NOTIZIARIO Neutroni e Luce di Sincrotrone - Issue 5 n.2, 2000

Cover photo:

The study of protein-membraneinteractions

Il è pubblicato a

cura del C.N.R. in collaborazionecon il Dipartimento di Fisicadell’Università degli Studidi Roma “Tor Vergata”.

Vol. 5 n. 2 Dicembre 2000Autorizzazione del Tribunale diRoma n. 124/96 del 22-03-96

DIRETTORE RESPONSABILE:

C. Andreani

COMITATO DI DIREZIONE:

M. Apice, P. Bosi

COMITATO DI REDAZIONE:

L. Avaldi, F. Carsughi, G. Ruocco, U. Wanderlingh

SEGRETERIA DI REDAZIONE:

D. Catena

HANNO COLLABORATO

A QUESTO NUMERO:

G. Cicognani, M. Zoppi,P. Bosi

GRAFICA E STAMPA:

om graficavia Fabrizio Luscino 7300174 RomaFinito di stamparenel mese di Dicembre 2000

PER NUMERI ARRETRATI:

Paola Bosi, Tel: +39 6 49932468Fax: +39 6 49932456E-mail: [email protected].

PER INFORMAZIONI EDITORIALI:

Desy Catena, Università degli Studidi Roma “Tor Vergata”, Dip. di Fisicavia della Ricerca Scientifica, 100133 RomaTel: +39 6 72594364Fax: +39 6 2023507E-mail: [email protected]

Vol. 5 n. 2 Dicembre 2000

NOTIZIARIONeutroni e Luce di Sincrotrone

SOMMARIO

Rivista delConsiglio Nazionaledelle Ricerche

EDITORIALE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2C. Andreani

RASSEGNA SCIENTIFICA

The Tomography Experiment at the SYRMEP Beamline at ELETTRA . . . . . . . . . . . . . . . . . . . . . . . . . 3S. Pani et al.

Neutron Reflectivity Studies of ImmisciblePolymer/Polymer Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9M. Sferrazza

Neutron-Electron Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18E. Balcar and W. Lovesey

Chemical-Shift Normal Incidence X-Ray StandingWave Determination of Adsorbate Structures . . . . . . . . . . . . 25V. Formoso

DOVE NEUTRONI

A Second Target Station at ISIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39J. Penfold

VARIE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

CALENDARIO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

SCADENZE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

FACILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

NOTIZIARIONeutroni e Luce di Sincrotrone

2

NOTIZIARIO NEUTRONI E LUCE DI SINCROTRONE • Vol. 5 n. 2 Dicembre 2000

EDITORIALE

A partire da questo numero il prof. GiancarloRuocco entra a far parte del Comitato diRedazione in sostituzione del prof. FedericoBoscherini. Ringrazio il collega Boscherini

per l’attività svolta in questi anni e mi auguro che lasua collaborazione con il Notiziario continui, sia conarticoli scientifici che di rassegna e riflessione sullapolitica italiana in questi settori. Per quanto riguardale iniziative internazionali di spettroscopianeutronica vorrei ricordare il progetto che prevede lacostruzione di una stazione per una secondatarghetta, che affiancherà quella già operante ad ISIS(sito web http://www.isis.rl.ac.uk/targetstation2/).La targhetta permetterà la produzione di intensi fascidi neutroni “freddi”, con l’obiettivo di effettuarestudi complementari a quelli attualmente possibili adISIS; in particolare si prevede un programma diricerca specificatamente dedicato allo studio dellamateria soffice, la scienza della vita e i materiali bio-molecolari. Questo progetto rappresenta unimportante occasione, in vista dell’auspicabilerinnovo dell’accordo tra CNR e CLRC previstonell’anno 2001, per la definizione di nuovecollaborazioni scientifiche tra la comunità italiana equella inglese, sia nell’ambito della ricerca di base sianello sviluppo di strumentazione e di formazione delpersonale. Ricordiamo che dal 23 Settembre al 3 Ottobre si èsvolta, in Sardegna presso l'Hotel Capo D'Orso(Palau, SS) la V edizione della Scuola diSpettroscopia Neutronica intitolata da quest’annoalla memoria di Francesco Paolo Ricci, che dellaScuola era stato il sostenitore ed il primo Direttore.Questa edizione ha avuto come tema la DiffusioneAnelastica dei Neutroni, ed è stata diretta dal Prof. A. Deriu (Università di Parma) e dal Dr. M.Zoppi (Istituto di Elettronica Quantistica del CNR -Firenze). Alla Scuola hanno partecipato docenti siaitaliani che stranieri, questi ultimi provenientidall’ILL e da Los Alamos, e diciotto studenti che sisono impegnati sia nel seguire le lezioni teoriche, chenelle attività seminariali e nelle esercitazioni praticheche hanno costituito una parte molto rilevante delleattivita svolte. È stato assegnato un premio al gruppoche aveva svolto la migliore esercitazione, tutti idocenti hanno particolarmente apprezzato il grandeimpegno e l’entusiasmo dimostrato dai partecipanti.Infine, per quanto riguarda il progetto TOSCA, negliultimi mesi è stata completata l’istallazione ed ilcommissioning di TOSCA II presso ISIS. Lospettrometro, realizzato presso l’Istituto diElettronica Quantistica del CNR di Firenze, è oradisponibile per la comunità internazionale di utentidella spettroscopia neutronica, ed arricchisce il parcostrumenti della sorgente ISIS.

S tarting from this issue the Editorial committee willbenefit from the scientific expertise of Prof.Giancarlo Ruocco, who substitutes Prof. F. Boscherini. I would like to thank Federico

Boscherini for his work in these years and I hope that hiscollaboration with Notiziario will continue, both withscientific papers and with contributions on theorganisation aspects of Italian Research. As far as theinternational initiatives in neutron scattering areconcerned I want to recall the project of a second targetstation at ISIS (web site http://www.isis.rl.ac.uk/targetstation2/),which aims to double the number of neutron instrumentsand will be optimised for neutron techniques such asreflection small angle scattering and high-resolutiondiffraction and spectroscopy. The instruments on the newtarget will be designed for the growing fields of soft-condensed matter and bio-molecular science. This projectrepresents an important opportunity, in view of therenewal of the international agreement between CNR andCLRC planned for the year 2001, for the definition ofscientific collaborations among the Italian and Britishneutron scattering communities, both in basic researchand in development of neutron instrumentation; We recall that the fifth edition of the Italian ‘Scuola diSpettroscopia Neutronica’ was held in Palau, SS, at theHotel Capo D’Orso, in the period 23rd September - 3rdOctober. The neutron school, from this year entitled inmemory of its first Director Francesco Paolo Ricci, wasaddressing the theme of the Neutron Inelastic Scatteringand was directed by Prof. A. Deriu (Universita' di Parma)and Dr. M. Zoppi (Istituto di Elettronica Quantistica delCNR - Firenze). Teachers were from Italy, from ILL andfrom Los Alamos and 18 students were attending theschool, partecipating to both theoretical seminars andexperimental activities with great enthusiasm. As far as the TOSCA project is concerned, in the lastmonths the installation and commissioning of TOSCA IIhave been completed at ISIS. The spectrometer, realised atth Istituto di Elettronica Quantistica of the CNR inFlorence, is now available for allocation of beam time aspart of the instrument suite at ISIS.

Carla Andreani

3

Vol. 5 n. 2 Dicembre 2000 • NOTIZIARIO NEUTRONI E LUCE DI SINCROTRONE

IntroductionObtaining the best possible information with theminimum risk for the patient: this is the challenge ofmodern radiology. Radiologic examinations represent themain cause artificial exposure to the population: theexposure level due to such exams is about one half thetotal exposure from the natural radioactivity background(100-200 mrem/year). On the other hand, a late diagnosis of a tumour candramatically decrease the probability of full recovery;breast tumours, particularly, need to be detected inasymptomatic women as soon as possible. Early detection of breast cancer is very difficult due tothe characteristics of early signs of tumour: these may benodules, whose X-ray absorption characteristics are veryclose to those of breast tissue, thus resulting in a very lowimage contrast, or microcalcifications, whose dimensionsare about 100 µm and require therefore a very highspatial resolution of the detection system [1].Furthermore, breast is one of the most radiosensitiveorgans, and a high delivered dose can increase the risk ofcarcinogenesis.Therefore, it is necessary to optimize radiologicexaminations, and in particular mammography, on twosides: first, it is necessary to reduce the dose delivered ina typical exam; second, the image quality must beincreased, giving a higher information content, in termsof spatial resolution and contrast resolution, at the samedelivered dose. The main limitations of a conventionalmammographic apparatus lie in the limited possibility ofvarying the X-ray beam energy and in the low dynamicrange of the radiographic film, which may not allow asimultaneous visualization of very dense and very softtissues.The radiation source for a conventional apparatus is anX-ray tube, giving in output a bremsstrahlung spectrumin which the maximum energy corresponds to the tubevoltage, and featuring two or more superimposed peaks,due to the characteristic emission of the anode material.In the case of a mammographic tube, the anode materialis usually Molybdenum, whose characteristic energiesare 17.5 and 19.6 keV. The effective energy of the outputspectrum, about 17 keV, is regarded as an optimal energyfor mammography. The low flux from a clinical tube doesnot allow the use of monochromatization devices.

On the other hand, the high flux from a synchrotronradiation machine allows the use of a monochromator, inorder to select the most suitable energy for eachexamination, depending on organ thickness andcomposition. In this way, there is no dose delivery due tolow energy components, which are almost fully absorbedin the organ and thus do not contribute to theradiographic information, as it happens with aconventional X-ray tube.The SYRMEP (Synchrotron Radiation for MedicalPhysics) experiment at the synchrotron radiation facilityElettra in Trieste consists in developing a mammographicapparatus with monochromatic synchrotron X-ray beamsand with a linear Silicon pixel detector coupled to asingle photon counting electronics. The combination ofmonochromatic beams and high efficiency detector hasalready given very promising results in mammographicphantoms and breast tissue imaging [2,3,4].Planar imaging, however, does not allow thedetermination of the depth of a tumour inside an organ,and may also not allow the detection of very lowvariations in attenuation coefficients. These are thetypical features of Computed Tomography (CT), atechnique capable of providing the attenuationcoefficient map by reconstructing a section of an objectfrom attenuation profiles acquired along 180°.Synchrotron radiation CT is currently investigated byseveral researchers [5,6,7], because a monochromatic andhighly collimated beam obtainable from a SR machinewould be an optimum tool for tomography: SRtomograms are not affected by artifacts due to beamhardening, which occur in tomograms acquired withpolychromatic beams [8]; furthermore, reconstructionalgorithms for divergent beam tomograms needgeometrical corrections, which are not required for SRtomograms: this results in the use of the simplest possiblereconstruction algorithms. A feasibility study of SR tomo-mammography, combining the advantages of SRmammography and SR tomography, has been carriedout. Moreover, we have applied to tomography aninnovative technique, called Diffraction EnhancedImaging (DEI), in order to verify the velocity vIP

possibility of detecting very low contrast details, whichwould be undetectable by means of conventionaltomography.

THE TOMOGRAPHY EXPERIMENT AT THE SYRMEP BEAMLINE AT ELETTRA S. Pani1,2, F. Arfelli1,2, D. Dreossi1, R. Longo1,2, R. Menk3, A. Olivo1,2, P. Poropat1,2, L. Rigon1,2, G. Tromba3, E. Castelli1,2.

1 Dipartimento di Fisica, Università di Trieste.2 INFN Sezione di Trieste.3 Sincrotrone Trieste SCpA.

Articolo ricevuto in redazione nel mese di Ottobre 2000


4



Materials and methods: the beamline and the detectionsystemsThe radiation source of the SYRMEP beamline is thebending magnet n. 6 of the Elettra storage ring. Aschematic of the beamline is shown in figure 1. The onlyoptical element of the beamline is the monochromator: itconsists of a channel-cut Si (111) crystal, which providesX-ray beams in the energy range 15-35 keV. The beamwidth at the experimental hutch is 10 cm; the FWHM inthe vertical direction is 4 mm at 20 keV.

Upstream the monochromator, a copper slits systemallows one to select the dimensions of the beamimpinging on the crystal. At the entrance of theexperimental area, a tungsten slit system is used to matchthe beam size to the detector sensitive area, thus avoidingunnecessary dose delivery. Several stepping motorsallow the positioning and the movimentation of thesample and of the detection systems. The translationstages allow an accuracy of 1 micron, while the rotationstage provides an accuracy of 10-3 degrees. After thetungsten slits, a parallel plate ionization chamber is usedas a flux monitor to evaluate the dose delivered to thesample.The detector designed by the SYRMEP collaboration issketched in fig 2. It consists of a silicon microstripdetector used in "edge on" geometry, with the radiationimpinging on the side rather than on the surface of thechip. The chip depth is 1 cm, the strip pitch (determiningthe pixel width) is 200 µm and the chip thickness(determining the pixel height) is 300 µm. The "edge on"

configuration allows a very high absorption efficiency inthe mammographic energy range; the detection efficiencyis actually limited by the presence of a "dead volume" infront of the strips; in this volume, the charge collectionefficiency is reduced. Due to this fact, the detectionefficiency ranges from 70% to 90% in the mammographicenergy range [9]. The overall detector width is 5.12 cm,corresponding to 256 pixels. The detector is coupled to asingle photon counting VLSI readout chain, namedCASTOR (Counting and Amplifying sySTem fOrRadiation detection) [10]; each CASTOR chip consists of apreamplifier, a shaper-amplifier, a discriminator and a 16bit counter. A photon interacting within the detector activevolume produces a signal, which is processed bypreamplifier and amplifier. Its amplitude is then comparedto the discriminator threshold, and, if it is higher, thecontent of the counter is incremented by one unit.When acquiring a tomographic image with the SYRMEPdetector, the detector is kept stationary with respect tothe beam, while the sample is rotated in discrete steps infront of the detector; for each angular position, a profileof the object is acquired and is stored in a row of a matrix(the sinogram); the map of the attenuation coefficientsinside the sample is then reconstructed by means of thestandard CT reconstruction technique called filteredbackprojection [8].Furthermore, a commercial photostimulable phosphorimaging plate (IP) BAS-MP2025 has been used; the platearea is (20x25)cm2. The IP is read out by a BAS-1800reader. The reader scanning step, corresponding to theimage pixel size, can be 50, 100 or 200 µm.Each sinogram is acquired by simultaneously rotating thesample and translating the IP; the ratio of the samplerotation speed to the IP translation speed must be smallenough to avoid artifacts on the reconstructed image. As

shown in figure 3, if the sample cannot be regarded asstationary during an IP translation corresponding to thepixel dimension, the image may appear blurred. Twoimages, corresponding to different ratios of the sample

Fig. 1. Sketch of the SYRMEP beamline.

Fig. 2. Sketch of the SYRMEP detector.

Fig. 3. Artifacts due to the motion of the imaging plate during the angularscan. a) vθ/vIP=1/2 degrees/mm; b) vθ/vIP=1/8 degrees/mm.


5


angular velocity (vθ) to the IP translation (vIP), are shown. In the present study, velocities where chosen in order toscan a 7.2 cm long portion of the plate during a 180°angular scan, thus providing an angular step of 0.125degrees per image line when the IP is read out with a 50µm step.

Absorption tomographyAs previously underlined, an ideal mammographicapparatus should provide both a high spatial resolutionand a high contrast resolution. Furthermore, a goodtomographic apparatus should provide a precisedetermination of attenuation coefficients, in order toallow one to determine the composition of tissues and todetect neoplastic formations. Therefore, the first tomography tests at the SYRMEPbeamline have been carried out on custom designed testobjects, developed in order to check the fulfillment of theabove conditions. Due to the small size of the SYRMEPdetector, the test objects diameters are 3 or 4 cm. Furthermore, a procedure was adopted to give areasonable estimate of the dose which would bedelivered to a full breast to obtain a comparable imagequality in a real breast tomography. The best indicator of biological risk in mammography isthe mean glandular dose (MGD), which is defined as thedose delivered to the glandular component of the breasttissue. An average breast is in fact composed, accordingto NCRP [11], by 50% glandular tissue and 50% adiposetissue, the latter being non radiosensitive.In planar mammography, MGD is evaluated for acompressed breast and thus the thickness along the beamdirection is constant. In the present study, MGD for acircular breast section, with a variable thickness, wasevaluated as the average of the MGD’s delivered to

infinitesimal width slices along the beam direction. The MGD needed to image small objects is much smallerthan the dose delivered to a standard size breast; in orderto conduct a meaningful dose estimate, we evaluated the

dose needed to obtain a comparable image quality of a 12cm diameter sample, assuming 12 cm as the maximumdiameter for a breast section in tomographic geometry.For each image of small objects, the average outputphoton fluence is calculated; then, the entrance fluenceN12, which would give the same average output from a 12cm diameter breast section, is computed. The MGDdelivered by a photon fluence N12 to a 12 cm diameterbreast, defined MGD12, represents, to a goodapproximation, the maximum MGD that would bedelivered to a real breast.The spatial resolution test has been carried out with a 4cm diameter object, with five series of holes. Thediameter of the holes ranges from 500 µm to 3 mm. Figure 4 shows the image of the test object acquired at 25keV with both detectors with the minimum dosenecessary to obtain a signal-to-noise ratio equal to 5 onthe 500 µm details. According to Rose’s criterion [12], 5 isthe minimum SNR necessary to visualize a detail. The MGD12’s corresponding to these images were 0.16and 1 mGy for the SYRMEP detector and for the IP,respectively. These values are both lower than the dosedelivered for a conventional mammography, whichranges between 1.2 and 1.8 mGy. The relevant differencebetween the MGD12 delivered in the IP and in theSTRMEP image is due to two reasons: first, the SYRMEPdetector has a much higher efficiency at 25 keV;moreover, the angular sampling step used for IP imagesis much smaller than the step actually needed to detect500 µm details. This was done in order to avoid themotion artifacts previously described. This effect can besolved by moving the IP in discrete steps, and using a fastshutter to avoid IP irradiation during the translation.Figure 5 shows the 500 µm details from images acquiredwith the SYRMEP detector with an increasing angularstep size and a correspondingly increasing entrancefluence on the sample. The details can be detected evenwith a 4 degrees angular step, but their shape is betterresolved with a fine angular scan.A second test object has been built to evaluate thecontrast resolution of the detection systems. It consists ofa 2 cm diameter BR12 [13] cylinder with embedded chalk,silicone, wax, polyethylene and water details. All detailsare 5 mm in diameter. The images of this test object at 34keV are shown in figure 6.Table I shows the theoretical and the measuredattenuation coefficients for images acquired with bothdetectors for water, polyethylene (PE) and BR12 at 34keV, compared to the theoretical values. The MGD12 required to obtain a SNR equal to 5 for thewater detail, simulating a nodule embedded in breasttissue, were 0.1 mGy and 1.0 mGy for the SYRMEPdetector and for the IP, respectively.All attenuation coefficients are correctly reproduced.

Fig. 4. Images of the spatial resolution test object acquired at 25 keV. a)SYRMEP detector; b) Imaging plate.

6



water PE BR12(cm-1) (cm-1) (cm-1)

Theory 0.31 0.23 0.27

SYRMEP 0.28±0.04 0.21±0.04 0.24±0.05

IP 0.33±0.03 0.24±0.03 0.27±0.05

Table I. Theoretical and measured attenuation coefficient from imagesshown in figure 6.

Diffraction enhanced tomographyA widely applied technique in synchrotron radiationimaging is the so called Diffraction Enhanced Imaging(DEI). DEI consists in placing a Silicon crystal, calledanalyzer, between the sample and the detector. DEI

would be useful in biological imaging, because thetypical scattering angle of biological tissues is very closeto the angular acceptance of a typical monochromator.

Figure 7 shows a schematic explanation of the DEIworking principle. The radiation transmitted through the sample consists ofboth primary (undeflected) and scattered photons.When the analyzer is perfectly aligned with respect tothe beamline monochromator, i.e. when the lattice

planes are parallel, only undeflected photons will betransmitted by the analyzer, which thus acts as a scatterrejecter.On the other hand, when a slightly misalignment isintroduced between the two crystals, other photons willmeet the Bragg condition and different scatteredcomponents will transmitted with different weights, asdescribed by the crystal rocking curve. This techniqueallows one to almost completely remove the primaryradiation and to acquire scatter images. Several studieshave already given good results in the visualization ofsoft tissues [14,15].

In this study, we applied DEI to tomography.Reconstruction algorithms in conventional tomographyare based on the assumption that most of the scatteredradiation is removed by means of collimators; in DEItomography, one is not dealing with line integrals as in

transmission tomography. Nevertheless, to someapproximation the same reconstruction algorithms canbe used. Due to geometrical configuration of the apparatus, DEItomograms can only be acquired with the imaging plate,because the beam transmitted after the analyzer is notparallel to the entrance beam, and the SYRMEP detectorcannot be aligned with respect to the beam.Two possible approaches are possible: the first one isbased on contrast enhancement effects, while the secondone is based on edge enhancement effects.

Fig. 6. Images a BR12 test object containing chalk (center), water (top),polyethylene (left), wax (right), silicone (bottom): a) SYRMEP detector; b)Imaging plate.

Fig. 7. Working principle of Diffraction Enhanced Imaging.

Fig. 9. X-ray deflection in the paraxial approximation: a) beam impingingon the top of the wire; b) beam impinging on the center of the wire.

Fig. 5. 500 µm details fromimages acquired at a constantdelivered dose and withdifferent angular steps: a) 180steps, 100 ms/step; b) 90 steps,200 ms/step; c) 45 steps, 400ms/step.


7


a) Contrast enhancement effects.Contrast enhancement effects are based on the differencein the shape of the scattering curves of different

materials: low scattering materials have a high scatteredintensity in a very small angular interval, while a highintensity can be obtained from highly scattering materialsat larger angles. In figure 8 we show the reconstructed tomograms of thesame test object shown in fig. 6. The images wereacquired at 17 keV. The ratio of the transmitted intensityby the analyzer over the primary intensity,corresponding to the position on the rocking curve, isindicated below each image. The delivered MGD12 is 1.5mGy for all images.The water detail, which is not visible in the scatter-free

tomogram (on the top of the rocking curve), can bedetected when working on the slope of the rocking curveand becomes more visible as the transmitted/primary

intensity ratio is made smaller: since BR12 scatters morephotons at a wider angle, if we introduce a largemisalignment between the two crystals, a high intensityfrom BR12 can still be detected, while the intensity fromthe water detail is very low.

b) Edge enhancement effects.When imaging a sample with a very poor absorptioncontrast but very sharp edges, this can be detectedbecause of the phase shift introduced on the incidentwave by the sharp refractive index variationscorresponding to the detail edges: the so called phase

Fig. 8. DEI tomograms acquired at 17 keV: a) "top" of the rocking curve; b) 3% transmitted intensity; c) 0.3% transmitted intensity.

Fig. 10. Detail from images a test object showing a 100 µm wire: a) transmission; b) positive misalignment; c) negative misalignment.

8



contrast imaging is based on this principle. In theparaxial approximation [16], the phase shift for a ray paththrough an object is proportional to the integral of therefractive index along the beam direction, and theangular deviation is directly proportional to the gradientof the phase shift in the direction transverse to thedirection of propagation.DEI images can also be described by means of thisprinciple.As shown in figure 9, a thin beam impinging on a nylonwire can give rise to different effects, since the angulardeviation becomes as large as the distance from thecenter of the wire increases: when the beam is impingingon the top of the wire (a), a high intensity from the wirecan be detected when the misalignment of the analyzer ispositive, a small intensity is detected when the twocrystals are pefectly aligned and the signal is completelyremoved by a negative misalignment. On the other hand,when the beam strikes the centre of the wire (b), a highintensity is detected when the two crystals are aligned,and the detected intensity is smaller when moving on theslopes of the rocking curve of the crystal.The effects are shown clearly in figure 10: figure 10.ashows the detail of a transmission image of a custommade test object with a 100 µm nylon wire. The wire isvisible, but no information about its orientation can beobtained.Figure 10.b and 10.c are acquired with a 100 µm beamand two opposite positions of the analyzer with respectto the beamline monochromator: in figure 10.b themisalignment is positive, while in figure 10.c it isnegative. The total intensity diffracted by the crystal is10% of the primary one. A higher intensity comes fromthe zone called "A" with a positive misalignment of theanalyzer (figure 10.b), and the signal is completelyremoved in the region named "B".The opposite situation takes place with a negativemisalignment (figure 10.c). This demonstrates that thebeam is impinging on the upper part of the wire in A, andon the lower part in B. Hence, this gives us aninformation about the orientation of a structure, notobtainable from the transmission image.The delivered dose is the same for the transmission andfor the DEI images.

ConclusionsAt the SYRMEP beamline we have built a setup whichallows the acquisition of both conventional anddiffraction enhanced tomographic images, with aparticular attention to possible applications tomammography. Conventional (absorption) tomogramsallow a good visibility of small size and low contrastdetails, while diffraction enhanced tomograms give

further information about the scattering properties ofmaterials and about the orientation of structures. Thedelivered dose is comparable to the one delivered inclinical mammography, thus opening the way to clinicalapplications of both techniques.

References[1] Shaw De Paredes E, Radiographic breast anatomy: Radiologic signs of

breast cancer, in Syllabus: a categorical course in physics - Technicalaspects of breast imaging, M Yaffe ed., Oak Brook, IL, RSNAPublications, 1993, 35-46.

[2] Arfelli F et al. The digital mammography program at the SR lightsource in Trieste. IEEE Trans Nucl Sci 1997: 44; 2395-2399.

[3] Arfelli F et al. Low dose phase contrast x-ray medical imaging. PhysMed Biol 1998: 43; 2845-2852.

[4] Arfelli F et al. Improvements in the field of radiological imaging at theSYRMEP beamline. SPIE 1999: 3770; 2-12.

[5] Salome M et al. A synchrotron radiation microtomography system forthe analysis of trabecular bone samples. Med Phys 1999: 26; 2195-2204.

[6] Dilmanian FA et al. Single-and dual-energy CT with monochromaticsynchrotron x-rays. Phys Med Biol 1997: 42; 371-387.

[7] Beckmann F et al. X-ray microtomography (µCT) using phase contrastfor the investigation of organic matter. Journal of Computed AssistedTomography 1997: 21(4); 539-553.

[8] Kak AC, Slanley M. Principles of Computerized TomographicImaging. New York. IEEE Press.

[9] Arfelli F et al. At the frontiers of digital mammography: SYRMEP. NuclInstr Meth in Phys Res A 1998: 409; 529-533.

[10] Colledani C et al. CASTOR: a VLSI CMOS analog-digital circuit forpixel imaging application. Nucl Instr Meth in Phys Res A 1997: 395;435-442.

[11] Mammography: a users’ guide. NCRP Rep. No. 85. Bethesda, Md.NCRP 1985.

[12] Rose A. Vision : human and electronic. New York. Plenum 1973. 21-23.[13] White DR et al. Epoxy resin based tissue substitutes. Br J Radiol 1977:

50; 814-821.[14] D. Chapman et al. Diffraction enhanced X-ray imaging. Phys Med

Biol 1997: 42; 2015-2025.[15] Arfelli F et al, Mammography with Synchrotron Radiation: Phase

Detection Techniques. Radiology 2000; 215: 286-293.[16] S. W. Wilkins et al. Phase-contrast imaging using polychromatic hard

X-rays. Nature 1996: 384; 335-338.

9


NEUTRON REFLECTIVITY STUDIES OF IMMISCIBLEPOLYMER/POLYMER INTERFACESMichele SferrazzaDepartment of Physics, University of Surrey, Guildford,Surrey GU2 5XH, U.K.

Articolo ricevuto in redazione nel mese di Ottobre 2000

AbstractIn this paper, a detail study of the interfacial properties betweentwo immiscible polymers is reported. The structure of theinterfacial width between immiscible polymer pairs is studiedusing neutron reflection. Neutron reflectivity, probingdifference in the density profile perpendicular to the plane of thesample, is ideal for this type of studies. Deuterated polystyrene(d-PS) on poly(methyl methacrylate) (PMMA) over a range ofthickness and molecular weights were used. A logarithmicdependence of the interface width on film thickness is observed,characteristic of an interface broadened by thermal inducedcapillary waves, whose spectrum is cut-off by dispersion forcesacross the polymer layer. If the system is inverted, a thin d-PMMA layer on a thick PS film, the Hamaker constant of thesystem becomes negative and dispersion forces will tend todestabilise the interface. In this case, thermally capillary wavesare amplified and grow until the top film dewets. This case wasstudied with both specular and off-specular neutron reflections.

IntroductionPolymer surfaces and interfaces play an essential role inmany commercial applications of polymers, such ascoatings, adhesives, blends and resists. Understandingthe microscopic processes taking place at interfaces is,therefore, increasingly important for the wide ranginguses of polymers in many industrial applications. Thenature of the interface between immiscible polymers isimportant to investigate, both because these interfacesprovide model systems to elucidate fundamentalproblems of the statistical mechanics of surfaces andinterfaces, and because an understanding of themicroscopic structure of the polymer interface will help toaddress technological questions connected to theadhesion of polymers and the properties of multiphasepolymer systems.The interface between two immiscible polymers is notatomically sharp. This is because the unfavourableenthalpy of mixing that occurs at a diffuse interface isoffset by a gain in chain entropy. The self-consistent fieldtheory predicts the width and the interfacial tension foran immiscible polymer pair. The volume fraction profileof one components through the interface is predicted totake the hyperbolic tangent profile form, the interfacialwidth varies as χ--1/2 for small χ (χ is the Flory-Hugginsinteraction parameter), and the surface energy isproportional to χ1/2 [1]. The experimental interfacial

width obtained for incompatible polymers is in generalbroader than the value predicted by the self-consistentfield theory developed by Helfand and Tagami [1]. Anunderstanding of the reason for the discrepancy isfundamental for developing a better understanding of theproperties of polymer interfaces.The experiments performed on the PS/PMMA bilayersystem described in this paper show the importance oflong-range dispersion interactions across a thin polymerfilm in modifying the structure of the interface betweenthe two polymers. While for a thin PS film on a thickPMMA substrate, the long range van der Waals forcestend to stabilise the interface and an equilibriuminterfacial width is obtained, for some cases thedispersion forces will instead tend to destabilise theinterface and the dewetting of the top layer on thesubstrate film will take place. This process is caused bythe instability of the film against thermally excitedcapillary waves at the interface and/or surfaces. For theliquid/liquid system, the instability of the films againstcapillary wave fluctuations at the interface is importantand difficult to study in detail, since the interface ofinterest is buried. We have studied this case of dewettingusing specular and off-specular neutron reflection.

Neutron reflectionReliable experimental measurements to test in detail thetheoretical prediction for the interfacial width betweentwo polymers have become available quite recently, withthe technique of neutron reflection providing the mostaccurate information. The wavelength of cold neutrons isof the order of a few tenths of manometers, and this setsthe length-scales probed by reflection experiment.Typically the experiment is sensitive to structural featuresperpendicular to the plane of the film with length scalesbetween 0.5 and 50 nm [2,3,4]. Two other advantages ofneutrons make them particularly suited for studyingorganic film: their penetration power is larger than forexample, X-rays, and so it is straightforward to studyburied interfaces, and the great difference in neutronscattering length density between deuterium andhydrogen. When neutrons propagate through a mediumin which the scattering centres are small compared withthe wavelength of the neutrons, the effect of the mediumcan be represented by a pseudo-potential. This pseudo-potential has a magnitude that is related to the scattering


10


length density of the material by

where b is the sum of all the scattering lengths in avolume V. The neutrons obey to the Schrödingerequation and if the potential is a function of only onedirection, z, the perpendicular component of the motionthen obeys a one-dimensional equation

where E = (h–2 k2)/em, and k is the perpendicularcomponent of the wavevector of the incoming beam. Thisequation can be solved to give the amplitude reflectancer and then the reflectivity R = |r|2. If neutrons propagatefrom vacuum into a uniform material of scattering lengthdensity b/V, the perpendicular component of thewavevector within the material ki is expressed by

In the case 4πb/V > k2 , ki is imaginary and then theneutrons propagate into the material only as anevanescent wave, giving rise to total external reflection(reflectivity is unity). In the case of 4πb/V < k2 and a sharpinterface (between vacuum and medium), then thereflectivity is given by the Fresnel expression

that, for high value of k, has the limiting form of

If the interface is rough or graded, the reflectance of theinterface is modified and a roughness factor is introducedthat describes the interface, given by

where σ is the Gaussian width of the interface. Thereflectivity of a multilayer of stack of thin films can becalculated by the following recursive scheme. For eachslab i and i-1, the reflectance denoted by ri-1,i , is given bythe Fresnel expression:

The combined reflectance of the interfaces between thesubstrate and layer n-1, and layer n-1 and n-2, denoted byrn-2,n, is given by the combination of these individualinterfaces:

where d is the thickness of the slab. By combining thisreflectance with the reflectance of the interface betweenlayers n-3 and n-2 to yield the reflectance of all interfacesfrom n-3 to the substrate, the process can be continuedrecursively until the top interface is reached andtherefore the reflectivity is obtained. This algorithmic canbe applied to calculate a general profile. Byapproximating a continuous profile by a stack of suchthin layers (normally one chooses layer thickness to givea constant increment in scattering length density betweeneach layer), one can calculate the reflectivity of anyprofile to the accuracy required within a given k range.The continuum limit of this expression, in the limit ofhigh k, is

Thus, the reflectivity is simple related to the Fouriertransform of the derivative of the scattering lengthdensity profile.In practice data analysis has a number of difficulties.Since only the potential V determines the density profile,if the sample contains more than two unknowncompositions then the composition profile cannot beuniquely determined.Also, there is no reliable way of inverting a neutronreflectivity profile to recover a unique potential profile V.Instead a trial function needs to be used as a startingpoint and its parameters refines until a best fit is achievedbetween the experimental and predicted reflectivityprofiles. This implies that neutron reflectivity is notsuitable to analyse films of unknown composition; itshould instead be considered as a powerful tool forcharacterising the structure of thin films at highresolution when their unknown morphology is alreadyknow at some coarse level and one has some knowledgeof the materials from which they are composed.Neutron reflection is at its most powerful when samplesare specifically made for the technique, using inparticular a scheme of labelling selected componentswith deuterium substituted for hydrogen. For neutronreflectometry, in fact, the contrast can be generated by thedeuteration of one component. This consists of exchange,for example in polymers, of the hydrogen in themolecule's chains with deuterium.Figure 1 shows a schematic diagram of the CRISPneutron reflectivity instrument at the ISIS spallationfacility, Rutherford Appleton Laboratory (UK). A whiteneutron beam, with a wavelength range of order 0.5-6.5Å is incident on a sample at grazing angle of incidence,which is typically between 0.25 and 1.5 degree. These

r

k

d b V

dze dzikz( )

( / )0 2

2= ∫π

r

r r e

r r en n

n n n nik d

n n n nik d

n n

n n−

− − −

− − −=

++

− −

− −2

2 1 12

2 1 12

1 1

1 11,

, ,

, ,

r

k k

k ki ii i

i i−

−

−= −

+11

1,

r r eFresnelkki= −2 2σ

Rb

V k~ π2

2

41

R rk k

k ki

i= = −

+2

k kb

Vi = −2 4π

−

+ =h2

2

2

2m

d

dzVz E

Φ Φ

Veff =2 2πh m

bV


11


give a q range between 0.008 Å-1 and 0.3 Å-1. The neutronbeam is inclined at 1.5o to the horizontal, which allowseasier study of liquid surfaces. Fine collimation isobtained by two slits before the sample, which define theresolution and illuminated area of the sample position.There are another two slits before the detector position tosuppress the background noise. Neutrons specularly

reflected are detected and sorted by wavelength usingtheir time-of-flight. If one is interested only in thespecular reflection a single 3He detector can be used, butoften a liner or two dimensional position sensitivedetectors is used which has the advantage that theintensity of neutron in the off-specular direction can bestudied.This allows the investigation of in-plane correlation.After data processing, the intensity of the reflectedneutrons are displayed as a function of the wavelength,or more normally as a function of the change in wave-vector on reflection – the scattering vector q, which isrelated to the wavelength of the neutron λ and theincident angle ϑ by the expression

A study of the PMMA/PS interfaceThe main result of the self-consistent theory is that theinterfacial width between two immiscible polymersvaries as χ--1/2 . Measured values of the polymer-polymerinterfacial width, obtained with the neutron reflectiontechnique for various types of polymer interfaces,ranging from block copolymers to polymer brushes inpolymer matrices, are typically higher than the valuesextracted from the self-consistent field theory. For thesystem of a polystyrene (PS) and poly(methyl

methacrylate) (PMMA) blend, the discrepancy betweenthe theory and the experimental result is still present. Forthis case, the theoretical interfacial width w between

PS/PMMA can be estimated using the expression .

With a reference volume equal to the volume of a PS

repeat unit, with χ=0.04, and with the statistical segmentlengths for PS and PMMA as a(PS) = 6.7 Å anda(PMMA)=7.5 Å, the calculated interfacial width is 29 Å[1,5]. Experiments by different groups gave a value forthe interface width between PS and PMMA of around 50Å, independently of the molecular weight (Mw) and filmthicknesses used [6,7]. This result for the interfacial widthbetween immiscible polymers is not reproduced by theself-consistent field theory. A possible explanation hasbeen suggested recently: the experimental result agreeswith the theoretical prediction if a correction to theinterfacial width due to capillary wave fluctuations isconsidered [5].To study in detail the contribution to the interfacial widthgiven by the capillary wave term, neutron reflectionexperiments using the reflectometers CRISP and SURF atthe Rutherford Appleton Laboratory have beenperformed [8]. Bilayers of d-PS on h-PMMA wereprepared by spin-casting a PMMA layer onto a siliconsubstrate, producing a thick film between 4000 Å to 9000Å. The silicon substrates used were disks polished on oneside, of diameter 5 cm and thickness 0.5 cm withorientation (111). The d-PS film was first spun-cast on aglass slide and then floated onto the PMMA. Thethickness of the top d-PS layer was varied between ~50 Åand ~20000 Å. All the film thicknesses were alsomeasured using ellipsometry. The samples prepared in

q = 4π ϑ

λsin


Fig. 1. Schematic diagram of a neutron reflectometer - CRISP, at the Rutherford Appleton Laboratory.

12


this way were then dried at 60 oC in a vacuum oven forsome time and then annealed at 120 oC for 6 hours: thistime was found to be sufficient to reach equilibrium, asneutron reflectivity experiments at different annealingtimes have shown. All the polymers were obtained fromPolymer Laboratories (UK) and were prepared byanionic polymerisation with Mw/Mn of 1.1 or less. TheMw of the polymers used was ranging from 200K to400K. Neutron reflectivities were measured for both theunannealed and the annealed samples.Figure 2 shows the reflectivity curves for a system as anexample. For all pairs used, the film thickness of thebottom PMMA layer was quite thick and the filmthickness of the d-PS layer was varied: the reflectivitycurves in fact showed interference fringes characteristicof the d-PS layer thickness and not of the PMMA layer.The solid lines in figure 2 are fits to the reflectivity curves

obtained by a least-squares fit to a multilayer model withGaussian roughness introduced at the surface and filminterfaces. The parameters involved in the fit aremeasured independently, except for the roughnessbetween the d-PS and PMMA layer, which is theparameter fitted. The silicon dioxide layer thickness,PMMA layer and d-PS layer thickness were alsomeasured by spectroscopic ellipsometry. The surfaceroughness was kept constant during the fitting at a valueof 7 Å: this value has been extracted by independentmeasurements and is in line with literature data [8]. Theresults of the interfacial roughness extracted by the fittingof the reflectivity curves for all pairs measured, areplotted in figure 3 as a function of the thickness of thed-PS layer.

The interfacial width parameter ∆ plotted in figure 3 isconnected to the hyperbolic tangent profile 2w by

, therefore a ∆ value of 20 Å corresponds toa hyperbolic tangent value of 50.1 Å, in very satisfactoryagreement with the values obtained by other workers. From figure 3 it is clear that for thinner d-PS films thereis a thickness dependence of the value of the interfacialwidth with the film thickness, and it is at leastapproximately logarithmic. This logarithmic dependencepersists to up d-PS layer thickness of ~1000 Å, afterwhich the data clearly levels off to a value of ~20 Å. Toanalyse the result expressed by figure 3 morequantitatively, the assumption that the contributions tothe interfacial width due to the intrinsic interface and tothe capillary wave broadening add in Gaussianquadrature is taken [9,10].The theoretical total interfacial width is thereforeexpressed by [8]:

where ∆0 is the intrinsic width, λcoeh is the neutroncoherence length and adis is a dispersive capillary length.The second term is the capillary wave contributionimaging the interface as if it were a membrane in a stateof tension characterised by a bare interfacial tension γthat sustains a spectrum of waves, each of whose averageenergy is determined by equipartition of energy. Sinceone of the polymer is rather thin, the dispersive or Vander Waals forces across the film are important and areresponsible of the dispersive capillary length in the aboveexpression given by [8]

where l is the top film thickness and A is the Hamakerconstant for the interactions between the substratepolymer and the air across the thin film. The Hamakerconstant can be estimated from refractive index anddielectric constant data using an approximation based onLifshitz theory [11].For our system a value of 2⋅10-20 J is obtained. For film inthe thickness range of 50 Å to 500 Å, the dispersivecapillary length may be estimated as falling between 300Å and 3 microns. Thus for films in this range of thickness,this dispersive capillary length rather the neutroncoherence length dominates the capillary wavesexpression. This explains the logarithmic behaviour ofthe width up to ~1000 Å and the level off of the widthforth thicker d-PS thickness, the region where the neutroncoherence length is dominant. Using the expression, thedata of figure 3 have been fitted varying the intrinsicinterface width ∆0 and the interfacial tension γ. The solid

al

Adis2

44= πγ

∆ ∆ ∆ ∆

∆2 2 2 22

2 24

2

2 2= + = + ( )

( ) + ( )o oB o

coeh dis

k T

aζ

πγπ

π λ πln

/

/ /

∆ = 2 2w / π


Fig. 2. Neutron reflectivity curves as a function of the transfer momentumQz for d-PMMA/d-PS pairs. The bottom layer was thick ~ 5000 Å and thed-PS layer was varied from 60 Å to 20000 Å. The solid lines are fits to thedata as described in the text. The curves are shifted down by factors of 100from each other for clarity.

13


line in figure 3 is the fit obtained. The value of theneutron coherence length has been fixed at 20 µm. Thefitted parameter values are ∆0= 9.3 ± 1.4 Å and γ=2.7 ±0.3 mJm-2. It should be recognised that the estimate of the Hamakerconstant A is quite uncertain: in fact over the range of thethickness used in our experiments, retardation effects arelikely to become important, and so the effective Hamakerconstant will also depend on the thickness. Luckily sinceA appears in the logarithm, the results are ratherinsensitive to the value chosen for A: this is demonstratedby the fact that a change of 50 % on the value for theHamaker constant results in a change of 10 % on the bestfit of ∆o, and has even less effect in the best fit value of γ. The values of the interfacial width ∆0 and interfacialtension γextracted from the best fit can be compared withthe ones predicted by the self-consistent field theory:

where a is the statistical segment, χ the interactionparameter, and 1/ρ is the volume of the monomer repeatunit. For the temperature used in the experiments described inthis section, the value of the Flory-Huggins interactionparameter is χ = 0.037 ± 0.002 [12]. Using this value for χ,

a statistical segment length a=7 Å and a volume of themonomer repeat unit which takes a value of 174 Å3, ∆0

and γ as predicted by the self-consistent field theory are∆0= 11.8 ± 0.6 Å and γ=1.7 ± 0.05 mJm-2. Therefore thecomparison between the deduced bare interface width∆0 extracted from the experimental data and the valuedetermined from the self-consistent field theory is

excellent, while our deduced interfacial energy issomewhat larger than the one predicted. However theapproximation implicit in using the capillary waveexpression to fit the data has to be considered. Forexample in using the equation, the assumption that thefree surface of the d-PS film behaves like a rigid wall istaken, whereas it will have its own spectrum of capillarywaves that will be coupled to the waves at the interface. But, since the surface energy of d-PS is about an order ofmagnitude larger than the d-PS/PMMA interfacialenergy, the rigid wall assumption is a reasonablestarting point.

Early stage of spinodal dewetting for the PS/PMMA systemAs we have seen in the previous section, if the bottomlayer is a thick PMMA film on top of a silicon substrateand the top layer is a thin PS film, the Hamaker constantfor this system is A=2⋅10-20 J. The inverted system, the PSis a thick layer now at the bottom (on a silicon substrate)and the top layer is a thin PMMA layer, has a negativeHamaker constant of A=–1.7⋅10-20 J. In this case,dispersion forces amplify thermally excited capillarywaves at a fluid interface until dewetting takes place: thisprocess is known as spinodal dewetting. The study of theprocess of dewetting in thin polymer films, because of theimportance of thin films and coatings in technology, hasreceived much attention recently [13-19]. The excess freeenergy per unit area due to a sinusoidal perturbation ofwave vector q and amplitude ζq is [20]. Thus ifthe Hamaker constant is positive capillary waves lead toan increase in the system's free energy - dispersive forcesin this case stabilise the interface against capillary wavefluctuations.On the other hand, if the Hamaker constant is negative,the growth of long wavelength capillary waves leads to alowering of the system's free energy and the onset ofinstability, with the wave vector for marginal stability qc

given by qc=( A /2πh4σ)1/2 where σ is the interfacialtension.At early times the dynamics of the instability can besolved in the linear approximation [21]; for a thin liquidfilm A on a liquid substrate B [22] the mode that leads todewetting is called 'peristaltic mode': in this case thedisplacements of the free surface and the fluid/fluidinterface are in antiphase.Moreover the polymer/polymer interfacial tensions aretypically an order of magnitude smaller than the polymersurface tensions, so for a polymer film on a polymersubstrate it is the interfacial tension that dominates.Similarly, in a pure peristaltic mode the ratio of theamplitudes of the displacement of the surface and thedisplacement of the interface is equal to the ratio of thesurface tension to the interfacial tension so in apolymer/polymer system we expect the displacement of

A hqζ π2 48/

γ ρ χ= a k TB 6


Fig. 3. The interfacial width parameter ∆ for the d-PS and PMMA interfaceas a function of the d-PS film thickness, as measured by neutronreflection. Different symbols are for different systems.

14


the surface to be much smaller than the displacement ofthe polymer/polymer interface. In the liquid/liquid case,the expression for the rise time of the unstable mode isdifferent to that for a film on a solid substrate (dissipationin the substrate becomes important or dominant); it takesthe form

where L the thickness of the substrate film [22].Spinodal dewetting has been predicted to be the maindewetting process for ultra-thin films with a negativeHamaker constant. The main predictions of the theoryare that the wavelength of the fastest growing wave atthe interface or surface is proportional to the square ofthe thickness of the film; and that the rising time of theamplitude of the wave increases with the inverse of thenth power of the thickness, where n depends upon thetype of substrate on which the thin film is deposited.For the Si/PS/PMMA system, we then expect that thefluctuations at the interface between the two polymersshould now be amplified until rupture of the film takesplace. The characteristic wavelength of the fast mode is:

If the expression for a=( A /6πσ)1/2 is recalled and athickness of 100 Å is considered, the fast mode has awavelength of the order λ~1µm, where a value of 2mJm-2

for the interfacial tension between PS/PMMA is taken. The interface under interest is buried, and the only wayof monitoring the development of the unstable mode atthe interface between the two polymers is the study ofoff-specular neutron scattering. Off-specular scattering isin fact sensitive to the in-plane structure of the surface orinterface, whilst specular scattering is sensitive to thechange of refractive index in the direction perpendicularto the interface.The characteristic size of the fluctuations in the plane ofthe interface can be extracted from off-specular neutronanalysis. The change of its intensity with time can beconnected with the characteristic rising time of theunstable mode. From the theory introduced previously,the wavelength of the unstable fastest mode and itscharacteristic rising time depend strongly on thethickness of the top layer. To study this effect in detail,samples with different top film thickness were used. Theneutron reflection experiments were performed using thereflectometer D17 at ILL (Grenoble, France) and SURF atthe RAL (Didcot, UK) [23]. Bilayers of PS/PMMA wereprepared: the top d-PMMA layer had a thickness in therange of 95 to 150 Å, while the bottom h-PS layer had athickness around 1300 Å. The annealing temperature wasfixed at 155 oC. For each sample, the reflectivity was

measured first. After measuring the reflectivity, for eachsample long scans were taken for study of the off-specular scattering at the same qz value. The resolutionwas relax in the vertical direction and very well definedin the horizontal direction. The intensity of the off-specular scattering as a function of the detection anglefrom the specular position was extracted. Reflectivity andoff-specular scattering measurements were measured fordifferent annealing times. On D17 the diffuse scattering

was acquired for each sample at a fixed qz value, while on SURF the off-specular neutron scattering was measuredat a fixed angle of 1.2o. The qz value at which the longmeasurements on D17 were performed was qz=0.0127Å-1. For each annealing time, the reflectivity was measuredfirst. Figure 4 shows the reflectivity curves for a samplewith top d-PMMA thickness of 95 Å for differentannealing times. The solid lines for the first 4 curves arefits to a 3-layer model, with roughness parameters both atthe surface and interface. Clear features were visible fromthe reflectivity measurements: for the unannealed sampleand up to 6 hours of annealing time, interference fringesdue to the bottom layer were visible, corresponding to athickness of around 1300 Å; fringes corresponding to thetop d-PMMA layer were also clearly visible. The intensityof these fringes decreased with increasing annealing timeand they were still present up to 6 hours, although quitereduced in intensity.After 6 hours, a clear difference in the reflectivity wasvisible, with the top d-PMMA fringes disappearing. Thefits with a multilayer model were good up to 6 hours ofannealing time. The solid line for the 6 hours reflectivity

q

a

lm ~ 2

τ η σ

mBh

L A~

6

2


Fig. 4. Reflectivity curves for the 95 Å samples for different annealingtimes, from unannealed (top curve) to 24 hours (bottom curve). Fits to thedata for some curves are also shown and described in the text. The datahave been shifted by factors of 100 with respect to each other for clarity.In the inset the surface and interface roughness are reported as well thesurface roughness of a contrast match experiment (see text for details).

15


was the best fit obtained with the model used. From thefitting of the reflectivity data, the values of the surfaceand interface roughnesses were extracted and arereported in the inset to the figure. It is possible to observethat the top surface roughness was around 8 Å up to 2hours of annealing.The reflectivities for the 3 and 6 hours annealing timesamples were fitted if the top roughness was increased to13 Å and 18 Å respectively. The surface and interfacialwidths, as compared with a hyperbolic tangent profilewidth, assumed values of 45 Å and 48 Å for the 6 hoursannealing time sample, comparable to the total thicknessof the top d-PMMA film. In the fitting procedure thesurface roughness and interface roughness are correlated,so to be confident that it was the interface roughness thatwas undergoing the largest increase we performed ananalogous experiment in contrast matched conditions.We made a PS/PMMA bilayer where the top 100 ± 6 Åfilm was a mixture of h-PMMA and d-PMMA whosecomposition was such that the top layer had the samescattering length density as PS.For this sample no reflection at all is expected at thepolymer/polymer interface and the data fitting issensitive only to the surface roughness, not to theroughness of the polymer/polymer interface. The fittedsurface roughness values as a function of time (shown in

the inset to figure 4) confirmed that the loss of visibilityin the fringes seen for the non-contrast match sample isdominated by the roughening of the polymer/polymerinterface. This was also confirmed by an Atomic ForceMicroscopy (AFM) study [23]. Thus even at early times,

long before gross dewetting is visible using opticalmicroscopy and while the surface remains essentiallysmooth, the growing capillary waves that ultimately leadto dewetting manifest themselves in a continuallyincreasing roughness of the polymer/polymer interface.Figure 5 shows as another example the reflectivity curvesfor samples with top d-PMMA film of thickness of 110 Åcorresponding to the unannealed, 6, 12, 24 and 72 hoursof annealing time. The fringes due to the top 110 d-PMMA layer were clearly present up to 12 hours ofannealing, and still visible after 24 hours of annealingtime. However, they instead disappeared for the 72 hourscase. This was different from the previous case, where thefringes due to the 15 Å smaller d-PMMA filmdisappeared after just 6 hours of annealing. In the inset offigure 5, the surface and interface roughnesses are plottedas a function of the annealing time. In the inset also thesurface roughness for the contrast match experiment isalso reported.The interface roughness that is probed by specularneutron reflectivity represents a combination of theintrinsic diffuseness of the interface with a contributionfrom capillary waves integrated over all possiblewavelengths up to an upper limit imposed by the lateralcoherence length of the neutron beam, which in thisgeometry is of the order of 20 µm.Clearly in order to provide a more searching test of thetheory it would be desirable to study the growth ofcapillary waves in a way that discriminates betweenmodes of different wavelength. To do this we turned tomeasurements of the intensity of neutrons scattered outfrom the specular beam. A quantitative analysis of thiscan be achieved if a long scan at a fixed qz is performedfor all the samples, which have been annealed fordifferent lengths of time. The intensity of the diffusescattering as a function of the angle from the specularposition, or the in-plane transfer momentum, can beplotted for the long scan measurements.Figure 6 reports the diffuse scattering for differentannealing times as a function of the angle from thespecular position for the samples with top thickness of 97Å. The measurements reported in figure 6 were taken ata constant qz=0.0127 Å-1, and the intensities werenormalised to the specular intensity. The data for theunannealed sample shows a rapid and monotonic fall ofintensity from the specular peak; after annealing,however, a prominent shoulder appears which grows insize with increased annealing time. The top inset plots infigure 6 shows the excess scattering relative to theunannealed sample plotted against the scattering angle.These show that there is a peak in intensity at a specificscattering angle, which for this case correspond to ascattering object with a wavelength of ~1 µm.The intensity of the scattering peak grows with annealing


Fig. 5. Reflectivity curves for the 110 Å samples for different annealingtimes, from unannealed (top curve) to 72 hours (bottom curve). Fits to thedata for some curves are also shown and described in the text. The datahave been shifted by factors of 100 with respect to each other for clarity.In the inset the surface and interface roughness are reported as well thesurface roughness of a contrast match experiment (see text for details).

16


time, and there are indications that at later times the peakmoves to smaller wavevectors. We interpret the data asarising from diffraction from capillary waves at thepolymer/polymer interface, with the peak arising fromscattering from the fastest growing capillary wave. Theanalogy with bulk spinodal decomposition is obvious.Once again, we used contrast matching to confirm our

supposition that the scattering is dominated by thepolymer/polymer interface; when the PMMA film iscontrast matched to the PS film, so no scattering occurs atthe polymer/polymer interface, we saw no particularchange in the off-specular intensity as a function of theannealing time.The off-specular neutron scattering peaks observed atearlier times, when the surface was still flat, weretherefore connected to the instability of the PS/PMMAinterface. In the case studied here, considering that anestimation of the wavelength characteristic of the fastestunstable mode for the PS/PMMA case is 1 µm (as hasbeen shown previously), the case where a dependence ofthe top PMMA thickness of l6 is expected for the risingtime of the unstable wave. It is worthwhile to mentionthat the theoretical approach is valid for the early stagewhich leads to the rupture of the film, and off-specularscattering is a powerful technique to investigate this for aburied interface. The characteristic wavelength for the instability at earlytimes, extracted from the off-specular neutron scattering

for samples with various PMMA thickness is reported inthe bottom inset to figure 6. The dashed line correspondsto a quadratic dependence of the characteristicwavelength λm on the film thickness h:λm =bh2 , where thebest fit value of b is 1.1(± 0.1) 1010 m. We expect such aquadratic dependence from theory; this also predicts forthe value of the pre-factor b=a(π3σ/ A )1/2 [22]. The value of the PS/PMMA interfacial tension calculatedfrom self-consistent field theory is 2.0mJm-2 and thesurface tension of PMMA at 150 oC is 31mJm-2, so we candeduce a value of the Hamaker constant A= -0.8⋅10-20 J. This is of a plausible order of magnitude, though it issomewhat smaller that our estimate of A= -1.7⋅10-20 Jderived from an approximation of Lifshitz theory.Nonetheless, given the uncertain inherent in making thiskind of estimate this degree of agreement is satisfactory.One point that may be important is that for these filmthicknesses retardation effects should begin to besignificant, decreasing the effective Hamaker constantwith increasing thickness, and producing a slightlyweaker than quadratic dependence of characteristicwavelength on film thickness.Finally we consider the time dependence of the

interfacial instability. An important features of spinodaldewetting is the very strong dependence of the rise timeof the interfacial instability on film thickness; it is thisstrong thickness dependence of the kinetics of dewettingthat clearly distinguishes spinodal dewetting from, forexample, nucleation and growth of holes. Figure 7 showsthe relative intensity at the maximum of the off-specularpeak as a function of film thickness for three filmsmeasured on D17. At early times the growth of intensity


Fig. 6. Diffuse scattering for different annealing times as a function of theangle from the specular position for the sample with top d-PMMAthickness of 95 Å. The legends in the figures distinguish differentannealing times. The Qz value for these measurements was 0.0127 Å-1.The top inset shows the excess scattering for the different measurementsshowing a characteristic wavelength that grows fastest. In the bottominset the characteristic wavelengths as a function of the d-PMMAthickness extracted form the diffuse scattering are also shown (see text fordetails).

Fig. 7. Diffuse scattering intensity as a function of annealing time for the3 different thickness systems: 95 Å, 110 Å, and 147 Å. In the inset thecharacteristic rise times of the interfacial instability are shown as afunction of the d-PMMA film thickness. The line is the dependence of therise time to the sixth power on the thickness.

17


is exponential, saturating at later times. The rise times areshown in the inset. They show the very strongdependence of rise time on film thickness characteristicof spinodal dewetting; a 50 % increase in film thicknessleads to a 770 % increase in rise time. Of course, given thesparsity of the data it is not possible to test predictions forthis dependence quantitatively, but the inset shows thecurve for dependence on the sixth powers of thickness, aspredicted for liquid/liquid dewetting. In the future moreexperiments and a more detailed analysis of the off-specular scattering should be possible, allowing one totest the predictions of a linear theory for the growth ratesof capillary wave modes of arbitrary wavevector and toprobe the breakdown of the linear approximation.

ConclusionsIn conclusion we found that the interfacial widthbetween two immiscible polymers has a logarithmicdependence on film thickness, indicating that capillarywaves contribute substantially to the interfacial width asmeasured by neutron reflectivity.When corrected for the effect of capillary waves, there isreasonable quantitative agreement between thepredictions of the self-consistent field theory andexperimental measurements. We have also identified andcharacterised the earlier stages of the interfacialinstability that leads to the dewetting of one polymer filmby another. Dispersive forces amplify thermally excitedcapillary waves at the polymer/polymer interface; wehave monitored the growth of these waves both bymeasuring a continual increase in the overall diffusenessof the polymer/polymer interface and by detectingneutrons diffracted in grazing incidence from the fastestgrowing capillary wave modes.Our results are consistent with the predictions of a lineartheory of spinodal dewetting at a liquid/liquid interfaceboth in respect of the length-scale and the characteristictime of growth of the unstable capillary wave.

AcknowledgementThe results reviewed in this paper were the outcome ofenjoyable collaborations with R.A.L. Jones (University ofSheffield) and C. Xiao (University of Surrey).The author also would like to thank R. Cubitt (ILL), D.Bucknall, J. Webster and J. Penfold (RAL) for their helpduring the experiments.

References[1] E. Helfand and Y. Tagami, J. Chem. Phys. 56, 3592 (1972)[2] T.P. Russell et al., Mater. Sci. Rep. 5, 171 (1990)[3] R.W. Richards and J. Penfold, Trend Poly. Sci. 2, 5 (1994)[4] M. Stamm and D.W. Schubert, Ann. Rev. Mater. Sci. 25, 325 (1995)[5] K.R. Shull et al., Macromol. 26, 3929 (1993)[6] M.L. Fernandez et al., Polymer 29, 1923 (1988)[7] S.P. Anastasiadis et al., Macromol. 92, 5677 (1990)[8] M. Sferrazza et al., Phys. Rev. Lett. 78, 3693 (1997)[9] F.P. Buff et al., Phys. Rev. Lett. 15, 621 (1965)[10] J.S. Rowlinson and B. Widom, Molecular Theory of Capillarity,

Claredon Press, Oxford, (1992)[11] J. Israelachvili, Intermolecular and Surface Forces, Academic Press,

(1992)[12] T.P. Russell et al., Macromol. 23, 890 (1990)[13] G. Reiter, Phys. Rev. Lett. 66, 715 (1991)[14] G. Reiter, Macromol. 27, 3046 (1994)[15] G. Krausch, J. of Phys.: Condens. Matt. 37, 7741 (1997)[16] Lambooy et al., Phys. Rev. Lett. 76, 1110 (1996)[17] S. Qu et al., Macromol. 30, 3640 (1997)[18] Sharma and G. Reiter, J. Colloid. and Interf. Scie. 178, 383 (1996)[19] R. Xie et al., Phys. Rev. Lett. 81, 1251 (1998)[20] T.E. Faber, Fluid Dynamics for Physicist, Cambridge University Press,

Cambridge 1995[21] F. Brochard-Wyart and J. Daillant, Can. J. Phys. 68, 1084 (1990)[22] F. Brochard-Wyart et al., Langmuir 9, 3682 (1993)[23] M. Sferrazza et al., Phys. Rev. Lett. 81, 5173 (1998


18


Inelastic scattering of neutrons by electrons in a solid is anestablished technique by which to investigate processes thatoccur on a scale of energy up to about 0.2 eV, and the processesinclude spin-wave excitations and crystal-field states. Thearticle looks at some examples of what can be gained frominvestigations conducted above this energy and up to severalelectron volts.

OrientationThe aim of this article is to stimulate, or possibly helprenew, interest in the use of energetic neutrons to studyproperties of electrons in solids. To this end, we bringtogether basic features of the scattering of neutrons byelectrons and look at what is expected if the electrons arelocalized in space (e.g. electrons in the f shell of alanthanide ion) or mobile and band-like.Of course, the use of thermal neutrons to study electronsin solids is very well established. Physical properties inthis energy range include crystal-field states andcollective spin excitations, often called magnons or spinwaves.There are at least two practical problems to be facedwhen it comes to using neutrons to investigate propertiesat higher energies, above about 0.2 eV; one problem, asufficient supply of neutrons, in principle is largelyresolved and the second, satisfying kinematics in anexperiment, is handed down by Nature and we mustlearn to live with it. First, one needs a copious supply ofenergetic neutrons, and for these we turn to spallationsources. In the not too distant future, the supply will bevery much improved by the operation of the SpallationNeutron Source under construction in Oak Ridge, USA(Mason et al. 2000) and perhaps also the AUSTRONsource if the current project in Austria is brought tofruition (Rauch et al. 2000). Existing sources of energeticneutrons are more than adequate to make somemeasurements but new sources will surely open newvistas of research. The second problem stems from the huge difference inthe masses of the electron and the neutron. The challengewhich then arises is to satisfy the twin objectives of alarge transfer of energy, exceeding 0.2 eV, say, and only amodest change in the wavevector so as not to undulysuffer a shortfall in scattered intensity that will comeabout because of a monotonically decreasing form factorin the scattering length, which at a wavevector around

4πÅ-1 might achieve a magnitude that is a few percent ofits maximum value. With regard to this second problemone can realistically be optimistic about future activitiesbecause of what has been achieved with existinginstrumentation. In what follows, we do not dwell on thechallenge to the experimentalist of detecting neutronsdeflected through small angles.The emphasis in this article is on processes that occur atan energy of about 0.2eV and beyond. In particular, wehave little to say about crystal-field states even thoughthey are an active field of research (Mesot 1995, Loveseyand Staub 2000, and Staub and Soderholm 2000).

Neutron-electron interactionThe contribution to the scattering length of the neutrondue to its interaction with an electron has a size set by theclassical radius of an electron, re = α2ao = 0.282 x 10-12cm,which is similar to the magnitude of many nuclearscattering lengths. Not surprisingly, the huge differencebetween the spatial sizes of a nucleus and an orbital foran electron in a solid leads to significant differences in thescattering profiles for nuclei and electrons.There are two physically different components in theneutron-electron interaction. One is due to the spin of anelectron and the second, quite often called the orbitalinteraction, is due to the magnetic field created by amoving charge.Hence, the interaction involves two degrees of freedombelonging to the electron, namely, its spin and itsmomentum (or velocity). The momentum can be re-expressed in terms the orbital motion of the electron,much as one does in a multipole expansion of a photonwave function to expose electric and magneticabsorption events. For elastic neutron scattering there isa one to one correspondence between the momentum-dependent interaction operator and orbital angularmomentum; the correspondence was demonstrated byTrammell (1953) in a landmark paper. For inelasticscattering processes, of interest to us here, thecorrespondence is not one to one and, in fact, there is nobenefit in using orbital angular momentum operators todescribe truly mobile electrons. Instead, calculations ofthe cross-section, also called a profile, for itinerantelectron systems are best done in terms of the linearmomentum operator.In discussing localized electron systems, we will not

NEUTRON-ELECTRON SPECTROSCOPYE. BalcarAtominstitut, TU Vienna, Stadionallee 2, A-1020 Vienna,Austria.

S. W. Lovesey ISIS Facility, Rutherford Appleton Laboratory, Oxfordshire OX11 0QX, England, UK.

Articolo ricevuto in redazione nel mese di Settembre 2000


19



consider the algebra involved in relating the linearmomentum operator to the orbital angular momentumoperator for the algebra is quite involved (Balcar andLovesey 1989) and a discussion of it adds little to themain aim of this article. Instead, we settle for quoting aresult for the neutron-electron interaction operator forone ion expressed in terms of the total spin, S, and totalorbital angular momentum, L, of the valence electrons (amatrix element of the interaction with electrons in corestates is zero because electrons in these states are paired).Restricting our attention, for the moment, to scatteringevents in which the primary beam is deflected througha modest angle the neutron-electron interactionoperator is,

(1)

Here, ⟨jn(k)⟩ is a radial integral for electrons in the valenceshell created by weighting the radial density of electroncharge in the valence shell by the spherical Besselfunction jn(kr) where k is the magnitude of thewavevector. It follows from the properties of jn(kr) that⟨jo(k)⟩ tends to a non-zero value as k → 0, and it iscustomary to arrange normalization of the radial chargedensity such that ⟨jo(o)⟩ = 1. For n > 0 radial integralsevaluated with k → 0 tend to zero. Hence, in the extremelimit of forward scattering (i.e. no deflection of the beam)the interaction operator (1) is proportional to themagnetic moment carried by the valence shell.Let us consider a matrix element of (1) taken betweenstates of different total angular momentum withquantum numbers J and J′. The situation J ≠ J′ which weconsider is realized by a scattering event that induces atransition between states with different J values in amultiplet, and between J-multiplets belonging todifferent term energies.The key result is that, non-diagonal matrix elements ofthe operator J = L + S are zero, and

(2)

The characteristic dependence of this matrix element on kis a signature of an inelastic event, and the dependence isquite different from that encountered with elasticscattering, e.g. Bragg diffraction.To complete the present discussion let us consider thecross-section which arises from the result (2). Non-zerooff-diagonal matrix elements of L must have J′ = J ± 1.Hence, for modest deflections of the beam inelasticevents are subject to the dipole selection-rule. Consider J= L + S and the event J′ = J - 1 where the states labelled byJ and J′ are separated in energy by an amount ∆. The

cross-section corresponding to this situation derivedfrom (2) is found to be proportional to,

(3)

where hω is the change in energy of the neutrons; k = q -q′ and hω = h2(q2 - q′2)/2m with h2/2m = 2.072 meVÅ2.The result for J = L - S and J′ = J + 1 is very similar tothis, of course, and the dependence on the atomicquantum numbers is a bit more complicated.In general, the dipole selection rule does not limit theevents observed in neutron-electron spectroscopy. Theforegoing expressions actually refer to a very special caseand experimental evidence for this is discussed later. Foran arbitrary value of k there are no simple and generaltheoretical expressions for the cross-section, and each ionneeds to be examined on an individual basis. Results forall the rare-earth (tripositive) ions are tabulated byOsborn et al. (1991).As the last topic in the present discussion of inelasticscattering by an ion we wish to emphasize that, theintensity distribution as a function of k can be verydifferent in inelastic and elastic scattering events. Inthe latter case the k-dependence of the scatteringlength is embodied in a so-called atomic form factor.To illustrate the difference in the k-dependence ofintensity for inelastic and elastic events we show inFig. 1 calculated results for Sm3+. The inelastic event isdipole-allowed so the intensity for k = 0 is differentfrom zero. The k-dependence of elastic scattering isunusual in so far that it is not monotonicallydecreasing, and this feature can be traced to a nearcancellation of the spin and orbital contributions of themagnetic moment which also leads to a small value ofthe gyromagnetic factor, namely, g = 2/7. The calculated elastic and inelastic structure factorsdisplayed in Fig. 1 are in accord with experimentalfindings (Moon and Koehler 1979, Williams et al. 1987).We next consider mobile electrons and find that the cross-section for inelastic scattering is likely to be profoundlydifferent from results we have just discussed for localizedelectrons. One model we can treat completely is a gas ofelectrons which do not interact, the so-called idealelectron fluid or jellium model of electrons in a solid.Even though the Coulomb interactions betweenelectrons, and between electrons and ion cores in thecrystal are not included in the ideal electron fluid theelectrons are correlated because of the quantummechanical exchange force originating from the Pauliexclusion principle.Since the electrons in the ideal fluid are identical andcorrelated it is not correct to consider scattering by asingle electron. Instead, to calculate the profile one must

16

22

2re o( ){ ( ) ( ) } ( ),LSJ j k j k< > − < > −δ ωh ∆

< ′ > = − < ′ > < > − < >J J J J j k j k| | ( ) ( ) }Q | L |12 2{ o

Q k L S( ) ( ) ( ) } ( )[{= > + < > + < ><12 2 2j k j k j ko o ]

employ a formalism that describes all electrons in thefluid on an equal footing and accounts for the Pauliexclusion principle. There are various ways of handlingthe calculation; perhaps the simplest is to use secondquantization of the spin density and momentum densityof the electrons. One finds that the cross-section isproportional to,

(4)

In this expression, fq is the Fermi occupation function foran electron with energy ε(q) = (hq)2/2me and no is the density of electrons. Theexpression (4) is the sum of contributions due to the spinand the orbital interactions, with the latter distinguishedby the vector product k x q.The Fermi functions in (4) are a signature of quantummechanics, and their influence is profound at lowtemperatures (the so-called degenerate Fermi fluid). Onegeneral feature that merits comment is the form of thecross-section for large values of k. On taking the limit k →∞, the contribution from the orbital interaction is

negligible and f k+q → 0. The corresponding value of (4) isoften referred to as the Compton limit of the cross-section, namely,

(5)

Although this result has been derived for ideal electrons,in fact, it is correct in the general case when fq is theappropriate momentum distribution and the bare massof the electron, m e, is replaced by an effective mass.Let us return to (4) and consider its behaviour in the limitof zero temperature. In this case,

where the Fermi wavelength pf satisfies,

In the description of the profile it is quite convenient touse reduced variables for the energy transfer x = hω/εf

and wavevector transfer y = k/pf where the Fermi energy εf = (hpf)2/2me. Scattering is restricted to a domain in x-yspace as a direct consequence of the Fermi functions.One finds the intensity profile for a degenerate idealelectron fluid is different from zero in the domainspecified by the following conditions (the domain is oftencalled the particle-hole continuum);

0 ≤ x ≤ (2y + y2) : 0 ≤ y ≤ 2 (y2 – 2y) ≤ x ≤ (y2 + 2y) : y > 2.

The spin and orbital profiles, calculated from (4) with T =0 K, pf = 0.96 Å-1, εf = 3.5 eV, and y = 0.57, are displayed inFig. 2 as a function of hω = xεf. Two features meritattention. First, the profiles extends over a wide range ofenergies and, secondly, the orbital contribution exceedsthe spin contributions. The latter feature is due to thefactor 1/k2 in the orbital contribution to (4). In a realmaterial and k → 0 this contribution will saturate due todiamagnetic screening, for example.

Intermultiplet transitionsFig. 3 shows experimental data collected on thuliummetal at 20K. The incident neutron energy = 2.14 eV andthe deflection of the beam = 5°. Calculations (Osborn etal. 1991) for tripositive thulium ion predict the lowestCoulomb transition 3H6 → 3F4 at 693.5 meV, withtransitions to the 3F3 and 3F2 levels at just under 2eV.Coulomb transitions, which are designated by ∆L ≠ 0,have energies that depend on the Coulomb repulsionbetween electrons in the valence shell. Since the Coulombintegrals are likely to be more sensitive to changes in thelocal environment than spin-orbit interactions, Coulombtransitions are a useful probe of the way intra-atomic

p nf o= ( ) /3 2 1 3π

f p f pq qq q= < = >1 0,| | : ,| |f f

rde

o

2

32nf

( )[ ( ) ( )]

πδ ω ε εq q k qq∫ + − +h

rd xe

o

2

3 42

21 1

2

n kf f

( ){ ( ) } ( ) [ ( ) ( )].

πδ ω ε εq k q q k qq k q∫ + − + − ++ h

20



Fig. 1. Calculated structure factors for Sm3+, showing the significantdifferences to be found for elastic and inelastic (dipole-allowed) events.

21



correlations are influenced by forming the metallic state.There are strong deviations from the Landé interval rulebecause of intermediate coupling. Looking at Fig. 3, theCoulomb transition 3H6 → 3F4 is lower in energy than thedipole-allowed transition 3H6 → 3F5 and it is more intense.Measured and calculated intensities are gathered in Fig.4. At a sufficiently small value of k the intensity of thedipole-allowed transition exceeds the intensities of allother transitions because these intensities vanish in thelimit k → 0.

Comparison of the transition energies for thulium metal,obtained by neutron-electron spectroscopy, withcorresponding energies for thulium in LaF3, obtained byoptical studies, shows no appreciable differences. Thisfinding suggests that for thulium metal there is noadditional screening of the Coulomb interaction to thatoccurring in the isolated ion. A similar finding has beenreported for

EuBa2Cu3Ox ( x = 6.1 and 7) by Staub et al. (1997). Thereappears to be a more efficient screening of the Coulombinteraction by conduction electrons in Pr metal. Lookingto the future, it will be valuable to measure thedependence on wavevector of structural factors for(intermultiplet) transitions split by the influence of thecrystal field. Calculations (Staub et al. Privatecommunication) using all the current information on thismaterial predict different dependencies for crystal-fieldsplit terms belonging to a multiplet.Significant departures from isolated-ion behaviour canbe expected in metallic systems where the 4f moment isinherently unstable. These include intermediate valencecompounds, where two nearly degenerate electronicconfigurations are hybridized, and heavy-fermioncompounds where the valence is nearly integral butstrong band-4f hybridization suppresses the 4f moment.Similar phenomena are found in actinide compounds. Inall cases, intermultiplet spectroscopy can providevaluable information, including the mixing ofconfigurations and the degree of hybridization of thevalence shell. A review of early work can be found inOsborn et al. (1991).

The density of an electron gas and rs are related through,

where ao is the Bohr radius. In units of eV, the Fermienergy and plasmon energy are, and

, and pf = (εf /3.80)1/2 Å-1.

Metal rs pf (Å-1) εf(eV) hωp(eV)

Li 3.24 1.12 4.78 8.08

Rb 5.23 0.69 1.83 3.94

Zn 2.30 1.58 9.48 13.52

Al 2.07 1.75 11.70 15.83

Sn 2.22 1.63 10.17 14.25

Table I.

Neutron scattering from mobile electronsAs already mentioned, the physical processes weconsider occur on an energy scale beyond about 0.2 eV.No experimental studies of mobile electrons in thisdomain of energy have been reported. In contrast, thereis a multitude of studies of processes occurring at lowerenergies and, in particular, we have in mind the manystudies of spin wave excitations in metallic systems.

hωp s= 47 15 3 2. / /r

ε f s= 50 13 2. / r

43

133

πr

ns

o oa=

Fig. 2. Spin and orbital profiles for scattering from mobile electrons.Smooth dotted curves are derived from the spin and orbital contributionsto (4). The reduced wavevector y = 0.57, and for this relatively small valuethe orbital contribution exceeds the spin contribution. The full and brokencurves are the spin and orbital contributions obtained from a band-structure model of sodium and k = (1/4, 1/4, 0) in units of (2π/a) = 1.55Å-1.Results are taken from Blackman et al. (1987).

Energetic neutrons might be used to conduct Comptonscattering experiments. Because the scatteringwavevector in these experiments is by design made verylarge the scattering is (spatially) incoherent. Results areessentially to do with individual particles and revealnothing about spatial properties. The value of Comptonscattering for studying the motion of nuclei in condensedmatter is well established (Watson 1996 and referencestherein). We have seen that the Compton cross-sectionsfor nuclei and electrons are the same; the information in

them is to do with distributions in momentum space, andin the case of electrons the distribution in question is thatof the spin density. The same physical information isavailable from the Compton scattering of x-rays, andmany successful experiments using synchrotron sourceshave been reported (Sakai 1996).Let us turn to the inelastic and coherent scattering ofneutrons by mobile electrons. Our starting point is thediscussion in section 3 of scattering by an ideal Fermifluid. First, we consider the modification to this resultbrought about by switching on the Coulomb interactionbetween electrons. The plasmon is a collective oscillationin the density of electrons. The plasmon frequency at k =0 is ωp = (4π n o e2/m*)1/2 where m* is the effective massand some values calculated with m* = me are given in Table I. The dispersion of theplasmon is quite weak. For k beyond about pf theplasmon is subject to strong damping from particle-holestates, which is also known as Landau damping.The plasmon is invisible in neutron scattering, in zeroapplied field. The origin of this effect, essentially aselection rule, is orthogonality of the particle density thatcarries the oscillation and the spin density to whichneutrons couple; the first density is n↑ + n↓ and thesecond is n↑ - n↓ . Application of a magnetic field breaksthe selection rule, and thereby a magnetic field is an

excellent switch to use in picking out in the profile thecontribution due to the plasmon.Secondly, we look at the effect due to electron scatteringby ion cores in a crystal on the spin and orbital profiles.Figs. 2 and 5 display profiles obtained from band-structure models of sodium and paramagnetic iron. In allthe cases illustrated, spin underperforms orbitalscattering.The unit of intensity used for the profiles is such that onmultiplying by 0.29 the intensity is barns sr-1 eV-1.(Noise arises in the profiles calculated from band-structure models from band degeneracies at the zoneboundaries and the handling of the delta function thatexpresses conservation of energy, cf. (4).)Fig. 2 includes profiles calculated for an ideal degenerateFermi fluid. These are found to be a good guide for hωless than about 2 eV. Dramatic differences betweenresults for the ideal fluid and band-structure modelsappear just above 3 eV. The origin of the distinctivefeatures from the band-structure model is the splitting ofthe free electron degeneracy at the zone boundary, whichis perpendicular to the (1, 1, 0) direction. This has beendiscussed extensively in the literature under the title ofzone-boundary collective state.An estimate of the profiles in terms of the density ofelectronic states G (ε) can be obtained by averaging thecross-section over the directions of k. For the spin profileone arrives at,

(6)

where f(ε) is the Fermi occupation function for anelectron with energy ε. The expression (6) has the form ofa joint density of states. For simple metals εf is typically afew eV, as can be seen by reference to Table I. Hence, atroom temperature it is appropriate to replace f(ε) by astep function at εf, and in this case (6) reduces to,

or (7)

In the limit hω << εf one obtains as anestimate of the profile. The density of electronic statesdepends on the spatial dimension of the system. Thisobservation leads one to anticipate that, cross-sections forhighly anisotropic systems, which are quasi-one or –twodimensional, will be quite different from those featuringin Figs. 2 and 5.

re f2 hω εG2( )

r de

of

f2

εε ε ω ω ε∫ + >G G( ) : .ε) ( h h

r de

of f

2h

h hω

ε ε ω ε ε ω ε∫ + − + <G G( ) ( ) : ,ε f

r de2 1

−∞

∞∫ − + +ε ω ε ε ω εf f G G( ){ ( )} ( ) ( ) ,ε h h

22



Fig. 3. Neutron scattering cross-section for thulium metal showingintermultiplet transitions from a ground state 3H6. The peaks are labelledby the final state of the transition. After Osborn et al. (1991).

23



For a ‘top hat’ electron density of states of total width εo

the profile derived from (7) is zero except within aninterval of energy between 0 and ε0, where it is a triangle

of height . This finding indicates that scatteringfrom mobile electrons is particularly intense for materialswith a narrow band width.Calculated spin and orbital profiles displayed in Fig. 5are for a d-band metal, and the potential used in theband-structure calculation models paramagnetic iron. Atthe high energies of interest here enhancement of thespin response, required at low energies to reproduce thespin wave, is not significant and calculations reported inFig. 5 contain no such enhancement. As noted already,the orbital outweighs the spin contribution to scattering.Another feature to note is the effect on a profile ofincreasing the magnitude of the scattering wavevector.In Fig. 5 the two panels correspond to wavevectors thatdiffer by a factor = 4.3, while the scale for the intensitydiffers by an order of magnitude. Thus, increasing kreduces the signal from inelastic scattering by mobileelectrons, and in this respect the scattering is not unlikeintermultiplet transitions discussed in section 3. Thecross-section to be observed in an experiment is the sumof the spin and orbital profiles separately displayed inFigs. 2 and 5.Spin wave excitations in metallic systems continue to

( / )re o2 2ε

Fig. 4. Inelastic structure factors for intermultiplet transitions in thulium.Experimental data is obtained from spectra such as the one displayed inFig. 3, which is taken at an angle = 5°; 3F4 filled circles, 3H5 filled triangles,3H4 open circles and 3F3 filled squares. The continuous curves are resultsobtained from theory based on an isolated ion. After Osborn et al. (1991).

Fig. 5. Calculated spin and orbital profiles obtained from a band-structure model of paramagnetic iron (Blackman et al. 1987). In the upper panel k = (1/2,1/2, 0) and the bottom panel k = (3, 0, 0) and the unit is (2π/a) = 2.3Å-1. For k = 1.6Å-1 (upper panel) and k = 6.9Å-1 the spin profile (solid curve) is weakerthan the orbital profile.

pose a major theoretical challenge, even elementalmagnets, and there is a need for experimentalinvestigations. For example, there remain basic questionsabout the spin waves in iron and nickel (Karlsson andAryasetiawan 2000). Some calculations for nickel predicta spin wave branch that extends up to an energy of 0.5eVat the zone boundary, in addition to a softer spin wavethat achieves half this energy. The dispersion, andtemperature dependence, of the spin waves in nickel andother so-called simple metallic magnets are notunderstood and interplay between theoretical andexperimental studies are much needed.Magnetic materials very much at the centre of currentresearch also support high-energy spin excitations.Measurements on La2CuO4 (Coldea et al. 2000) possiblyinclude the most energetic excitations to have beenobserved. By using a spectrometer with many counterssupplied with neutrons from a pulsed spallation source(the ISIS Facility), the investigators were able toconvincingly demonstrate dispersion at the level of 10%in an excitation centred around 300 meV. Thisobservation of dispersion has a very direct bearing onproperties of the appropriate spin Hamiltonian.

DiscussionIn bringing the article to a close it is fitting to add a fewwords about spectroscopic techniques applied to metalsthat utilize beams of x-rays. Synchrotron sources of x-rays developed over the past decade or so have madethese techniques much more valuable than before(Margaritondo 1988). X-ray absorption lineshapes and photoemitted electronenergy distribution curves contain a wealth of usefulinformation. The distribution curves measured inphotoemission spectroscopy, for example, containinformation on the initial electronic states and it appearsin the curves convoluted and mixed with other effectsthat could be the prime interest. Effects in questioninclude bulk and surface plasmons, secondary electronsfrom inelastic scattering processes, and the orthogonalitycatastrophe which is a many-body effect that arises withphotoelectrons excited from core levels of an ion. Theabsorption of x-rays contains another many-body effect,namely, excitons which arise from the addition ofelectrons to the conduction band. Excitons and theorthogonality catastrophe cause power law behaviour atthe absorption edge. Because of these and other effectsmeasurements using x-ray spectroscopy are indeed richin information content. On the other hand, theinterpretation of data is both subtle and demanding, andvery much more so than one anticipates with datagathered using neutron-electron spectroscopy.One attraction of neutron-electron spectroscopy has to be

the delicate nature of the neutron as a probe of condensedmatter. The many-body effects in x-ray spectroscopy justmentioned are a measure of the disruptive nature of x-rays in this mode of investigation. If the goal of anexperiment is to measure properties of electrons in thesample without a violent disturbance the method ofchoice is neutron-electron spectroscopy.

AcknowledgementWe are grateful to Dr. R. Coldea and Dr. U. Staub foruseful comments that improved a first version of thearticle.

ReferencesBalcar E and Lovesey SW (1989) Theory of Magnetic Neutron and Photon

Scattering (Clarendon Press: Oxford)

Balcar E and Lovesey SW (1993) J. Phys.: Condens. Matter 5, 7269

Blackman JA et al. (1987) J. Phys. C: Solid State Phys. 20, 3887

ibid 20, 3897

Coldea R et al. (2000) Physica B 276 – 278, 592

Karlsson K and Aryasetiawan F (2000) J. Phys.: Condens. Matter 12, 7617

Lovesey SW and Staub U (2000) Phys. Rev. B61, 9130

Margaritondo G (1988) Introduction to Synchrotron Radiation (OxfordUniversity Press : New York)

Mason TE et al. (2000) Proceedings of the LINAC 2000 Conference, paperFR203

Mesot J (1995) in Magnetic Neutron Scattering edited by A Furrer (WorldScientific:Singapore)

Moon RM and Koehler WC (1979) J. Mag. & Mag. Mat. 14, 265

Osborn R et al. (1991) in Handbook on the Physics and Chemistry of RareEarths, Vol. 14, Chapter 93 (Elsevier Science Publishers: Amsterdam)

Rauch H et al. (2000) Physica B 276-278, 33

Sakai N (1996) J. Appl. Cryst. 29, 81

Staub U et al. (1997) Phys. Rev. B55, 11629

Staub U and Soderholm L (2000) in Handbook on the Physics andChemistry of Rare Earths, Vol. 30, Chapter 194 (Elsevier SciencePublishers : Amsterdam)

Trammell GT (1953) Phys. Rev. 92, 1387

Watson GI (1996) J. Phys.: Condens. Matter 8, 5955

Williams WG et al. (1987) J. Phys. F17, L151

24



25



The technique of X-ray standing wavefield (XSW) absorptionis a very sensitive tool to probe the position of implanted oradsorbed atoms on the sample surface. Its application to thedetermination of adsorbate structures at the solid-vacuuminterface is now very well established. For many years, the useof this technique was limited because of the need of highlyperfect samples. The angular range of total reflectivity is, infact, typically of a few of arc-seconds. As the XSW rangearound a Bragg reflection is very narrow, not only the incidentx-ray beam must be highly collimated and monochromatic, but,also the crystal sample must be perfect at a similar degree. Thislatter requirement had the effect to restrict studies onsemiconductor materials, which typically do have the necessaryhigh degree of crystalline perfection. Therefore, relatively fewXSW experiments were done on metal surfaces in UHVconditions. For 10 years Woodruff [1] and co-workers pursuedsuch aim by using the Synchrotron Radiation Source (SRS) atDaresbury, and developed the technique of Normal IncidenceX-ray standing waves (NIXSW), a new variant of the wellestablished XSW technique. They demonstrated that atincidence normal to the Bragg scattering planes, NIXSWbecomes insensitive to the crystal mosaicity, to the incident-beam collimation, and to the exact incidence angle, makingmetal single crystals perfectly adequate.At the European synchrotron radiation facility (ESRF), whereI worked as scientist for five year on the ID32 beamline (theSurface-EXAFS and X-Ray Standing Waves Beamline), anUHV system well adapted for several surface sciencetechniques (SEXAFS, Resonant Photoemission, XPS, etc.),and also for NIXSW studies, was installed. ID32 offered anexcellent compromise between a good flux gain and a goodresolution. For this reason, Woodruff and co-workers started aprogram to exploit the special performances of the ESRFsynchrotron as far as the capabilities of the NIXSW technique.I had the big chance to be involved and collaborate to thatproject. The general aim of the project was to develop theNIXSW method and to apply this to a range of adsorbatestructures on metal single crystal surfaces of general relevanceto the area of heterogeneous catalysis. The high flux and highspectral resolution available on ID32 opened new possibilities.First of all, it was possible to exploit ‘chemical shifts’ inphotoelectron binding energies and perform NIXSW structuredeterminations of coadsorbed molecular fragments resultingfrom a surface reaction. Second, we could study the local

structure of adsorbates at low coverages. Third, we couldinvestigate low Z (C, N and O for example) adsorbate atomswhich have low photoionization cross-section at the X-rayenergies used in the NIXSW experiment. The potentiality of the chemical-shift NIXSW technique (CS-NIXSW) will be discussed and recent results will be presented.We were able to demonstrate that this technique offers thepossibility to shed new light on the determinations of severalsurface structures.

IntroductionThe study of the nature of the chemical bond, of theabsorption site and, in general, of the geometry ofadsorbed atoms is the starting point for a quantitativeknowledge of a well characterised metal single crystalsurface. Many peculiar techniques have been developedand applied to investigate the proprieties of solid-vacuum surfaces and interfaces. A large part of the wellestablished surface structures was obtained by methodsusing coherent interference of elastic scattered electrons(Low Energy Electron Diffraction (LEED), photoelectrondiffraction (PhD), and surface extended X-ray absorptionfine structures (SEXAFS).SXRD (Surface X-ray Diffraction) has been alsoextensively employed to study surface and interfacestructures. It is a very powerful technique for structuralstudies on crystalline surfaces, but it presents,nevertheless, some restrictions: a1) It is not trivial to achieve the necessary surfacesensitivity because atoms are weakly scattered from X-rays. a2) It has no elemental sensitivity.a3) It probes the long-range order part of a surface. Onlywhen the surface has an excellent long-range order of theadsorbate layer, this technique is able to supply detailedlocal adsorbate-substrate structural information. Overthe last few years scanning probe microscopes havehighlighted the poor average quality of many surfacesshowing that surface patches with a lack of long-rangeorder, coexist with long-range ordered domains.We will show that surface structure information may beobtained from X-ray Standing Wavefield technique(XSW), even when a surface contains regions of goodlong-range order as well as regions of local order or

CHEMICAL-SHIFT NORMAL INCIDENCE X-RAY STANDINGWAVE DETERMINATION OF ADSORBATE STRUCTURESVincenzo FormosoI.N.F.M., Unità di Cosenza, Dipartimento di FisicaUniversità degli Studi della CalabriaVia P. BucciI-87036 Arcavacata di Rende (CS)

Articolo ricevuto in redazione nel mese di Novembre 2000

26



complete disorder. We would like to stress this realcrucial point because structural data obtained fromdiffraction methods, may give us the misleading view ofa surface structural perfection. Commonly, there is nolong-range order in the adsorbate layer.

Basic physics of the XSW techniqueThe existence of a X-ray Standing Wavefield within acrystal was demonstrated by Borrman2 (Bragg reflectionin the Laue case geometry) and by Batterman3 (Braggreflection in the Bragg case geometry). In this latter case,the incident, , and the reflected,

, coherent X-ray plane wavesinterfere to set up a XSW field. This XSW field is parallelto the diffraction planes of the crystal and has the samespatial periodicity.ν0 and K0 are the frequency and the propagation vector ofthe radiation, respectively. Because the two waves arecoherent, we can define, a phase factor between the twoamplitudes E0 and EH: . It is independenton time and space. The intensity of the X-ray standingwavefield in the crystal is the squared modulus of thesum of the incident and reflected X-ray amplitudes andcan be written as:

.We note that the time dependence vanishes out in theexpression for the intensity. H=KH - K0 is the reciprocal lattice vector associated withthe Bragg reflection, r is a real-space vector defining theposition of the atomic absorber at which the intensity ismeasured, R is the amplitude factor given by

and Φ is the phase factor written as :

where

and are given in terms ofthe geometrical structure factors FH and for the Hand reflections.

In the standard form of XSW experiments, the ηparameter is related to the incident angle θ, and it isgiven by

.

Here F0 is the structure factor for the (000) reflection, P is

a polarization factor and Γ is expressed by

,

where V is the volume of the unit cell, λ is the x-raywavelength, m and e the mass and charge of the electron,ε0 the permittivity of the free space, and c the speed oflight.It fig.1 we plot the calculated reflectivity R, the phase Φand the XSW profile (eq. 1) in terms of the dimensionlessη parameter.In case of a non absorbing crystal, when the incidentangle θ is scanned through the nominal Bragg angle, θB ,we have that in the Bragg diffraction condition (-1< η < 1)the reflectivity is equal to 1 and the phase changes of π .The superposition of the two coherent plane wavescreates a planar standing wave. The intensity (seeequation 1) at a particular location r in space isdetermined by H and by Φ and is spatially modulated:minima (nodes) or maxima (antinodes) of the wavefieldintensity lie on scattering planes. The XSW is set upinside the crystal and it extends well outside. The

Γ = e

mc V

2

02

2

4πελπ

ηθ θ θ

=− ( ) +( )B B

H H

sin F

P F F

2 0ΓΓ

HF

H

E E F FH H H02 1= − ± −

η η

ϕ = ( ) ( )[ ]arc E E E EH Htan Im Re0 0

Φ =( ) >

+ ( ) <

ϕ

ϕ π

for E E

for E E

H

H

Re

Re

0

0

0

0

RI

I

E

EH H= =0

2

02

IE

R ei= = + − •( )ε ε π*

( )0

22

21 1Φ H r

E E R eHi= 0Φ

ε π νH H

i tE e H= − •2 ( )K rH

ε π ν0 0

2 0= − •E e i t( )K r0

Fig. 1. Calculated reflectivity and phase for a non-absorbing crystal versusthe η parameter. The shape of the XSW profile is strongly dependent onthe adsorber position, z, measured with respect to the scattering planes.

27



Fig. 2. Relationship of (111) and (-111) layer spacing for atop, hcp, fcc adsorption sites.

Fig.3. The angular distribution of electrons ejected from an atom in the XSW field. In our experiment, the angle θ between the detector and the x-ray Avector was 45°, β =2, δ =0, γ =1.

28



scattered intensity by a certain atom depends on itsposition with respect to the maxima or minima of the X-ray interference field.The intensity of the planar wavefield,

is a periodic functionwhose period is given by where

is the spacing between thenodes of the standing wave. Adjusting H and/or θ, it ispossible to choose a well-adapted wavefield spacing forany structural problem. The XSW method provides information on the spatialdistribution of absorbers perpendicular to specificscattering planes by measuring the X-ray absorptionprofile for an adsorbate atom. When we scan through aBragg reflection, the X-ray standing wave shifts its phasein a systematic and predictable fashion. One way tochange the phase, Φ, is just rocking the crystal around theBragg condition. The maximum scattered intensity isobserved when the maxima of the intensity of the XSWcoincide with the atomic position. In the case of a singleabsorbing atom at a single high symmetry adsorptionsite, the XSW measured intensity is:

(2)

By fitting the intensity versus photon energies, wedetermine the structural parameter z, i.e. the distancebetween the adsorbed atom and the scattering planes. Inorder to take into account some distribution of zpositions, due to vibrational or static disorder, or todifferent adsorption sites, with a probability f(z)dz, theXSW profile can be written

(3)D is the coherent position, the adsorbate position withrespect to the diffraction planes. The coherent fraction,fC0, provides a measure of the local order of adsorbateatoms on the surface.

The NIXSWThe Normal Incidence X-ray Standing Wave (NIXSW)technique retains not only the basic capabilities ofstandard XSW experiments to determine adsorbatestructures, but also allows us to use commercial metalsingle crystals due to the relaxed constraint on thecrystalline perfection of the sample. For a symmetric Bragg reflection of σ polarized X-raysthe Darwin width, Wθ ,is expressed as[4]:

A different expression for angles close to 90° has to beused [5]. Let us suppose E=15KeV is the energy of theincident X-rays striking on Si(111), we obtain Wθ =3.6 arcsec=17.4 µrad, being the lattice constant a=5.431 Å-1, thediffraction plane spacing dH=3.14 Å the Bragg angleθB=7.6° , and the structure factor FH=6.53.Commercial single metal crystals have a mosaic structurein the substrate of order 0.5°. On a Cu(111) metal crystal,for example, the standard XSW technique cannot beapplied because the total angular range is only 0.016°, ifthe incidence angle is 45° and the photon energy is 4.2KeV. But, if we set the incidence angle at 90°, the angularrange increases considerably. It has a value close to 1°,providing the desired insensitivity to crystalline quality. The use of an impinging beam normal to scatteringplanes, however, gives wide rocking curves and, thus, alow sensitivity to the crystal mosaicity, making metalsingle crystals routinely usable. In fact, at θB =90° theenergy width of the Cu(111) reflection is 0.87eV,comparable with the resolution of double crystal X-raymonochromators used on synchrotron radiation sources.Since in NIXSW technique the scattering angle is keptfixed while the X-ray energy is scanning, it is importantto give a different expression for the η parameter:

and, thus, we note that it remains linear in E even atθB=90°.Moreover, the photoemission or the Auger emissionelectrons associated with the photoabsorption ensuressurface specificity in the absorption profile. At incidencenormal to (111) planes of fcc metals, the photon energy istypically in the 2.5-3.5 KeV range. This photon energyrange ideally matches to the first harmonic of a standardESRF undulator, such as that on ID32.It may be worthwhile to illustrate the procedure we used(see fig.2) to extract quantitative structural informationby means of this technique:a1) We selected two independent Bragg reflections andfor each, we measured the XSW absorption profile. Forfcc crystal surfaces, we used the (111) and the (-111)reflection planes.a2) The measured lineshape was determined by twostructural parameters, the coherent position, D, and thecoherent fraction, fC0 . These parameters were obtained byfitting the NIXSW profiles.a3) Once we had the coherent position for the (111)reflection, we could deduce the coherent position value forthe (-111) reflection, D(-111). See fig.2 for the highsymmetric adsorption sites. If a bridge site is occupied,the coherent position is D(-111)=( d(111)+1.5 D(111) )/3.

ηθ

=

− − +2 20

( )E E

Esin F

P F F

BB

H H

Γ

Γ

W Åd F

aH H B

θθ

=( )− −3 6 10 5 1

2

. *tan

I R R fD

dCH

= + + −

1 2 20 cos Φ π

I R Rz

dH

= + + −

1 2 2cos Φ π

H Hd H K Ksin

= = − =− −10

1

2

λθ

H r• = =H dH 1

I R R= + + − •( )1 2 2cos Φ π H r

29



a4) Then, we compared the D(-111) measure with thepredicted one. If the two values are quite close, theadsorption site is determined.a5) The coherent fraction values were very importantbecause an high value of fC0 (-111) is consistent with afully symmetric adsorption site, while a low value isconsistent with lower or mixed symmetry sites.

CS-NIXSWIn a surface chemical reaction several co-adsorbedmolecular moieties are involved. The coexistence ofseveral different adsorbates on a surface at the same time,a mixture of reactants and products, make intermediatereactions also possible. Structural studies of suchmultiple chemical state species can supply quantitativeinformation on chemical and physical phenomena atsurface. Little progress has been made so far indetermining their local adsorption geometries.As well known, X-ray fluorescence, Auger electronemission and electron photoemission can monitor X-rayabsorption in atoms near the surface. By photoemission,then, it is possible to measure the absorption in a specificchemical state. Small changes in the photoelectron yieldas well as in binding energies, as measured at a specificcore level, are due to changes in the local electronicenvironment of the atom.Several surface structural techniques are based onphotoabsorption. One successful approach is theChemical-Shift Photoelectron Diffraction (CSPhD)[6].Photoelectron diffraction is intrinsically elementselective, and, thus, by measuring the photoelectrondiffraction signal related to a specific chemically shiftedcomponent of the core-level photoemission yield, onemay obtain chemically specific structural information. Another successful experimental procedure to determinelocal adsorption geometries of adsorbates is the use of theX-ray standing wavefield (XSW) technique [7]. This, infact, selects chemical species at a crystal surface orinterface by measuring the photoabsorption yield. A firstattempt to resolve a chemical state by X-ray standingwave analysis using chemical shifts was performed [8]recently, but their spectral resolution was not enoughhigh to distinguish among several chemically shiftedstates. It was found, in fact, that for adsorbates of lowatomic number, or when chemical state resolution isrequired, photoemission from the adsorbed atom is thebest tool to monitor the x-ray absorption.The new third generation of synchrotron radiationsources offer several impressive opportunities because oftheir very high spectral brilliance, their increased spatialand spectral resolution. Such combination of powerfulmeans at ID32 of ESRF allows the researcher to monitorlocal adsorption sites on surfaces using normal incidenceX-ray standing waves, not only in an element-specific

fashion, but also with chemical state specificity, bymeasuring the intensity of 'chemically shifted' core levelphotoemission signals from the different states. Then, theenhanced monochromator resolution and output flux,combined with a suitable electron energy analyser givethe possibility to deal with the problem of chemically-resolving coadsorbed species containing the sameelements, and allow to study low-Z elements which haveonly shallow and weakly-absorbing core levels.Chemical-Shift NIXSW (CS-NIXSW) studies can, thus, beperformed routinely at ESRF by using the wide

potentiality of that light source both for the creation ofthe X-ray standing wavefield and for the detection of thephotoabsorption signal. In other words, if a particularelemental species is present on the surface in two or moredifferent states, either in a different local geometry withrespect to the substrate or to other atoms to which it maybe bonded in a molecular species, the core levelphotoemission from these atoms will show differentphotoelectron binding energies associated with thesedifferent states.The detected signal is rather surface specific and can beused to provide information on the structure of theoutermost layer especially if photoelectron energies arereasonably low, below 1 KeV. The photoemission process,then, involves only primary photoionisation events. Theelemental and surface specificity is, in fact, not clouded

Table 1. Chemical shifts associated with the fragmentation of PF3

adsorbed on Ni(111) and Ru(0001) surfaces at different temperatures.

Table 2. Coherent fraction and coherent position values obtained fromfitting the XSW profiles associated with the PF3 fragmentation on Ni(111).

30



by secondary electron ionisation events because they arenot detected.

Non-dipole photoemission angular effects in NIXSWThe differential photoelectric cross section for an atom inan external X-ray interference field (XSW) is given by [9]

where

andare the matrix elements, k is the photon wave vector, A is

the photon polarization vector, r the electron position , pthe electron momentum operator, f is the final state(outgoing electron), i is the initial state (core level), O theincident wave, and H the reflected wave.Let consider the expansion:

For sufficiently low photon energies, Kr << 1, the so-called dipole approximation is generallyadopted in photoemission. Note also that in the NIXSWcase the incident and reflected X-rays are collinear,therefore, we will have . With theseconsiderations we can write the differential cross sectionas

because the matrix element factorises out.

The conclusion is that, in the dipole approximation, thephotoemission intensity measured in NIXSW doescorrectly monitor the XSW absorption profile.The experimental geometry of NIXSW measurements is:the photon beam hits the sample at incidence normal to

the scattering planes and the photon polarization vectorA of the incident and reflected waves are collinear. Thephotoemission signal is detected by an electron energyanalyser placed at an angle θ with respect to the photonpolarisation vector A, such that the incident x-raysprovide backward photoemission and, the reflectedx-rays forward photoemission.Because of the angle resolving nature of typical electronspectrometers, the measured photoemission signal is thedifferential cross section given by [10, 11]

(4)

where β is a dipolar asymmetry parameter. Theparameter γ represents the correction term correspondingto the dipole-electric quadrupole interference and theparameter γ is the magnetic-electric-dipole term, which ispresent only if core relaxation occurs. Finally, θ is theangle between the photoemission direction and thephoton polarization vector, ϕ is the angle between thephoton propagation direction and the projection of theelectron wave vector in the plane perpendicular to A.For many years, the dipole approximation has beenwidely used to interpret photoemission from atoms,molecules, and solids if the photon beam energy waslower than 10–20 keV. Quadrupole and magnetic dipoleeffects can, however, substantially influence the angulardependence of the photoemission even at much lowerphoton energies as recent theoretical [4] andexperimental data [12,13] from inert gas atoms showed. Inaddition, it was proved that quadrupole and magneticdipole effects are important also for the photoemissiondetection in x-ray standing wave determinations ofsurface structure [14]. If a forward/backward asymmetry exists in thephotoemission, and this depends only on the geometry ofthe experiment, the detected photoemission signal doesnot monitor the true absorption profile [7] because themeasured signal will not monitor the two components ofthe XSW in an equivalent fashion. In fact, in this case thecoherent fraction f obtained from the standard fittingprocedure of the data would be too high, physicallymeaningless because greater than one.Thus, in order to take into account of quadrupole andmagnetic dipole effects we introduced aforward/backward asymmetry parameter Q, defined(see equation 4) as the ratio of the photoemission signaldetected in the forward direction to that detected in thebackward direction:

.

d

d

forward d

d

backward Q

Q

σω ϕ

σω ϕcos cos=

= −

= +

−1 1

1

1

d

d

σωσπ

β θ δ γ θ θ ϕ

=

+ −( ) + +( )

4

12

3 12 2cos cos sin cos

f iA p• 2

d

dR R

σω

π≈ + + − •( )1 2 2cos Φ H r

A p A p0 • = •H

exp( )iK r• ≈ 1

exp( )i iK r K r K r• ≈ + • − •( ) + ⋅ ⋅ ⋅11

22

M f i iifH

H H= • •A p K rexp( )M f i iif0

0 0= • •A p K rexp( )

dσdω

=2

Mif0 +

EHE0

ifH

exp i Φ −2πH ⟨ r( )( )M

Table 3. Coherent fraction and coherent position values obtained fromfitting the XSW profiles associated with the SO2 fragmentation on Ni(111).

31



Fig. 5. CS- NIXSW technique. The (111) XSW absorption profiles obtained from thefour distinct peaks of the photoemission energydistribution curve. Clearly, they differ significantly,implying different local adsorption sites for the fourspecies.

Beyond the dipole approximation and includingmagnetic dipole and electric quadrupole contributions,( ), the differential cross section can bewritten in term of the Q parameter [15]

Let us write this equation in the standard form defining

an apparent reflectivity, ,

For positive Q values the apparent reflectivity and theapparent amplitude of the interference term areenhanced with respect to the dipole case where Q=0.

d

dR R

Da a C

H

fd

σω

ϕ π∝ + + −

1 2 2

0cos

R RQ

Qa = +−

1

1

d

dR

Q

Q

RQ

Q

Dc

H

fd

σω

ϕ π

∝ + +−

+

+ +−

−

11

1

21

12

0cos

exp( )i iK r K r• ≈ + •1

32



This means that when we use the measuredphotoemission signal as a monitor of the intensity of theX-ray standing wave, we obtain a profile withanomalously large modulation amplitude as comparedwith a true XSW absorption profile. These effects stronglyinfluence the NIXSW measurements because their typicalutilisation range is 2.5-3.5 KeV and are still important,even at these photon energies, if we study low atomicnumber absorbers. In this case, in fact, the beam photon

energy is far above threshold for the ionisation of the 1score level and quadrupolar contribution tophotoemission becomes increasingly important as thephoton energy rises above the threshold. Using these photoemission XSW profiles without the Qcorrection, significant errors in both coherent positionand coherent fraction can occur, in particular, ananomalously large coherent fraction is obtained. Recently, a new simple method for measuring thequadrupole-dipole interference asymmetry factor forphotoemission from core levels, was presented in a veryinteresting paper [16].

Recent resultsWe applied the chemical shift XSW technique to thestudy of three different model systems. The local X-rayabsorption of the adsorbed atoms on a surface in aspecific chemical-state was investigated by resolving thedifferent chemically shifted components of the core levelphotoemission. First of all, we set up a demonstrative experiment of thenew capabilities of the CS-NIXSW technique when usedin conjunction with a third generation X-ray synchrotronsource. The chosen system was a very well known modelsystem, PF3/Ni(111). The investigation was extensivelyperformed with several techniques. [17] Then, the same technique was applied to two differentsystems, SO2/Cu(111) and CH3SH/Cu(111), to learn onthe structure of coadsorbed reaction products, yieldingquantitative information on their local adsorptiongeometry. We demonstrated the utility and the capabilities of theCS-NIXSW technique, particularly for thecharacterisation of coadsorbates produced by a chemicalreaction and, moreover, that it is a quite useful andpowerful tool for the investigation of adsorbates withlow atomic numbers.The experiments were performed at the EuropeanSynchrotron Radiation facilities (ESRF) on beamlineID32. The samples, Ni(111) and Cu(111), were preparedby the usual combination of in situ argon-ionbombardment and of annealing cycles until a clean andwell ordered surface was obtained. Auger ElectronSpectroscopy (AES), Low Energy Electron Diffraction(LEED) and the width of the substrate XSW profile, wereused to check the cleanness and the quality of the crystal.

Coadsorption site determination of PFx fragments on Ni(111) by CS-NIXSW [18]The coadsorbed PFx speciess on Ni(111) were producedby X-ray induced fragmentation of an initially adsorbedPF3 overlayer. PF3 was introduced into the chamber to the

Fig. 4. Photoelectron energy spectra in the P 1s peak region recorded:- at140 K immediately after PF3 dosing;- after exposure to a monochromatic incident X-ray beam for 10,50,90 min;- at 300 K, soon after the exposition for an extended period to radiationdamage at 140 K; - a freshly prepared PF3 layer at 300 K.

33



typical pressure of 10-8 mbar. The total exposure of thesurface was 10-5 mbar s which was enough to produce asaturation coverage at room temperature and at 140K. Our results revealed some significant and unknowndetails in the surface chemistry of the PF3 photo-dissociation on Ni(111). The incident X-ray beam causeda reduction of the intensity of the P 1s peak associatedwith the initial intere PF3 adsorbate, and induced newchemically shifted components. An important distinction between the behaviour at roomand low temperatures was observed for the photon-induced decomposition of PF3 (fig.1). At low temperatureexposure, four resolved components were seen (PF3, LT1,LT2 and LT3 labels). In the case of the room temperatureexposure or warming up of the sample from low-temperature, instead, only one new peak (RT2 label)appeared. It was clear that, in addition to the peak

associated with the intere PF3 species, the spectraobtained after low temperature radiation exposureshowed three additional components.In a previous photoemission study of P on Ru(0001) [19]the four components of the P 1s, corresponding tosuccessively larger chemical shifts , were assigned to PF3,PF2, PF, and P. We showed that these suggestions werenot consistent with our measurements. On the basis of the measured chemical shifts (see table 1)we could remark the following points:b1) The values were almost similar but not identical.

Fig. 6. NIXSW profiles of (111) and (-111) obtained by fitting theindividual components of the P 1s, the F 1s and Ni 2p3/2 photoemissionpeaks. The Ni(111)- PF3 surface was exposed to the X-ray beam at atemperature of 140 K.

Fig. 7. NIXSW profiles of (111) and (-111) obtained by fitting theindividual components of the P 1s, the F 1s and Ni 2p3/2 photoemissionpeaks. The Ni(111)- PF3 surface was exposed to the X-ray beam at atemperature of 300 K.

Table 4. Values for the NIXSW fitting parameters of the coherent position(D) and coherent fraction (fco) for the four distinct chemically shifted S 1sstates associated with the CH SH interaction with Cu(111).

34



b2) The close similarity of the LT1 chemical shift (2.13 eV)with that attributed to PF2 on Ni(111) (2.1eV ), suggestedthat LT1 fragment was very likely PF2.b3) Even the 1s and 2p core levels of P on the Ni(111)surface, do have a significantly different degree oflocalisation.

b4) Since none of the chemical shifts associated with theother P 1s states reproduces the amount of the P2p shiftattributed to PF ( 3.3 eV), any doubt on the assignment ofPF is ruled off. b5) It is worthwhile to notice, moreover, that thechemical shift (4.02 eV ) of the single additional state

Fig. 8. Schematic diagram of the adsorption geometry of PF3 on Ni(111).

35



produced at room temperature (RT2) is significantlydifferent ( by more than 0.3 eV ) from its closest-energypeak (LT2) seen after low temperature exposure. It opensthe very real possibility that these do not correspond tothe same surface species. The main conclusions for the different coexistent P sitesas implied by our chemical-shift P 1s NIXSW data, aresummarised in Table 2 and figs. 6 and 7. The differentchemically shifted peaks, that is PF3, LT1, LT2, LT3, showdifferences in their XSW profiles for low and room-temperature surface preparations.This is consistent with the occurrence that each P speciesoccupies a different site on the surface. The NIXSWprofiles for the P 1s peaks having the closest chemicalshifts (LT2 and RT2) correspond to different localadsorption sites for their associated P atoms, supportingthe suggestion that these two peaks do not correspond tothe same species. It turns out that after low temperatureirradiation the NIXSW profile of the peak with the largestchemical shift, LT3, appeared to be very similar to theRT2 peak. Of course, while the consequent implication isthat the P atoms of the LT3 and RT2 surface species have

the same local geometry, the large difference in theirphotoelectron binding energies made clear that they didnot correspond to the same adsorbed species.A powerful aid for the assignments of these states wasthe additional information regarding the local geometryof the associated P atoms provided by the NIXSW data:c1) In the case of PF3 species, the NIXSW coherent positionvalues were consistent with atop adsorption at bothtemperatures (see Table 2). The coherent fraction values inboth (111) and (-111) NIXSW for these species were high,implying a single high-symmetry adsorption site.c2) The coherent position values found for LT1 specieswere in agreement with the occupation of bridge sites.The low value of the (-111) coherent fraction was alsoconsistent with single lower symmetry sites. We thereforeassigned LT1 to a bridge-bonded PF2 species (see Fig. 8b).c3) The second species observed at room temperature(RT2) corresponds to P atoms that occupy fcc hollowsites (with a possible minority co-occupation of hcphollow sites) (see Table 2). On the other hand, the trendof the results was consistent with the possibility that RT2states actually correspond to P atoms (see Fig. 8a).c4) The structural parameters for the LT3 species wereclearly correspondent to the fcc hollow site occupation. c5) The structural parameter values associated with theLT2 species were not consistent with any single highsymmetry site. The (-111) coherent fraction value wasvery low and could even be zero, implying very poorlateral coordination of the adsorption site on the surface.On the other hand, the (111) coherent fraction was high,implying good order perpendicularly to the surface.Thus, the (111) coherent position was easily identified asa true layer spacing. Since that the two LT2 and LT3species were seen together on the surface prepared at lowtemperature, we suggested that these two chemicallyshifted states would correspond to the two distinct Patoms in some P2Fx surface species. The LT3 state wouldbe associated with the lower atom bonded to the Ni(111)surface in the hollow site, whereas the LT2 state would bethe upper atom. The P-P distance in this species wasexpected to be approximately 2.20 Å, so that implied thatthe P-P bond was tilted at an angle of 50° relatively to thesurface normal. Notice that if this tilted bond had nopreferred azimuthal orientation, as seems likely, theupper P atom would occupy a multitude of lowsymmetry sites.

Interaction of SO2 with Cu(111) studied by CS_NIXSW[20].Sulphur dioxide, SO2, is an important air pollutantcreated by the burning of sulphur-contaminated fossilfuels. It does not dissociate on noble and transition metalsurfaces at low temperatures. Several coadsorbedmolecular moieties may coexist on a surface at the same

Fig. 9. S 1s photoelectron energy spectra recorded from the Cu(111)surface at a sample temperature of 140 K, and after heating to 300 K.

36



time. A structural change of the adsorbed SO2 occurs atlow temperature as a result of the temperature increaseand/or of the molecular cracking under monochromaticundulator radiation.In the case of the X ray irradiation, we could see theradiation induced state conversion of the S 1s spectra; O1s spectra, instead, did not show significant changes intotal oxygen coverage on the surface and in the binding

energy of the O 1s peak. As the temperature was raised,chemical interaction occurred, and coadsorbed SOx

species appeared on the surface via moleculardissociation. In fact, the spectrum taken at 100K wasdominated by a single peak identified as adsorbed SO2.Fig. 9 shows that increasing the temperature , threedistinct S1s chemical states appeared. Heating thesurface had the result to convert much of the SO2 speciesto a second state with a binding energy about 1 eV higher,while further heating gave rise to a third state. Itappeared at a binding energy about 4 eV smaller. The SOx

species, characterised by the largest S 1s binding energystate seen in Fig. 9, were associated with adsorbed SO3

produced by the reaction (3SO2 --> 2SO3 +S). The highesttemperature (lowest binding energy) state was associatedwith atomic sulphur.Atomic sulphur appeared to occupy a mixture of face-centred cubic and hexagonal close-packed hollow siteson the surface. The measured value of D[111]=1.75Åallowed us to find the values of D[-111] for atop, hcphollow and fcc hollow: they were respectively 0.58Å,1.28Å and 1.97Å (with near-unity value expected for D[-111] ). The measured value of D[-111] is actually 1.68Å,essentially at midway between the values for the twohollow sites.So, sulphur atoms occupy these two hollow sites withequal probability: in this case the predicted value of d [-111] is 1.63Å, in agreement with the measured values. SO2 was found to adsorb with its molecular planeessentially perpendicular to the surface, and the datawere most readily interpreted in terms of a bridginggeometry bonding through the oxygen atoms, althoughthe CS-NIXSW data indicated that oxygen atoms couldnot occupy only static near-atop sites.SO3 species adsorbed with its axis perpendicular to thesurface, atop a surface copper atom with S-O bonds outof plane such that oxygen atoms were closer to thesurface. The polarisation angle dependence of theNEXAFS data indicates that the SO2 molecular plane isessentially perpendicular to the surface, as is the axis ofadsorbed SO3.

A surface chemical reaction, CH3SH on Cu(111),investigated by CS-NIXSW[21]The interaction of alkane thiols with surfaces is of interestfor desulfurization catalysts and catalyst poisoning.These species form self-assembled thiolate monolayers(through deprotonation at the interface) which are ofpotential interest for molecular electronics. Methanethiol CH3SH, is the simplest of such species. The chemical shifts of the S 2p photoelectron bindingenergies of the intact molecular thiol, the LT thiolate, andthe HT thiolate, relatively to the atomic sulphur, were

Fig. 10. Local adsorption geometries at low temperature and experimentalenergy spectra in the region of the S 1s binding energy.

37



identified as 2.3, 0.6, and 1.4 eV, respectively. Thesephotoemission experiments showed that four distinct S-containing surface species resulted from the interactionwith the Cu(111) surface. They were assigned to:- an adsorbed methyl thiol molecule (present only at thelowest temperatures ≤150 K);- two distinct states, both attributed to methyl thiolate(CH3S

-) adsorbed in different geometries. The twothiolate species were conveniently labelled low and hightemperature (LT and HT) thiolates; both thiolate specieswere present at the lowest temperature, but the LTthiolate transformed to the HT thiolate as thetemperature was raised;- all species present on the surface transformed onatomic sulphur when the temperature was highest than400 K. For the S atom in the intact thiol, the experimentald(111) S-Cu layer spacing of 2.38 Å given by the coherentposition, provides a (-111) coherent position of 0.79, 1.48,or 2.18 Å, depending on whether the S site occupation,that is, to atop, hcp hollow or fcc hollow sites.Thus, the NIXSW data for the adsorbed intact thiol,suggested a single high-symmetry adsorption sitebecause of the relatively high value of D(-111)=0.8 Å forthe sulfur absorption signal. The experimental value D(-111) =0.79 Å was in perfectagreement with the prediction for the atop site, so wededuced that the intact thiol was bonded to the Cu(111)surface through its S atom, which is directly atop anoutermost Cu atom layer. If the LT thiolate would occupy a high symmetry site, themeasured (111) layer spacing of 1.88 Å would give (-111)coherent positions of 0.63 Å (atop), 1.32 Å (hcp hollow)and 2.01 Å (fcc hollow). The measured value of 1.55 Åwould be intermediate between the last of these two, andthe most likely interpretation would be a mixedoccupation of the two hollow sites.The RT thiolate showed a much smaller (111) coherentposition (D(111)=1.07 Å), and its (-111) coherent fractionwas close to zero (D(111)=0.15 ).The data could be only reasonably interpreted with alocal reconstruction of the outermost Cu atom layer to amore open packing structure, allowing the S atoms of theRT thiolate to penetrate deeper into the top layer. Thevery low (-111) coherent fraction was consistent with thisreconstruction provided be incommensurate or have alarge unit mesh with many nonequivalent local siteregistries.The measured coherent positions for the atomic S,triangulated quite well to the occupation of fcc hollowsites. In fact, the measured (111) coherent position of 1.56Å would lead to an expected (-111) coherent position forthis site of 1.91 Å, close to the measured value of 1.97 Å(see table 4). However, the coherent fraction for thesespecies were too low for a single high symmetry

adsorption site. For a single hollow site the Cu-S bondlength, is 2.15Å, significantly shorter than that measured,2.30Å, in a pure atomic S on Cu(111) [22]. Sulphur is known to cause complex reconstruction of theCu(111) surface. A mixture of an unreconstructedoverlayer and a partially penetrated reconstructedsurface could account for the apparent bond length andfor the low coherent fractions.

ConclusionsA brief review of the theory of the XSW absorption, andof the NIXSW technique was presented. Previously, theXSW method for the determination of adsorbatestructures on surfaces was applied only on high perfectsemiconductor sample using dedicated and specialdesigned beamlines. These restrictions were removedusing the normal incidence Bragg reflection geometry:conventional SEXAFS beam line became appropriatedand single metal crystals having a high degree ofmosaicity could be investigated. The aim of this paper is to highlight some developmentof the study of adsorbates on metal surfaces conseguentthe use of third-generation synchrotron radiationfacilities. Recent experimental results showed thatchemical-shift XSW technique provides complementaryand quantitative information on several chemicallyimportant problems involving the surface chemistry andthe structure of adsorbates. These results showed that:- Surface sample temperature and X-ray incident beaminduce fragmentation of the adsorbate.- Reactions between coadsorbed fragments aresuppressed at low temperature but occur at enough hightemperature.- CS-NIXSW is able to determine the local adsorptiongeometries of all coadsorption reaction products on thesurface. These local adsorption sites determination canbe done at different temperatures allowing us to followthe local adsorption sites through a surface chemicalreaction.

AcknowledgementI wish to acknowledge especially the collaboration withD. P. Woodruff, R. G. Jones, and B. C. C. Cowie sincetogether with them I started to explore the possibilities ofthe chemical-shift normal-incidence X-ray standingwaves technique. I would like also to acknowledge G. J.Jackson, N. K. Singh, J. McCombie, J. Ludecke, A. S. Y.Chan, and C. Fisher. Finally, I wish to acknowledge E.Colavita for useful conversation and critical reading ofthe manuscript.

References1. B. Krassig, M. Jung, D. S. Gemmell, E. P. Kanter, T.LeBrun, S. H.

Southworth, and L. Young Phys. Rev. Lett. 75, 4736 (1995).

2. M. Jung, B. Krassig, D. S. Gemmell, E. P. Kanter, T.LeBrun, S. H.Southworth, and L. Young Phys. Rev. A 54, 2127 (1996).

3. C. J. Fisher, R. Ithin, R. G. Jones, G. J. Jackson, D. P. Woodruff, and B.C. C. Cowie J. Phys. Condens. Matter 10, L623 (1998).

4. D. P. Woodruff, Prog. Surf. Sci. 57, 1 (1998).5. G. J. Jackson, B. C. C. Cowie, D. P. Woodruff, R. G. Jones, M. S.

Kariapper, C. Fisher, A. S. Y. Chan, and M. Butterfield, Phys. Rev. Lett.84 (2000) 2346

6. K.U. Weiss, R. Dippler, K. M. Schindler, P. Gardner, V. Fritzsche, A.M.bradshaw,D. P. Woodruff, M..C. Acensio, and A.R. Gonzales-ElipePhys. Rev. Lett. 71 (1993) 581

7. G. J. Jackson, J. Ludecke, D. P. Woodruff, A. S. Y. Chan, N.K. Singh, J.McCombie, R. G. Jones,B. C. C. Cowie , and V. Formoso SurfaceScience 441 (1999) 515

8. S.A.Joyce, J.A. Yarmoff, T.E. Madey Surface Science 245 (1991) 17829. G. J. Jackson, S.M. Driver, D. P. Woodruff, N. Abrams, R. G. Jones,

M.T. Butterfield, M.D. Crapper, B. C. C. Cowie , and V. FormosoSurface Science 459 (2000) 231

10. G. J. Jackson, D. P. Woodruff, R. G. Jones, N.K. Singh, A. S. Y. Chan, B.C. C. Cowie , and V. Formoso Phys. Rev. Lett. 84 (2000) 2346

11. N.P. Prince, D.L. Seymour, M.J. Ashwin, C.F. McConville, P.Woodruff,and R. G. Jones Surface Science 230

12. B. Krassig, M. Jung, D. S. Gemmell, E. P. Kanter, T.LeBrun, S. H.Southworth, and L. Young Phys. Rev. Lett. 75, 4736 (1995).

13. M. Jung, B. Krassig, D. S. Gemmell, E. P. Kanter, T.LeBrun, S. H.Southworth, and L. Young Phys. Rev. A 54, 2127 (1996).

14. C. J. Fisher, R. Ithin, R. G. Jones, G. J. Jackson, D. P.Woodruff, and B. C.C. Cowie J. Phys. Condens. Matter 10, L623 (1998).

15. D. P. Woodruff, Prog. Surf. Sci. 57, 1 (1998).16. G. J. Jackson, B. C. C. Cowie, D. P. Woodruff, R. G. Jones, M. S.

Kariapper, C. Fisher, A. S. Y. Chan, and M. Butterfield, Phys. Rev. Lett.84 (2000) 2346

17. K.U. Weiss, R. Dippler, K. M. Schindler, P. Gardner, V. Fritzsche, A.M.bradshaw,D. P. Woodruff, M..C. Acensio, and A.R. Gonzales-ElipePhys. Rev. Lett. 71 (1993) 581

18. G. J. Jackson, J. Ludecke, D. P. Woodruff, A. S. Y. Chan, N.K. Singh, J.McCombie, R. G. Jones,B. C. C. Cowie , and V. Formoso SurfaceScience 441 (1999) 515

19. S.A.Joyce, J.A. Yarmoff, T.E. Madey Surface Science 245 (1991) 178220. G. J. Jackson, S.M. Driver, D. P. Woodruff, N. Abrams, R. G. Jones,

M.T. Butterfield, M.D. Crapper, B. C. C. Cowie , and V. FormosoSurface Science 459 (2000) 231

21. G. J. Jackson, D. P. Woodruff, R. G. Jones, N.K. Singh, A. S. Y. Chan, B.C. C. Cowie , and V. Formoso Phys. Rev. Lett. 84 (2000) 2346

22. N.P. Prince, D.L. Seymour, M.J. Ashwin, C.F. McConville, P.Woodruff,and R. G. Jones Surface Science 230 (1990) 13


38


39


Neutron scattering has made a unique and fundamentalcontribution to the understanding of the structure anddynamics of condensed matter. Although much of theearly work was based on reactor sources; in recent years,through the contributions of ISIS, IPNS. LANSCE andKEK, the potential of accelerator based pulsed neutronsources has been firmally established. The major newpulsed source developments, the SNS project in the USA,the Joint Hadron Facility in Japan, and ultimately theEuropean Spallation Source, demonstrate the importanceof pulsed neutron sources to the future of neutronscattering.ISIS is currently one of the most intense and extensivelyinstrumented pulsed neutron sources, and the effective

and growing exploitation of ISIS over a broad range ofcondensed matter research has powerfully demonstratedthe specific benefits of the time of flight technique inneutron scattering on pulsed sources. The Second TargetStation will build on the success of ISIS, and extend itscapability into new areas.The proposed Second Target Station at ISIS is a lowfrequency, low power target station, which wouldoperate at 10 hz, taking 1 – 5 pulses from the ISISSynchrotron (see figure 1).The low power dissipation and low frequency will enableit to be optimised for the production of cold neutrons, ina way not possible on the existing high power targetstation. Substantial gains in performance of greater than

A SECOND TARGET STATION AT ISISJ. PenfoldISIS Facility, Rutherford Appleton Laboratory, CCLRC, Chilton, Didcot, Oxon, UK

Articolo ricevuto in redazione nel mese di Novembre 2000

Fig. 1. Schematic representation of the Second Target Station at ISIS

DOVE NEUTRONI

40


DOVE NEUTRONI

an order of magnitude over the existing target station atISIS will be achieved for cold neutrons and highresolution instrumentation with a broad spectral range.These potentially impressive gains in the capability ofcold neutron scattering and high resolution studies willprovide exciting new opportunities in the technologicallysignificant areas of Soft Condensed Matter, AdvancedMaterials and Bio-Molecular Sciences. Just as the adventof the high flux reactor at the Institute Laue Langevin,Grenoble in the early 1970’s, with its dedicated coldneutron source, broadened considerably the appeal ofneutron scattering, so the development of the SecondTarget Station at ISIS, optimised for long wavelengthneutrons, will have a major impact on the study ofcomplex condensed matter systems. The combination ofthe cold neutron flux, the simultaneously available broadspectral range and the potential for high resolution, willprovide facilities that are not available elsewhere.Within the UK a number of major technological themesand priorities have been identified (a number of similarexercises have been carried out in many other Europeancountries), many of which are closely associated with thenew research opportunities that will be afforded by thedevelopment of the Second Target Station at ISIS, and areclosely associated with the broad areas of Soft CondensedMatter, Bio-Molecular Science and Advanced Materials.(See Table 1).

Chemicals Bio-chemical technology, advanced materials,

polymers, processing, sensors.

Energy Enhanced oil recovery, waste management.

Materials Sensors, advanced materials, adhesion, surface

Engineering.

Manufacturing Processing, product formulation.

Defence and Aerospace Sensors, advanced materials, process

technology.

Retail and Distribution Intelligent or smart packaging.

Health and Life Sciences Drug delivery, drug creation, pharmaceuticals,

immune response manipulation, metabolic

pathways.

Food and Drink Material changes during processing.

Agriculture Pesticides, environmental controls.

Table 1.

In the areas of Advanced Materials and Soft Matter thenew scientific opportunities are in the study of complex,multi-component or multi-phase systems, the use ofcomplex sample environments, and the investigation ofnon-equilibrium systems. In such systems the dimensionscales of importance often range from molecular to meso-scale. This dictates the need for a broad wavelengthrange with a particular emphasis on cold neutrons.Kinetic studies (ranging from chemical reactions to

probing dynamic surface tension) require the same broadspectral range with higher fluxes of cold neutrons. Multi-component or multi-phase systems are only tractablewith the parametric studies possible using enhanced fluxand high resolution. The Life Sciences are currentlymaking an immense impact, with many health-relatedissues underlying this importance. The success of X-raycrystallography in solving complex proteins and virusstructures to high resolution has been critical forunderstanding structure-function relationships, and thishas been further boosted by the Human Genome Projectand the prospects of post genome research on a largenumber of newly identified protein sequences. The roleof neutron crystallography is secondary to these highresolution x-ray studies. However, there are importantcontributions that neutrons will make in, for example,determining water or hydrogen location at lowerresolution. In the broader context the post Genome erawill provide exciting and important new opportunities inthe broader bio-molecular sciences remit, in fields ofpharmacy, food science, sensors, bio-compatibility andbio-functionality. In these particular areas there is muchoverlap with the areas of interest identified in Soft Matterand Advanced Materials, and their specific requirements.The main scientific areas in which significantdevelopments are envisaged are summarised as follows,

Soft Condensed MatterSurface, interfacial and bulk properties of complex fluids(polymers, surfactants, colloids).• Interfacial studies: self-assembly and ordering of

complex mixtures of surfactants, polymers and proteinsat interfaces; with emphasis on kinetic processes andmulti-component systems at technologically relevantinterfaces (liquid-liquid and liquid-solid), thin filmdevices.

• Processing of soft solids: relationship betweenmicroscopic structures and bulk properties (rheology)in industrially relevant fluid fields.

• Self-Assembly: structure of lyotropic meso-phases,micro-emulsions and vesicles, with emphasis ondynamics of structural phase changes, association anddisassociation, and self-assembly in super-criticalfluids.

• In-situ electrochemistry.

Bio-molecular SciencesPharmaceuticals, drug delivery formulations,membrane-protein, interactions, bio-compatibility andfunctionality, food technology.• Interfaces and membranes: structural organisation of

membranes and membrane-protein systems.• Macromolecular assemblies: low resolution studies on

macromolecular assemblies, in systems not tractable by

DOVE NEUTRONI

41


high resolution crystallography, viruses, glyco-proteins, protein folding, protein-nucleic acidinteractions. Solvent structure. Meso-scale structure.

• Pharmaceuticals: determination of new drug structures,where the role of hydrogen atoms is essential inunderstanding drug-receptor interactions. Molecularengineering.

• Food Technology: study of solvent distribution andstructural changes in complex assemblies (for example,starches) during processing. Protein fouling.

Advanced MaterialsCrystalline, magnetic, disordered and engineeringmaterials; including complex inorganic and organicassemblies, clathrates, intercalates, zeolites, nano-structured materials, high temperature superconductors,giant magneto-resistance materials, magnetic films andmulti-layers, spin valves, glasses, complex fluids, porousmedia.• Structural details of giant/collossal magnetoresistive

materials, non-stoichiometric oxides, piezoelectric,ferroelectric and negative thermal expansion materials.

• Composite multi-crystalline materials, such as mixedgeological phases, magnetoresistive manganates,toughened engineering materials and high Tc materials.

• Studies under extreme conditions: catalysis, in-situchemical reactions (including electrochemistry andbattery function), ultra-high pressures, novelprocessing routes.

• Structure of nano-scale materials: sol-gel processing,ceramics.

• Structure of complex materials: Role of zeolites in ionexchange materials and catalysis, novel electrodes andelectrolytes, clathrate formation, nano-structuredmaterials.

• Determination of new magnetic structures: materials withnovel ground states.

• Magnetic thin films, multilayers, spin electronics.• Nucleation and growth processes, molecular clustering,

glasses.• Complex fluids in porous media; oil-recovery; waste

disposal; supercritical fluids.• Stress/strain mapping in engineering materials.To realise this broad area of inter-disciplinary research apreliminary instrument suite has been identified, in theareas of Large Scale Structures, Diffraction andSpectroscopy (see figure).The specific benefits of the time of flight method onpulsed neutron sources include improved instrumentalresolution, wide simultaneous coverage of momentumand energy transfer, intrinsically low backgrounds, theease of use of fixed geometry sample environmentequipment, the high fluxes of epithermal neutrons, andthe broad spectral range. The second target station willextend and reinforce those benefits for theinstrumentation described above to which include:

•A wider dynamical range in momentum transfer forcrystalline and non-crystalline diffraction, SANS andreflectometry, and in energy and momentum transferfor spectroscopy, by virtue of the increase of the timeframe from 20 to 100 ms.

•Enhanced flux for long wavelength/low energyneutrons, arising from a fully optimisedtarget/moderator assembly.

•A coupled cold moderator will provide a furtherenhancement of flux for those instruments that do notrequire a sharp pulse structure.

•The availability of new technologies provides theopportunity to enhance further the performance of newinstruments on the Second Target Station. The largertime frame will allow resolution improvementsthrough the use of supermirror guides for the efficienttransport of neutrons along longer flight paths.

The key areas of instrumentation that will benefit mostare neutron reflectivity, small angle neutron scattering(SANS), very high resolution spectroscopy, non-crystalline diffraction, high resolution crystallinediffraction and large scale crystallography.Further details of the preliminary instrument suite can befound on www.isis.rl.ac/target station 2/instruments/.The second target station will be situated on the southernside of the existing high power target (see figure 6), andone pulse in five from the ISIS synchrotron will be

directed along a new proton beam line. The relatively lowpower transfer tantallum will have two cryogenicmoderators in “wing” geometry (a 25K decoupled solidmethane moderator, and a 25K coupled liquid hydrogenmoderator), and surrounded by a beryllium reflector. 9

Fig. 3. The study of protein-membrane interactions.

42


DOVE NEUTRONI

Fig. 4.

Fig. 2.

Fig. 6. Schematic representationof the Second Target Station

DOVE NEUTRONI

43


Fig. 5. Possible Instrument suite for theSecond Target Station

Fig. 7. The target, moderator and reflector assemblies.

instrument beam points will view each moderator (seefigure 7).The relatively low proton power will allow target andmoderator designs which are optimised for theproduction of long wavelength neutrons, and hence more

efficient than those on the existing target station. Inparticular, the use of a pre-moderation and a coupledhydrogen moderator will provide a significantenhancement in cold neutron flux (with a relaxed pulsestructure, ∆t ≤ 300 msec) (see figure 8).It is now recognised that neutron flux does not simplyscale with proton power/current, and that at modesttarget power levels there are gain factors and efficiencyfactors which are not available at higher power levels.This is particularly true for for cold neutron production;and this coupled with the broad wavelength band moreeasily available at low source frequencies, is an attractiveproposition.The scientific case for the Second Target Station at ISIShas been developed, and was enthusiastically endorsedby the current (and potential) UK user community at ameeting at RAL in May (see “Second Target Station atISIS: New opportunities for interdisciplinary researchusing neutrons”, RAL report, RAL-TR-2000-032, andhttp://www:isis.rl.ac.uk/ targetstation2/). Formalapproval and funding is now being sought. In themeantime detailed programme of calculations isunderway; in order to optimise and refine the neutronicperformance of the target/moderator assembly and to

provide the detailed design criteria. With an estimatedconstruction period of 3 to 4 years, it is anticipated thatthis project could be completed by 2005/6. With thefunding of the new synchrotron radiation source,Diamond, secured, it is CCLRC’s highest priority. Thesetwo new facilities will provide a powerfulcomplementary combination for inter-disciplinaryresearch in Soft Condensed Matter, Bio-MolecularSciences and Advanced Materials over the next 5 to 10years. Further details can be obtained from AndrewTaylor ([email protected]) or Jeff Penfold([email protected]).

Fig. 8. Some recent Monte Carlo calculations of the target / moderatorassembly. The triangles are for a broad pulse coupled moderator, and thecircles for a sharp pulse moderator suitable for high resolutionapplications.

DOVE NEUTRONI

44


45


The INFM Nucleo di valutazione1

(Assessment Committee) voicedtheir complete satisfaction with therelationship between ILL and INFMduring an informal visit to thefacility in July. They wereaccompanied by Roberto Felici andMatilde Bolla, respectively theresponsible and the administrator ofthe on-site INFM Outstation whichacts as a gathering point for ItalianUsers.The Committee was personallywelcomed by the ILL British Director,Colin Carlile, who described the roleof the Institute focussing particularlyon its partnership with Italy. He gavean overview of the type of neutronbeam instrumentation available atthe ILL using an interactivecomputerised model emphasisingthe investment programme recentlylaunched to upgrade the instrumentsuite at the ILL - the MillenniumProgramme. Italian use of the ILLbeam time over the past three yearswas detailed by GiovannaCicognani, former INFM employeeand now Head of ScientificCoordination at the ILL. Thestatistics show that Italy makes fulluse of its 3% budget allocation and,in fact, demand exceeds currentavailability demonstrating vividlythe health of the Italian neutronscattering community.The Assessment Committee weretreated to a tour of the swimmingpool of the reactor Level D, wherethey looked at the reactor pool and

saw the Cherenkov effect, which isalways an impressive sight. Theysaid that they considered it to be areal privilege to be guided by theHead of the Reactor Division,Ekkehardt Bauer, who took the timeto answer in detail their manyquestions.They finished the tour by a visit tothe experimental hall, where theymet Giovanna Fragneto (instrumentresponsible of the ILL diffractometerD16), Francesca Natali (post-doc onthe Italian CRG IN13 from Parma),Ferndinando Formisano (localresponsible of the BRISP BrillouinSpectrometer), Claudia Mondelli(PhD at the ILL) and Italian users onsite.The visit complemented the scientificpresentations which had been givenon the IN13 backscatteringspectrometer and BRISP BrillouinSpectrometer by Francesco Sacchettito the Panel at its meeting held inGenova in June this year.

Many thanks to everyone at the ILLwho helped us in the organisation ofthe day. Both the ILL team and theINFM Assessment Committeeconsidered the visit as very positiveand one which underlined theexcellent relations between the twoparties.

Giovanna CicognaniILL-SCO

Assessment Committee supports INFM at ILL

1 Members of the Committee are: Prof. S. Carra(Milan, Chairman), Dr. L. Scarpitti (ENEA,Rome), Prof. M. Calderini (Turin), Prof F Toigo(Padua), Prof. C. Zannoni (Bologna) TheCommittee was accompanied by Dr. F. Calvi(INFM, Genoa), Dr E. Narducci (INFM,Genova). Ekkehardt Bauer (second from the left) explaining the Cherenkov effect at the reactor pool.

VARIE

A glance at the ILL future: Colin Carlileexplaining the ILL interactive mock-up duringthe INFM Assessment Committee visit to thefacility (from the left: C.M. Bertoni, N. Narducci,F. Calvi, S. Carra, C. Carlile and C. Zannoni.

Il CNR ha approvato nel 2000 un

nuovo Progetto di Ricerca e Svilup-

po nel campo delle applicazioni del-

le sorgenti di luce di sincrotrone di

quarta generazione basate su sorgen-

ti free electron laser (fel).

Scopo del progetto è la realizzazione

di esperimenti su sorgenti di luce di

Sincrotrone di quarta generazione,

basate su sorgenti Free Electron La-

ser (FEL). Tali sorgenti permettono

di raggiungere un flusso di fotoni ed

una brillanza di diversi ordini di

grandezza superiori a quelli otteni-

bili nelle migliori sorgenti attuali.

Se a tali eccezionali caratteristiche si

aggiunge la coerenza tipica delle sor-

genti laser, si capisce come sia possi-

bile pensare esperimenti completa-

mente nuovi ed esplorare nuovi

campi di ricerca della fisica, della

chimica, della scienza dei materiali

in generale, della biologia e della

medicina.

Attualmente sorgenti di tipo FEL so-

no state realizzate nell’infrarosso e

sono iniziati progetti per la realizza-

zione di FEL nel vicino ultravioletto

e nella regione dei raggi X molli. Le

sorgenti nell’ultravioletto e nei raggi

X si basano sull’effetto SASE (Self

Amplified Spontaneous Emission),

in cui la radiazione coerente origina

dall’interazione tra la radiazione

spontanea, emessa dagli elettroni

iniettati in un ondulatore, ed il pac-

chetto stesso di elettroni.

E’ stato recentemente dimostrato,

presso i laboratori di Stanford, che

l’effetto SASE è possibile nell’infra-

rosso e si sta procedendo con nuovi

progetti verso la realizzazione di

sorgenti FEL ad energie più alte

Uno di tali progetti è in corso di rea-

lizzazione ad Amburgo presso i la-

boratori DESY.

Il programma del Progetto si svilup-

perà su due linee parallele: la prima

prevede l’utilizzo di una sorgente

FEL nell’infrarosso già operativa a

Nashville (USA); la seconda l’utiliz-

zo della sorgente DESY.

In particolare il programma di ricer-

ca si articolerà su tre diverse classi di

esperimenti:

1. accoppiamento di un microscopio

ottico a campo vicino (SNOM)

con il FEL di Vanderbilt per utiliz-

zarlo in modo spettroscopico. So-

no previsti esperimenti su films

sottili; esperimenti su fluttuazioni

laterali di barriere di potenziale

presenti in particolari interfacce

di interesse per la microelettroni-

ca; esperimenti su campioni bio-

logici.

2. spettroscopia di assorbimento a

due fotoni presso la sorgente di

DESY su campioni di alogenuri

alcalini, bromuri e floruri, quarzo

e ad alcuni tipi di ossidi.

3. fotorisoluzione di mescolanze di

aminoacidi con luce polarizzata

circolarmente presso il FEL di

Amburgo.

Alle attività sperimentali suddette

verrà aggiunta una attività di forma-

zione nel campo della cristallografia

delle macromolecole. Tale attività

verrà per ora realizzata presso le sor-

genti di luce di sincrotrone attuali :

ELETTRA di Trieste ed ESRF di Gre-

noble, con l’intenzione di continuar-

la, appena possibile, su una sorgente

X-FEL.

Il Progetto, che ha una durata di tre

anni, é coordinato dal Dr. P. Perfetti

ed ha come Unità operative:

- Istituto di Struttura della Materia,

CNR, Tor Vergata (Dr. P. Perfetti)

- Istituto di Chimica dei Materiali,

CNR, Montelibretti (Dr. T. Prosperi)

- Istituto di Biologia Cellulare, CNR,

Campus Buzzati Traverso, Monte-ro-

tondo (Prof. G. Tocchini Valentini)

P. Bosi

Segretaria Scientifica

Commissione Luce di Sincrotronedel CNR

VARIE

46


Premessa

Nel contesto delle diverse ipotesi che

circolano attualmente sul futuro del-

la neutronica in Italia, anche in rela-

zione alle diverse iniziative che si

stanno sviluppando su scala Euro-

pea, si ritiene opportuno che la So-

cietà Italiana di Spettroscopia Neu-

tronica (SISN) assuma il ruolo che le

spetta, a livello di consulente nazio-

nale, quale portavoce della comunità

neutronica italiana.

In generale, basandosi sulle attuali

prospettive, si può affermare che le

sorgenti termiche (reattori) non sem-

brano destinate ad espandersi, con

l'unica eccezione, forse, del reattore

di Monaco. Viceversa, si prevede che,

se ci sarà un'espansione, questa sarà

orientata verso le sorgenti pulsate. In

particolare, tenuto conto delle diver-

se ipotesi attualmente in discussione

su scala europea (ESS, AUSTRON,

ISIS-2) la SISN ritiene doveroso dare

una valutazione obiettiva degli inte-

ressi della comunità nazionale.

A questo scopo, è importante che la

comunità neutronica italiana, infor-

mata di quanto sta accadendo sulla

scala europea, e delle possibili azioni

che possono essere intraprese nell'in-

teresse della comunità nazionale,

faccia sentire la sua voce tramite la

SISN che viene quindi ad assumere

attivamente quel ruolo di consulen-

za che è previsto dallo statuto di co-

stituzione dell'associazione.

Un precedente forum di discussione,

tenutosi in occasione del convegno

nazionale dell'INFM a Genova, ha

evidenziato alcuni punti che posso-

no rappresentare una chiave di lettu-

ra per quanto concerne le future sor-

genti neutroniche in ambiente euro-

peo. In tale riunione sono state di-

scusse sia le prospettive di nuova

scienza, accessibile con le nuove sor-

genti pulsate, sia le informazioni tec-

niche sulle varie ipotesi attualmente

aperte (ESS, AUSTRON, ISIS-2, LE-

GNARO). Dalla discussione di Ge-

nova sono stati individuati due pun-

ti importanti:

1.prospettive per le nuove sorgenti

europee

2.ottimizzazione dell'accesso alle

sorgenti esistenti per la comunità

italiana

Su questi punti, l'Assemblea della

SISN, riunitasi a Roma il 19 Ottobre

2000, ha ulteriormente discusso rag-

giungendo un'opinione comune, che

viene descritta nel presente docu-

mento programmatico, e che costi-

tuisce il punto di vista ufficiale della

SISN. L'Assemblea della SISN ritiene

che questo documento debba essere

opportunamente pubblicizzato e so-

stenuto sia presso il MURST che

presso gli Enti che attualmente fi-

nanziano le attività di spettroscopia

neutronica italiana (INFM e CNR).

Prospettive per le nuove

sorgenti europee

Fermo restando che ESS rappresenta

l'obiettivo principale per la comunità

neutronica italiana, in quanto realiz-

zerebbe la creazione di una large-sca-

le facility europea, occorre considera-

re che i tempi di realizzazione di tale

obiettivo sono diluiti nel tempo e so-

no attualmente stimabili in circa un

decennio di ulteriore studio e pro-

gettazione. Inoltre, occorre tener

conto che nel breve periodo (5-10 an-

ni) sono prevedibili sensibili diminu-

zioni del flusso medio di neutroni

accessibile alla ricerca, anche in con-

seguenza dell'invecchiamento, ed al-

la conseguente obsolescenza, di sor-

genti storiche che presumibilmente si

stanno avvicinando alla fine del loro

ciclo produttivo. In questo contesto,

deve essere considerato sintomatico

lo shut-down definitivo del reattore

di Risø e la crisi momentaneamente

intervenuta negli USA a seguito del-

la chiusura di BNL e del lungo shut-

down di Oak Ridge.

Nel contempo, la costruzione del se-

cond target station ad ISIS, denominata

ISIS-2, risulta un'ipotesi più che reali-

stica che, secondo la programmazio-

ne prevista, potrebbe essere realizzata

sulla scala di qualche anno. Il proget-

to prevede, per questa realizzazione,

un aumento della corrente nell'accele-

ratore da 200 a 300 A e lo splitting del

SISN: Documento programmatico della SISN sulle futuresorgenti neutroniche su scala europea

VARIE

47


48


VARIE

fascio di protoni su due distinti bersa-

gli di spallazione. Il primo dovrebbe

mantenere le stesse caratteristiche che

ISIS presenta attualmente (a 50 Hz di

frequenza), mentre per il secondo è

prevista una frequenza più bassa (10

Hz) che permetterebbe, con l'allunga-

mento del frame, l'utilizzo più effica-

ce dei neutroni freddi ed il conse-

guente trasferimento su ISIS-2 della

strumentazione che li utilizza (basso

angolo, riflettometria, diffrazione ad

alta risoluzione, spettroscopia quasi-

elastica, etc.). Ovviamente, questo

comporterà la contemporanea messa

a disposizione di beam-lines attual-

mente occupate su ISIS-1 e permet-

terà lo sviluppo di nuova strumenta-

zione che viceversa utilizza più effi-

cacemente i neutroni termici e/o epi-

termici.

Dal punto di vista realizzativo, non

si prevedono problemi sostanziali in

quanto la tecnologia necessaria è già

acquisita ed il problema si riduce al-

la reperibilità di mezzi finanziari. I

tempi caratteristici dell'operazione

sono stimati in 2-3 anni. Infine, oc-

corre tener presente che il RAL è sta-

to scelto per la costruzione di un sin-

crotrone (DIAMOND Project, in coo-

perazione con la Francia) che ren-

derà il polo di Oxford equivalente a

quello di Grenoble per quanto con-

cerne la disponibilità, nello stesso si-

to, di neutroni e luce di sincrotrone.

Ottimizzazione dell'accesso alle sorgenti

esistenti per la comunità italiana

Stante la mancanza di sorgenti neu-

troniche nazionali, la comunità ita-

liana ha attualmente accesso, tramite

accordi o convenzioni internazionali,

alle seguenti facilities europee: ILL

(Grenoble), ISIS (Oxford), LLB (Sa-

clay). Inoltre, pur non essendo ope-

ranti specifici accordi o convenzioni

esistono collaborazioni scientifiche a

vari livelli che permettono l'utilizzo

da parte dei ricercatori italiani di al-

tre sorgenti neutroniche europee

(Svezia e Germania) o negli USA

(Brookhaven, Oak Ridge, Los Ala-

mos, etc.). La partecipazione attuale

dell'Italia presso ILL prevede un ac-

cesso a livello del 3% del tempo

macchina totale disponibile. Tale ac-

cordo è gestito da una convenzione

con l'INFM. Per quanto concerne

ISIS, all'Italia è garantita una parteci-

pazione del 5% che viene gestita da

un accordo internazionale con il

CNR. La convenzione con LLB è ge-

stita anch'essa da INFM.

La costruzione di un second target ad

ISIS, con la conseguente liberazione

di un discreto numero beam-lines, e

la realizzazione del millennium pro-

gramme dell'ILL, che prevede l'up-

grading di diversi strumenti e la co-

struzione di nuovi, rappresentano

un'occasione che vale la pena coglie-

re per accrescere le competenze ita-

liane nel campo della strumentazio-

ne neutronica. Inoltre, si sottolinea

che la comunità italiana vanta nume-

rosi e proficui rapporti di collabora-

zione scientifica con LLB che potreb-

bero essere ulteriormente potenziati

da valide proposte nel campo del

rinnovo del parco strumenti.

Per quanto concerne ILL, il millen-

nium programme prevede, in una pri-

ma fase, il potenziamento di 5 stru-

menti:

- Stress residui

- Thermal LADI

- Polarizzazione su IN20

- Potenziamento del D3

- Time resolved SANS

L'Italia ha proposto che IN13 (CRG

con la Francia) venga incluso nel MP

e, probabilmente, questo rientrerà

nel quadro più generale del miglio-

ramento delle guide per neutroni

con un conseguente incremento di

flusso al campione. Un contributo

italiano per il miglioramento delle

guide di IN13, che venisse incluso

nel millennium programme, sarebbe

altamente auspicabile.

Inoltre, per venire incontro alle cre-

scenti esigenze scientifiche della co-

munità italiana sarebbe anche auspi-

cabile un incremento della quota di

partecipazione italiana ad ILL. A

questo proposito, si consideri che, a

fronte di una richiesta ben più eleva-

ta, la quota di utilizzo italiana non

può mai superare, secondo i termini

dell'accordo, la percentuale negozia-

ta del 3% del tempo macchina totale

disponibile.

Per quanto concerne ISIS, si deve os-

servare che la scelta dello strumento

inerente il progetto TOSCA, che fa

parte integrante dell'accordo di coo-

perazione tra il CNR ed CRLC, è sta-

ta condizionata dalla mancanza di

una beam-line disponibile ed è stata

quindi, in qualche maniera, imposta

da ragioni oggettive ma contingenti.

L'accordo di cooperazione, che giun-

48


ge a scadenza con il 2001, dovrà es-

sere rinnovato e si reputa necessario

proseguire sulla strada oramai trac-

ciata di includere la costruzione di

uno strumento nei termini dell'ac-

cordo. Ovviamente, la costruzione di

ISIS-2 apre molte prospettive e tutto

lascia presagire che il prossimo stru-

mento potrà essere scelto avendo a

disposizione un maggior numero di

gradi di libertà.

Il CNR ha erogato, nel corrente anno

2000, un primo finanziamento a

fronte della costruzione di una sta-

zione sperimentale italiana a valle

dello strumento TOSCA. La Com-

missione Neutroni del CNR si è già

pronunciata affinché tale stazione

sperimentale sia dotata di un diffrat-

tometro. Non si esclude, comunque,

che questa possa essere dotata di al-

tra strumentazione (p. es. anelastica)

per scopi di test e/o formazione. In-

fine, è doveroso ricordare che la di-

sponibilità (se pur parziale) di una

sorgente di neutroni epitermici per-

metterebbe alla comunità neutronica

italiana di presentarsi con altre cre-

denziali, che non quelle attuali, al-

l'appuntamento con ESS.

La convenzione con LLB ha reso

possibile la realizzazione di proget-

ti sperimentali che, per le loro ca-

ratteristiche, trovano difficile accet-

tazione presso facilities di maggiori

dimensioni, ha facilitato l’instaurar-

si di proficue collaborazioni scienti-

fiche con centri di ricerca industria-

li italiani ed ha permesso l’adde-

stramento di numerosi giovani ri-

cercatori italiani. Si auspica pertan-

to che la convenzione venga rinno-

vata alla sua scadenza naturale di

fine anno 2000.

Infine, per quanto concerne l'assenza

cronica di sorgenti neutroniche na-

zionali, che indubbiamente potrebbe-

ro facilitare il processo di formazione

dei giovani ricercatori, appare oggi

chiaro che il reattore della Casaccia

risulta inaccessibile e che, forse, l'uni-

ca alternativa nazionale è rappresen-

tata da un'ipotesi di accordo tra

l'INFM e l'INFN per realizzare, pres-

so i Laboratori Nazionali di Legnaro

(PD), una piccola sorgente pulsata di

neutroni. Questa sorgente, per la

quale sono state effettuate positiva-

mente alcune prove di fattibilità, po-

trebbe fornire un flusso di neutroni

sufficiente per costituire un ottimo

laboratorio finalizzato alla formazio-

ne del personale ed alla realizzazione

di tests su rivelatori e strumentazio-

ne neutronica in generale.

Conclusioni e raccomandazioni

In conclusione, si può affermare che

la via ottimale per uno sviluppo ar-

monico della neutronica in Italia

passa per un coinvolgimento italiano

in ISIS-2, che permetterebbe l’acces-

so in tempi brevi ad una sorgente di

nuova generazione, lo sviluppo delle

esistenti linee di ricerca e l’acquisi-

zione del know-how (sia per il lato

tecnico che per quello delle risorse

umane) necessario per una efficace

partecipazione ad ESS.

La rafforzata partecipazione italiana

ad ILL, con le iniziative di completa-

mento dei CRG già avviati e nell'am-

bito del millennium programme, per-

metterebbe alla ricerca italiana di

crescere ulteriormente avvicinandosi

ai livelli europei.

La continuazione della partecipazio-

ne italiana ad LLB consoliderebbe il

proficuo rapporto instaurato da an-

ni, estendendo a nuovi giovani ricer-

catori possibilità di formazione già

sperimentate nel passato.

Un eventuale accordo INFM-INFN

per la realizzazione di una piccola

sorgente pulsata a Legnaro permet-

terebbe, infine, di disporre di una

sorgente nazionale presso la quale,

parallelamente alla stazione italiana

ad ISIS, all'OGG presso ILL, ed al-

l’auspicata SANS facility presso LLB,

potrebbe essere esteso il progetto di

formazione di personale giovane,

che riveste un'importanza strategica

per la nostra comunità, in attesa del-

lo sviluppo delle sorgenti di nuova

generazione.

Marco Zoppi

Segretario SISN

VARIE

49


TMRTMRTraining and Mobility of Researchers

Support for activitiesin the field of neutron scattering is availablefrom the neutron round-table.

The neutron round-table is funded by the EC (DGXII)with approximately 100.000 Euro per year. The mission of the round-table is:

1. To actively

encourage co-

ordination and collaboration

between user facilities - such

that the European users will

benefit through a better quality

and an increased quantity of

access to the European neutron

scattering facilities.

2. To spread the

knowledge about the

potential of neutron scattering,

and support studies on future

prospects with neutron

scattering.

3. To support training

of young scientists

and other scientists, new to the

field of neutron scattering about

the potential of the method.

4. The round-table

supports non-

national access to summer

schools, workshops, training

courses, co-ordination activities

etc. Detailed information on

how and when to apply for

support can be found on the

round-table web page:

http://www.risoe.dk/fys/TMR.

htm

5. The round-table

consist of

representatives from all major

European neutron user facilities,

from EC supported networks

developing novel

instrumentation and techniques

for neutron scattering plus 5

user representatives appointed

by ENSA (European Neutron

Scattering Association). The

name of all contact persons can

be found on the web page

mentioned above. The present

chairman/co-ordinator of the

round-table is Kurt Nørgaard

Clausen, and can be contacted

as [email protected]

50


VARIE

18-19 gennaio 2001 FOLGARIA (TN), ITALY

1O Convegno Utenti GILDAPaolo Fornasini, Dipartimento di Fisica, Università diTrento, 38050 Povo (TN), Italiae-mail: [email protected]://www.science.unitn.it/~rx/gilda01/bando.html

23 maggio -3 giugno 2001 ERICE, ITALY

International School of CrystallographyPaola Spadon, Dipartimento di Chimica Organica,Università di Padova, Via Marzolo 1, 35131 Padova,Italia.Tel: +39 049 8275275; Fax: +39 049 8275239e-mail: [email protected]

23-27 luglio 2001 TRIESTE, ITALY

13O International Conference on Vacuum UltravioletRadiation Physics (VUV-XIII) e-mail: [email protected]://vuv13.elettra.trieste.it/vuv13/

30 luglio - 2 agosto 2001 ROME, ITALY

VUV-XIII Satellite Meeting"Decay Processes in Core-Excited Species"M.N.Piancastelli, Dipartimento di Chimica, University"Tor Vergata", Rome, [email protected]://www.uniroma2.it/eventi/coredec/

9-13 settembre 2001 MUNCHEN, GERMANY

International Conference on Neutron Scattering 2001(ICNS 2001)Physik Dept. E13, Technische Univ. München , D-85747Garching, GermanyTel: +49 89 28912452; Fax: +49 89 289 12473e-mail: [email protected]://www.icns2001.de

maggio 2002 NIST, USA

American Conference on Neutron Scattering

51


CALENDARIO

52


SCADENZE

Scadenze per richieste di tempo macchina presso alcuni laboratori di Neutroni

ISISLa scadenza per il prossimo call for proposalsè il 16 aprile 2001 e il 16 ottobre 2001

ILLLa scadenza per il prossimo call for proposalsè il 15 febbraio 2001 e il 15 agosto 2001

LLB-ORPHEE-SACLAYLa scadenza per il prossimo call for proposalsè il 1 ottobre 2001per informazioni: Secrétariat Scientifique du LaboratoireLéon Brillouin, TMR programme, Attn. Mme C. Abraham, Laboratoire Léon Brillouin,CEA/SACLAY, F-91191 Gif-sur-Yvette, France.Tel: 33(0)169086038; Fax: 33(0)169088261 e-mail: [email protected]://www-llb.cea.fr

BENSCLa scadenza è il 15 marzo 2001 e il 15 settembre 2001

RISØ E NFLLa scadenza per il prossimo call for proposalsè il 1 aprile 2001

Scadenze per richieste di tempo macchinapresso alcuni laboratori di Luce di Sincrotrone

ALSLe prossime scadenzesono il 15 marzo 2001 (cristallografia macromolecolare)e il 1 giugno 2001 (fisica)

BESSYLe prossime scadenzesono il 15 febbraio 2001 e il 4 agosto 2001

DARESBURYLa prossima scadenzaè il 30 aprile 2001 e il 31 ottobre 2001

ELETTRALe prossime scadenzesono il 28 febbraio 2001 e il 31 agosto 2001

ESRFLe prossime scadenzesono il 1 marzo 2001 e il 1 settembre 2001

GILDA(quota italiana) Le prossime scadenzesono il 1 maggio 2001 e il 1 novembre 2001

HASYLAB(nuovi progetti) Le prossime scadenzesono il 1 marzo 2001, il 1 settembre 2001e il 1 dicembre 2001

LURELa prossima scadenza è il 30 ottobre 2001

MAX-LABLa scadenza è approssimativamente febbraio 2001

NSLSLe prossime scadenzesono il 31 gennaio 2001, il 31 maggio 2001e il 30 settembre 2001

ALS Advanced Light SourceMS46-161, 1 Cyclotron Rd Berkeley, CA 94720, USAtel:+1 510 486 4257 fax:+1 510 486 4873http://www-als.lbl.gov/Tipo: D Status: O

AmPS Amsterdam Pulse StretcherNIKEF-K, P.O. Box 41882, 1009 DB Amsterdam, NLtel: +31 20 5925000 fax: +31 20 5922165Tipo: P Status: C

APS Advanced Photon SourceBldg 360, Argonne Nat. Lab. 9700 S. Cass Avenue,Argonne, Il 60439, USAtel:+1 708 252 5089 fax: +1 708 252 3222http://epics.aps.anl.gov/welcome.htmlTipo: D Status: C

ASTRIDISA, Univ. of Aarhus, Ny Munkegade, DK-8000 Aarhus, Denmarktel: +45 61 28899 fax: +45 61 20740Tipo: PD Status: O

BESSY Berliner Elektronen-speicherring Gessell.fürSynchrotron-strahlung mbHLentzealle 100, D-1000 Berlin 33, Germanytel: +49 30 820040 fax: +49 30 82004103http://www.bessy.deTipo: D Status: O

BSRL Beijing Synchrotron Radiation Lab.Inst. of High Energy Physics, 19 Yucuan Rd.PO Box 918,Beijing 100039, PR Chinatel: +86 1 8213344 fax: +86 1 8213374http://solar.rtd.utk.edu/~china/ins/IHEP/bsrf/bsrf.htmlTipo: PD Status: O

CAMD Center Advanced Microstructures & DevicesLousiana State Univ., 3990 W Lakeshore, Baton Rouge,LA 70803, USAtel:+1 504 3888887 fax: +1 504 3888887http://www.camd/lsu.edu/Tipo: D Status: O

CHESS Cornell High Energy Synchr. Radiation SourceWilson Lab., Cornell University Ithaca, NY 14853, USAtel: +1 607 255 7163 fax: +1 607 255 9001http://www.tn.cornell.edu/Tipo: PD Status: O

DAFNEINFN Laboratori Nazionali di Frascati, P.O. Box 13,I-00044 Frascati (Rome), Italytel: +39 6 9403 1 fax: +39 6 9403304http://www.lnf.infn.it/Tipo:P Status: C

DELTAUniversität Dortmund,Emil Figge Str 74b,44221 Dortmund, Germanytel: +49 231 7555383 fax: +49 231 7555398http://prian.physik.uni-dortmund.de/Tipo: P Status: C

ELETTRASincrotrone Trieste, Padriciano 99, 34012 Trieste, Italytel: +39 40 37581 fax: +39 40 226338http://www.elettra.trieste.itTipo: D Status: O

ELSA Electron Stretcher and AcceleratorNußalle 12, D-5300 Bonn-1, Germanytel:+49 288 732796 fax: +49 288 737869http://elsar1.physik.uni-bonn.de/elsahome.htmlTipo: PD Status: O

ESRF European Synchrotron Radiation Lab.BP 220, F-38043 Grenoble, Francetel: +33 476 882000 fax: +33 476 882020http://www.esrf.fr/Tipo: D Status: O

EUTERPECyclotron Lab.,Eindhoven Univ. of Technol, P.O.Box 513,5600 MB Eindhoven, The Netherlandstel: +31 40 474048 fax: +31 40 438060Tipo: PD Status: C

HASYLABNotkestrasse 85, D-2000, Hamburg 52, Germanytel: +49 40 89982304 fax: +49 40 89982787http://www.desy.de/pub/hasylab/hasylab.htmlTipo: D Status: O

INDUS Center for Advanced Technology, Rajendra Nagar,Indore 452012, Indiatel: +91 731 64626Tipo: D Status: C

L U C E D I S I N C R OT R O N ESYNCHROTRON SOURCES WWW SERVERS IN THE WORLD(http://www.esrf.fr/navigate/synchrotrons.html)

FACILITIES

53


54


KEK Photon FactoryNat. Lab. for High Energy Physics, 1-1, Oho,Tsukuba-shi Ibaraki-ken, 305 Japantel: +81 298 641171 fax: +81 298 642801http://www.kek.jp/Tipo: D Status: O

KurchatovKurchatov Inst. of Atomic Energy, SR Center,Kurchatov Square, Moscow 123182, Russiatel: +7 95 1964546Tipo: D Status:O/C

LNLS Laboratorio Nacional Luz SincrotronCP 6192, 13081 Campinas, SP Braziltel: +55 192 542624 fax: +55 192 360202Tipo: D Status: C

LUREBât 209-D, 91405 Orsay ,Francetel: +33 1 64468014; fax: +33 1 64464148E-mail: [email protected]://www.lure.u-psud.frTipo: D Status: O

MAX-LabBox 118, University of Lund, S-22100 Lund, Swedentel: +46 46 109697 fax: +46 46 104710http://www.maxlab.lu.se/Tipo: D Status: O

NSLS National Synchrotron Light SourceBldg. 725, Brookhaven Nat. Lab., Upton, NY 11973, USAtel: +1 516 282 2297 fax: +1 516 282 4745http://www.nsls.bnl.gov/Tipo: D Status: O

NSRL National Synchrotron Radiation Lab.USTC, Hefei, Anhui 230029, PR Chinatel:+86 551 3601989 fax:+86 551 5561078Tipo: D Status: O

PohangPohang Inst. for Science & Technol., P.O. Box 125Pohang, Korea 790600tel: +82 562 792696 f +82 562 794499Tipo: D Status: C

Siberian SR CenterLavrentyev Ave 11, 630090 Novosibirsk, Russiatel: +7 383 2 356031 fax: +7 383 2 352163Tipo: D Status: O

SPring-82-28-8 Hon-komagome, Bunkyo-ku ,Tokyo 113, Japantel: +81 03 9411140 fax: +81 03 9413169Tipo: D Status: C

SOR-RING Inst. Solid State PhysicsS.R. Lab, Univ. of Tokyo, 3-2-1 Midori-cho Tanashi-shi,Tokyo 188, Japantel: +81 424614131 ext 346 fax: +81 424615401Tipo: D Status: O

SRC Synchrotron Rad. CenterUniv.of Wisconsin at Madison, 3731 SchneiderDriveStoughton, WI 53589-3097 USAtel: +1 608 8737722 fax: +1 608 8737192http://www.src.wisc.eduTipo: D Status: O

SRRC SR Research Center1, R&D Road VI, Hsinchu Science, Industrial Parc,Hsinchu 30077 Taiwan, Republic of Chinatel: +886 35 780281 fax: +886 35 781881http://www.srrc.gov.tw/Tipo: D Status: O

SSRL Stanford SR LaboratoryMS 69, PO Box 4349 Stanford, CA 94309-0210, USAtel: +1 415 926 4000 fax: +1 415 926 4100http://www-ssrl.slac.stanford.edu/welcome.htmlTipo: D Status: O

SRS Daresbury SR SourceSERC, Daresbury Lab, Warrington WA4 4AD, U.K.tel: +44 925 603000 fax: +44 925 603174E-mail: [email protected]://www.dl.ac.uk/home.htmlTipo: D Status: O

SURFB119, NIST, Gaithersburg, MD 20859, USAtel: +1 301 9753726 fax: +1 301 8697628http://physics.nist.gov/MajResFac/surf/surf.htmlTipo: D Status: O

TERAS ElectroTechnical Lab.1-1-4 Umezono, Tsukuba Ibaraki 305, Japantel: 81 298 54 5541 fax: 81 298 55 6608Tipo: D Status: O

UVSORInst. for Molecular ScienceMyodaiji, Okazaki 444, Japantel: +81 564 526101 fax: +81 564 547079Tipo: D Status: O

D = macchina dedicata; PD = parzialmente dedicata; P = in parassitaggio.

O= macchina funzionante; C=macchina in costruzione.

D = dedicated machine; PD = partially dedicated; P = parassitic.

O= operating machine; C= machine under construction.

FACILITIES

55


BENSCBerlin Neutron Scattering Center, Hahn-Meitner-Institut,Glienicker Str. 100, D- 14109 Berlin-Wannsee, GermanyRainer Michaelsen;tel: +49 30 8062 3043 fax: +49 30 8062 2523E - Mail: [email protected]://www.hmi.de

BNLBrookhaven National Laboratory, Biology Department,Upton, NY 11973, USADieter Schneider;General Information: Rae Greenberg;tel: +1 516 282 5564 fax: +1 516 282 5888http://neutron.chm.bnl.gov/HFBR/

ESSEuropean Spallation SourcesAndrea FournierTel: + 49 2461 61 2184E-mail: [email protected]://www.kfa-juelich.de/ess/

GKSSForschungszentrum Geesthacht, P.O.1160, W-2054Geesthacht, GermanyReinhard Kampmann; tel: +49 4152 87 1316 fax: +49 4152 87 1338E-mail: PWKAMPM@DGHGKSS4Heinrich B. Stuhrmann;tel: +49 4152 87 1290 fax: +49 4152 87 2534E-mail: WSSTUHR@DGHGKSS4

IFEInstitut for Energiteknikk, P.O. Box40, N-2007 Kjeller,NorwayJon Samseth; tel: +47 6 806080 fax: +47 6 810920 telex: 74 573 energ n E-mail: Internet [email protected]

ILLInstitute Laue Langevin, BP 156, F-38042, GrenobleCedex 9,FranceHerma Büttner; tel: +33 76207179 E-mail: [email protected]: +33 76 48 39 06 http://www.ill.fr

IPNSArgonne National Laboratory, 9700 South Cass Avenue,Argonne, IL 60439-4814, USAP.Thiyagarajan,Bldg.200,RM. D125;tel :+1 708 9723593 E-mail: THIYAGA@ANLPNSErnest Epperson, Bldg. 212;tel: +1 708 972 5701 fax: +1 708 972 4163 or +1 708 9724470 (Chemistry Div.)http://pnsjph.pns.anl.gov/ipns.html

ISISThe ISIS Facility, Rutherford Appleton Laboratory,Chilton, Didcot Oxfordshire OX11 0QX, UKRichard Heenan; tel +44 235 446744 E-mail: [email protected] King; tel: +44 235 446437 fax: +44 235 445720; Telex: 83 159 ruthlb gE-mail: [email protected]://www.isis.rl.ac.uk

JAERIJapan Atomic Energy Research Institute, Tokai-mura, Naka-gun, Ibaraki-ken 319-11, Japan.Jun-ichi Suzuki (JAERI); Yuji Ito (ISSP, Univ. of Tokyo);fax: +81 292 82 59227 Telex: JAERIJ24596http:// neutron-www.kekjpl

JINRJoint Institute for Nuclear Research, Laboratory forNeutron Physics, Head P.O.Box 79 Moscow, 141 980Dubna, USSRA.M. Balagurov;E-mail: [email protected] M. Ostaneivich;E-mail: [email protected]: +7 095 200 22 83 Telex: 911 621 DUBNA SUhttp://www.jinr.dubna.su

KFAForschungszentrum Jülich, Institut fürFestkörperforschung, Postfach 1913, W-517 Jülich,GermanyDietmar Schwahn; tel: +49 2461 61 6661; E-mail: [email protected] Maier; tel: +49 2461 61 3567; E-mail: [email protected]: +49 2461 61 2610 Telex: 833556-0 kf d

N E U T R O N INEUTRON SCATTERING WWW SERVERS IN THE WORLD(http://www.isis.rl.ac.uk)

FACILITIES

56


FACILITIES

LLBLaboratoire Léon Brillouin, Centre d’Etudes Nucleairesde Saclay, 91191 Gif-sur-Yvette Cédex FranceJ.P Cotton (LLB); tel: +33 1 69086460 fax: +33 1 69088261 telex: energ 690641 F LBS+E-mail: [email protected]://bali.saclay.cea.fr/bali.html

NISTNational Institute of Standards and Technology-Gaithersburg, Maryland 20899 USAC.J. Glinka; tel: + 301 975 6242 fax: +1 301 921 9847E-mail: Bitnet: GLINKA@NBSENTHInternet: [email protected]://rrdjazz.nist.gov

ORNLOak Ridge National Laboratory Neutron ScatteringFacilities, P.O. Box 2008, Oak Ridge TN 37831-6393 USAGeorge D. Wignall, Small Angle Scattering GroupLeader; tel: +1 423 574 5237 fax: +1 423 574 6268E-mail: [email protected]://neutrons.ornl.gov

PSIPaul Scherrer InstitutWurenlingen und VillingenCH-5232 Villingen PSItel: +41 56 310 2087 fax: +41 56 310 2939E-mail: [email protected]://www.psi.ch/sinq

RISØEC-Large Facility Programme, Physics Department, RisøNational Lab.P.O. Box 49, DK-4000 Roskilde, DenmarkK. Mortenses; tel: +45 4237 1212 fax: +45 42370115E-mail: [email protected] or [email protected] in SwedenE-mail: [email protected]

SNSSpallation Neutron SourceSNS Project Office, 701 Scarboro Road, Oak Ridge, TN37830tel: +1 865 574 0558E-mail: [email protected]://www.sns.gov/

NOTIZIARIO Neutroni e Luce di Sincrotrone - Issue 5 n.2, 2000

Documents

Transcript of NOTIZIARIO Neutroni e Luce di Sincrotrone - Issue 5 n.2, 2000