fixed and moving plates. Lubricantsoften deviate from this behaviour. Either the lubricant may stickwell to the surface of the plates, while the interior layers areweakly coupled, yielding a velocity profile as shown in the toppanel; or the interaction between the liquid layers is strong,leaving only a thin liquid layer with an intermediate velocity. TheNewtonian profile is preferred for good lubrication. Howeverunder start up conditions, profiles resembling the top or bottomsituation have been observed with high resolution inelasticneutron scattering [3]. Neutron scattering experiments test thevelocity profiles and provide simultaneously information on theinternal dynamics of the molecules and their frictional losses.

Mechanism of deformation and damageDescribing materials deformation and understanding themechanisms involved are a vital part of engineering science.Neutron scattering provides unique micro-mechanical data.The neutron sampling size is particularly well matched to thescale of engineering stress-strain concepts. However, goingbeyond model systems and timescale to real fatigue cyclingconditions and components requires data acquisition rates,which are inaccessible today. ESS will enable the assessmentof real scale components on realistic time scales. Forexample, new solid state joining techniques require moreaccurate information about the generation of residual stressesthat will add to in-service stresses to foreshorten life.Furthermore, finite element models have become the mainmethod for design and assessment of engineering structures[4]. These models cannot be developed reliably withoutaccurate information to validate them. Neutron diffraction isthe only technique that can do this, providing measurementsdeep inside most engineering materials. Model validation hasbecome vital to modern industry, as timescales for techniqueand component design process, from drawing board to finaluptake, become shorter.

Neutron scatteringprovides unique micro-mechanical data fromdeformed materials underreal conditions

Neutron diffractionprovides deep insideinformation from mostengineering materials

50

Figure 2: Crack tests are performed on engineering steels that aresubjected to fatigue cycling with a varying load. This modifies thedimensions of the plastic region close to the crack tip. The figureshows the geometry of a crack test sample and the measurementdirections. The plastic zone near the crack tip is schematised inblue. Neutron measurements have been performed in the middleof the sample (at a depth of 5mm) along the crack openingdirection y [Ref. 5]. These type of experiments can in the futurebe carried out in real time under real fatigue cycling conditionswith the pulsed spallation source.

Energy storage and conversion devicesSociety is becoming increasingly dependent on energystorage and conversion devices, such as batteries, fuel cellsand solar cells. New rechargeable cathode and anode batterymaterials should provide a higher energy density, beenvironmentally friendly and cheap in order to facilitate large-scale application of renewable energy. Optimisation of suchmaterials and devices relies on the in-situ characterisationunder operating conditions for improving the performance anddurability of such devices.

Figure 3: Schematics of a Li - ion battery. Neutrons provide the essentialinformation on the position and mobility of the Li+ ions inenvironmentally more friendly and high energy density LiMn2O4

batteries. Neutron scattering results have recently proven thecoupling between structural transitions and ordered ionic chargelocalisation in the LiMn2O4 spinel structure [Ref. 6,7]. At the ESSthe charge and discharge of the battery could be monitored underreal operating conditions, which will help improving rechargeablebattery lifetimes.

cathode electrolyte anode

e-

e-

Li+Li+

51

ESS will take structural and dynamical information fromtailored laboratory model systems to real time scales andconditions. In the drive to replace LiCoO2 with moreenvironmentally friendly and high energy density LiMn2O4

batteries, neutrons provide information on the movement ofthe Li ions and associated structural changes. Such studiestake advantage of the neutron´s inherent ability to monitor lightatoms, even in the presence of heavy atoms.

Information technologyGreater demands are being placed on new magneto-electronic devices for faster data handling and higher densitynon-volatile data storage. The device applications (MRAM,GMR heads, spin transistor) depend on the control over theswitching behaviour, hysteretic losses, and bit life times [8].These properties have their origin in the magnetic domainstructure of ultra-small artificially shaped mesas. Advancedhigh intensity neutron scattering techniques are required toprovide this information on a nanometer length scale [9].Presently magnetization profiles can be gained only fromlaterally extended films. The ESS will contribute to theunderstanding of spin structures and fluctuations in magneticdot arrays embedded in heterostructures not achievable withcurrent neutron sources.

Figure 4: Design of a non-volatile magnetic random access storage device.The device relies on the switching of magnetic domains withcurrent pulses in the read and write lines. Spin valves structuresare used for controlling the magneto-resistance depending on therelative orientation of the magnetisation vectors in the top andbottom ferromagnetic electrodes. Neutron scattering is necessaryfor understanding the magnetic domain structure and themagnetic roughness deep inside of spin valve systems.

Spin valves for readingheads and for non-volatilemagnetic accessinformation storagedevices

Process monitoring and optimisationBecause of the non-destructive deep penetration of neutrons,they are particularly well suited for imaging embeddedfeatures inaccessible by any other means [10]. Combined withthe inherent ability of time-of-flight methods to discriminatebetween different structural components, including magneticphases, ESS will add a new dimension to real scale

Imaging of embeddedobjects

52

tomography and radiography. Relatively little neutrontomography has been undertaken at present because evenradiographs take a significant time to acquire. The new sourcewill enable large field (200x200mm) sub-millimetre resolutionimages to be obtained in seconds and thus tomographyexperiments will become feasible for the first time.Furthermore the pulsed nature will allow the development of anew area of science, such as structural mapping neutrontomography.

Figure 5: Neutron tomographic imaging of an engine part. Neutrons notonly provide a high contrast from hidden parts inside ofengineering components but also a chemical contrast due tophase sensitivity. At the ESS the data acquisition rate will be highenough to monitor a running engine.

Tomography andradiography of hiddenobjects become availablein real time

Monitoring protonsThe presence of hydrogen in the atomic lattice can switch theoptical and magnetic properties of materials. Neutrons canmonitor where the hydrogen atoms go and how they changethe structure. With the ESS not only the steady state structurecan be determined, but it will also become possible to monitorhydrogen penetration and diffusion in real scale fuel cells atoperating temperatures. Yttrium metal coatings have recentlybeen discovered, which can be switched between reflectingand transparent by the charging and discharging of hydrogen[11]. Figure 6 displays the coating in the reflecting state andafter hydrogen charging in the transparent state. Neutronscattering has played a crucial role in determining the locationof the hydrogen atoms and the understanding of the basicmechanism of the switching behaviour [12,13]. Newcompounds have recently been discovered which promisefaster switching times, as well as easy and safe hydrogenhandling through electrolytic cycling. In the future hydrogenwill be crucial for energy storage materials and for energyconversion devices. Neutron scattering will continue to play avital role for the understanding and optimisation of materialsand of devices for future hydrogen based technologies.

Neutrons tell where thehydrogen atoms go andwhat they do inside of thehost lattice

53

(a) (b)

Figure 6: Two states of a Yttrium film are shown by their optical propertiesbefore and after hydrogen loading. The 500 nm thick yttrium filmis covered with a 20 nm thick Pd protection layer and placedbehind a white knight and before a chess-board. In the α phasethe film shows metallic reflectivity (a), while in the trihydride γphase the film becomes transparent for the visible light (b) [11].

IV. Instrumentation and target choicesIn order to capitalize on the opportunities the ESS offers forneutron research in the area of Materials Science andEngineering, the following instrument suite has been definedas necessary, in an approximate order of importance:• Engineering diffractometer for stress-strain analysis, able

to receive large samples and complex loadingenvironments.

• High resolution and focussing small angle scattering for in-situ kinetic measurements of microstructure and poregrowth.

• Polarized neutron reflectometer with high Q-resolution andwith high intensity for the analysis of magneticheterostructures and nanopatterns.

• Tomography and radiography instrument for theabsorption and time-of-flight imaging of large industrialcomponents.

• High resolution powder diffractometer for the analysis ofcrystal structures and phases.

• High resolution backscattering spectrometer for theinvestigation of hydrogen motion and for the flow dynamicsof lubricants.

• Variable resolution cold chopper spectrometer for theanalysis of hydrogen diffusion, molecular rotation andconfirmation, and for the internal dynamics of lubricants.

• Diffuse scattering diffractometer with full polarisatonanalysis for the investigation of disordered and hardmagnetic materials.

Instrument suite formaterials science andengineering

For most of the instruments specified above the optimal targetoptions are• 50 Hz short pulse from decoupled and poisoned moderator

for high resolution experiments, in particular for highresolution powder diffraction, for the engineeringdiffractometer and the radiography beamline.

• 50 Hz short pulse from a cold coupled moderator forneutron work requiring long wavelength neutrons such ashigh resolution backscattering spectroscopy, quasi-elastic

Target options

54

neutron scattering, and neutron reflectometry.• 16.7 Hz long pulse target with a cold coupled moderator

for high intensity work and for high Q resolution but withrelaxed wavelength resolution, in particular forexperiments with high intensity reflectometry, focussingsmall angle scattering, and diffuse scattering.

V. SummaryMaterials Science and Engineering encompasses many sub-disciplines, and also overlaps, in part, with solid state physics,with chemical structures and kinetics, with soft matter, andwith liquids and glasses, described in other parts of thisdocument. Unique to Materials Science and Engineering is,however, the bridging function from the fundamentalunderstanding of materials properties to device applicationsand development of engineering components. Neutronscattering has played a recognized key role in this field duringthe past 50 years. With the increased flux and intrinsic timestructure the ESS is expected to have a huge impact onmaterials science and engineering developments in the futurebecause of:• the higher flux that enables real time experiments,

providing movies of structural and dynamicaldevelopments of materials;

• the higher resolution in respect to time, real space andreciprocal space that supplies a much improved data basefor finite element models of engineering parts;

• the real operating conditions of pressure, temperature, andliquid or gas environments that materials can be tested in;

• the real scale components that are accessible to neuronbeam studies.

Thus neutron scattering at the ESS in combination withcomputer simulations will make a very large impact on studiesof materials and engineering components in real time, on realscale, and under real conditions.All future trends in materials science and Engineering requirenot only a strong increase in neutron flux but also advancedand innovative instrumentation at a short pulse 50 Hz and along pulse 16.7 Hz target station

References

[1] B.N.J. Persson: Sliding Friction, Springer Verlag, Heidelberg Berlin (1998)

[2] S.M. Clarke, A.R. Rennie, and P. Convert, Europhys. Lett. 35, 233 (1996)

[3] A. Magerl, H. Zabel, B. Frick, and P. Lindner, Appl. Phys. Lett. 74, 3474 (1999)

[4] JWL Pang, G Rauchs, PJ Withers, NW Bonner and ES Twigg, “Engineering Aeroengines for the 21st

Century”, ISIS Annual Report 2000, p66-67

[5] K. Hirschi, M. Ceretti, B. Marini, J.-M. Sprauel, LLB Scientific Report 1997-98, p. 69

[6] J. Rodriguez-Carvajal, G. Rousse, Masquelier C, Hervieu M, Phys. Rev. Lett. 81, 4660 (1998)

[7] H.G. Schimmel, W.Montfrooij, G.J. Kearley, V.W.J. Verhoeven, I.M. de Schepper, Phys. Rev. B 63,214409 (2001)

[8] G.A. Prinz, Device physics – Magnetoelectronics Science 282, 1660 (1998)

[9] A. Schreyer, T. Schmitte, R. Siebrecht, P. Bödeker, H. Zabel, S.H. Lee, R.W. Erwin, J. Kwo, M. Hong, andC.F. Majkrzak, J. Appl. Phys. 87, 5443 (2000)

55

[10] E.H. Lehmann, P. Vontobel, "Neutron Imaging for Industry Related Inspections", PSI Annual Report 1999;Annex VI, p. 50-54

[11] J.N. Huiberts, R. Griessen, J. H. Rector, R.J. Wijngaarden, J.P. Dekker, D.G. de Groot, and N.J. Koeman,Nature (London) 380, 231 (1996)

[12] T. J. Udovic, Q. Huang, and J. J. Rush, J. Phys. Chem. Solids 57, 423 (1996)

[13] A. Remhof G. Song, Ch. Sutter, A. Schreyer, R. Siebrecht, H. Zabel, F. Güthoff, and J. Windgasse, Phys.Rev. B 59, 6689 (1999)

56

7.3 Biology and Biotechnology

T. Bayerl1, O. Byron2, J.R. Helliwell*3, D. Svergun4, J.-C. Thierry5, J. Zaccai6

1Lehrstuhl für Experimentelle Physik V, Universität Würzburg, Am Hubland, D-97074 Würzburg, Germany2Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow, G12 8QQ, Scotland, UK3Department of Chemistry, University of Manchester, M13 9PL, UK4EMBL, c/o DESY, Notkestrasse, D-22603 Hamburg, Germany5IGBMC-CNRS, Illkirch, Strasbourg, France6Institute Laue Langevin, 6, rue Jules Horowitz, F-30842 Grenoble 9, France*john.helliwell@man.ac.uk

AbstractNeutrons have a unique role to play in determining the structure and dynamics of biological macromolecules andtheir complexes. The similar scattering signal from deuterium, carbon, nitrogen and oxygen allows the fulldetermination of the positions and dynamics of the atoms “of life”. In addition the negative scattering length ofhydrogen allows the well-known contrast variation method to be applied to dissect the component parts ofmultimacromolecular complexes. In the post genomic era structure determining techniques are reaching towardshigh throughput and a high number of proteins to be investigated. Thus, the ESS will offer major gains in neutroncapability over the current technical frontier with reactor source technology whereby smaller samples, smallerquantities and lower concentrations all become viable. Thus, the major structure and dynamics techniques ofprotein crystallography, small angle neutron scattering, inelastic scattering and membrane reflectometry will allbenefit in a major way. The considerably reduced measuring times will allow native rather than artificialmembranes to be probed by reflectometry, including membrane bound proteins at the surface of actual cells.From the membrane biophysical studies via such native state reflectometry new nano-composites can beenvisaged and designed. Structural biology, as well as biotechnology, will benefit from the powerful ability ofneutrons to contribute to the location of hydrogen atoms and water molecules in biological systems. Thus it willcontribute to the production of missing complementary data relevant for molecular modelling and to the strategy ofrational drug design, in synergy with other biophysical approaches.

I. IntroductionThe present major driving forces of life science at themolecular and cellular scale are functional genomics andproteomics. Information on the specific functions of most if notall proteins encoded in human and other genomes isdesirable. Major obstacles are the vast complexity of theindividual proteins and the even more delicate interaction ofdifferent proteins and other biomolecules to form (transient)functional complexes. Neutrons are a unique non-destructivetool for probing precious biological molecular samples [1].

We are in the post genomicera (Figure 1)

Biology strives tounderstand molecularrecognition

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Evolution of GenBank1990-2000

2.10 9

10 9

3.10 9

4.10 9

5.10 9

6.10 9

7.10 9

8.10 9

9.10 9

1 000 000

2 000 000

3 000 000

4 000 000

5 000 000

6 000 000

7 000 000

8 000 000

Num

berof

bases

Num

berof

sequences

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Evolution of GenBank1990-2000

2.10 9

10 9

3.10 9

4.10 9

5.10 9

6.10 9

7.10 9

8.10 9

9.10 9

1 000 000

2 000 000

3 000 000

4 000 000

5 000 000

6 000 000

7 000 000

8 000 000

Num

berof

bases

Num

berof

sequences

Figure 1: We are in the post genomic age of vast quantities of genesequence information being available.

57

Hydrogen and water are involved in all the molecularprocesses of life. Very seldom is information on these aspectstaken into account and in any case this experimentalinformation is mostly incomplete. If only neutron capabilities inprotein crystallography could reach the throughput of X-raywork then the impact of neutrons would radically alter thissituation. In fact the research and development has largelybeen done with some very fine scientific examples e.g. on thedetailed dissection of enzyme mechanism involving hydrogenlocation. Since many enzyme reactions involve hydrogenthere is great potential for wide application IF the technicalcapability can be found. Moreover the role of water inmolecular recognition is pivotal as, for example, the lubricantof protein ligand interactions or the bonding mediator.Structural definition of bound water by X-ray diffraction is verysensitive to mobility especially at room temperature. Neutronshave the advantage that the scattering of deuterium is equalto that of the water oxygen. The incorporation then of the fullstructural detail of bound water can radically alter themodelling of proteins in silico for the improved discovery ofnew compounds, for example as leads in drug design. Thereare two major hurdles for wide application of neutron proteincrystallography; firstly the size of crystals routinely availableand secondly a molecular weight ceiling of about 30 kDa; themolecular weight histogram in the yeast genome for examplepeaks at 30 kDa and so at least half of all proteins in thegenome are out of range of current neutron proteincrystallography capabilities even if big crystals can be grown.This situation should change radically with ESS.

In the modern trend towards high throughput, as well as highresolution structure determination, many proteins will notreadily crystallise. The field of solution scattering can offer thechance to help determine the fold of a protein in solution.Efforts have started in this direction and show promise but it isstrongly felt that the use of deuterated specific amino acidscan, with small angle neutron scattering (SANS), provide themuch needed breakthrough in this field. Major gains inneutron flux to reduce especially the quantity of proteinrequired will be of interest for higher throughput folddiscovery. SANS at ESS could do this.

Neutrons provide unique possibilities in solution scatteringstudies of biological macromolecules, primarily thanks tocontrast variation by H/D exchange. For multi-componentsystems (e.g. the ribosome (Figure 2), nucleoproteins orlipoproteins) information about the distribution of thecomponents has been obtained by variation of the D2Ocontent in solution. Moreover, specific deuteration permits tohighlight and analyse selected parts of macromolecularstructures in situ.

The results provided by neutron scattering are highlycomplementary to other analysis techniques (small angle X-ray scattering (SAXS) and diffraction, electron microscopy(EM) and analytical ultra-centrifugation (AUC)). Moreover,structural models built on the basis of neutron scattering data

Neutrons are well suited toreveal water and hydrogen

Molecular recognitioninvolves water

The small crystal challenge

A molecular weight ceiling

Not all proteins crystallise

Helping protein folddiscovery

Dissecting complexes viacontrast variation

Neutrons are crucial tobiophysics

58

are able to incorporate information from different high and lowresolution techniques and possibly also reconcile concurrentmodels.

Figure 2: Comparison between the 50S subunit in the map of the 70SE.coli ribosome obtained from solution scattering [2,3] (left,resolution 3 nm), and the crystallographic model of the 50Sribosomal subunit H.marismortui (right, resolution 0.24 nm,Figure adapted from [4]. Yellow, ribosomal proteins, grey,ribosomal RNA.

Despite its acknowledged importance for biological functionand activity, the dynamics dimension in molecular structuralbiology remains difficult to characterise and poorlyunderstood. Neutron scattering is perfectly and uniquelysuited to the space-time window of biologicalmacromolecules. Protein function and activity, includingenzyme catalysis, ligand binding, receptor action, electronand proton transfer, are strongly dependent on internaldynamics. It is now known that structural information alone isnot sufficient to explain specific drug binding effects and thatthe dynamics dimension should be taken into account. Werecall that chemical dynamics provides a much more stringenttest of a model than structure. Practically no information isavailable at present on the dynamics of complexes ororganised systems. In an original approach bridgingmolecular and cell biology, neutron experiments haveprovided data on the dynamics of proteins, in vivo, within theircellular environment. Neutron scattering in fact providesunique opportunities to probe the natural cellular environment,which because of its molecular crowding properties, is verydifferent from the usual conditions of laboratory biochemistry.The field of neutron applications to study biological moleculardynamics is wide open, with hydrogen-deuterium labellingallowing to focus on the dynamics of amino acid groups withina protein, or protein domains within a complex. Samplerequirements, however, are at present unrealistic for thetechnique to accomplish a definitive impact. The ESS shouldprovide a factor of at least 100 gain due to the source andoptimised instrumentation.

Membrane biophysics studies on the molecular scale arecrucial for understanding the self-organisation processeswhich underlie many functional aspects of the membrane, inparticular membrane transport, molecular recognition on

Macromolecular dynamicsare poorly understood

Neutrons can revealmolecular vibrations

The dynamics ofmolecular complexesremains understudied

The cell is a crowdedenvironment

ESS is a route to the realcell

59

surfaces and adhesion between cells and substrates. Whilethese problems have been studied in the past, largely inmodel systems, there is now a strong tendency towardsstudying far more complex native membranes. Newpreparation techniques allow native membranes (e.g. plasmamembranes from eukaryotic cells) to be deposited onto solidsubstrates while maintaining their functional integrity. Withfaster data acquisition ESS should allow, via neutronreflectometry (NR), the study of membrane behaviour in situincluding membrane transport and membrane damage byinvasive toxins.

Functional membranes willbe studied

II. Forecast of novel opportunitiesIn high throughput structural biology research the best samplesize is rarely above 100 x 100 x 100 µm. It is essential forneutron protein crystallography to find source, instrument andsample (deuteration) combinations to face this challenge.Since we know from genomics that the average molecularweight of, for example, yeast gene products is 30 kDa, thereare many projects that would become amenable for study byneutron protein crystallography methods if the sample sizerequirement could be relaxed.

There is a barrier to the application of high resolution neutronstructural study posed by molecular weight, which determinesthe unit cell volume, of large biological complexes. Suchweakly scattering crystals cannot be studied currently (Figure3). If, however, we combine the ESS source and instrumentimprovements, and improved knowledge of the proteinpreparation and crystallogenesis for the growth of largecrystals, the unit cell size capability could reach 250Å. Thus,for example, the small subunit of the ribosome could beamenable to study by neutron protein crystallographymethods.

Recently, a new method has been developed to analyse wideangle X-ray solution scattering data up to 5Å resolution interms of the approximate positions of dummy amino acidresidues. With neutron scattering, this approach can beextended to gain additional information about the positions ofindividual residues. This can be achieved either by thepreparation of proteins with specifically deuterated residuesor, for native samples, by making use of the change incontrast of residues during H/D exchange (e.g. after placing ahydrogenated protein in D2O). This requires a neutron sourcewith high flux and high dynamic range. This should contributeto the determination of the protein fold in solution fromexperiments on native samples and it would be an approachrelevant for high throughput fold definition for proteins whichdo not easily crystallise e.g. detergent solubilised membraneproteins.

Small biocrystals areusually the reality

ESS should allow the studyof smaller samples

Large complex neutroncrystallography at highresolution is a new frontier

Extending the resolution ofSANS

60

Figure 3: The determination of the hydrogen atoms for this tetramericcomplex of concanavalin A with glucose is out of range of bothSR Xray and neutron techniques. The crystals are not stable incryoprotectant and room temperature SR undulator diffraction(ESRF Quadrigia) is only at 1.9Å resolution. On the LADI ILLinstrument with a D2O soaked crystal and a crystal size of4x3x2mm the neutron diffraction resolution is 3.5Å, insufficientfor atomic analysis. An optimised ESS ‘Large Molecular Weight’nPX instrument could bring the 50 kDa in the asymmetric unitcrystal in range to find the sugar recognition hydrogen atoms (asdeuteriums) [5,6].

SANS is also used to determine the kinetics, stoichiometryand organisation of large macromolecular complexes(viruses, molecular machines such as chaperones, etc).According to its flux specification, ESS will reduce thequantities needed for these studies by at least ten. SANS willbe used to study complexes identified in ever greaternumbers by proteomic analysis. The products of structuralgenomics (i.e. atomic resolution structures) will be combinedwith the SANS results to deconvolute the complexes into theirfunctional components (Figure 4).

The long measuring times at current source flux levels restrictreflectometry and scattering studies to non-labile systems.Thus native membranes are not generally amenable to study.Major gains in peak and average fluxes open up suchtechniques to native, biologically relevant systems. This willwiden the applicability of the knowledge gained tobiotechnologies such as biosensors and nano biostructures.

Taken together the above methods help create a cell model insilico; a virtual cell. The seamlessness and speed with whichmodern biologists can interrogate vast databases of genomic,proteomic and structural data was inconceivable to most inthe field 15 years ago. How will we continue to improve thecomputer-user interface so that life scientists can maximisetheir use of these expanding reserves of knowledge? How

SANS can help to definethe make-up of molecularmachines

Neutron reflectometry is akey to membrane-basedbiosensor development

The virtual cell

61

can we ensure that all experimental data are being used, notjust published? The integration of all existing databases into avirtual cell framework may accomplish this. The virtual cell willbe a graphical, interactive, "functional" representation of agiven cell (be it bacterial or even perhaps mammalian). Thebiologist will design investigations based on the response ofthe cell to imposed stimuli; the response will be the result ofcalculations made by the cell on the basis of literature dataaccessed from existing databanks. Neutrons should besignificant contributors to the knowledge data base not only atthe atomic level but more generally at the cellular and thusfunctional level.

The role of neutrons in thevirtual cell project

Figure 4: Schematic diagram of a typical eukaryotic cell signallingpathway. When the B cell receptor TLR2 binds its ligand itrecruits a cascade of proteins in the cell cytoplasm: Myd88,IRAKs 1 and 2 and Traf6. This leads to the activation ofMAPKKK and the phosphorylation and activation of theMAPkinase cascade and the IkBK signalsome. The resultingphosphorylation of IkB induces its ubiquination and degradation.Degradation of IkB results in the release of active p50/p65components of NF-kB. NF-kB then translocates to the nucleusand transactivates immunomodulatory genes. Similarly,MAPkinase (MAPK) can translocate to the nucleus and activatetranscription factors such as Elk, AP-1 and ATF-2. The resolutionof structures for the constituent molecules (and their complexes)varies from zero to atomic. Neutron scattering andcrystallography can help to fill in the gaps in this knowledge.

III. Assessment of science caseHow will the opportunities with neutrons at ESS map ontointernationally agreed tasks and trends in the biosciences aswell as open up new fields? In biology the post-genomic era,created by the speed and efficiency of gene sequencing,provides a huge stimulus to and the radical need fordevelopment of the structure determining techniques.

62

It is a paradigm that structure is based on function. Howeverthis paradigm has been recast largely by neutron inelasticscattering techniques to include dynamics. Thus structure anddynamics determines function. The major gains in capabilityof ESS for inelastic scattering should greatly widen theapplicability to many more systems.

Concerted action projects exist in structural genomics at worldlevel. Notable cases include projects (i) in USA e.g. theStructural Genomic Program supported by the NationalInstitute of Health (NIH) for high throughput structuredetermination of proteins from complete genomes(pathogens, arcahebacteria,…) and from human, mouse andother higher organisms (ii) in Europe on human proteins andalso on pathogens and yeast and (iii) in Japan on variousgenomes including human and mouse proteins supported inparticular by the Riken, the MITI and the Ministry ofEducation. Greatly expanded provision of synchrotronradiation beamlines for structural biology in the USA, Europeand Japan have been made for these projects and a generalexpansion of the field. Japan has a large new NMR park forstructural genomics in Yokohama. The unique role ofneutrons as a non-destructive probe of the structure anddynamics of biological macromolecules have been assessedby the USA and Japan who are now building state-of-the-artspallation neutron sources at the megawatt power level. TheESS reaches beyond even those power levels thus making atechnical and scientific capability that is compelling.

Finally the structure and dynamics results obtained will beapplicable for industrial and biotechnology exploitation (seesection 7).

Biological membranes are worthy of a special mention. Theextreme sensitivity of neutron reflection makes it uniquelysuitable for the study of labile biological structures. Theinternal reflection at the solid/liquid interface combined withcontrast variation allows the exact determination of themembrane structure and of the crucial polymer layerseparating the biological membrane and the solid support.This will provide deep insight into the role of the soft polymercushion for maintaining membrane integrity and function,crucial knowledge for the design of advanced biosensors.Moreover, using 2-D detection and in-plane Bragg scattering,there is the chance to study at molecular resolution theassociation and self-assembly of functional clusters in theplane of the membrane. This knowledge is crucial for theunderstanding of how proteins and lipids temporarilyassociate in a functional membrane. As a flagship experimentin this field we propose the study of native membranes andwhole cultured cell layers on polymer cushioned solidsubstrates (Figure 5). This will permit the measurement of thecellular membrane response on the action of external stimuli(drugs, stress, pressure...) at molecular resolution. It will alsoshed light on the extremely poorly understood interactionbetween different membrane constituents under conditions of

New fields

Global concerted actionsmap well onto ESS frontiercapabilities

The USA and Japan haveapproved the constructionof state-of-the-artmegawatt level spallationneutron sources

Advanced biosensors

63

membrane transport, ligand receptor binding and celladhesion.

Figure 5: Neutron reflectometry. ESS fluxes will allow the study of nativeplasma membranes extracted from living cells deposited toplanar substrates. A specially designed polymer cushion [9]between solid substrate and membrane will provide the softinterface required for keeping the membrane spanning proteins(ion channels, receptors, transporters) in their active state duringthe experiments.

IV. Relation to ESS instrumentationIn neutron protein crystallography we forecast two mainfrontiers. The first is to considerably reduce the crystal size,including to reduce the data collection time from weeks todays, and the second frontier is to considerably extend themolecular weight capability to study complexes at highresolution (Figure 3, refs 5,6). In the first case a brighterneutron source, well-focussed beams and smaller-pixeldetectors will be in the design. To meet the second challengewe have to continue to harness the expertise of thecrystallogenesis community to produce big crystals. Thuslarger beams and bigger-pixel detectors are needed. This is adifferent instrument altogether. Also we should harness longerwavelengths to enhance the scattering-efficiency-with-wavelength-effect as well. A methane moderator tailored towavelengths 1.5 to 5 A should be investigated. A thirdinstrument should be developed at ESS for contrast variationat low resolution to dissect the intermolecular interactions inmembrane crystals so as to better understand suchcrystallogenesis cases. There is a need here because thebest current source, ILL, still has long measuring times(months). ESS will bring a big benefit but perhaps demandwill be low. Perhaps this third instrument could be shared withthe higher resolution SANS beamline (see the next section).

In SANS we forecast an increase in the number ofexperiments that will be undertaken, not least because ofgenomics and initiatives in high throughput samplesproduction, but also where experiments in solution will be at apremium (over techniques like crystallography wherecrystallisation is a recognised bottleneck to high throughput).The range of SANS experiments, tailored to molecular weightranges that will be encountered, will be wide. Also theharnessing of as high an intensity as possible, to reduce thequantity of sample and the concentration needed, will be very

Three neutron proteincrystallographyinstruments are proposedfrom the outset

High throughput SANSneeds high flux

64

important. Thus the long pulse target station in particularoffers extra opportunities for realising such intensity gains forSANS.

In dynamics experiments the instrument and target optionsneed to match the time-space windows of interest. These arefrom the picosecond to the nanosecond or longer timescales,and from 1 to 10Å length scales. Gains with respect topresent ILL instruments will be >100 at ESS. The 50 Hz targetstation is appropriate for these instruments. Instrumentalpriorities are for the variable resolution cold neutron chopper,multichopper spectrometers, the 1.5 µeV backscatteringinstrument and an 8 µeV back scattering instrument with largeQ range (up to 5Å-1, similar to IN13). As a benchmark theESS 8 µeV instrument will be >500 x IN13 performance atILL. These instruments would cover the elastic, quasi-elasticand inelastic resolutions implicated in the dynamics ofbiological systems of various sizes, from diffusing watermolecules to domain motions in large complexes. Neutronspin echo (NSE) may in principle present interestingapplications for studying slow coherent diffusive motions, butan optimistic estimate is that a gain of a factor of at least 100on present ESS instrument conceptual designs would berequired. The potential to use NSE to study the dynamics ofsmall drug molecules or hormones in the context of theirbiological interactions should nevertheless be studied andassessed.

The membrane and cell surface projects identified in sectionsI-III are largely not feasible with today’s reflectometersbecause of the flux which requires accumulation times thatare long for labile biological structures. The planned ESSreflectometer with its more than an order of magnitude higherflux will for the first time enable such studies. The desirableQ-range for these types of studies is up to 0.5Å-1 whileresolution is not essential. Crucial will be the availability of 2-D detection and the option to measure Bragg scattering in themembrane plane.

Sample production is as important as the source and theinstrument for biology, especially in order to properly exploitthe unique ability of neutrons to distinguish between majorbiological components. We must be able to produce and tolabel the constituents of our systems, be this individual atomswithin a protein, proteins within a complex or complexeswithin a cell. Current, commonly-used, overexpressionsystems must be optimised for growth in deuterated media orin minimal media supplemented with deuterated amino acids(or sugars, co-factors etc). This must include not just thework-horse of protein production, E. coli, but also yeast,insect and mammalian cell systems. The future of molecularlife sciences demands that we are able to successfullyoverexpress, purify and characterise fully post-translationallymodified proteins. Further, if we are to achieve broadcoverage of the cellular world in the virtual cell program,identified in section II, we must be able to characterise thesecells in their entirety.

Molecular dynamics servedby the 50Hz target station

Matching the energies oflife

Observing function in themembrane plane

Samples production is veryimportant

Labelling biomolecules

65

V. Uniqueness of neutrons and complementary aspectsCrystallography, SANS and inelastic scattering of biologicalmolecules constitute major components of the future needs atESS for 3D structure analysis. In addition reflectometry onnative systems will allow 2D systems to be studied. Hence inall these methods the role of neutrons as a non-destructiveand non-absorbing probe, with contrast variation involvingH/D exchange, place neutron scattering as a uniqueapproach. X-ray scattering and, in a different range ofaccuracy as well as time scale, NMR, provide, of course,important results in structure-function biology. Howeverneutron studies more efficiently locate functional importantwater molecules and hydrogen atoms, especially attemperatures which are near physiological, rather than atcryo-temperature by SR X-ray protein crystallography [5].The structural precision of neutron room temperature data isroutinely better than NMR atomic position data.

Multi-dimensional studiesSuccessful characterisation of biological systems depends ondata from several orthogonal, complimentary studies.Increasingly, in order to ensure the reproducibility ofconditions under which a system is studied andcorrespondence of the data so obtained, these studies arebeing carried out simultaneously. This is also efficient in termsof experimental time - a key factor in high-throughput post-genomic studies. Thus, in order to make a significant andrapid contribution to the databanks of the virtual cell (or eventhose in use currently) SANS can be coupled to capillaryelectrophoresis. Proteins within a cell extract will beseparated electrophoretically in quartz capillaries as they flowslowly past an orthogonal, non-destructive neutron beam.Their meso-resolution structures will be restored automatically(using programs which will have evolved from DAMMIN,GASBOR etc [2,3]) and related directly to the electrophoreticband (and the identity of the protein obtained from MALDI-TOF analysis of digestion fragments thereof).

The combination of neutron reflectometry with eitherfluorescence microscopy (single molecule), plasmonspectroscopy or infrared spectroscopy can provide muchadditional information, in particular for the assessment ofbiomolecule interaction from the bulk with the membrane.Neutron reflectometry is a natural partner for surface plasmonresonance (SPR) studies, although currently the timescalesfor these experiments are rather different. Nonetheless, SPRgives crucial information on the on- and off-rates of molecularinteractions (even within deposited model membranes). Thecoupling of NR and SPR would permit the through-bilayer andin-plane structural characterisation of the equilibrium system(once binding saturation or de-saturation is reached). Thedesign of a new reflectometer on a high intensity ESS shouldbe integrated with other biologically relevant andcomplementary techniques in a single setup.

Neutrons providephysiologically relevant,and yet precise, structuraldata

Combining neutron as anon-destructive probe ofmatter with othertechniques

Unique role of neutrons

66

Complementary aspectsThere are of course other structure definition techniques suchas X-ray crystallography and scattering, NMR, electronmicroscopy, mass spectroscopy, atomic force microscopy andlight scattering. Data obtained with these techniques presentkey data to describe biological molecules and their complexesat different levels of detail. Neutrons provide a unique role indiffraction and inelastic scattering as a probe of biologicalstructures because of the near equal scattering lengths ofdeuterium, carbon, oxygen and nitrogen and for dynamicmeasurements because the momenta of neutrons arematched to atomic vibrations. The additional possibility forharnessing the negative scattering length of hydrogen ofcourse makes the unique contrast variation approachfeasible. Thus the results provided by small angle neutronscattering are highly complementary to other analysistechniques (X-ray scattering and diffraction, electronmicroscopy, analytical ultracentrifugation).

Adding movement to staticstructures

H +

cytoplasm

extra-cellularmediumk'' = 0.12 N/m

k' = 0.55 N/m

k = 1.7 N/m

k' = 0.33 N/m

k = 1.4 N/m

labelled sample

unlabelled sample

C terminal loop

0

0.1

0.2

0.3

0.4

0.5

0 40 80 120 160 200 240 280 320

T (K)

<X 2> (Å 2)

T (K)

<X 2> (Å 2)

0

0.1

0.2

0.3

0.4

0.5

0 40 80 120 160 200 240 280 320

Figure 6: Dynamics-function relation in bacteriorhodopsin, the light drivenproton pump in purple membranes, characterised by neutronscattering and specific isotope labelling of amino acids in theactive core of the protein. An effective force constant, <k'>, foreach part of the structure was calculated from the mean squarefluctuations measured as a function of temperature. The protein,globally, is quite soft at physiological temperature with a <k'>value of 0.12 N/m. The core, however, is significantly stiffer witha <k'> value of 0.33 N/m, suggesting it ensures a valve functionin the pump mechanism [10]. For a review see [11].

The structural models built on the basis of neutron scatteringdata and all other presently available methods allow toincorporate information from different high and low resolution

67

techniques and possibly also reconcile divergent models. Asa demonstrative example we can cite the determination byinelastic neutron scattering of the role of vibrational states inthe function of the bacteriorhodopsin membrane proton pump(Figure 6), for which a high resolution structure is known.

VI. Target stationsThe 5 MW ESS targets are world beating. The 50 Hz shortpulse target station is considered to be the most favourablechoice for protein crystallography, dynamics, neutronreflection and in-plane Bragg scattering.

A SANS station should provide high flux and high dynamicrange which would permit the use of smaller amounts andconcentrations of material. The values of Qmin down to 5x10-4 and Qmax up to 1.0 are needed for detailing largecomplexes in solution and for protein fold definition,respectively. These requirements would be met using the longpulse target station, desirably with an option of focussinggeometry, which would provide the necessary intensity andrange. The long pulse target station is thus a primary choicefor SANS.

As to the 50 Hz target station focussing-SANS for very highresolution, the big sample requirement of 25x25x1mmdimensions will hamper its wide uptake. Further, highwavelength resolution is not the most important concern ofbiological SANS.

The 10Hz station might have been interesting for longerwavelength utilisation (e.g. large molecular weight proteincrystallography, low resolution protein crystallography andSANS). It appears however that the energy per pulserestriction of 1 MW cannot be overcome and thus this loss isworse than versus 5 MW on the 50Hz target station with lessflux per MW at longer wavelengths; the development of amethane moderator for the 50 Hz station would then help withthat optimisation.

50 Hz target is a popularchoice for biology

A role for long pulses

VII. Industry: implications in biotechnology and industryThe major advances in the fields of cell culture, softening andbiocompatibilization of solid surfaces, protein reconstitutiontechniques and nanostructuring methods can now providenative membrane samples of different protein and lipidcomposition in geometries that are uniquely suited for neutronstudies. Membrane based biochips will become the key foradvanced biosensors, as well as screening and bioseparationdevices. In particular, the combination of nanostructuredsemiconductor surfaces with native membranes via a softpolymer cushion represents a crucial technology for manyappications in diagnostics and also in proteomics. Researchand development in these fields amounts presently to $20billion world-wide with an annual growth of 25 %.The enablingtechnologies that will emerge in the next years from thisresearch will provide the tools for finding molecular markers

Membrane science forbiosensors and health

68

for the early stage detection of illness and for the unravellingof the human proteome. Neutron data, if neutrondevelopments cope with the need of the methods, on suchcomplex systems can become a prerequisite in the design ofeven more advanced combinations of biological matter withsolid surfaces for biochips including biosensors.

The role of structural data in drug discovery in thepharmaceutical industry will increase when it is much moreroutine that hydrogen atoms, bound water deuteriums and thedynamics information can be incorporated. Thus thediscovery of new pharmaceuticals and of enhanced efficiencycompounds will accelerate. Also, because neutrons are non-destructive, unlike X-rays, room temperature structure anddynamics data can be provided.

Rational drug design

Neutrons provide structureand dynamics data atphysiologically relevanttemperature

References

[1] O. Byron and R.J.C. Gilbert, Current Opinion in Biotechnology 11, 72-80 (2000)

[2] D.I. Svergun, K.H. Nierhaus, J. Biol. Chem. 275, 14432-14439 (2000)

[3] D.I. Svergun, M.V. Petoukhov, M.H.J. Koch, Determination of domain structure of proteins from X-raysolution scattering. Biophys. J. (in press 2001)

[4] N. Ban, P. Nissen, J. Hansen, P.B. Moore, and T.A. Steitz, Science 289, 905-920 (2000)

[5] J. Habash, J. Raftery, R.Nuttall, H.J. Price, C. Wilkinson, A.J. Kalb (Giloboa), J.R. Helliwell, Acta. Cryst.D56, 541-550 (2000)

[6] A.J. Gilboa (Kalb), D.A.A. Myles, J. Habash, J. Raftery, J.R. Helliwell, Neutron Laue diffractionexperiments on a large unit cell: concanavalin A complexed with methyl alpha D glucopyranoside J. Appl.Cryst. (in press 2001)

[7] C. Naumann, C. Dietrich, A. Behrisch, T. Bayerl, M. Schleicher, Bucknall, E. Sackmann, BiophysicalJournal 71 (2), 811-823 (1996)

[8] G. Fragneto, T.J. Su, J.R. Lu, R.K. Thomas, A.R. Rennie, Physical Chemistry Chemical Physics 2 (22),5214-5221 (2000)

[9] J. Schmitt, B. Danner, T.M. Bayerl, Langmuir (1), 244-246 JAN 9 (2001)

[10] G. Zaccai, Science 288,1604-1607 (2000)

[11] J.C. Smith, Quarterly Review of Biophysics 24, 227-291 (1991)

69

7.4 Soft Condensed Matter

A. Arbe1, F. Boué2, J. Colmenero1,3, S. Janssen4, K. Mortensen5, D. Richter6, J. Rieger7,P. Schurtenberger8, R. K. Thomas9

1Unidad de Física de Materiales (CSIC- UPV / EHU), Apartado 1072, 20080 San Sebastián, Spain2Laboratoire Léon Brillouin, CEA Saclay, F-91191 Gif-sur-Yvette Cedex, France3Departamento de Física de Materiales, Universidad del País Vasco, and Donostia International Physics Center, Apartado1072, 20080 San Sebastián, Spain4Laboratory for Neutron Scattering, ETHZ & PSI, CH-5232 Villigen PSI, Switzerland5Danish Polymer Centre, Risø National Laboratory, DK-4000 Roskilde, Denmark6Institut für Festkörperforschung, Forschungszentrum Jülich, D-52425 Jülich, Germany7BASF Aktiengesellschaft, Polymerphysik, Festkörperphysik, ZKM-G201, 67056 Ludwigshafen, Germany8Physics Department, Soft Condensed Matter Group, University of Fribourg, Perolles, CH-1700 Fribourg, Switzerland9Department of Chemistry, University of Oxford, South Parks Road, Oxford, OX1 3QZ, United Kingdom

AbstractNeutron scattering techniques play a unique role in the study of both the structural and dynamical properties ofthe wide range of substances categorised as “soft matter”. Among the advantages presented by thesetechniques, two are of crucial relevance in the soft matter field: the suitability of the length and time scalesaccessed by neutrons, and the capability to manipulate the contrast by specific deuteration of any constituent ofthe system. Neutron scattering is the only tool for unravelling the molecular morphology and motions in soft mattersystems at the different relevant length scales. On the other hand, the understanding of structural properties anddynamics at a molecular level is the key for advancing in this field: the envisaged trends move towards the studyof increasingly complex, often multi-component materials tailor made for industrial applications. The combinationof neutron scattering techniques, advanced chemistry and molecular modelling will be essential. Experiments onvery dilute components, or on very small amount of matter (e.g. particular topological points, at the interfaces...)are demanded. Moreover, in-situ studies will investigate time dependent and transient phenomena, non-equilibrium situations and so on. Such experiments will become possible by the orders of magnitude increase insensitivity offered by the next generation neutron sources. In particular, the availability of a 16.7 Hz long pulsetarget station would allow to optimise most of the instruments devoted to soft-matter studies, such as small angleneutron scattering instruments, reflectometers and high resolution neutron spin echo spectrometers.

I. IntroductionThe concept of “soft matter” subsumes a large class ofmolecular materials, including e.g. polymers, thermotropicliquid crystals, micellar solutions, microemulsions and colloidalsuspensions, and also includes biological materials, e.g.membranes and vesicles. These substances have a widerange of applications such as structural and packagingmaterials, foams and adhesives, detergents and cosmetics,paints, food additives, lubricants and fuel additives, rubber intyres etc., and our daily life would be unimaginable withoutthem. In spite of the various forms of these materials, many oftheir very different properties have common physicochemicalorigins such as a large number of internal degrees of freedom,weak interactions between the structural elements and adelicate balance between entropic and enthalpic contributionsto the free energy. These properties lead to large thermalfluctuations, a wide variety of forms, sensitivity of theequilibrium structures to external boundary conditions,macroscopic softness and various metastable states.

Soft Matter includes alarge variety of daily usedmaterials with a widerange of properties basedon common physico-chemical origins

The structural units of soft matter systems are large moleculesor aggregates of molecules showing different structural anddynamical properties depending on the length scale ofobservation. This implies a need to cover large ranges in theexperimental space / time windows for a completeunderstanding of their characteristic features. Furthermore,aggregation in these systems may lead to a large internal

Spatial scales range fromnanometers to micro-meters; characteristictimes from picoseconds tothe macroscopic times

70

interfacial region, which can then make a dominantcontribution to the overall properties.

0 0.5 1 1.5 2 2.5 3

S(Q)

Q (Å-1)

0.01 0.1 1Q (Å-1)

Polymers

Large Scale DynamicsDiffusionReptation

Rouse Dynamics...

Intermolecular Rangeαααα -Relaxation

GLASSY BEHAVIOR

CC

O C

H

H

H

HH

H

Molecular Length Scaleββββ -Relaxation

VibrationsSide Groups ...

DynamicsStatic

Lengthscale T

imescale

1 Å

1000 Å

1 ps

1 ms

COMPLEXITY

ChainForm Factor

0 0.5 1 1.5 2 2.5 3

S(Q)

Q (Å-1)

0.01 0.1 1Q (Å-1)

Polymers

Large Scale DynamicsDiffusionReptation

Rouse Dynamics...

Intermolecular Rangeαααα -Relaxation

GLASSY BEHAVIOR

CC

O C

H

H

H

HH

H

CC

O C

H

H

H

HH

H

Molecular Length Scaleββββ -Relaxation

VibrationsSide Groups ...

DynamicsStatic

Lengthscale T

imescale

1 Å

1000 Å

1 ps

1 ms

COMPLEXITY

ChainForm Factor

Figure 1: The different static and dynamic scales relevant for soft mattersystems are schematically shown in the simplest case of apolymer.

II. The Role of NeutronsAmong the experimental techniques used for the investigationof the structure and dynamics of soft matter, neutronscattering (NS) plays a unique role for several reasons:

i) The suitability of the length and time scales accessed,especially by Small Angle Neutron Scattering (SANS)and Neutron Spin Echo (NSE), allows the exploration oflarge scale properties (for instance, the conformation ofa large macromolecule, its diffusion in the embeddingmedium and its entropy driven dynamics) as well asfeatures characteristic of the local scales (e.g. the inter-and intra-chain correlations in a glass forming polymerand their time evolution, the rotational motions of methylgroups, vibrations...).

ii) By variation of the contrast between the structural unitsor molecular groups, complex systems may be studiedselectively. In particular, the large contrast achieved byisotopic substitution of Hydrogen (one of the maincomponents of soft matter) by Deuterium constitutes themost powerful tool for deciphering complex structuresand dynamic processes in these materials.

iii) Neutron reflectometry constitutes a unique technique forthe investigation of surfaces and interfaces in softmatter.

iv) The high penetration of neutrons in matter allows thestudy of the influence of external fields or parameters,e.g., the evolution of the system under processingconditions.

v) The space-time resolution of these techniques revealsthe molecular motions leading to the viscoelastic and

Space-time resolution atproper scales, variation ofcontrast and highpenetrability qualifyneutrons as a unique toolfor the study of structuraland dynamic properties ofsoft matter at a molecularlevel

71

mechanical properties of soft matter systems. Thisknowledge is valuable for the design of tailor madematerials.

The unique power of neutron scattering for revealing essentialfeatures in the field of soft matter can be exemplified with twopioneering experiments that can already be considered as“classic”. The first one is the experimental proof of the randomcoil shape of polymer chains in the melt or in the glassy stateas proposed in the 50's by Flory. This confirmation was onlypossible in the 70's [1] with the development of SANS. Sincein the bulk a given macromolecule is surrounded by similarunits, only by using contrast variation and deuterating singlemolecules could Flory's proposition be demonstrated. Thismeasurement of the single chain form factor by SANS wasone of the first applications of NS to polymer science. Thedynamic counterpart of this experiment could only be solved25 years later, neutron spin echo (NSE) investigations on thelong time chain dynamics recently allowed the confirmation ofde Gennes predictions on the mechanism of reptation inpolymers [2].

Seminal experimentsunravelled chain structureand dynamics in polymermelts

III. Future OpportunitiesSoft condensed matter systems in the future will increase incomplexity both in structure as well as in the number andspecific role of their components, e.g. multi-component softand soft / hard materials tailor made for industrial applications.Such complexity will cover a wide range of length and timescales, posing challenging problems to basic science.Desirable systems show complex interaction potentials withseveral minima, generating different structures according tothe mechanical and thermal history. The understanding of theinterplay of geometry and topology, and the characterisation ofinterfacial features are of the utmost importance for the futuredevelopments and design of novel materials. Finally, thestructural changes induced by external fields such as shearplay a crucial role in the outcome of industrial processing.These issues have to be dealt with as a necessaryprecondition for achieving controlled improvement in thefabrication of future materials.

General future trend:Tailor made multi-component materials forindustrial applications.This implies increasingcomplexity!

Neutron scattering in combination with advanced chemistry isthe necessary tool for facing the new challenges in the field ofsoft matter. Here the focus is on linking chemical architectureto microscopic and macroscopic properties. The interplaybetween computer simulation and neutron scattering alsopromises to become particularly effective because of thecommon ability of neutron scattering and computer simulationto home in on a key structural unit.

Future trends will require a wide variety of experiments,including investigations on dilute components, or on very smallamounts of matter such as particular topological points or atinterfaces. Sometimes these experiments involve polarisationanalysis, short time measurements or in-situ studies. In allthese cases, very high intensities of the neutron beam are

The interplay betweenneutron scatteringadvanced chemistry andcomputer simulations willbe of great importance

High intensity neutronbeams are essential forfuture developments onsoft matter

72

required.

Assuming the availability of a high flux neutron source with thecharacteristics of ESS some flagship areas in this field canalready be envisaged and are now listed.

Molecular rheologyThe application and industrial processing of many softcondensed matter systems strongly depends on theirrheological properties, which are determined by theinteractions and motions of the constituent structural unitssuch as chain molecules, aggregates, colloidal particles,surfactants, etc. Their understanding is one of the greatchallenges of basic soft condensed matter and would certainlyfacilitate the molecular design of new materials. As anexample we consider the rheology of polymer melts which iscurrently described in terms of the reptation model. A verylong polymer chain in a melt suffers local restriction of itsmotion by topological entanglements with other chains alongits length. The polymer chain can be envisaged being confinedin a “tube” formed by the neighbouring chains. The snakelikemotion along the tube (reptation) is the main mechanismcontrolling the dynamics of a highly entangled linear chain.

Figure 2: Schematic representation of the reptation mechanism of abranched polymer. The topological constraints imposed by thesurrounding chains are modelled by the imaginary forked redtube. The relaxation of the internal modes and the diffusion of thepolymer have to take place inside and along the tube prior to freediffusion. The yellow portion in the selected macromoleculerepresents the labelled branching point which dynamics wouldreveal the detailed molecular mechanism for reptation in thissystem.

The time is now ripe for the investigation of more complextopologies such as branched polymers (stars, H-polymers,combs, trees). Only a small number of branches in a polymer

A rheology based onmolecular understandingwould allow to designtailor made materials

73

molecule may substantially alter the industrial processingalthough they have little effect on the solid-state properties ofthe final material. Neutrons are uniquely suited to achieve thisgoal. Hydrogen atoms would be replaced with deuterium inspecific topological points of the molecule and the dynamics ofthese points followed by NSE techniques. The high dilution ofthe labelled topological points makes such experimentsinaccessible to currently available neutron fluxes.

Buried InterfacesNeutron reflectivity gives unique structural and compositionalinformation about buried interfaces. At present, only arestricted range of materials are sufficiently transparent topermit the experiment. By allowing a reduction of illuminatedarea of about two orders of magnitude, next generationneutron sources will relax the conditions for transparency toan extent that will give neutron reflection access to almost anyinterface. Two areas of particular importance are the liquid /liquid interface, where adsorbed polymers or amphiphiles playa crucial role in determining the stability of emulsions, andbiolubrication, where the delicate control of environmentalfactors (pH, ion concentration, etc.) is used to manipulate theconformation of polyelectrolytes at the lubricated interface.

A key experiment here is the response of the layer structure tocompression and / or shear under different solution conditions,an experiment that would become feasible at significantlyhigher neutron fluxes.

appliedvertical force applied

lateral force

aqueoussolution

neutronbeam

fixed block

Figure 3: The study of biolubrication by neutron reflection. The highintensity neutron beam will allow the illuminated area of interfaceto be reduced to a size (less than 10mm2) that should bemanageable in conjunction with a force balance. This will allowthe direct study of the conformation of adsorbed polyelectrolyteby reflection while various forces are applied to the system.

A large reduction inilluminated area givesaccess to new classes ofburied interface

74

Molecular self-assemblyMolecular self-assembly is increasingly being used as thedriving force to generate many new materials. The control ofthe self-organisation by means of applied chemical andphysical stimuli is the key to the successful creation ofstructures on the chosen nano- or micro-meter scale. The useof self-assembly ranges from the morphological control ofadhesion of bio-materials to plastics and composites, to thedevelopment of molecular electronics. Combination of nano-meter scale processing and molecular self-assembly is ascientific and technological challenge where neutronscattering will make key contributions. Specific labelling andthe large penetration depth of neutrons are both essential inrevealing the detailed structure and mutual interactions. Inmolecular electronics the precise control of surface structure isof crucial importance for the performance. High intensityneutron sources can investigate the molecular organisation ona surface-size relevant to electronic components.

With the next generation neutron sources, it will be possible tofollow the response of self-organized systems toenvironmental change. An example is the phasetransformation dynamics exhibited either in an oscillatoryelectrical excitation, which could be studied using stroboscopicobservation linked to the pulse structure of the beam, orfollowing shear or temperature quenches, which is bothfundamental and highly relevant to the processing andapplication of polymer components. Another exampleconcerns materials whose structures spontaneously respondto external changes, such as temperature or pH (smartmaterials), or systems which restore their microstructure afterdamage (self-recovery), each because of prescribed selfassembly.

Figure 4: Block copolymer self-association into micellar form can be usedfor drug delivery systems, where the molecular associationdependence on temperature or pH is utilised.

Combination of nanometerscale processing andmolecular self-assembly isa scientific andtechnological challenge

75

New Materials Produced by External ConstraintsExternal fields can be applied close to phase boundaries togenerate new materials. For example blending via shearand/or chemical reaction two polymers above their separationtemperature will produce an anisotropic structure, much finerthan classical blends, which can be quenched to produce thedesired material. Studying the successive stages following theonset of shear allows us both to select the best stage forapplication, and to understand the process. For example, asshown in Figure 5, flow enhances fluctuations of concentrationperpendicular to flow and, due to the vicinity of the phasetransition, these fluctuations diverge into layers alternately richand poor. Separation before shear would give a much coarsergrained mixture.

Figure 5: Snapshots of the evolution of a polymer-deuterated solventmixture close to phase transition under shear: a) early stages:enhancement of concentration fluctuations, accessible only bySANS; b) later stages (microscopy): separation occurs also atlarger scale (mm), the system is a multi-scale material.

These experiments absolutely require a powerful neutronbeam. The control of the final stage system needs extremelyprecise tuning because of the complex temporal evolution. In

Combining phasetransitions and externalfields generates novelmaterials

5µm50nm

z

v

∆∆∆∆v

q

Couette shear

z

v

θ

Early stage :scattering - >Model

Late stage : directimaging by Microscopy

76

turn, this demands fast, in situ and simultaneous observationsover a full range of length scales in conjunction withmeasurements with other techniques. The latter are alsoessential for both industrial and scientific reasons:measurement of the stress, light scattering, or birefringence ormicroscope measurements are easy through a SANS shearcell. For example in the ultimate stages, a strong change instress might result from phenomena at the micron scale.Without multiscale observation with light and neutrons,understanding is hopelessly complicated.

Soft-hard nanocompositesHybrids of crystalline inorganic particulates and polymers havebeen developed to combine the advantages of both classes ofmaterials and widen their range of application. Enhancedmechanical, dielectric, fire-resistance, heat deflection,permeation-barrier properties have been reported or can beexpected. The properties of particulate-filled composites aregenerally determined by the properties of the individualcomponents, composition, structure (spatial distribution),particle-particle interaction and particle-matrix interaction.Apart from the processing conditions (shear forces), the extentto which particles agglomerate depends on the balancebetween the attractive and repulsive forces between theparticles as well as between them and the matrix. Neutronscattering is a unique tool for studying the structure,interactions and dynamics of such complex hybrid materials.Partial structure factors in colloidal mixtures and colloid-polymer mixtures are for example the key quantities for adecisive test of the newly emerging theoretical description ofthe effects of interactions in these systems. However, themeasurement of partial structure factors requires much higherneutron flux than presently available. Moreover, the lengthscales found in many of the application-oriented studiesrequire a substantial extension of current instrumentation.Key topics in such complex mixtures are kinetics and non-equilibrium aspects such as aggregation, gelation and/orphase separation. The temporal evolution of the structure ofsystems that undergo aggregation and eventually a sol-geltransition is a very interesting problem that is also directlyrelated to recent progress in material sciences (novel non-metallic materials, ceramics processing, food science andtechnology etc.). However, this requires time-resolved SANSexperiments in the second and millisecond time scale. Thiscan only be achieved at much higher neutron flux thancurrently available, possibly combined with stroboscopicsampling and an extension of the accessible range of lengthscales by at least one order of magnitude.

The temporal evolution ofnon-equilibrium structuresis a key for understandingprocessing andperformance of hybridsystems

Complex Liquids in Porous MediaUnderstanding the behaviour of complex liquids in porousmedia is a particular challenge for the science of soft matter.By complex liquids we mean multicomponent systems ofpolymers, colloids, micelles and surfactants, whosecharacteristic length scales are frequently identical to those of

Contrast variationstrategies can reveal thestructure and flow ofcomplex liquids in porousmedia

77

porous materials. Complex liquids in porous media are ofgreat practical significance. They are found in oil productionwhere water comes up against petroliferous rocks, inprocesses separating materials through membranes, inremediating contaminated soils, in cleaning powders andpastes, and even in cosmetics.

The coincidence of the characteristic length scales of theliquid and the geometrical constraints due to the pores has aprofound influence on the thermodynamics, the phasebehaviour as well as on the transport properties of thecomplex liquid. For example, a network of small pipes couldbe partially plugged by the oil droplets of a microemulsion.The ability to predict the behaviour of oil/surfactant,polymer/surfactant, polymer/protein or protein/surfactantsystems in microenvironments based on a scientificunderstanding would enable effective control of manyindustrial processes.

Without adjusting the mean contrast of the liquid to that of theporous material it is practically impossible to obtain directinformation on the morphology and structure of liquids in theporous medium. In order to perform successful experiments aprecise background control is indispensable. Furthermorecontrast variation strategies involving several componentsneed to be envisaged. This is only feasible with at least oneorder of magnitude higher fluxes than that available.

Probing Molecular Dynamics in Non-Crystalline MatterThe unique capability of neutrons to tell "where the atoms areand what they are doing" (Nobel prize citation 1994,Brockhouse and Shull) derives from the fact that inelasticneutron scattering can reveal the motion of atoms on the timeand length scale of the atomic order and motion in crystallinematerial. The periodicity of the lattice allows us to drawconclusions on aspects of the atomic motion on virtually anylength scale, including collective motion over distances muchlarger than the elementary cells, such as propagating phononmodes.

Currently much technological and fundamental researchinterest has shifted from crystalline materials to non-crystallinematter. In soft materials the structure, macroscopicmechanical properties, and phase changes are determined toa high degree by the motion of atoms, i.e. the moleculardynamics. This is quite understandable in the absence of awell defined crystalline order because the quasi infinitenumber of different local atomic configurations will includesituations in which atomic rearrangements can easily takeplace by overcoming low barriers. The collective motion ofatoms, correlated over distances of several atomic spacingswill be of particular importance for determining the materialfunctions. To explore and understand this intermediate scaledynamics neutrons remain the unique probe for deliveringdirect information in space and time, although the task will bemuch more tedious in this case. Without the crystalline

Dynamics of collectivemodes in non-crystallinematerials provide a linkbetween the motion ofatoms and the mechanicalproperties

78

periodicity we have to observe directly the intermediate lengthscales between atomic and mesoscopic distances. For thisreason we will have to record and analyse with precision 2-3orders of magnitude smaller signals than those in crystallinematerials. This will become possible for the first time with the 3orders of magnitude gain in sensitivity ESS will provide forinelastic scattering spectroscopy with cold neutrons. Motionson the intermediate, typically nanometer scale are naturallymuch slower than on the atomic time scale of 10-13 s, so theseslow neutrons are best suited to reveal them.

The interplay between neutron scattering, which is the onlyprobe we have to explore space and time behaviour at theintermediate microscopic/mesoscopic scale boundary, andadvancing molecular dynamics model calculations will allowus to tell "where the atoms are and what they are doing" forthe first time in atom by atom detail and not only in simplecrystalline matter.

IV. Instrument and target requirements at ESSThe instruments at ESS relevant for soft matter research canbe put into three categories:

i) Top priority- High intensity small angle neutron scattering (SANS)- Focussing low Q SANS- High intensity reflectometer- High resolution neutron spin echo

All flagship areas require large increases of intensity overprevious sources in order to be successful. The 16

2/3 Hz long

pulse target station will offer the required intensities with gainfactors between 10 and 100 for our highest priorityinstruments. For the high intensity SANS and thereflectometer resolution in wavelength is not crucial.

ii) Second priority- Variable chopper cold time of flight- Wide angle NSE- Polarised diffuse scattering instrument of the D7 type

These instruments are also best placed at the long pulse 162/3

Hz target station, where they will have gains of up to 1000over the best of their class currently available. This will createnew opportunities for the study of local dynamic processessuch as primary and secondary relaxations in polymers, softvibrations and fast collective motions, all of which to a largeextent determine the mechanical properties of polymers andcomposite systems. The D7 type instrument will allow theinvestigation of partial correlations and a more selectiveapproach to local order in disordered materials.

iii) Third priority- 1.5µeV backscattering instrument- Liquids diffractometer

Soft condensed matterrequires as top priorityHigh intensity SANS,Low Q SANS,High intensityreflectometer,High resolution NSE,at the long pulse targetstation

79

The backscattering instrument is complementary to the wideangle NSE, but should be extended to a larger Q-range byproviding Si 311 crystals. Whilst NSE gives time dependentrelaxation functions directly, backscattering measures in theFourier domain. For incoherent scattering backscatteringnevertheless has the advantage that it does not pay thepenalty of polarisation loss. The liquid diffractometer willsupplement the D7-type instrument for the determination oflocal structures.

In summary, soft condensed matter research definitelyrequires the high intensity and large dynamic range of the longpulse slow target station. The 50 Hz short pulse target stationis the second but less favoured choice.

V. SummaryThe science of soft condensed matter constitutes a field oflarge breadth and a richness of phenomena, many with closelinks to technological applications. As a consequence of theirunique properties neutrons must play a key role for theexploration of this field. The decisive advantages of theneutron are the simultaneous accessibility of the proper lengthand time scales together with the possibility of changing thescattering contrast at will.

Future developments in soft matter science will move towardsthe study of increasingly complex, often multicomponentsystems, where time dependent phenomena in real timeexperiments, non-equilibrium situations and transientphenomena will be the focus.

Neutron scattering in combination with computer simulationswill make a very large impact in all these future scientificendeavours. Finally there remains the possibility that theprinciples of soft condensed matter science may also have animpact on the better understanding of a number of biologicalfunctions.

All the future trends require a strong increase in the availableneutron intensity. The prime instruments for soft condensedmatter research all require the power of the long pulse targetstation.

References

[1] R. G. Kirste, W. A. Kruse and J. Schelten, Makromol. Chem. 162, 299 (1973)

[2] P. Schleger, B. Farago, C. Lartigue, A. Kollmar, and D. Richter, Phys. Rev. Lett. 81, 124 (1998)

80

7.5 Chemical Structure, Kinetics, Dynamics

W.I.F. David1, H. Gies2, H. Jobic3, M. Latroche4, M. Prager5, A. R. Rennie6, C. Wilson1

1ISIS Facility, Rutherford Appleton Laboratory, Chilton, Didcot OX11 0QX, UK2Institut für Mineralogie, Universität Bochum, Universitätsstr. 150, D-44780 Bochum, Germany3Institut de Recherches sur la Catalyse, 2 Avenue Albert Einstein, F-69626 Villeurbanne Cedex, France4CNRS, Laboratoire de Terres Rares, 2-8 rue Henri Dunant, F-94320 Thiais Cedex, France5Institut für Festkörperforschung, Forschungszentrum Jülich, D-52425 Jülich, Germany6Department of Chemistry, King’s College, Strand, London WC2R, UK

AbstractOur informed understanding of the materials world around us is based upon a detailed knowledge of the structureand dynamics of materials on the atomic and molecular level. This contribution reports on the molecular andsupramolecular structures for sizes from tenths to hundreds of nanometers as well as dynamic processes studiedby inelastic and quasi-elastic spectroscopy in the field of chemistry.

Neutron and X-ray scattering are complementary processes: X-rays scatter from the atomic electrons while theneutrons probe nuclei. This important difference means that neutrons have the ability to accurately locate lightelements in the surrounding of heavy atoms. Since the chemistry of a mixed-metal oxide is determined principallyby the location of the light oxygen atoms (key examples include high temperature superconductors and colossalmagnetoresistance materials), neutron diffraction is the technique of choice for such measurements. Scatteringfrom the nucleus in non-magnetic systems avoids any effect due to charge transfer and thus gives valuableinformation on the chemical bond. Isotopic effects and contrast between neighbouring elements is also anadvantage compared to X-rays. Neutrons scatter relatively weakly but the cross-section is well understood andthe high penetration depth of neutrons allows measurements under extreme conditions of pressure andtemperature or in-beam chemical reaction measurements on large components. The ESS will give substantialintensity gains over current sources removing much of the flux-limited problems of current neutroninstrumentation. Furthermore, the time-of-flight technique at the ESS combined with short pulses and long pathswill allow high resolution and high flux to solve more difficult structural problems. The high flux and fixed scatteringgeometry inherent at ESS will facilitate a new generation of complex sample environments for in-situ experiments.

Chemists respond to the present demands for higher performance materials, cleaner environments and improvedefficiency in use of chemicals in a wide variety of ways. These include use of smart materials that respond to theirenvironment, use of thin films to build devices and the exploitation of pharmaceuticals and other agents such ascatalysts that are active in much smaller quantities than previously used. These developments require extensionof analytical tools to study chemistry and chemicals in small quantities, in complex mixtures and under theconditions of imposed external environments of stress, temperature, chemical environment and other fields orconstraints. One of these tools is vibrational spectroscopy, where neutrons have unique properties compared withother techniques. With the larger neutron flux available at the ESS, it will be possible to follow in situ catalyticreactions. One will be able to record vibrational spectra above room temperature with a spectrometer covering awide range of energy transfers, at low momentum transfers. The reaction pathways will be tracked down byobserving the reaction intermediates. With such an instrument, it will be possible to measure spectra in aqueoussolutions, which is the natural medium of biological molecules. Chemistry also involves local and diffusivetransport processes which give rise to incoherent (single particle) and coherent (collective) quasielastic scattering.High-resolution neutron spectroscopy yields the microscopic information in space and time. The increased flux atESS will extent such studies to lower concentrations, to systems with large inherent background, to more complexmotions and parameter dependent studies.

I. IntroductionNeutron diffraction gives a unique structural fingerprint of thecrystalline state. Light atoms are detected with high precisioneven in the presence of heavy atoms such as transition metalsand actinides. In materials science, archetypal examplesinclude hydrogen storage materials such as metal hydrides[1], and mixed metals oxides such as high temperaturesuperconductors [2] and battery materials [3]. In organicchemistry, the precise location of hydrogen atoms free fromcharge transfer effects contributes to our detailedunderstanding of hydrogen bonding from simple modelpeptide systems to supramolecular chemistry. The possibilityof isotopic replacement, and in particular H/D substitution, can

Neutrons probe nuclei andgive a better contrast forlight elements, isotopicsubstitutions or neighbourelements in the periodictable

81

be used to great advantage in elucidating specific structuraldetails (Fig. 1).

Figure 1: Hydrogen bonding schemes in nitroanilines with non-linear opticalactivity.

II. The impact of high flux sourcesWith increasing neutron fluxes and particularly with the adventof the ESS, neutron diffraction experiments typically involvethe collection of a large series of diffraction profiles. Thesetime-resolved, parametric experiments enable structuraltrends to be analysed as a function of physical parameterssuch as temperature [4], pressure [5] and magnetic field. Inturn, this leads to a fuller understanding of phase diagramsand structural transitions, and to deeper insights into structure-property relationship. Time-resolved neutron diffraction studiesare also very powerful means of following chemical reactions.Neutron powder diffraction allows bulk analysis of materials in“real-life” industrial configurations yielding importantcrystallographic, thermodynamic and kinetic information aboutreaction behaviour. Recent studies include investigations intoconcrete ageing, silicate compounds, hydrothermal syntheses,self-propagating chemical reactions, amorphisation, hydrideformation and decomposition and the charge/dischargebehaviour of batteries (Fig. 2).

Neutron penetration depthallows in situ experimentsto follow bulk reactions incomplex environment

III. Future opportunities and flagship areas

Energy storage and conversionEnvironmental problems such as the green house effect, leadto research of new solutions for energy management. Fuelcells will probably be the cleanest and the most versatilepower source of this century. However many scientificproblems remain to be solved: efficient catalytic processes atelectrode surfaces, ionic diffusion in solid state electrolytes,chemical reaction kinetic optimisation. In all these cases,neutrons will be useful to probe structural aspects andchemical mechanisms.

82

Materials for cryo-coolers used in space, tritium storage units,fuel cells and alkaline batteries are now metallic hydrides.From a fundamental point of view, competition betweenmagnetic and hydrogen ordering in rare earth-transition metalhydrides has been subject to much recent research. Therecent discovery of switchable mirrors also gives an insight tothe physics than can be studied through these compounds. Inthe surrounding of heavy metals, neutron diffraction is aunique tool to locate the proton in metallic hydrides: during theabsorption process, the electron is transferred from the protonto the conduction band and only neutrons, probing thenucleus, can give accurate structural data. Moreover, for theforthcoming applications, most of them related to clean energystorage and energy conversion, time resolved experimentswould give valuable information on the chemical and kineticprocesses involved during charge/discharge cycles allowingoptimisation of these materials.

Figure 2: Three dimensional view of the neutron diffraction patterns of ametal hydride electrode during an electrochemical charge in 10hours. The different phases involved in the reaction are given onthe plot [3].

An increase of flux is essential to study faster processes likevery fast discharge phenomenon or short circuit batteries.Materials used in modern batteries often include H and Li ascharge carriers. The cations are distributed over various sitesin the crystal structure and show high mobility. Their detailedanalysis is most important in order to understand theconduction pathway and requires neutron diffractionexperiments. In addition, the dynamics can be studied withquasi-elastic neutron scattering. On the other hand,crystallographic data accuracy is often affected by thecomplexity of the experimental environments. A betterresolution associated with a larger Q range will be necessaryto solve new problems.

Flagship area 1:Clean and efficient energystorage and conversion.

83

In situ studies of catalystsIn catalysis, one can study with neutron spectroscopy thesurface species which result from molecular adsorption,dissociation, or chemical reaction even for some systems thatcannot be studied by diffraction. The technique is well suitedto determine the different adsorption sites for atomic hydrogenon metal or sulfide catalysts [6], to identify the active speciesin catalytic reactions, and to understand deactivation (Fig. 3).A limitation of the technique is the quantity of catalyst whichhas now to be prepared (tens of grams). A neutron flux one ortwo orders of magnitude larger would also allow to study theadsorption of non-hydrogenous compounds, such as CO, SO2

or NOx. Kinetic studies would then be possible.

Figure 3: Observed (solid line, recorded at 20 K) and calculated (dashedline) spectra of an industrial palladium catalyst after reaction. Thesurface is covered with methyl groups, which explains thedeactivation [7].

In order to make catalytic processes cleaner and moreefficient, one must identify the active species and the reactionintermediates. To follow in situ catalytic reactions, one needsa spectrometer that can measure the whole vibrationalspectrum at small momentum transfer. There are numerouscatalytic reactions that would benefit from such an instrumentin hydrogenation, oxidation or desulfurisation. For example inthe conversion of n-butane to maleic anhydride one could findout if the intermediate is an olefin or an alkoxide. This wouldpermit the building of a kinetic model for the reaction.

A catalytic cycle involvesadsorption, diffusion, andreaction steps. Neutronscattering techniques canplay a major role in allthese aspects of catalysis

Flagship area 2:Improving catalysts byspectroscopic studies ofreagents andintermediates

Hydrogen bonding and proton dynamics in AdvancedMaterialsThe analysis of the structure of molecules of biologicalrelevance is important to understand their functions. To studythe conformational flexibility of these molecules, a knowledgeof the intra- and inter-molecular forces is required. To achievethis task, vibrational data can be used jointly with quantummechanical calculations. Hydrogen vibrational dynamics can

Neutron scattering is apowerful mean to analysethe proton dynamics ofmolecules of biologicaland pharmaceuticalinterest

84

only be accessed by neutron scattering [8]. A better energyresolution is needed to separate the numerous modes ofthese complex molecules. So far, the measurements havebeen performed in the solid phase, aqueous solutions wouldbe a more natural medium. The analysis of organic moleculesinteracting with the surface of a substrate as in biomineralsand drug supports offers new opportunities. The details of thebonding interaction would lead to a basic understanding of thebiomimetic processes of formation in biomineralization.

There has been a recent increase in exciting work in the areaof developing molecular materials with useful and tuneablephysical properties such as magnetism, superconductivity,non-linear optical activity, polymorphism, etc, bearing fruit.This area is likely to expand dramatically in the next decades.Much of this work is focused around understanding theintermolecular interactions holding 3D arrays of moleculestogether, often weak hydrogen bonding interactions. Thedirectionality of these interactions is crucial, and the accuratedefinition of hydrogen atom positions by single crystal neutrondiffraction vital. For example, in pharmaceutical materials, theunderstanding of polymorphism can often rely on small energydifferences between molecular configurations, while insupramolecular chemistry accurate quantification of weaklybonded motifs will allow for more rational crystal engineeringallowing chemists to tailor properties by designing structures.ESS will advance this expanding area of molecular science byallowing routine characterisation of all atoms in suchstructures. Specifically, ESS will allow such studies to becarried out on smaller crystals, will offer faster structuralcharacterisation and allow more systematic examinations ofthe phase space of candidate molecular materials; mappingmost appropriately onto the needs of the science.

Proton transfer along a hydrogen bond is the simplestexample of a chemical reaction, a covalent bond is broken andthe hybridisation of the acceptor and donor atoms isexchanged. The potential energy barrier to proton transfer,separating the 2 wells, which correspond to the stablepositions of the hydrogen atom, is therefore high and themechanism for transfer entails tunnelling through the barrier.However, in view of the modulation of the electronic state ofthe molecular skeleton during proton transfer, molecularvibrations also participate, promoting or hindering protontransfer. New theoretical methods are being developed tohandle coupled tunnelling and vibrational dynamics. Neutronscattering is a uniquely powerful tool for precisely locatinghydrogen atoms in these systems and then measuring theirdynamics. The tunnelling dynamics in hydrogen bonds isprobed directly by quasi-elastic scattering and molecularvibrations are measured by inelastic scattering.

Flagship area 3: Advancedmolecular materials

Hydrogen bonds play afundamental role in thestructure and reactivity ofchemical and biologicalsystems

Diffusion in porous materialsMolecular diffusion in porous materials, such as zeolites, isimportant in catalysis or in separation processes. In addition totheir fundamental character in elucidating confinement effects,

Molecular diffusion can befollowed, on a microscopicscale, with neutrons

85

the aim of these studies is to create new diffusion models validfor complex systems. When the size of the molecule iscomparable to the pore size, this leads to diffusion limitations,and diffusion coefficients are from 3 to 12 orders of magnitudelower compared to the gas phase. Various experimental andtheoretical techniques (Fig. 4) are used to determine diffusioncoefficients. In several cases, it has been found that quasi-elastic neutron scattering is the only technique which is able toderive reliable intracrystalline diffusivities [10]. The neutronspin echo technique allows to probe much longer time scales,this has been demonstrated for the diffusion of deuteratedmolecules in zeolites.

Figure 4: Minimum energy path for benzene in NaY zeolite, between acation site (Na in blue) and a window site (the molecule at thetransition state is in red [9]).

Rotational tunnellingRotational tunnelling is among the simplest quantumdynamical processes. With high resolution neutronspectroscopy entering the field the understanding of propertiesand importance of quantum motion in solids has beensignificantly promoted. This is largely due to the extreme(exponential) sensitivity of tunnelling on the intermolecularpotentials in combination with its unambiguous identification.Combined with theoretical chemistry programs rotationaltunnelling is especially suited to obtain precise rotationalpotentials. They allow determination of the intermolecularinteraction potentials [11] via pressure, disorder in molecularalloys and glassy systems, the influence of time dependentperturbation of the environment (coupling with phonons),deviations from single particle dynamics by coupling to otherdegrees of freedom. This last field of multidimensionaltunnelling processes will be one of the most exciting ones inthe future. Rotation-translation coupling is already established[12]. The coherent counterclockwise rotation of a methyl groupand its centre of mass in a four-fold environment imposes asurprising fourfold proton density distribution of a three-foldrotor (Fig. 5) confimed by neutron diffraction.

Rotational tunnelling is auniquely sensitive probeof fundamental andsystem properties.Neutrons yield completemicroscopic information

86

Figure 5: Four-fold proton density distribution of a three-fold rotor due torotation-translation coupling in a four-fold environment [12].

With ESS tunnel splittings of excited rotational states, welldispersed molecules isolated in a matrix, extended coherentsurfaces, single crystals, disordered materials, time-dependent effects as spin conversion, new forms ofmultidimensional tunnelling and new non-hydrogenous rotorswill become routinely accessible to rotational tunnellingspectroscopy.

Electrochemistry at surfacesNeutron reflection experiments are relatively new but are nowbeing applied to a wide range of chemical studies. A goodexample is the application of this technique toelectrochemistry. The interesting chemistry happens atinterfaces (electrodes) and a wide-range of different chemicalspecies are present. Neutron reflection is an excellent tool forthe determination of the distribution of various ions andmolecules near an interface and for the determination of thecomposition and structure of deposited layers. The underlyingscientific problems arise from important technologies such asthose of energy storage, analytical and microanalyticaldevices and biological sensors. At present relatively fewexperiments have been conducted [13,14] but considerableprogress can be foreseen. The uniform surface areas ofsamples that are available for study is often very limited and afew mm2 is more common than the few cm2 usual for reflectionexperiments with current instrumentation and sources. Higherflux instrumentation will allow experiments on these realisticsamples. The changes that occur in electrochemical deviceswill also be followed in real time or by application of cyclic dataacquisition phased with external potentials or currents.

Neutron reflection is apowerful tool for the studyof interfaces such asthose in electrochemicalsystems. ESS with highflux, by reducing theaccessible area, will bringforward many newsystems for study

87

82.0 84.5 87.02-θθθθ / degrees

0

Time

/

secondss

84

168

Figure 6: Diffraction from a dispersion of Ni(OH)2 particles in D2O showingthe change in intensity of the 004 peak in one direction when flowstarts. Each time slice is two seconds [16].

An example of an experiment that is of interest to the chemicalindustry is one in which alignment of plate-like particles underflow is studied [15]. This is shown in the Fig. 6. Particles with~90 nm diameter and approximate aspect ratio of 5:1 wereprepared as model systems for experiments. Each particle is asingle crystal. The alignment can be followed by diffraction ofneutrons from the particles in the dispersion. This can befollowed dynamically by use of a cyclic data acquisitionprocedure. Experiments are limited by the small signal inrelation to the background. At present only relatively largevolumes can be studied. This type of data, in conjunction withsmall-angle neutron scattering, has already been used toidentify phase changes under shear. This type of experiment,including magnetic and electric orientation, could becomecommon if higher flux were available to study small samplesthat could be subject to more uniform fields.

Stroboscopic dataacquisition has manypossibilities to studydynamic processes underthe influences of externalfields such as flow, stresselectrical and magneticforces

Polymer synthesisSmall-angle neutron scattering has become a pre-eminent toolin the characterisation of polymers, colloidal particles and avariety of mesophases. It is to be expected that work to followthe synthesis of these materials in-situ will become morecommon [17]. This has been difficult up to now because of thepoor time resolution available and the need to study samplesdilute in the component of interest surrounded by many otherdifferent molecules. It is to be expected that investigationsaimed at deepening the understanding of polymerisationmechanisms will develop. For example these could study thelocation of initiator in emulsion polymerisation reactions or theconformation of the polymer molecules as they form. SANShas been widely used to look at the morphology of resultinglatex particles [18] but only a little work has been viable in this

SANS can be used tofollow polymerisationreactions in real time; todetermine reactionmechanisms and theinfluences of synthesisconditions on structure ofmaterials

88

area with present flux [19].

Chemical kineticsESS with a high flux and instruments optimised for resolutionfor particular experiments will allow a much wider range ofkinetic experiments. Cyclic data acquisition permits timeresolution of 1 ms. At present experiments have beenperformed on pulsed flow and ultrasonic excitation of a crystal.For example studies of reorientation dynamics in concentratedcolloids have already been made. Future applications wouldinclude cyclic electrochemical processes. This would giveinformation about the distribution of ions in both solid andliquid phases: this would be studied by neutrons reflection,SANS, and neutron diffraction.

Flagship area 4: Timeresolved studies ofchemical kinetics withstroboscopic dataacquisition

IV. Instrumentation requirementsChemistry has wide-ranging requirements for neutroninstrumentation. These include elastic scattering, quasi-elasticscattering and inelastic scattering.

In determining more complex structures or looking at moresubtle microstructure, high resolution powder diffraction iscrucial to providing a sufficiently large number of well-definedBragg peaks. With ESS fluxes, a high-resolution machine willbe able to study small samples, such as isotopically enrichedmaterials or compounds synthesised in small quantities bynovel synthetic routes.

Parametric studies provide a wealth of information intostructure-property relationships. Such a powder diffractometerfor rapid time-resolved experiments should have gooddetector stability, short acquisition time, large Q range, bestachievable resolution and a versatile sample environment.Data accumulation times should be of order of 10 ms.

To increase pressure beyond the actual limit (about 50 GPa) itis necessary to reduce the sample sizes to about 0.01 mm3.This type of study obviously needs a powder diffractometeroptimised for extreme environments with high flux and largebeam time to get valuable results. Possibility to combine HP(>50GPa) and HT (>2000K) is a challenging goal to reach.

A high throughput small-molecule single crystal diffractometerwill offer high quality data from sub-mm3 samples of bothorganic and inorganic materials in around 1 hour. Withsufficiently large crystals, parametric studies as a function oftemperature and/or pressure should be routine. The capabilityfor single crystal experiments on larger molecules will alsoprovide detailed insights into the behaviour of complexsupramolecular assemblies and large inorganic systems.

Small-angle scattering will be important with a wide range ofmomentum transfer in a single configuration to follow dynamicprocesses. A choice of resolution appropriate to the problemis desirable. A Q range of at least 0.001 to 0.5 A-1 should be

Relation to ESS-instrumentation

Powder and single crystaldiffractions, small anglescattering andreflectometry

89

accessible.

Neutron reflection to study interfaces will need high flux and awide dynamic range to cover reflectivities as small as 10-8. Arange of Qmax/Qmin of 30 in a single configuration is needed.

Vibrational spectroscopy for in situ catalysis or biologicalmolecules in aqueous solutions requires measurements ofspectra at low momentum transfer values, Q<2 Å-1 , up to 400meV, at intermediate resolution. A high resolution instrument,∆E/E ≈ 1%, is also required.

Transport processes, intramolecular vibrations, hydrogenbond dynamics or rotational tunnelling are all influenced orperturbed or coupled to lattice vibrations. A direct geometryhigh performance time-of-flight instrument for the energyregime up to 100meV is necessary to control or exploit thephonon density of states.

In the quasi-elastic domain, spectrometers with well-definedline shapes are needed. Only an instrument able to accesslarge Q values at intermediate resolution will allow todifferentiate between different models of motion. Polarisationanalysis would be useful to separate coherent from incoherentscattering (to eliminate Bragg peaks, or to study collectivephenomena).

Backscattering spectrometers should be available with thebest possible energy resolution for very slow motions or strongpotentials and with large momentum transfers.

A neutron spin echo spectrometer with a Fourier time limit of 1µs is required to follow slow diffusion.

Most of the diffractometers and spectrometers listed above fitbetter on the 50 Hz short pulse target station.

Apart from the general requirements for improvement in fluxand resolution in energy and momentum transfer that willbenefit most experiments, there are some interestingpossibilities for a new, pulsed neutron source. Instrumentscould be built that have some available configurationsoptimised for cyclic data acquisition with appropriate shortsample to detector distances and high spatial resolution on thedetector.

Inelastic and quasi-elasticscattering

References

[1] M. Latroche, J.-M. Joubert, A. Percheron-Guégan, and P. H. L. Notten. J. of Solid State Chem. 146, 313(1999)

[2] R.J. Cava, A. W. Hewat, E. A. Hewat, B. Batlogg, and M. Marezio, Physica C 165, 419 (1990)

[3] F. Cuevas, J.-M. Joubert, M. Latroche, and A. Percheron-Guégan, Applied Physics A 72, 225 (2001)

[4] H. Lehnert, H.Boysen, F. Frey, A. W. Hewat, and P. Radaelli, Z. Kristallogr. 212, 712 (1997)

[5] I. N. Goncharenko, I. Mirebeau, A.V. Irodova, and E. Suard, Phys. Rev. B 56, 2580 (1997)

[6] H. Jobic, G. Clugnet, M. Lacroix, S. Yuan, C. Mirodatos, and M. Breysse, J. Am. Chem. Soc. 115, 3654(1993)

90

[7] P. Albers, H. Angert, G. Prescher, K. Seibold, and S. F. Parker, J. Chem. Soc. Chem. Commun. 1619(1999)

[8] M. P. Gaigeot, N. Leulliot, M. Ghomi, H. Jobic, C. Coulombeau, and O. Bouloussa, Chem. Phys. 261, 217(2000)

[9] S. M. Auerbach, Int. Rev. Phys. Chem. 19, 155 (2000)

[10] H. Jobic, J. Mol. Catal. A 158, 135 (2000)

[11] M. Neumann and M. R. Johnson, J. Phys. Chem. 107, 1725 (1997)

[12] P. Schiebel, G. Amoretti, C. Ferrero, B. Paci, M. Prager, and R. Caciuffo, J. Phys.: Condens. Matt. 10,2221 (1998)

[13] R. W. Wilson, R. Cubitt, A. Glidle, A. R. Hillman, P. M. Saville, and J. G. Vos, Electrochim. Acta. 44, 3533(1999)

[14] A. R. Hillman, P. M. Saville, A. Glidle, R. M. Richardson, S. J. Roser, M. J. Swann, and J. R. P. Webster,J. Am. Chem. Soc. 120, 12882 (1998)

[15] A. B. D. Brown and A. R. Rennie, Phys. Rev. E 62, 851 (2000)

[16] A. B. D. Brown, S. M. Clarke, A. R. Rennie, P Convert, and T. Hansen, Annual ILL Report (1997)

[17] J. Stellbrink, L. Willner, O. Jucknischke, D. Richter, P. Lindner, L.J. Fetters, and J.S. Huang,Macromolecules 31, 4189 (1998)

[18] R. H. Ottewill, S. J. Cole, and J. A. Waters, Macromol. Symp. 92, 97 (1995)

[19] M. F. Mills, R. G. Gilbert, D. H. Napper, A. R. Rennie and R. H. Ottewill, Macromolecules 26, 3553 (1993)

91

7.6 Neutron Scattering in Mineral Science, Earth Science,and related fields with the European Spallation Source

R. Rinaldi1, G. Artioli2, M.T. Dove3, W. Schäfer4, P.F. Schofield5, B. Winkler6

1Dipartimento di Scienze della Terre, University of Perugia, I-06100 Perugia, Italy2Dipartimento di Scienze della Terra, University of Milano, Via Botticelli 32, I-20133 Milano, Italy3Department of Earth Science, Downing Street, Cambridge CB2 3EQ, University of Cambridge, UK4University of Bonn and Forschungszentrum Jülich, D-52425 Jülich, Germany5Department of Mineralogy, Natural History Museum, London, SW7 5BD UK6Institut für Geowissenschaften, Kristallographie/Mineralogie, Universität Kiel, Olshausenstr. 40, D-24098 Kiel, Germany

AbstractThe objects under study in the Earth Sciences (and related fields) pertain to all aggregation states of matter, i.e.:solid, glass, molten, liquid and gas. Hence the methods of investigation are those adopted in all other branches ofScience dealing with such states. In the case of the Earth Sciences however, we must consider a further degreeof complexity introduced by the non restricted nature of the chemical and physical parameters underlying theformation and transformation of these objects (e.g. chemical composition, temperature and pressure).Neutron scattering has been added only recently to the Earth Sciences methods of investigation, mainly thanks tothe latest generation of diffractometers and spectrometers at modern neutron sources, allowing the accuratedetermination of subtle, yet very important, structural details in minerals also as a function of temperature andpressure.

Still, many areas of Earth Science research remain out of reach of present-day neutron instrumentation. Of thesewe have selected a small, representative number, which in turn is represented by a few “keynote experiments”that could be tackled with an ESS-type of instrumental set-up and could therefore provide significantadvancements in the Earth Sciences and related fields. The three broad themes identified are:1. In situ measurements of structure-properties relations in mineral phases under geological conditions.2. The structure, reactivity and physical properties of multi-component melts and fluids under pressure and

temperature conditions representative of the Earth’s interior.3. Texture and stress analysis of polymineralic rocks for the reconstruction of tectonic processes and modelling

of rock anisotropies.

Within these themes, four “flagship experiments” were envisaged, namely:1. - a) In situ diffraction and spectroscopic studies of molecular components in methane clathrates.

- b) Spin dynamics of iron-containing deep Earth phases.2. - Space- and time-resolved tomography of volatile-containing crystallising magmas.3. - Time-resolved simultaneous structure, texture, and stress analysis of polymineralic systems under

variable P/T conditions.

These areas of study and the relative experimental proposals are set against coloured backgrounds in the text forclarity. For the majority of experiments envisaged at present in this field, the general preference goes for a targetstation allowing maximum space and time resolution, i.e. a 50Hz, 5MW short pulse spallation neutron source.

I. IntroductionThe use of neutron scattering by the Earth Sciencescommunity has a relatively short history, but it is now clearthat the potential of neutron scattering methods for thesolution of Earth Sciences problems, including manyenvironmental problems, is enormous. Many of the problemsencountered in Earth Sciences have, until recently, simplybeen too complicated for earlier neutron sources andinstrumentation. Only with the advent of the latest generationpowder diffractometers at modern spallation sources such asISIS, has it become possible to study the crystal structures ofminerals as a function of temperature and pressure withsufficient accuracy to be really useful in solving subtleproblems such as cation ordering. However, there are manyareas in the Earth and environmental sciences for which thepresent sources and instrumentation are still inadequate.Examples include measurements of the structural changes in

The Earth Sciences cangreatly profit from neutronscattering studies withlatest and futuregeneration neutronsources

92

minerals at very high pressures and simultaneous hightemperatures, locations of light elements in complexstructures, strain measurements and scanning ofpolycrystalline aggregates under non ambient conditions, andstudies of the dynamical properties (neutron spectroscopy)also at non-ambient conditions. This information will enablethe modelling of fundamental processes in the Earth, rangingfrom large scale phenomena such as deep-focus earthquakesand volcanic activity, through to the transport (and disposal) ofpollutants in the Earth's crust and stone preservation inmonuments.

Recent experience with state-of-the-art sources, such as ISISand ILL, has shown that the outlook for applications of neutronscattering in Earth Sciences is extremely promising. There aremany features of neutron scattering that find ready applicationto the study of natural materials. One clear example is thatmany natural materials contain hydrogen. Hydrogen is virtuallyinvisible to X-rays, but it will scatter neutrons reasonablystrongly, both coherently and incoherently. This means thatneutron scattering from hydrogen can act as a probe of bothsingle-atom dynamics and collective excitations. Hydrogen iscentral to so many problems in geology and environmentalscience that there are countless important applications ofneutron scattering in these areas. As the facilities for collectinghigh-quality data are further developed, so our ability toresolve these scientific issues increases.

Typical examples mayinclude: hydrogencontaining phases etc.

The ESS will enable us to tackle many long standing issuesrelated to geological and environmental processes. The abilityto construct sample environments that will reproduce thetemperature and pressure conditions of the deep Earth, andthe increase in neutron intensity allowing reduced interactionvolumes and shorter data collection periods, will allow us toperform many "in situ" studies of mineral behaviour, which willgreatly increase our understanding of the behaviour of theconstituent materials of the Earth. The ability to probe thestructures and motions of relatively-complex minerals willprovide many new insights and allow us to understand thenumerous interactions that govern the behaviour of Earthmaterials in their natural environment. We anticipate beingable to study both solids and fluids, as well as the interfacesbetween these two phases, for the first time by neutrontechniques.

In situ studies on dynamicprocesses andphenomena

II. Points of merit for neutron scattering in the Earth andEnvironmental Sciences

Hydrogen in mineralsNeutrons, as opposed to X-rays, are efficiently scattered byhydrogen 1H and deuterium 2H atoms. Many minerals containhydrogen, often in the form of bound or free hydroxide ions, orin the form of bound or free water molecules within eitherstructurally active sites, or interstitial cavities in the crystal

“Water” in minerals androcks is of paramountimportance in determiningtheir properties

93

structure. Water in minerals and rocks is extremely importantin regulating a large variety of behaviours and properties ofinterest to the Earth Sciences spanning from the atomic to thecontinental scale. Because hydrogen has an extremely largecross section for incoherent scattering, neutron scattering canbe used to study the motions of individual hydrogen atoms.Slow motions of the hydrogen atoms, such as diffusion orreorientational motions, can be probed by quasi-elasticscattering, and fast motions by high-energy spectroscopy. Onthe other hand, since deuterium has a reasonable crosssection for coherent scattering and no appreciable crosssection for incoherent scattering, deuterated samples can beused in diffraction studies for the location of hydrogen sites incrystal structures and their modifications under inner earthconditions [1].

Scattering cross sectionThe fact that the scattering cross section for neutrons does notchange with scattering vector, whereas with X-rays it falls offmore-or-less as the inverse of the atomic radius, means thatneutron scattering allows to collect diffraction data to largescattering vectors. This is useful for a number of reasons.First, for investigating complex crystal structures and crystalchemistries, evident in many minerals, it allows for asignificant increase in the amount of information available in adiffraction pattern. Second, for information about thermalmotion a wide coverage of scattering vector is essential. Third,to extract information about site occupancies, and to decouplethis information from the thermal motion, it is again essentialto have data over a wide range of scattering vectors.Furthermore, in crystals with considerable disorder, or inamorphous materials or liquids, there is a lot of informationabout short-range order contained within the total scattering,S(q). The Fourier transform of S(q) provides information aboutthe pair distribution function g(r), the resolution of which willdepend directly on the range of the scattering vector in thedata. Thus one can obtain better data for g(r) from neutronscattering than from X-ray scattering, although it is mostprofitable, sometime essential, to combine data from bothtechniques especially where complex systems are concerned.

Thermal motions, siteoccupancies short- andlong-range orderinformation are moreeasily detectable byneutron scattering

Iso-electronic speciesThe contrast between the neutron scattering cross sections ofmineralogically common atoms or cations which have equal orsimilar numbers of electrons, such as Ti4+, Ca2+, K+; or K+ andCl- or Na+, Mg2+, Al3+ and Si4+ or Fe2+ and Mn2+, allows neutrondiffraction to be used for the direct determination of their siteoccupancies and order-disorder distributions. Untangling theordering of these cations by X-rays can only be achievedindirectly by the analysis of bond lengths, but these are notdefinitive since bond lengths are affected by factors other thanthe specific site occupancy. Neutron diffraction allows for thedirect determination of site occupancies for these frequentlycoexisting cations in minerals. Furthermore, althoughsynchrotron X-ray resonant scattering can certainly be

Scattering contrast is notelectron-dependent

94

achieved to enhance scattering contrast in favourable cases, itcannot be performed systematically being dependent onavailable edges and bonding features.

In-situ experimentsTo study the behaviour of minerals requires the reproductionof their “natural” environment and thus the need forsimultaneous high temperatures and high pressures. “In situ”studies are best suited for a thorough knowledge of therelations between thermo-baric variables and structuralproperties such as phase transitions, cation partitioning, bondvalence, electronic structure, etc..Traditionally high pressures have been easier to work withusing X-ray diffraction and diamond anvil cells, but the use oftime-of-flight neutron techniques has lately allowedconsiderable progress in high pressure mineralogy. The lowattenuation of neutron beams by many materials caneffectively make extreme sample environments (HT, HP,Reaction Cells, differential loading frames, etc.) easier tohandle for neutron scattering than for other experimentaltechniques.

Examples of frontier applications of neutron diffraction arenowadays mostly in the field of in situ studies where mineralstructures are investigated while the sample is kept at hightemperature (Figure 1) [14], [13] and/or high pressure [5].

0 200 400 600 800 1000 1200 14000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.00 200 400 600 800 1000 1200 1400

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

KD

Temperature (°C)

Figure 1: Example of data obtained by high temperature in situ neutrondiffraction on single crystals. The plot shows the variation ofordering with temperature between Fe and Mg in Fa12 and Fa10natural olivines. Up triangles: Fa12 ISIS-SXD data (880, 1060°C);stars Fa12 ISIS-SXD data (25, 960, 1030°C); down trianglesFa12 ISIS-SXD data (800, 1050, 1300°C); filled circles Fa10 ILL-D10 data (25, 900, 1070°C); line at KD=1 marks total disorder;points above 1 indicate Fe2+ segregation into site M1; pointsbelow 1 indicate Fe2+ segregation into M2. The crystals undergoa peculiar, previously undetected, ordering reversal withtemperature which is non-quenchable (Rinaldi et al., Phys. Chem.Min, 2000, 27, 623-629).

“Natural occurrence” for amineral often means hightemperature and highpressure

95

With the ESS, the foreseeable increment in neutron flux (afactor of 30 over ISIS) and detector efficiency, are expected toprovide a much wider scope for these studies, extending thecapabilities of pressure/temperature devices by a factor oftwenty fold in a conservative foreseable future, hence openinga whole new area of Earth Science studies (Figure 2).

Figure 2: Pressures and Temperatures achievable with present-day andfuture designs cells usable at Neutron Facilities.

Spectroscopy and modesNeutron scattering is extremely good for studying thedynamical properties of materials. Unlike spectroscopy withelectromagnetic radiation, inelastic neutron scattering is notsubject to tight selection rules on mode symmetries and wavevectors. For this reason neutron scattering can be used todetermine phonon dispersion curves and phonon densities ofstates, providing a fundamental understanding and theprediction of mineral behaviour and phase transformations ofminerals under pressures and temperatures of the Earth’sinterior.

Extension to the inelastic and quasi-elastic scattering of “insitu” techniques is very promising. Such measurements,although requiring highly sophisticated means of datainterpretation, offer a unique opportunity for solving finestructural details (atomic and protonic dynamics, soft modes,etc.) and allow better modelling and interpretation offundamental thermodynamic parameters. Limiting factors aremainly associated with the availability of large enough andhomogeneous natural single crystals of the phases of interest.Powder inelastic neutron scattering could also provide a viablecomplementary route, especially when associated with non-ambient techniques.

Fundamental structuralproperties of minerals canbe investigated by INS

96

The limitations in INS studies of minerals are essentiallycorrelated with the lack of large pure samples as both singlecrystals and powdered pure specimens. The limited scope ofthe studies carried out so far on minerals reflect this drawback in the use of present-day neutron sources. Here again,an increment in neutron flux, as expected from an ESS-typesource, would serve the purpose of extending INS to smallerpurer samples of many more mineral species and phases [2].

Magnetic propertiesNeutron scattering is the best probe of microscopic ordering ofmagnetic moments, and can be used to determine magneticstructures, collective magnetic excitations, and crystal fieldenergy levels. The magnetic structures and transitions of Feminerals present in high pressure environments in the deepEarth is of paramount importance to elucidate their physicalproperties and behaviour. Although magnetic X-ray scatteringcan certainly be performed with synchrotron radiation, it ispractically limited only to resonant species (i.e. Fe and a fewREE), therefore the use of the ESS neutron source will allowmuch better measurements, especially under pressure.

Neutrons are especiallysuited for studies of themagnetic properties andmagnetic structure ofminerals

Direct ImagingNeutron penetration and the time-structure of a pulsed sourcecan be advantageously exploited for time-resolved neutronabsorption measurements to determine viscosity and densityof magma-type melts at high pressure and temperature.Neutron imaging experiments at pressures up to 5-10 GPaand temperatures of 1300-1500°C in a cm scale cell wouldyield precise in situ measurements that could also beextended to the study of reaction fronts in silicatecrystallisation [17]. More readily available measurementswould be those related with the inner fabric of materials andartefacts, beyond the reach of less penetrating probes, forapplications in many fields including archaeology andpreservation of cultural heritage.

In situ physical propertiesof magma melts and directimaging of internal fabricsof rocks and historicalartefacts in bulk

Mineral SurfacesThe breakdown, weathering and transformation of minerals onthe Earth inherently involves the migration of hydrogenthrough the mineral surface and into the subsurface of thecrystals, thus changing the physical properties of the mineralsin the surface region. As these reactions occur at themineral/mineral, mineral/fluid or mineral/biota interface, thestudy of these protonation reactions is fundamental to ourunderstanding of weathering and mineral breakdown.Currently, X-rays are used in reflectivity mode to investigatemineral surfaces, but as previously mentioned, in order toinvestigate protons, neutrons are far superior to X-rays.

Study of protonationreactions responsible formineral alterations andsurface transformations

Texture and stress analysisTexture, defined as preferred orientation in a crystallinematerial, carries a fingerprint of the rock‘s history. Thecomplexity of geological texture analysis results mainly from

Using the penetratingpower of neutrons toextract the geological

97

the overprinting of different textures upon several mineralcomponents from different periods of geological activity.Quantitative texture analyses provide fundamental informationfor the modelling of rock anisotropies and reconstruction oftectonic events.

The high penetration capability of neutrons and the availabilityof wide beams allow the investigation of large specimenswhich produce global volume textures with high grain statisticseven on coarse-grained materials. Using position-sensitivedetectors and time-of-flight techniques, texture can beanalysed from reflection-rich diffraction patterns ofpolymineralic rocks containing low symmetry mineralconstituents [15].

Residual stress analysis of geological material is crucialbecause natural effects on rocks are orders of magnitudesmaller than in technological materials and drilling gives rise tostress relaxation. Furthermore, transient stresses and strainscan be directly observed through in-situ measurements atvarious pressures and temperatures.

Future prospects at a new high flux neutron source arepromising in performing simultaneous phase, structure,texture, and stress analyses.

history from rocks

Non-destructivenessIn general, the non-destructive nature of many neutronscattering experiments makes the technique well suited forhandling large, undisturbed samples, and/or rare and uniqueobjects, natural and man-made encompassing areas asdiverse as, for instance: sediment layers, meteorites, andhistorical artefacts .

Non-destructiveness

III. Prospects for advancement in the Earth Sciences(and related fields) with the ESS

Some of the most significant issues in the Earth Sciences arethose related to the prediction of earthquakes and volcaniceruptions. The reliability of the relevant models largelydepends on the knowledge of the physical and chemicalproperties of the materials involved (oceanic crust, uppermantle, continental crust). First and foremost among theseproperties is the role of water in these materials and in thebehaviour of the related magmas.

To draw an effective parallel, one may consider the problem ofweather prediction based upon atmospheric models. It is quiteevident that present day prediction of up to five days was noteven foreseeable two decades ago. As regards to theprediction of earthquakes due to plate subduction, if and whenit becomes possible, it will be entirely dependent on theaccuracy of the models that are currently being developed.

At present we can expect to obtain considerable knowledge inthis direction, in part by the use of neutron scattering

Examples of frontierapplications with an ESS-class neutron source aremostly in the field of insitu studies where mineralstructures and materialbehaviours areinvestigated whilesamples are kept at hightemperature and/or highpressure

98

techniques to study the structure and the properties ofminerals under mantle conditions. The main obstacle beingthe comparatively low flux of existing neutron sources.

Given the availability of an ESS-type neutron source, threeareas of key research activity may be envisaged for a leapforward in the aforementioned direction by the Earth Sciences:

1. In situ measurement of structure-properties relations inmineral phases under high pressure and temperatureconditions representative of the Earth's interior.This mineral physics project would be of great interest tomany fields of research in the areas of mantle rheology,subduction modelling, seismology, tectonophysics, etc.

2. The study of the structure, reactivity and physicalproperties of multi-component melts and fluids underpressure and temperature conditions representative of theEarth's interior.This petrology and mineral chemistry project would be ofgreat interest for: magmatology, volcanology (includingancient and present-day volcanic activity), rocks andminerals genetics, and many other related fields.

3. Texture and stress analysis of polymineralic rocks for thereconstruction of tectonic processes and modelling of rockanisotropies.The project concerns the characterization and theinterpretation of the textural and mechanical properties ofcomplex polyphasic materials, and their evolution duringgeological processes. These processes are oftenanalogous to those occurring during HP/HT industrialprocessing of materials. This project is of interest for:geology, geophysics, petrology, mineralogy, and materialsscience.

Flagship ResearchProgrammes

Some examples of novel representative experimentsWithin the framework of these wide spanning fields ofresearch activity, a number of representative novelexperiments can be proposed also in view of simulations to becarried out in the design stage of the instrumental set-up forthe ESS. Within the described research fields, 4 flagshipexperiments have been selected to show the futureexperimental possibilities.

Representativeexperiments

Study of the pressure-induced spin dynamics and spin-collapse in (FeXMg1-XO) and Fe2SiO4

In a first step, this includes the determination of the magneticand crystal structure under pressure up to 150 GPa. Evenmore advanced would be the determination of magnons underpressure up to 150 GPa. Rationale: The pressure-inducedspin collapse in 3d ions has long been proposed as amechanism for adaptation of simple crystal structures to highpressures. Violations of Hund's rule (spin maximisation) could

High pressure, Fe spin-dynamics in the Earth’smantle

99

reduce the 'volume' of an Fe2+ ion by around 25%. However,only recently has experimental evidence (Mössbauerspectroscopy [10], [3]) shown that this occurs at around 100GPa. The experiments did not allow a detailed understandingof the nature of the phase transition. From a theoretical pointof view, the problem is also very challenging, as currentlyused methods (DFT with generalised gradient approximation)only give semi-quantitative results.

Pressure-induced spin dynamics and spin-collapse in(FeXMg1-XO) and Fe2SiO4

Hence, what is required for this experiment is a study in whichfirstly the crystal and magnetic structure of (FeXMg1-XO) andFe2SiO4 are investigated under pressures up to 150 GPa,using a high neutron flux magnetic powder diffractometer.Then a study of inelastic magnetic scattering on a singlecrystal at the same pressures, using a HET like instrument(high energy chopper and cold chopper spectrometers) toobtain the transition energies between spin levels and (mostdemanding) a constant Q PRISMA like instrument with veryhigh flux to measure magnons at high pressures. A 50 Hz,short pulse, 5 MW source is advised.

FLAGSHIP EXPERIMENT 1

In situ high P/high T and high P/low T neutronspectroscopic investigation of molecular dynamics ofvolatile species (H2O, OH, CO2) in minerals andnanoporous compounds through inelastic and totalscatteringInner surface molecular interactions require studies at variableT and P (0-1500 K, up to 1GPa). High temperatures arerequired for deep Earth's materials, and low temperatures arerequired for Earth’s crust and planetary surface materials. Inaddition, total scattering at high P/high T could be used tostudy minerals under geological conditions. Currently availablefluxes and resolutions limit "molecular neutron spectroscopy"to energy transfers of about 1500 cm-1 and low temperatures.This prohibits the full utilisation of neutron molecularspectroscopy as a complementary tool to Raman and IR-spectroscopy. The need is for spectroscopic measurements ofthe molecular dynamics of H2O in nanoporous, hydrous, andnominally anhydrous (NAM) mineral compounds includingstudies of surface hydration and reactions.To understand the molecule-inner surface interaction typicallyencountered in nanoporous solids, experiments in aheatable/coolable high pressure cell (0-1500°C, up to 1 GPa)are required. Very low temperatures are important tounderstand the transition into the quantum regime (tunnelling),high pressures are important to be able to tune the strength ofthe host-guest interaction.

Molecular dynamics,vibrational spectroscopy,volatiles

High P/high T and high P/low T structural behaviour inmineralsCurrent technology has a foreseeable upper pressure limit ofabout 20 GPa achievable with a Paris-Edinburgh type cell HP/HT structures, mantle-

100

(PE) with heating. At present the max operating pressure is 7GPa and simultaneous P/T measurements can be carried outup to 1500°C (Fig. 2). An increase of the T limit to 2200°C canbe envisaged. Reduction of the sample volume, improvementsin pressure cell technology, and the higher neutron flux of ESSwould allow access to higher pressures more readily, possiblyabove the 150 GPa range.A few pertinent examples are given below.

a) In situ studies of mantle hydrous and nominally anhydrousMg-silicates (alphabet phases, wadsleyite, spinel) atpressure and temperature of the transition zone (~15 GPa;~1500°C) would help characterise the nature of the seismicdiscontinuity known to occur at a depth of 410 Km and thewater budget of subduction zones [11], [18]. The structuralbehaviour of protons in hydrous phases at high P/Tgoverns upper mantle melting, volcanic and earthquakeactivity, although little is known about the effect of pressureand temperature (20 GPa, 1300 K) on the stability or theequilibrium amounts of "water" incorporated in thesephases. This is presently possible but the low H content ofsuch minerals precludes neutron diffraction experimentswith existing sources. A gain of 10 in flux and a further gainof 5 to 10 with improved instruments (diffractometers,detectors, data processing) will make it possible.

b) Cation ordering at high P and T. The thermodynamicconsequences of cation order-disorder as a function of P, Tand time, must be investigated in order to understand thegeophysical and geochemical mechanisms involved.Important phases such as the pyroxenes, olivines andspinels are just beginning to be investigated. A largeamount of work is required to cover the chemical andphysical variants of the corresponding natural phases andthe kinetics of phase transitions encountered.

c) Accurate determination of the structure of pressure-stabilised micro-porous compounds (gas hydrates) and oftheir physico-chemical properties. The kinetics of theirphase transformations are needed to elucidate manyaspects of these poorly understood significant componentsof shallow geological environments in both oceanic andcontinental sediments (Figure 3). Gas clathrates have beenpostulated to be of societal relevance in at least threeways: resource, climate, and hazard.

Kvenvolden (1998) [7] reports on the immediate importance ofsubmarine geo-hazard aspects in considerations of humanactivities and installations subject to the instability of deepwater oceanic sediments (communications cables, oceandrilling rigs, etc.) which may be affected by slope failures,debris flows, slumps, slides and possible tsunamis and,perhaps, also ancient and historical events of global warmingassociated with the release of green-house effect gases. Theneed is for accurate phase diagrams, stability under variableP/T conditions and saline concentrations. Structural studiescan also elucidate the mechanisms responsible for the seismic

core phases, water inminerals, phasestransitions

101

reflectivity attributed to these compounds.

Figure 3: A 12-fold multichannel seismic reflection profile from the crestand eastern flank of the Blake Outer Ridge. The strong BSR(bottom-simulating reflection) is inferred to represent the base ofthe gas hydrate stability zone. (Kvenvolden, 1998 [7]).

In situ diffraction and spectroscopic studies of molecularcomponents in methane clathratesThis experiment may be indicative of the high level ofaccuracy required to investigate fine structural and vibrationaldetails in complex compounds.

Gas hydrates have hydrogen bonded rigid cage structures ofwater molecules entrapping gas molecules. The molecule, forexample methane, has a different symmetry point group withrespect to the cage symmetry and possibly rapid flipping of themolecule over several configurational states occurs. This isreflected both in the spectroscopically observable vibrationalmodes and in the long-range disorder of the moleculesobservable by diffraction, where anharmonic motion may bealso observed.

Both experiments need to be of high quality to yield finestructural details, and they must be carried out under lowT/high P conditions, although in this case the P range is easilyaccessible.

The experiment requires high resolution single crystal andpowder diffractometry for the structural part, and resolution-enhanced TOSCA-like molecular spectroscopy for thevibrational part.A 50 Hz, short pulse, 5 MW source is advised.

FLAGSHIP EXPERIMENT 2

Time-resolved neutron radiography and tomography ofthe behaviour of fluids and melts at HT/HP conditionsImaging based on neutron transmission measurements arepotentially an excellent tool to study macroscopic changesdown to length scales of 5 micrometers with a time resolutionof 1/10th of a second. An example are falling sphereexperiments at high pressures. Currently, synchrotron-basedfalling-sphere experiments for the in situ determination of theviscosity and density of melts are the only method to

Absorption, tomography,radiography, rheology,melts, magmas

102

determine viscosities at very high pressures. Theseexperiments suffer from problems with sample size andconstraints due to the high-pressure cell. From our ownexperience, we know that neutron imaging measurements canbe much more accurate, especially as convection can bemonitored by doping with highly absorbing compounds, suchas Gd-oxide. A set-up which would allow a falling sphereexperiment at pressures up to 5-10 Kbars / 1500°C in a cell ofabout 10 cm height would allow very precise measurements.In addition to a homogeneously illuminated area of about10x10 cm2, this would also need the development of positionsensitive detector with 5 micrometers resolution (current stateof the art is about 250 micrometer). In these dynamicsexperiments the pulse structure and high peak flux of ESS canbe fully exploited by synchronising it with the camera shutter;an advantage over a continuous source.

Neutron tomography is also a promising technique toinvestigate in non-invasive manner the internal structure ofmulti-component systems [8]. Time resolved neutron imagingcould also be used study reaction fronts, such as occur duringcrystallisation of silicates or alloys as well as magma mixingand mingling properties in order to model natural systems inmagma chambers involving the need for high temperature andmoderate pressures.

New techniques such as resonance absorption for precisetemperature measurements and transmission Bragg edgedetection for partially crystalline melts would also have to befurther developed to improve characterisation of samples insitu.

Space- and time-resolved neutron absorptionmeasurements of volatile-containing crystallisingmagmasNeutron imaging and tomography are used to study therheology and the processes (magma mixing, convection, gassegregation, reaction fronts, crystal growth, etc.) occurringduring the high-temperature crystallisation of silicate magmas.Parallel very intense beam is needed on an area of 10x10 cm2

for recording the image of the autoclave vessel. New highresolution position sensitive detectors are needed to resolvefine details of the evolving system. A 50 Hz, short pulse, 5 MWsource is necessary for time-resolved measurements.

FLAGSHIP EXPERIMENT 3

Strain partitioning during the deformation ofpolymineralic rocksOne of the major aims in the Earth Sciences is to refine ourunderstanding of the structure and composition of the Earth'sinterior using seismological data. The interpretation of thisdata is heavily reliant upon laboratory measurements of theelastic properties of the relevant rock types [6]. These rock-types are essentially polymineralic aggregates with a range incomposition and microstructure that is far too large for it to befeasible to determine the properties of every rock-typeindividually.

In situ measurements ofstress and strainpartitioning during rockdeformation

103

Consequently, it makes sense to seek a method that willpermit the elastic properties of polymineralic materials to bespecified in terms of the elastic properties of the componentphases. In order to formulate mechanical equations of statefor polymineralic materials that are based on some function ofthe properties of the component phases, it is necessary toestablish the relative contribution of each phase to theaggregate's properties. However, in rock deformationexperiments it is usually only aggregate properties that can bemeasured. There is a glaring dearth of experimental evidenceas to what happens to the strain partitioning between thecomponent phases in a composite in the elastic regime, andalso when one or more of the phases starts to yield plastically.Once yielding occurs there is the possibility of load transferbetween the phases, and the extent to which this occursexerts a profound influence on aggregate properties (Figure4).

Halite Microstrain

0 500 1000 1500

Calcite

Microstrain

0

500

1000

1500

60:4030:7070:3050:50

cc:hl

homogeneous strain

homoge

neous s

tress

HS+

HS-

Figure 4: Calcite and halite axial elastic strains at different applied loads.Also shown are the predicted phase strains assuminghomogeneous stress, homogeneous strain, and as given by theupper (HS+) and lower (HS-) Hashin-Shtrikman bounds on thecomposite properties. The dashed line shows the trend in thedata (calcite ≥ 50%) and the arrow shows the calcite strain at theelastic limit of calcite.The elastic limit of the halite was at a halite strain of about350µstrain. The elastic strain partitioning between the two phaseswas unaffected by the yielding of the halite. However, above acalcite elastic strain of about 550µstrain, the strain partitioningbetween the two phases started to tend towards homogeneouselastic strain, a condition that was attained by a total(elastic+plastic) strain of about 1%. The curve describing theelastic strain partitioning between the calcite and halite wasindependent of composition. The change in elastic strainpartitioning at a calcite strain of 550µstrain corresponds to thepoint at which the elastic limit of calcite is attained.

Neutron diffraction experiments at spallation sources,conducted on samples held under differential load in the

104

neutron beam line offer a solution to these two problems. Bydetermining the change in lattice parameters of eachcomponent phase as a function of applied load, the elasticstrain of each phase, and hence its contribution to the totaldeformation, may be ascertained [4]. These experiments areonly feasible due to the penetrating and polychromatic natureof the neutrons produced at spallation sources. Futuredevelopments foreseeable at the ESS would be an improvedstrain resolution from lower counting times, the developmentof high temperature equipment, and smaller beam sizesenabling greater spatial resolution. Such developments couldbe a large step beyond that available at ENGIN-X of ISIS.

Deformation mechanisms in polymineralic rocksDue to the high penetration capability of neutron volumetexture investigations can be performed on rather large naturalspecimens up to about 10 cm3. Such volumes are necessaryto ensure sufficient grain statistics for all mineral constituentsin polymineralic rocks especially with coarse grain sizes (e.g.about 1mm) [9].

Large suites of geological samples from different locationsmust be investigated to study mineral specific texturalchanges under different deformation conditions [16], or fordifferent minerals under similar deformation conditions in orderto understand the underlying deformation mechanisms [12]. Ahigh intensity source and a large beam are needed for texturalinvestigations on large samples suites in reasonable timescales. Time-of-flight techniques are essential because polefigure measurements can be performed without any samplescanning, and because simultaneous structure and texturerefinements become possible.

Comparative texturestudies of complexsystems, time evolution ofrock texture

Influence of stress and development of texture upondeforming geomaterialsLarge scale deformation of crustal and mantle materialsgenerates the development of microstructures, involvingtwinning, phase transitions, and mineral structural and texturaltransformations. To explore the influence of differential stressand strain partitioning on plastically deforming polymineralicmaterials and the development of textures, requires HT/HPconditions to simulate geological processes.

Texture analysis without sample rotation requires large banksof detectors (HIPPO-type instrument) and accurate detectionof lattice variations in low-symmetry materials requires high-resolution diffractometry (resolution-enhanced ENGIN-Xinstrument, with ample sample space). Preferably, in certaincases the two experiments should be performedsimultaneously, in order to follow the complete evolution of thesample. A 50 Hz, short pulse, 5 MW source is advised.

FLAGSHIP EXPERIMENT 4

Cultural heritagePhase and Texture analysis of natural materials such as stone

105

and ceramics, is still relatively new but the potentialapplication of such powerful techniques span many fields ofinterest within archaeological research, from standard finger-printing to complex conservation problems. Owing to the non-destructive character of the techniques, their applicability tolarge, undisturbed objects and large volume of interaction asopposed to the surface analysis of X-rays, neutron scatteringtechniques are easily predicted to find many new applicationsin the fields of study and conservation of historical artefacts.An example of intervention guided by such studies is given inFigure 5.

Figure 5: Foligno Cathedral; limestones, marbles and travertine (before andafter restoration).

Fingerprinting often helps identify the actual quarries utilised inhistorical times and hence discover the original source ofsimple Earth resources and thus investigate pre-historictrading and material exchange.

In the non-diffractive mode, information on the inner fabric oflarge scale materials and artefacts (a few µm to several cm)which is beyond the reach of X-rays, can be obtained bymaking use of recently developed neutron detectors whichlend themselves to neutron imaging and tomographicreconstruction. Applications of this technique to archaeologicalartefacts are already envisaged; the availability of improvedinstrumentation, especially in terms of detector capabilities,would definitely represent a major improvement for this area ofresearch.

Archaeological materials,conservation, non-destructive analysis

References

[1] M. Catti, G. Ferraris, S. Hull, A. Pavese, Static compression and H-disorder in brucite, Mg(OH)2, to 11 GPa– a powder neutron diffraction study. Phys. Chem. Minerals 22, 200-206 (1995)

[2] S.L. Chaplot, N. Choudhury, S. Ghose, M.N. Rao, R. Mittal, K.N. and Prebhatasree, Inelastic neutronscattering and lattice dynamics of minerals. Eur. J. Mineral, accepted (2001)

[3] Cohen, Mazina and Isaak, Magnetic collapse in transition metal oxides at high pressure: implications for theEarth, Science 275, 654-657 (1997)

[4] Covey-Crump, P.F. Schofield, Stretton, Strain partitioning during the elastic deformation of olivine +magnesiowustite aggregates, Geophysics. Rev. Lett., submitted (2001)

[5] M.T. Dove, M.S. Craig, D.A. Keen, W.G. Marshall, S.A.T Redfern, K.O. Trachenko, M.G. Tucke, Crystalstructure of the high-pressure monoclinic phase-II of cristobalite, SiO2, Mineralogical Magazine 64, 569–576(2000)

106

[6] Khazanedhari, Rutter, Brodie, High pressure- high temperature seismic velocity structure of the mid-crustaland lower crustal rocks of the Ivrea-Verbano zone and Serie dei Laghi, NW Italy. J. Geophys. Res. 105,13843-13858 (2000)

[7] K.A. Kvenvolden, Potential effects of gas hydrate on human welfare, PNAS proceedings on coll. "Geology,Mineralogy, and Human Welfare," Nov. 8-9, 1998

[8] Lehmann, Pleinert, Williams, Pralong: Application of new radiation detection techniques at the Paul ScherrerInstitut, especially at the spallation source. Nucl Inst. Meth Phys. Res. A424, 158-164 (1999)

[9] B. Leiss, K. Ullemeyer, K. Weber, Texture and physical properties of rocks. J. Struct. Geology. (Spec. Issue)22, no. 11/12 (2000)

[10] Pasternak, Taylor, Janeloz, Li, Nguyen and McCammon, High pressure collapse of magnetism in Fe.94O:Mössbauer spectroscopy beyond 100 GPa. Phys. Rev. Lett. 79/25, 5046-5049 (1997)

[11] S.M. Peacock, Fluid Processes in Subduction Zones. Science 248, 329-337 (1990)

[12] J. Pleuger, E. Jansen, W. Schäfer, N. Oesterling, N. Froitzheim, Neutron texture study of natural gneissmylonites affected by two phases of deformation. (submitted to ICNS2001)

[13] S.A.T. Redfern, G. Artioli, R. Rinaldi, C.M.B. Henderson, K.S. Knight, B.J. Wood, Octahedral cation orderingin olivine at high temperature. II: an in situ neutron powder diffraction study on synthetic MgFeSiO4 (Fa50),Phys. Chem. Min. 27, 630-637 (2000)

[14] R. Rinaldi, G. Artioli, C.C. Wilson, G. McIntyre, Octahedral cation ordering in olivine at high temperature. I:In situ single crystal neutron diffraction on natural mantle olivines (Fa12 and Fa10), Phys. Chem. Min. 27,623-629 (2000)

[15] W. Schäfer, Neutron diffraction applied to geological texture and residual stress analysis. Eur. J. Mineral.,accepted (2001)

[16] H. Siemes, B. Klingenberg, G. Dresen, E. Rybacki, M. Naumann, W. Schäfer, E. Jansen, Strength, textureand microstructure of hematite ores experimentally deformed in the temperature range 600 to 1000 C andstrain rates between 10-4 and 10-6 s-1, submitted (2001)

[17] B. Winkler, K. Knorr, A. Kahle, P. Vontobel, E. Lehman, B. Hennion, G. Bayon, Neutron imaging andneutron tomography as non-destructive tools to study bulk rock samples. Eur. J. Mineral., accepted (2001)

[18] B.J. Wood The Effect of H2O on the 410-Kilometer Seismic Discontinuity, Science 268, 74-76 (1995)

107

7.7 Liquids and Glasses

R.L. McGreevy1, T.C. Hansen2, A.K. Soper3, J.-B. Suck4 and R. Vacher5

1Studsvik Neutron Research Laboratory, Uppsala University, S-611 82 Nyköping, Sweden2Institut Laue-Langevin, BP 156, F-38042 Grenoble Cédex 9, France3ISIS, Rutherford Appleton Laboratory, Chilton, Oxon OX11 0QX, United Kingdom4Technical University Chemnitz, Institute of Physics, Materials Research and Liquids, D-09107 Chemnitz, Germany5Laboratoire des Verres, Université de Montpellier II, CNRS, F-34095 Montpellier, Cedex 5, France

AbstractNeutron scattering is a key experimental technique in the study of the atomic structure and dynamics of liquidsand glasses. Neutrons at ESS will be used as the central part of studies using multiple complementarytechniques, e.g. X-rays, light scattering and NMR, each providing information on specific aspects of the structureor dynamics of complex disordered materials. The data obtained will be simultaneously analysed withsophisticated modelling techniques or used as a stringent test of computer simulations. Such a coherentapproach will not only enable a radical step forward in our understanding of the basic physics of disorderedmaterials, but also in our ability to understand, control and eventually exploit the atomic scale structure anddynamics for the production of materials with optimised properties for technological and other applications.

Several 'flagship' areas of research can already be identified where major progress will be possible at ESS.Inelastic small angle neutron scattering, or 'neutron Brillouin scattering', will provide key information on, forexample, the transition from classical hydrodynamics to generalised hydrodynamics in liquids, or from extended tolocalised modes in glasses, or on the origin of the 'Boson peak'. Isotopic substitution and polarisation analysisshould become routinely used in inelastic scattering experiments to enable the determination of partial dynamicalstructure factors; this is nowadays only generally possible for structural work. Wide ranging studies of thestructural effects of different ions and molecules at different concentrations will be carried out in solutions, eitheraqueous or non-aqueous. There will be similar studies to explain the composition dependence of phase stabilityor ionic conductivity in complex materials for battery electrodes or electrolytes, fuel cells or sensors. Isotopicsubstitution, particularly H/D substitution, will also play an important role here. Studies will be made underextreme conditions of temperature and pressure, to mimic conditions deep in the core of the earth or otherplanets. Some of these may use pulse techniques or be 'single shot' experiments, taking advantage of the timestructure of the ESS neutron beam.

One of the strengths of neutrons for studies of liquids and glasses is the wide coverage of energy (ηω) andmomentum (ηω) transfer, so instrument and target options will be chosen to use this to fullest advantage. Themajority of instruments are best suited to a 50 Hz short pulse target. The 16 Hz long pulse target is the choice in afew cases and may also offer novel opportunities. A wide range (0.1 < Q < 600 nm-1) liquid/glass diffractometer isrequired for structural studies, together with a small angle scattering instrument whose range overlaps and thenextends to lower Q. A diffractometer optimised specifically for total scattering studies of structural disorder incrystalline materials should be considered. This is likely to be a growth area in the next decade. For dynamicalstudies a range of spectrometers is required providing as wide as possible a coverage of (Q,ω) space withparticular emphasis on the small Q, high ω, region that can only be accessed using detectors at small angles.While this region can be accessed more easily nowadays using X-rays it must be emphasised that, except in thecase of single component systems (i.e. elements), the information obtained is complementary and notcompetitive.

I. IntroductionDisordered materials play a central part in our daily life. Watercovers two thirds of the earths surface and is the majorcomponent of our bodies. Glasses are in our windows, inoptical fibres for communications and even eaten as candy orused as stable coatings on medicines. Ionic conductors are inbatteries in cars (electric cars in the future), mobile telephonesand computers. Yet our understanding of such materials,especially in relation to more ordered crystalline materials, isstill very limited.

Disordered materials arecommon in daily life;water, glasses, opticalfibres,batteries

The particular properties of the neutron make it a key probefor the study of liquids and glasses:• The ability to cover a large area of momentum/energy

transfer space (Q and ω), i.e. length and time scales, welladapted to the length scale of the atomic structure and the

The neutron is a key probefor the study of liquids andglasses: (Q,ωωωω) range,direct link to computer

108

adapted to the length scale of the atomic structure and theenergies of elementary excitations in liquids and glasses.Because liquids and glasses lack symmetry it is importantto cover as large an area as possible.

• Due to the simplicity of the coupling function (it is aconstant), the scattering can be measured on an absolutescale and directly related to the results of computersimulation/modelling, which considerably improves thepossibilities for data interpretation. Hence the importanceof information from neutron scattering increases ascomputing power increases.

• The isotope dependence of the scattering cross-sectioncan be used in a unique way to give detailed, elementspecific, information.

• The high penetrating power enables studies of materials incontainers or complex environments which is aconsiderable advantage, especially in the investigation ofliquids at high temperature.

• A magnetic moment, but no charge, enables detailedstudies of magnetic structure and dynamics and allowspolarisation analysis to isolate information on thecorrelations of single atoms.

simulation, isotopicsubstitution, highpenetrating power,magnetic moment

Other techniques, e.g. X-ray scattering, light scattering andNMR, can provide specific information and over a wider rangeof either Q or ω, but not both. Even when the range overlapsthe information obtained is nearly always additional to thatfrom neutrons, not the same. As the trend is towards the studyof more and more complex systems we envisage thatneutrons at ESS will be used routinely as the centralcomponent in a study using multiple complementarytechniques. Data will either be analysed simultaneously usingsophisticated modelling techniques, or used as stringent testsof computer simulations. Such a coherent approach to thestudies of liquids and glasses not only enables completeinformation from experiments, but also a detailedinterpretation of the data.

Neutrons at ESS will beused routinely as thecentral component in astudy using multiplecomplementarytechniques (X-rays, lightscattering, NMR etc.) withdata being analysedsimultaneously usingsophisticated modellingtechniques, or used asstringent tests ofcomputer simulations

II. Scientific directions and opportunities at ESS

Fundamental studies of the atomic dynamics ofdisordered matter.The determination of partial structure factors andtransformation to r-space to give the radial distribution functionwas the essential step forward which nowadays enables adetailed understanding of the atomic structure of disorderedmatter. ESS will give us the opportunity to take a similaressential step towards understanding the atomic dynamics.This is considerably more difficult because it requires anaccurate determination of the dynamic structure factor,S(Q,ω), in a very wide (Q,ω) range, with the additionalcomplication of isotopic substitution and possibly also ofpolarisation analysis, to obtain information on the dynamics ofindividual atom types in multi-component systems.Experiments of the completeness required here cannot be

In order to understand thedynamics of liquids andglasses at the same levelwe now understand thestructure will require athird generation neutronsource

109

performed at present sources.

This will open the way for a completely new approach tounderstanding the dynamics of disordered systems. Afterdouble Fourier transformation of S(Q,ω) to obtain G(r,t), whichis not possible now because of the limited range and statisticalaccuracy that can be obtained, we will be able to interpret thedynamics in real space and time. In addition there will be thepossibility for a direct comparison with molecular dynamicssimulations, or for the refinement of dynamical models in aniterative manner, as is routinely done nowadays for structuralmodels.

Analysis will be by directcomparison withmolecular dynamicssimulations or by iterativerefinement of dynamicalmodels

The collective excitations (i.e. phonons in harmonic crystals)carry essential information on the interatomic interactions. Forsuch investigations the Q range below the main peak in thestatic structure factor is the most important. There the staticand dynamic properties are most sensitive to the complicated,attractive part of the interaction. The ESS will give us thechance to investigate this region of dynamical space, which isvery difficult to access, with sufficient intensity and resolutionand over a sufficient energy range - in favourable cases evendown into the Q region where pure hydrodynamics applies.This is one of the two firm and basic dynamical theoriesapplicable to disordered systems. We will then have thepossibility to study the transition from this continuum theory togeneralized hydrodynamics where the parameters of thetheory become functions of Q and ω (in an unknown manner).Studies of this kind also have to be extended to the partialdynamical structure factors, presently possible only in specialcases. The partial collective dynamics, which depend on theelement in focus, are expected to be rather different from thesum over all partials contained in the total dynamic structurefactor.

The dynamics can befollowed from the wellunderstood region wherepure hydrodynamicsapplies into the poorlyunderstood region ofgeneralisedhydrodynamics

0,00,2

0,4

0,6

0,8

0,4 0,6 0,81,0

1,2 Q/nm

-1

ω/ps-1Figure 1: Dynamical structure factor S(Q,w) for liquid 36Ar at T = 300 K and

P = 200 bar, measured using neutron Brillouin scattering [1].

110

Another step towards a full understanding of the details of thedynamics will be the determination of the shapes and widthsof the inelastic ("Brillouin") and the central ("Rayleigh") line.These two quantities contain essential information on theprocesses leading to the decay of collective (in phase) atomicmotions. Up to now collective excitations in classical liquidshave never been observed as maxima or shoulders above ∼ 3nm-1, except for liquids with metallic pair interactions(screened Coulomb). In amorphous systems they contain inaddition information on the mixing of modes, which fromtheoretical investigations are expected to exist as pure modes(plane wave like) only in the very low Q region and not aboveabout 3 nm-1. Extremely reliable and specialized inelasticsmall angle scattering spectrometers are needed for this kindof investigation, sometimes known as neutron BrillouinScattering.

Specialised inelastic smallangle neutron scatteringinstruments will berequired

ESS will also allow us to address one of the unsolved"mysteries" in the dynamics of amorphous solids and defectcrystals: two level systems (TLS) or tunnelling states. Theseexcitations seem to have a broad distribution at rather lowenergies. Their density is very low (10-4 to 10-5) and thusinaccessible with present neutron sources, but they stronglyinfluence the thermal properties, especially below 1 K, theheat conduction around 10K (most likely causing the plateauregion characteristic of amorphous solids) and other materialproperties. To study TLS we will need a very strong neutronsource, isotopic substitution (to determine what is tunnelling),very high resolution (< 100 µeV) and low temperatures.

Two level systems, anunsolved "mystery" of thedynamics of glasses, canbe studied at ESS

Figure 2: S(Q,ω) for vitreous SiO2 measured on the MARI spectrometer atthe ISIS spallation neutron source, illustrating the wide range of(Q,ω) space that can be covered [2]. Important information is stillmissing in the region on the left hand side of the diagram.

Chemistry and life sciencesIn the past 25 years neutron diffraction investigations of theatomic structure of chemical liquids have assumed an

Absolute diffractionintensities and isotopic

111

increasingly pivotal role. A broad range of experimental andtheoretical techniques have been applied extensively to thesesystems, including several spectroscopies, X-ray and neutrondiffraction, EXAFS (X-ray absorption fine structure), and bothclassical and “first principles” computer simulations. Neutronsoften provide the fundamental benchmark data with which tocompare and sometimes normalise other data. Two particularqualities of the neutron have caused this revolution. Its simple,point-like, interaction with the scattering nucleus enablesabsolute diffraction intensities to be extracted, and the factthat different isotopes of some of the most relevant atoms(especially hydrogen) have sufficient contrast in theirrespective neutron scattering amplitudes to permit isotopiclabelling of those atoms within a complex fluid.

substitution are the keyfeatures of neutronscattering for the study ofsolutions

The earliest examples of this work were studies by Enderbyand coworkers of the hydration of ions in aqueous solution inthe 1970’s. They used the technique of isotope labelling tomap out the coordination structure around individual ions. Forthe first time it became possible to provide quantitativeanswers to the questions of, for example, how many watermolecules there are coordinating particular ions. Othertechniques, such as X-ray diffraction and computer simulation,had failed to provide unambiguous answers. The sameneutron data also provided early information about the likelyorientation of water molecules around specific anions andcations. However, the fundamental question of how ions areorganised in aqueous solution is still unresolved. Are they“charge ordered” as in a molten salt?

Ion-ion correlations insolution are still not wellunderstood

Later during the 1980’s it was established thathydrogen/deuterium isotope substitution could be used toprovide detailed information about water structure, initially inthe pure state, but also during the 1990’s, in both solutionsand liquid mixtures. It was demonstrated that the neutroncould be used to map out the fundamental correlations of thehydrogen atom on one molecule with atoms on neighbouringmolecules, thus providing a direct and quantitative probe of,among other things, the hydrogen bond.

Neutrons can be used tostudy hydrogen bondingin pure water andsolutions

The next step has been made only recently with the realisationthat neutron data with suitable isotope contrasts, particularlyfor dilute species if feasible and perhaps also combined withdata from other techniques such as X-ray scattering and NMR,can be used to build realistic 3-dimensional models of liquidsof interest, in much the same way as has been done for therefinement of crystal structures for many years. This processof liquid structure refinement enables the experimenter to mapout in unprecedented detail the way water molecules, ions andother molecular entities are arranged. The methods developedhave been applied to aqueous and non-aqueous systemsalike, and a variety of molecular liquids have been tackledsuccessfully [3].

Nowadays realistic threedimensional models ofatoms and molecules inliquids can be built basedon neutron diffraction andother data

112

Figure 3: Isodensity surface plot showing the coordination of methanolmolecules around a central molecule [4].

Against this background the ESS presents an excitingopportunity to extend this work in different directions. What isnow clear is that each ion or other molecular entity in solutionhas a subtly different effect on the solvent structure. Someions, such as sodium, cause profound structural changes tothe solvent itself, but others apparently have little effect. Thebizarre fact is that some ion combinations or molecularspecies in solution can induce protein folding, while otherscause protein denaturation. There is now little doubt that animportant factor in this behaviour arises from the effect theions or other molecular species have on the structure of thewater surrounding the macromolecule. It should beemphasised here that structural information at the level ofdetail that is attainable with neutron diffraction on thesesystems could not be achieved by other techniques, althoughthose other methods can often provide importantcomplementary data. The ESS will enhance this capacity evenfurther by allowing the aqueous environment of largermolecular entities, of interest to the chemistry and biologycommunities, to be examined for the first time.

ESS will give an excitingopportunity to study howions and dissolvedmolecules affect thestructure of the wateraround them and have aninfluence on biologicalfunction, for exampleprotein folding

The type of study envisaged here is therefore a wide-rangingstudy of the effect of ions and other molecular entities onsolvent structure. The parameter space to be explored isenormous, involving at least 100 different combinations of ionpairs in solution at several key concentrations. There is a largenumber of other molecular entities in solution which need tobe studied at the same time, both on their own in solution andin the presence of ions. Only by performing this systematicneutron study of an appropriate sample of the relevantsolutions will the characteristic structural trends with differentions and molecules in solution be identified. With currentfacilities the complete exploration of the dissolved speciesphase space (type and concentration) would take very manyyears to complete, particularly at the lowest soluteconcentrations where some experiments are hardly feasible atpresent.

An enormous parameterspace must be covered ina systematic study. This isnot feasible with currentfacilities

A further stage of this work would be the exploration of theaqueous environment of large molecules in solution in thepresence of different ions and molecular species, via isotope

Special facilities for thepreparation of isotopicallylabelled samples will

113

labelling of suitably synthesised macromolecules. The isotopecontrast available for such work is likely to be low, and willrequire extremely stable instrumentation as well as the highestneutron fluxes available. It will also require dedicated samplepreparation facilities, including the production of isotopicallylabelled molecules.

enhance the power of ESS

The methods described above are not restricted to aqueoussystems and a variety of other molecular systems can betackled. A lot of chemical processes take place in much morecomplex non-aqueous fluids. Some examples are roomtemperature electrolytes, catalytic fluids used inpolymerisation processes and “Lewis acids”. All of thesesystems can be explored at the atomic level with the methodsdescribed, in suitable cases before, during and after achemical or physical change has taken place (if a real timeexperiment is possible). The data obtained will provide theessential information for a detailed understanding of howmolecules interact in solution.

More complex non-aqueous fluids can also bestudied, or molecules asthey interact

The main feature of these liquids, that will enable the ESS torevolutionise the way we look at them, is their complex nature– several different molecular species mixed together, some atlow concentrations. Isotope or other labelling experiments willbe required to identify different components in the fluid. Inthese cases it is likely that combining the neutron data withthat available from a variety of other probes will be essential,since none of the available probes can identify all species withequal effectiveness. Obviously the neutron probe is mostrelevant when isotope substitution is available, but evenwithout this the neutron provides important new datacomplementary to that available by other methods. A featureof this kind of work would be the development of a chemicalreaction cell, with the ability to probe particular samples with avariety of radiations and techniques in addition to neutronscattering.

For studying suchcomplex fluids manyexperimental probes willbe needed, but the neutronremains cruciallyimportant

Extreme conditions, kinetics, melting and freezing,crystallisation and order-disorder transitionsStructural and dynamical investigations under extremeconditions will become possible at ESS thanks to theincreased count rate, allowing for smaller samples (as usedfor high pressure studies) and faster data acquisition ifrequired. If these extreme conditions are only present for avery short time, e.g. extreme temperature and pressure in ashockwave experiment, the diffracted neutrons from individualpulses could be used. In addition the instrument design of atime-of-flight diffractometer on a pulsed spallation source hasintrinsic advantages for the layout of complicated sampleenvironment, particularly pressure cells. Higher pressures,produced either statically or dynamically (shockwaves), areneeded in order to understand for example the fluids whichexist in the core of the earth and other planets at pressureswhich are up to now inaccessible. Even metallic hydrogenmight become observable this way.

Systems under extremeconditions, such as veryhigh temperature andpressure, can be studied,possibly in shockwaveexperiments

114

4060

80100

120

2 Theta / °

d20 25-Apr-98 09:36:47 UserDahlborg, Fedotov L.C.Hansen 62612:63146*d20_1998_04_058774.241

250

200

150

ln(cou

nts

/1min) 7.0

6.5

6.0

5.5

temperature / K

GaSb amorphisation

amorphous phase, room temperature

metastablecrystalline

phase (under liquid nitrogen)

Time/min

Coun

ts/4

min/0

.1°

2Theta/°140

2.6

400

250000

0

0

Figure 4: Amorphisation of crystalline GaSb [5] and crystallisation ofamorphous YFe [6] studied by neutron thermodiffractometry

The preparation of bulk amorphous materials, e.g. byannealing a quenched, crystalline high-pressure phase [4],can be investigated by neutron diffraction, averaging over thebulk sample. These materials can serve as precursors formaterials with tailored properties, e.g. nano-crystallinemagnetic materials. Alternatively the crystallisation processesof amorphous precursors can be followed in-situ in real timeby neutron diffraction [5]. Powder pattern line profilebroadening analysis, which requires a good resolutiondiffractometer, can provide information on the evolution ofgrain size. Time resolved small angle neutron scattering canprovide information about the formation of grains. However theamorphisation/crystallisation rates that can be followed arelimited by the power of current sources.

Crystallisation andamorphisation can bestudied kinetically

With ESS the investigation of phase transitions (e.g. the liquid-glass transition) will be a matter of interest not only underequilibrium conditions, but also under non-equilibriumconditions in real time, i.e. the kinetics. Topics of study includephase separations in supercritical systems - in particular thesolidification of super-cooled liquids - solution of solutes in asolvent, migration of liquids in porous media, ion migration inglasses, or glass formation using sol-gel processes. In somecases, e.g. fluid systems or migration in an electric field, theprocess to be studied is reversible and reproducible in a cyclicmanner, allowing for repetitive, stroboscopic, data acquisitionin order to achieve the required count rate. Performing suchexperiments at ESS will be an advantage where cycling is

Strobosocopicmeasurements of kineticprocesses will be possibledown to 10 µs, an intrinsiclimit

115

difficult to sustain over longer times. One can even envisagecomplex measurements where one utilises thetime/wavelength structure of the pulse; the time resolution thatcan be achieved is then intrinsically limited by the path lengthof the neutron in the sample (of order 10 µs for thermalneutrons).

In 'single shot' kinetic experiments a gain in intensity meansdirectly a gain in time resolution. The limit on ESS will be setby the frequency of 50 Hz, i.e. 20 ms. However, even withESS we are unlikely to reach such a short time scale whenstudying the structure of disordered systems, which require amuch higher count rate than in the case of powder diffractionof crystalline materials.

'One shot' measurementsof kinetic processes arelimited to 20 ms by thesource

Materials scienceMaterials science is increasingly concerned with a detailedunderstanding of the structure and dynamics of materials on amicroscopic level and its relation to macroscopic propertiesand function, with the eventual aim of being able to designnew or optimised materials on an atomic basis. Here we havechosen three examples concerning disordered materials,where the application of neutrons from ESS will play a keyrole. Brief demonstration measurements are possible now, butthe large number of components means that the wide rangingstudies required for a detailed understanding can only becarried out at ESS.

ESS can play a key role inproviding information atthe atomic level to aidmaterials design

Ultra-soft or ultra-hard magnetic metallic glasses can beprepared with direct technological applications (e.g.transformer cores, miniature induction coils in cars) or used asprecursors to obtain similar nanocrystalline materials. Foroptimisation we need to understand the structure, both atomicand magnetic, of systems with typically at least three atomicspecies. This will require the combination of neutron scattering(possibly including polarisation analysis to separate themagnetic contribution to the scattering), X-ray scattering,EXAFS and computer modelling for measurements on multiplesamples covering a three component phase diagram.

Ultra-soft or ultra–hardmagnetic metallic glassesor nanocrystallinematerials

Ion conductors have an enormous range of applications indevices for energy storage (e.g. batteries), energy production(e.g. fuel cells), sensors and smart windows. These areincreasingly important in modern society, with the demand forhigh efficiency, small size and weight, environmentalfriendliness and safe operation. Conflicting requirements leadto complex solutions, often without any real understanding.For example one needs to understand the apparentcompetition between favourable ionic conductivity andmechanical properties in polymer electrolytes, or the effectscaused by local atomic correlations in hydrogen storagematerials. This requires structural measurements ofmulticomponent systems and dynamical measurements of thediffusion of dilute ionic species or the relaxation of polymersover a wide time scale.

Batteries and fuel cells,sensors and smartwindows

116

Figure 5: Model of conduction pathways in a Ag based fast ion conductingglass, determined from the combination of neutron and X-raydiffraction data and reverse Monte Carlo modelling [7].

The requirements for optical fibres, waveguides, opticalamplification etc. can place extremely stringent requirementson the precise glass composition on an atomic scale. Tocontrol the transmission in a particular wavelength band, forexample, we may need to understand the factors thatdetermine the bonding of individual rare earth ions, or clustersof such ions, in a multi-component glass structure. These areusually at low concentrations and hence best studied withEXAFS. However the longer range correlations and thestructure of the host glass must be determined by thecombination of neutron and X-ray diffraction. Similar studiesare required for the characterisation of glasses used for longterm radioactive waste storage.

Waveguides, opticalfibres, opticalamplification

Techniques similar to those used to study the structures ofliquids and glasses can also be used to study structuraldisorder and local structural correlations in crystallinematerials. These are sometimes known as 'total scattering'studies to distinguish them from the elastic scattering studiesof conventional crystallography. This is a relatively new fieldand as yet no diffractometer has been built which isspecifically optimised for such work. The requirements arestringent; better resolution than a typical liquids diffractometer,but also better count rate and control of background since thediffuse scattering can be much weaker. ESS would offerconsiderable opportunities in this area. There are manypossible applications to technologically important materialssuch as high temperature superconductors, colossalmagnetoresistance (CMR) materials and ionic conductors.

Similar methods can beused to study disorder incrystalline materials, forexample high temperaturesuperconductors andcolossalmagnetoresistancematerials

117

-0.9

-0.6

-0.3

0.00.30.60.9

3.43.6

3.84.0

4.24.4

r / Å

cos

θ

Figure 6: Angular correlation of neighbouring magnetic moments as afunction of their separation (i.e. local spin-lattice correlations) forthe CMR material La0.8Sr0.2MnO3, determined by total neutronscattering [8].

III. Specific requirements for instruments and targetstations

High quality research on liquids and glasses requiresinstruments that are optimised for this purpose. In additionspecific, fast, software must be considered as an integral partof every instrument.

Diffraction studies of liquids and glasses, or total scatteringstudies of disorder in crystals, require very high precision inintensity measurement, so it is preferable that they areseparated from 'conventional' crystallographic studies, even ifthe instruments overlap in some of their specifications. Onlythis will assure the meticulous control of background andstability that is required, but is less important for most purelycrystallographic experiments.

High precisionmeasurement of diffractedintensity is crucial

One key point for inelastic scattering studies of liquids andglasses is the provision of small angle scattering detectors,with evacuated flight paths to reduce background scattering.Such detectors are vital if full advantage is to be taken of thepossibilities of ESS.

Most of the required - diffractometers and chopperspectrometers - are best placed on a short-pulse 50 Hz target.Only a few could work, with some decrease in performance,on a 10 Hz target. The liquids diffractometer could profit froma 25 Hz target, but this is not planned. However severalinstruments could benefit from a methane moderator on the 50Hz target, so the possibilities for this should be furtherinvestigated. Some instruments, such as the wide angleneutron spin echo spectrometer needed to complement the(Q,ω) coverage of the chopper and backscatteringspectrometers, and the SANS diffractometer, are best placedon the 16 Hz long pulse target.

50 Hz short pulse and 16Hz long pulse targets arethe preferred options

118

References

[1] U. Bafile, P. Verkerk, F. Barocchi, L.A. de Graaf, J.-B. Suck and H. Mutka, Phys. Rev. Lett. 65 2394 (1990)

[2] M. Nakamura, M. Arai, T. Otomo, Y. Inamura, S.M. Bennington, J. Non Cryst. Materials, in press (2001)

[3] A.K. Soper, Physica B 276-278 12 (2000)

[4] T. Yamaguchi, K. Hidaka, and A.K. Soper, Mol Phys. 96 1159 (1999)

[5] V.E. Antonov, O.I. Barkalov, M. Calvo-Dahlborg, U. Dahlborg, V.F. Fedotov, A.I. Harkunov, T. Hansen,E.G. Ponyatovsky, M. Winzenick, High Pressure Research 17, 261 (2000)

[6] S. Kilcoyne, P. Manuel, and C. Ritter, J. Phys.: Cond. Matter submitted (2001); ILL Annual Report 42 (1998)

[7] St. Adams and J. Swenson, Phys. Rev. Lett. 84 4144 (2000), Phys. Rev. B 63 (2001)

[8] A. Mellergård, R.L. McGreevy and S. Eriksson, J. Phys.: Cond. Matter 12 4975 (2000)

119

7.8 Fundamental Neutron Physics

H. Abele1, J. Byrne2, D. Dubbers3, W. Fischer4, M. Pendlebury2, H. Rauch5, J. Sromicki6

1Physikalisches Institut der Universität Heidelberg, Philosophenweg 12, D-69120 Heidelberg, Germany2School of Chemistry, Physics and Environmental Science, University of Sussex, Falmer, Brighton BN1 9RH, UK3Institut Laue-Langevin, Avenue des Martyrs, F-38042 Grenoble Cedex 9, France4Condensed Matter Research with Neutrons, Paul Scherrer Institut Villigen, CH-5232 Villigen PSI, Switzerland5Atominstitut der Österreichischen Universitäten, Schüttelstr. 115, A-1020 Wien, Austria6Institut für Teilchenphysik, ETH Zürich, CH-8093 Zürich, Switzerland

AbstractNeutrons appear both as composite particles and as quantum waves. Both features have been investigated withthermal, cold and ultracold neutrons at many neutron sources. The higher intensity and the pulse structure of theESS provide new possibilities for fundamental neutron physics experiments. The questions of small right-handedcontributions to our left-handed world, why is there much more matter than antimatter in the universe and howneutrons interact with each other can be tackled. Non-classical neutron states can be produced and used fornovel fundamental quantum optics investigations. Intensity gains of ultracold neutrons in the order of 1000 can beanticipated at a new ultracold neutron target station where the use of a spallation process for the production ofneutrons becomes especially obvious.

I. IntroductionNeutrons are known as a powerful tool for particle and nuclearphysics and they are ideal probes for quantum opticsinvestigations. The European Spallation Source is, therefore,of intense interest for fundamental studies in these fields.

During the past 25 years, our world-view of nature haschanged dramatically, ranging from the constituents ofelementary particles to the status of the universe. Neutronphysics has made major contributions to this evolutionaryprocess of understanding. On the grand scale, cosmology hasevolved into an exact science and neutron physics hascontributed to the understanding of element formation and ofphase transitions in the history of our universe. Various dataextracted from measurements of neutron beta decay havebeen used extensively to fix the number of particle families atthree. On the scale of the very small, neutron experimentshave made substantial contributions to our understanding ofstrong, electroweak and gravitational interactions. Neutroninterferometry and neutron spin-echo experiments have shownhow non-classical states of neutrons can be created and usedfor highly sensitive investigations in condensed matter andfundamental physics research.

Many crucial questions remain to be answered and theincreased flux from ESS will enable major progress in a rangeof areas. For example, unique experiments can be performedwhich will help (a) to determine the basic structure of thefundamental interactions acting in nature, (b) to elucidate thehistory of the universe and to predict its future, and (c) to studyfundamental questions of quantum and measurement theory.The community in this field is about 300 scientists strong, withmany young people starting new in this field.

Investigations of neutron'sproperties deliver informa-tion about strong, weak,electromagnetic andgravitational elementaryforces of nature

Quantum optics withneutrons opens new fields

II. Flagship experimentsThe following generic experiments will become feasible atESS. They use ultra-cold, cold and hot neutrons from the

120

source. The results of these experiments are intended to raisethe highest scientific interest and they can be published injournals with the highest impact factor but they are also ratherrisky.

The question of the origin of handedness of natureIn nuclear decay experiments it was recognized in the late1950s that one of the four fundamental forces - the weak force- is, as far as we have been able to discern so far, exclusivelyleft-handed. Most Grand Unified Theories, however, start witha left-right symmetric universe, and explain the evident left-handedness of nature through a spontaneous symmetrybreaking caused by a phase transition of the vacuum, ascenario, which, if true, would mean that the neutrinos todayshould carry a small right-handed component. Although limitson the right-handed currents have been derived from freeneutron and muon decay experiments, what is really neededis a clear-cut "yes" or "no" experiment. Such an experiment,planned for ESS, is the two-body β-decay of unpolarizedneutrons into hydrogen atoms and antineutrinos which occurswith a relative probability of 4.2⋅10-6 compared to the usual β-decay.

Usual decay mode: n→→→→ p + e- + νννν e

exotic decay mode: n→→→→ H + νννν e.

What is so interesting about this decay is that one of the fourhydrogen hyperfine states cannot be populated at all if theneutrinos are completely left-handed. A non-zero population ofthis substate would, therefore, be a direct measure of a right-handed component.

Figure1: Scheme for the measurement of the neutron decay into ahydrogen atom.

This experiment has severe background suppressionrequirements for which the pulsed structure of ESS, allied toits intensity, is well suited. Thus, with ESS it may be possible

The exotic decay of theneutron into a hydrogenatom and an antineutrinocan help to find pheno-mena beyond the StandardModel

121

to prove for the first time that nature does not possess anintrinsic handedness and that there is exciting new physicsbeyond today's Standard Model of particle physics.

In a second stage the experiment has to be done withpolarized neutrons where the transition probabilites betweenthe hyperfine levels can be changed drastically.

An intense pulsed cold neutron beam is required for theseexperiments, which are not tractable with current neutronsources. The long pulse ESS option is the most appropriatefor this project.

The origin of the baryon asymmetry of the universeThe big bang theory presumes that equal amounts of matterand antimatter were created in the primordial explosion. In thesubsequent process of annihillation of matter and antimatteronly very few heavy particles ("baryons") and an equalnumber of antiparticles from this early period could survive.Our mere existence contradicts this expectation; thereremained about 108 times more baryons in the universe thanpredicted and almost no antibaryons have survived. So far,the only viable solution of this problem is the violation ofcharge-parity symmetry (CP) which, on all reasonableexpectations, is equivalent to a violation of time symmetry (T)that could have led to a small excess of particles before theannihilation stage.

Violation of the CP-symmetry has been observed in the decayof kaons. However, this single positive result is not sufficient toverify the above conjecture, nor to identify the origin of CP- orT-violation. Grand Unified Theories (GUT) require T-violatingamplitudes that are orders of magnitude larger than can beaccommodated by the present Standard Model. Therefore,another generation of experiments is needed to obtaindecisive answers.

The most direct access to these questions lies in the detailedinvestigation of neutron decay and in measurements of itselectric dipole moment. Electric dipole moment measurementsstarted in the fifties and increased their sensitivity by one orderof magnitude every seven years. They are based on searchesfor a deviation equal to ±d⋅E from the well-known angularfrequency of

hω = 2 µB±d×E

of a neutron spin in a magnetic field B and a parallel orantiparallel electric field E. Current theories of the baryonasymmetry of the universe is related to an EDM of about 10-28

e cm, a limit that is accessable with the ESS. The currentupper limit is 6⋅10-26 e cm.

These experiments are most effectively done with ultra-coldneutrons (UCN) where recent developments on new UCNsources predict orders of magnitude intensity gains. The

Further measurements onthe electric dipole momentof the neutron can help forthe understanding ofmatter-antimatterasymmetry in the universe

122

possible arrangement of such a UCN station which consumesabout 1% of the beam power is shown in Figure 2.

Figure 2: General layout of ESS with a dedicated UCN target station (not inscale).

A separate UCN station served by 1% of the proton beampower and the long pulse option would give new perspectivesfor research with ultracold neutrons.

The question of charge independence of nuclear forcesThe strong or nuclear force is governed by the fundamentalquark-quark interaction described by Quantum Chromo-Dynamics (QCD). It is believed that the strong nuclear force isessentially the same for protons and neutrons or, moregenerally, for up and down quarks. In this respect, the nuclearpart of the singlet scattering length should be the same for theproton-proton and the neutron-neutron systems, and it shouldbe similar to the neutron-proton interaction. The neutron-proton scattering length is the only precisely known quantitywhereas the nuclear part in the proton-proton system ismasked by the Coulomb interaction and the neutron-neutronscattering length has only been extracted indirectly fromseveral three-body interaction processes. The best way tocheck, whether the deviations in the singlet scattering lengthsextracted from these experiments really signal a breakdown ofisospin invariance, is a direct scattering measurement of theneutron-neutron scattering at very low energy

n + n→→→→ n + ns s snp pp nna = a = a ?

tnna ≡≡≡≡ 0 ?

In a second stage a dedicated experiment using polarizedneutron beams could subject the hypothesis of the flavour

With a pulsed ESS for thefirst time a direct neutron-neutron scatteringexperiment becomesfeasible

123

independence of the quark gluon interaction to a precision testat the baryon level.

As the interaction rate for a neutron-neutron scatteringexperiment scales with the square of the neutron flux density,the high peak intensity of ESS has huge advantages. Theplanned ESS-experiments make use (a) of the time structure,by allowing the fast neutrons of one pulse to hit the slowerones of a preceding pulse and (b) coincidences in time andspace for the counts for each scattered pair of neutrons.

Figure 3: Sketch of the proposed neutron-neutron scattering experiments.

A well focused dense cold neutron beam merging with anintense hot beam would be optimal for such experiments. Theshort pulse option of ESS has additional advantages due to itshigher peak flux.

Neutron quantum opticsThe phase of a neutron wave has become a measurablequantity since the invention of neutron interferometry. Basictests of quantum mechanics have been performed in the pastand it has been shown how neutrons can be used as apowerful tool in quantum optics. Non-classical neutron states,which are extremely fragile against any dissipation, have beencreated in neutron interferometry and neutron spin-echoexperiments. Major interest concerns the verification oftopological phases which are determined by the geometricalform rather than by the strength of the interaction. A completequantum state reconstruction will become feasible by asimultaneous coherence function and momentum post-selection measurement procedure.

The coupling of the neutron magnetic moment to oscillatingmagnetic fields permits multi-photon exchange and dressedneutron states, while the quantization of neutron states insidemicroscopic structures facilitates new possibilities in basic andadvanced materials research. Pulsed beams can be trappedbetween perfect crystal plates forming narrow band neutronresonators that can be developed further as neutronaccumulator systems. Inside travelling magnetic fields anadvanced method of beam tailoring becomes feasible

Non-classical states ofneutrons can be producedand used in neutroninterferometry and neutronspin-echo systems

Neutron resonators andaccumulators becomefeasible

124

permitting intensity gains by another factor of ten. These newpossibilities have to be exploited as a step towards advancedquantum optical devices serving as resonators and phasespace transformers and compressors.

The long pulse option of ESS can surpass the existingpossibilities. A cold neutron beam line adaptable for travellingmagnetic fields and on vibration and thermally isolated andcontrolled experimental area would be desired for theseexperiments.

Figure 4: Wigner representation of a non-classical neutron state as it existsin neutron interferometry and neutron spin-echo arrangements.

III. Various other scientific achievements anticipated atESS

So far the flagship experiments for ESS have been discussed.There is a rich variety of other topics in the field offundamental neutron investigations of which we mention onlya few.

Neutron decay experiments, in particular measurements of theneutron lifetime and angular correlation coefficients determinecertain free parameters of the Standard Model complementaryto high energy physics research. The Vud parameter of thequark mixing matrix for the d-u transition in neutron decayplays a key role in testing the unitarity of this matrix, whichyields information on possible physics beyond the StandardModel. The experiments determine the strength and structureof the weak quark current and provide the possibility ofobserving new processes generated by scalar and tensorcomponents, with or without T violating terms or right handedcurrents.

Today all weak semileptonic phenomena with significance forcosmology, astrophysics and particle physics must be

Research on fundamentalphenomena is ratherpopular for youngstudents and researchers

125

calculated from neutron decay data. Certain neutron decayexperiments can make use of the pulse structure of ESS forbackground subpression.

Another topic of high interest is the investigation of the weakinteraction between nucleons may be carried out by means ofcoherent spin rotation of transversely polarized neutrons or bydifferential absorption of a longitudinally polarized neutronbeam interacting with unpolarized nuclei of hydrogen orhelium.

• The proposed ultracold neutron factory will host besidestwo long-term projects: the search for an electric dipolemoment of the neutron and measurements on free neutrondecay. Ultra-cold and very cold neutrons will be used forelastic and inelastic surface reflections and as probes fornano-structured materials. Quantum gravitational stateshave been measured and weak gravity effects becomeaccessable. New bunching, cooling and trapping systemswill be developed.

• Neutron quantum optical experiments will become feasiblewhere the time structure of the beam can be used toproduce a steady beam with an intensity governed by thepeak flux of ESS. Topological phenomena could betackled in a new way. The transition from a quantum to amixed state could be studied in detail contributing to ourunderstanding of a quantum measurement. QuantumZeno-effect experiments will show how a quantum statecan be frozen when a continuous measurement isperformed.

IV. Issue of target station and beam linesFor Fundamental Physics with ultracold neutrons the followingadditional target is needed:

(a) UCN-station accepting the whole beam power for about1% of the time (1 second on, 5 minutes off).

For Fundamental Physics with neutrons the following beamlines and experimental areas are needed:

(b) A beam line for producing high dense neutron gas at thecold moderator of the 16.6 Hz target station.

(c) A beam line for neutron optics at a thermal guideassociated with an experimental area with specialenvironmental conditions (vibration-free, air-conditioned,humidity-controlled etc.).

The proposed ultracoldneutron target station hasunique features and opensnew fields of research

126

A1 Members of the Science Groups (individual group convenors are marked with an asterisk)

Solid State Physics A. Furrer* PSI VilligenC. Vettier* Institut Laue LangevinR. Currat Institut Laue LangevinR. Cywinski University of LeedsC. Fermon LLB/CEA SaclayB. Keimer MPI für Festkörperforsch. StuttgartG. Lander Institut für Transurane, KarlsruheD.F. McMorrow Risø National LaboratoryH.R. Ott ETH Zürich

Material Science and Engineering H. Zabel* University of BochumT. Lorentzen* Danish Stir Welding TechnologyM. Ceretti LLB/CEA SaclayR. Cowley University of OxfordM. Daymond ISIS/Rutherford Appleton Lab.A. Magerl University of ErlangenF.M. Mulder Interfaculty Reactor Institute DelftP.J. Withers Manchester Material Science Center

Biology and Biotechnology J.R. Helliwell* University of ManchesterT. Bayerl University of WürzburgO. Byron University of GlasgowD. Svergun EMBL HamburgJ.-C. Thierry IGBMC-CNRSJ. Zaccai Institut Laue Langevin

Soft Condensed Matter J. Colmenero* University of the Basque Country & DIPCD. Richter* Forschungszentrum JülichA. Arbe University of the Basque CountryF. Boué LLB/CEA SaclayS. Janssen PSI VilligenK. Mortensen Risø National LaboratoryJ. Rieger BASF LudwigshafenP. Schurtenberger University of FribourgR.K. Thomas University of Oxford

Chemical Structure, Kinetics and Dynamics H. Jobic* CNRS/University of Lyon 1W.I.F. David* ISIS/Rutherford Appleton Lab.H. Gies University of BochumM. Latroche CNRSM. Prager Forschungszentrum JülichA.R. Rennie King’s College LondonC. Wilson ISIS/Rutherford Appleton Lab.

Earth Science, Environmental Science R. Rinaldi* University of Perugiaand Cultural Heritage G. Artioli University of Milano

M.T. Dove University of CambridgeW. Schäfer University of Bonn and FZ JülichP.F. Schofield National History Museum LondonB. Winkler University of Kiel

Liquids and Glasses R.L. McGreevy* University of UppsalaT.C. Hansen Institut Laue LangevinA.K. Soper ISIS/Rutherford Appleton Lab.J.-B. Suck TU of ChemnitzR. Vacher University of Montpellier II

Fundamental Physics H. Rauch* Atomic Inst. of the Austrian UniversityH. Abele University of HeidelbergJ. Byrne University of SussexD. Dubbers Institut Laue LangevinW. Fischer PSI VilligenJ. Sromicki ETH ZürichJ.M. Pendleburry University of Sussex

127

A2 Members of the Instrument Groups (individual group convenors are marked with an asterisk)

Powder Diffraction P. Radaelli* ISIS/Rutherford Appleton Lab.H.-J. Bleif Hahn-Meitner-Institut BerlinS. Hull ISIS/Rutherford Appleton Lab.J. Rodriguez Carvajal Laboratoire Léon BrillouinE. Suard Institut Laue Langevin

Direct Geometry Spectrometers R. Eccleston* ISIS/Rutherford Appleton Lab.R. Bewley ISIS/Rutherford Appleton Lab.F. Mezei Hahn-Meitner-Institut BerlinH. Mutka Institut Laue LangevinH. Ronnow Institut Laue Langevin

Indirect Geometry Spectrometers K. Andersen* ISIS/Rutherford Appleton Lab.P. Allenspach PSI VilligenB. Fak ISIS/Rutherford Appleton Lab.O. Kirstein Forschungszentrum JülichM. Zoppi CNRS

Neutron Spin Echo Spectrometer M. Monkenbusch* Forschungszentrum JülichB. Farago Institut Laue LangevinC. Pappas Hahn-Meitner-Institut Berlin

SANS R. Heenan* ISIS/Rutherford Appleton Lab.B. Cubitt Institut Laue LangevinK. Mortensen Risø National LaboratoryD. Schwahn Forschungszentrum JülichA. Wiedenmann Hahn-Meitner-Institut Berlin

Reflectometry H. Fritzsche* Hahn-Meitner-Institut BerlinC. Fermon Laboratoire Léon BrillouinJ. Webster ISIS/Rutherford Appleton Lab.

Single Crystal Diffraction C. Wilson* ISIS/Rutherford Appleton Lab.W. Jauch Hahn-Meitner-Institut BerlinG. McIntyre Institut Laue LangevinD. Myles EMBL Hamburg

Structure Factor Determination A. Soper* ISIS/Rutherford Appleton Lab.R. McGreevy University of Uppsala

Engineering P.J. Withers* Manchester Material Science CenterM. Daymond ISIS/Rutherford Appleton Lab.T. Lorentzen Risø National LaboratoryW. Reimers Hahn-Meitner-Institut Berlin

ESS Instrumentation Task Leader F. Mezei Hahn-Meitner-Institut BerlinESS Instrumentation Task Leader Deputy R. Eccleston ISIS/Rutherford Appleton Lab.ESS Instrumentation Assistant T. Gutberlet Hahn-Meitner-Institut Berlin

128

A3 Participants at the SAC/ENSA Workshop in Engelberg

Name Affiliation Country

H. Abele University of Heidelberg Germany

K. Andersen ISIS/Rutherford Appleton Laboratory United Kingdom

A. Arbe University of the Basque Country Spain

G. Artioli University of Milano Italy

G. Bauer PSI Villigen Switzerland

T. Bayerl University of Würzburg Germany

S. Bennington ISIS/Rutherford Appleton Laboratory United Kingdom

F. Boué LLB/CEA Saclay France

M. Braun Shea (Secretary) PSI Villigen Switzerland

J. Byrne University of Sussex United Kingdom

O. Byron University of Glasgow United Kingdom

M. Ceretti LLB/CEA Saclay France

K. Clausen ESS Project/Forschungszentrum Jülich Germany

J. Colmenero University of the Basque Country & DIPC Spain

R. Cowley University of Oxford United Kingdom

R. Currat Institut Laue Langevin France

R. Cywinski University of Leeds United Kingdom

M.R. Daymond ISIS/Rutherford Appleton Laboratory United Kingdom

D. Dubbers Institut Laue-Langevin France

R.S. Eccleston ISIS/Rutherford Appleton Laboratory United Kingdom

C. Fermon LLB/CEA Saclay France

D. Filges Forschungszentrum Jülich Germany

W. Fischer PSI Villigen Switzerland

H. Fritzsche Hahn-Meitner-Institut Berlin Germany

A. Furrer PSI Villigen Switzerland

H. Gies University of Bochum Germany

J.M.F. Gunn University of Birmingham United Kingdom

T. Gutberlet Hahn-Meitner-Institut Berlin Germany

T.C. Hansen Institut Laue Langevin France

R. Heenan ISIS/Rutherford Appleton Laboratory United Kingdom

J.R. Helliwell University of Manchester United Kingdom

S. Janssen PSI Villigen Switzerland

H. Jobic CNRS/Université Claude Bernard Lyon 1 France

H. Karow European Science Foundation France

B. Keimer MPI für Festkörperforschung Stuttgart Germany

G.H. Lander Institut für Transurane, Karlsruhe Germany

M. Latroche CNRS, Laboratoire de Terres Rares, Thiais France

129

Name Affiliation Country

T. Lorentzen Danish Stir Welding Technology Denmark

A. Magerl University of Erlangen Germany

T. Mason Spallation Neutron Source USA

R.L. McGreevy University of Uppsala Sweden

D.F. McMorrow Risø National Laboratory Denmark

F. Mezei Hahn-Meitner-Institut Berlin Germany

M. Monkenbusch Forschungszentrum Jülich Germany

K. Mortensen Risø National Laboratory Denmark

F.M. Mulder Interfaculty Reactor Institute Delft Netherlands

H.R. Ott ETH Zürich Switzerland

S. Oubenkhir (Secretary) Forschungszentrum Jülich Germany

J.-P. Pouget Centre National de la Recherche Scientifique, Paris France

M. Prager Forschungszentrum Jülich Germany

H. Rauch Atomic Institute of the Austrian University Austria

A.R. Rennie King’s College London United Kingdom

D. Richter (SAC Chairman) Forschungszentrum Jülich Germany

R. Rinaldi University of Perugia Italy

W. Schäfer University of Bonn & Forschungszentrum Jülich Germany

P.F. Schofield Natural History Museum London United Kingdom

P. Schurtenberger University of Fribourg Switzerland

A.K. Soper ISIS/Rutherford Appleton Laboratory United Kingdom

J. Sromicki ETH Zürich Switzerland

U. Steigenberger ISIS/Rutherford Appleton Laboratory United Kingdom

W.G. Stirling European Synchrotron Radiation Facility, Grenoble France

J.-B. Suck Technical University of Chemnitz Germany

D. Svergun European Molecular Biology Laboratory, Hamburg Germany

J.-C. Thierry IGBMC-CNRS France

R.K. Thomas University of Oxford United Kingdom

P. Tindemans ESS Project /Forschungszentrum Jülich Germany

R. Vacher University of Montpellier II France

C. Vettier Institut Laue Langevin France

N. Williams European Science Foundation France

C. Wilson ISIS/Rutherford Appleton Laboratory United Kingdom

B. Winkler University of Kiel Germany

A. Wischnewski Forschungszentrum Jülich Germany

P.J. Withers Manchester Materials Science Center United Kingdom

H. Zabel University of Bochum Germany

J. Zaccai Institut Laue Langevin France

130

9. Acknowledgements

We acknowledge the sponsorship of the European Science Foundation, the Swiss National

Science Foundation, the European Neutron Round Table, the Paul Scherrer Institute in

Villigen and the Forschungszentrum Jülich. Their generous support made the SAC/ENSA

workshop possible and led to this report. Furthermore, we would like to thank S. Oubenkhir

and M. Braun-Shea very much for their considerable secretarial help and hard work,

organising the workshop in Engelberg and editing the report.