Experiments on beam deflection by crystals Masataka IINUMADepartment of Quantum MatterGraduate School of Advanced Sciences of Matter Hiroshima University

CollaboratorsV. Biryukov(a), Yu. Chesnokov(a), I. Endo(b), M. Iinuma(b), H. Kuroiwa(c), T. Ohnishi(c), S. Sawada(d), S. Strokov(b), T. Takahashi(b), K. Ueda(b): IHEP-Protovino: Hiroshima University, ADSM: Hiroshima University, VBL: KEK

Recent activities proton beam 12 GeV proton beam : beam separation bent Si crystal ( length:10 mm, angle:32.6 mrad ) bent crystallographic plane : (111) plane Experiments at proton synchrotron facility in KEK ( KEK-PS )Scope : Application to J-PARC Beam separation in the slow extraction line Crystal collimation in the future

Recent activities electron beam 150 MeV electron beam : beam deflection straight Si crystal ( thickness : 16 mm ) crystal axis tilted to the beam axis Experiments at electron beam facility in Hiroshima UniversityScope Understanding of basic properties on the electron beam deflection Application to a beam collimation in ILC

Experiment at KEK-PS12 GeV Proton SynchrotronEast counter hallEP2 lineExperiments in EP2 line North counter hall

Experiment for beam separationFluorescence plate (10 x 10 cm)CsI plate(5 x 2.5 cm)GoniometerBent crystalMain beamDeflected beamTop viewCrystalDeflectionangle12 GeVprotonsq

Typical images Deflected beamCsI platePrimary beamraw imageimage after background subtractionfluorescence plate

Results intensity dependence Angle between crystal and beam axis ( q )intensity of deflected beamprimary beam 1012 ppsmradqpps

Dependence on beam divergence5 mrad1 mrad0.5 mrad0.3 mradSmulation by using CATCH code

Results intensity dependence angle between crystal and beam axis ( q )intensity of deflected beamprimary beam 1012 ppsmradqppsbeam divergence : 0.6 mrad simulation ( CATCH code ) experiment

Estimation of crystal efficiencyN deflected = Crystal Efficiency x Angle Efficiency x N incident protons upon the crystal.ratio of number of incident protons within Lindhard angle to number of total incident protons 0.3% of Intensity( 1012 )N incident upon the crystal = 3x109Crystal Efficiency 23%

Summary on proton experiment We demonstrated the beam separation of 12 GeV proton beam with the bent crystal.

Experimental results are quantitatively understood. The c2 analysis gives the beam divergence of 0.6 mrad. By using the obtained beam divergence and data, the crystal efficiency of 23 % was obtained.

Experiment with 150 MeV electronsREFER (Relativistic Electron Facility for Education and Research) ring at Hiroshima University

REFER (Relativistic Electron Facility for Education and Research)150-MeV electron beaminjection linebeam extraction lineQM3 magnet( control beam divergence )beam intensity: 1x104 s-1 REFER ring @ Hiroshima UniversitySetup area

Fiber Optics plate with a Scintillator (FOS plate)Experiment for beam deflectionStraight Si crystal thickness: 16mmbeam profile150-MeV electron beamfqdirection of axise 2.3 mObservation of a beam profile at the FOS plate in each combination of q and f angles Linhard angle : 0.7 mrad

Vertical beam divergence: 3.0 mrad q=0 mrad Results (1)crystal angle fdeflection anglebeam divergence > 0.7 mrad ( Lindhard angle )( mrad ) ( mrad )

Vertical beam divergence: 3.8 mrad q = 0 mradResults (2)Vertical beam divergence: 5.2 mrad q = 0 mradcrystal angle f, (mrad)deflection angle, (mrad)deflection angle, (mrad)crystal angle f, (mrad)The maximum deflection angle depends on the beam divergence.

Deflection vs. beam divergenceDerivation of D by fitting the plot with 1st derivative of Gaussian functionLarger beam divergence Smaller deflectionbeam divergencenormalized deflection magnitudeD( mrad )

Simulation : Comparison with Experiment Beam divergence: 5.2 mradBeam divergence: 3.0 mradcrystal angle f, (mrad)deflection angle, (mrad)deflection angle, (mrad)crystal angle f, (mrad)Rough behavior : Agreements with experimental results Details : Discrepancy between simulation and data Simulation includes only dynamics under the string potential.

Summary on electron experiments The results showed clear evidence of ability to use crystals for handling 150 MeV electrons under the following condition. Beam divergence > 0.7 mrad ( Linhard angle )

The deflection magnitude depends on the beam divergence. Larger beam divergence smaller deflection

On rough behavior, the simple simulation agrees with the data. On details, there are discrepancies between the simulation and the data.

Future prospect For application to J-PARC Studies for the beam separation in the slow extraction Learn practical know-how from Tevatron and LHC for the crystal collimation in future

Understanding of electron beam deflection aiming at the future application of a beam collimation in ILC Investigation on basic properties under the condition beam divergence < Lindhard angle Planning next deflection experiment at KEK-ATF ultra-low emittance, small beam size

Beam separation in J-PARC50 GeV proton beam : intensity of 1014 protons per second smaller beam profile (few mm2) and emittance compared with the conventional separation systems smaller beam losses

Deflection experiment at ATFBeam energy : 1.28GeVIntensity : 1x1010 /bunch Beam divergence < Lindhard angle Effect of multiple scattering with 1.28GeV beam smaller than that with 150 MeV beamLow Emittance X: 1x10-9 rad m Y: 1x10-11 rad mCharacteristic of electron beammultiple scattering ( 1.28 GeV ) : 0.14 mrad 16 mm straight Si crystalLindhard angle ( axis ) : 0.24 mradbeam divergence ( x ) : 0.1 mrad ( 10 mm size ) ( y ) : 0.001 mrad ( 10 mm size )We can see the deflection effect more clearly.

Layout at KEK-ATFtest for deflectionATF2 beam line

Proposed setup at KEK-ATFTaking account of radiation safety regulationVacuum condition : pressure less than 10-7 Torr small amount of material in the beamsimilar situation in ATF2 beam line

Conclusion We performed two kinds of experiments. Proton beam separation with the bent crystal for 12 GeV beam Electron beam deflection with the straight crystal for 150 MeV beam under the condition; beam divergence > Lindhard angle

Study for beam separation in the slow extraction line in J-PARC : the crystal collimation in future

Deflection experiment at KEK-ATF for understanding basic properties. These results can be linked to a feasibility test for the crystal collimation in ILC in future.

Backup slides

Trajectory of the 150-MeV electrons inside of the Si crystalSimulation: trajectoryInitial position : X=-2.5,Y=-2.5Initial position : X=0,Y=-2.5 axes axesX direction = 0.095 mradY direction = 0.09 mradX direction = 0.1 mradY direction = 0.01 mrad

QM3: 2.0 A, q = 0, f = -1.5 mradBeam divergence: 3.0 mradQM3: 2.6 A, q = 0, f = -1.5 mradBeam divergence: 5.2 mrad Beam profiles at FOS position

vacuum: 1.0x10-5 torrextraction lineQM3QM3: quadruple magnetthickness of crystal: 16mmSetupchange of beam divergence at the crystal positionFOS plate

Si crystalGoniometerBeamBeamIIT & CCDFOS plate+mirrorPhotos

REFER ringCreation of the system to extract 150-MeV electron beam Replace aluminium energy degrader by the crystal will reduce energy losses and increase the intensity of extracted beam.

String potentialLindhard continuum axial potentiala Thomas-Fermi radiusZ1e charge of incident particleZ2 atomic number of crystal materialC Lindhard constant ( ) distance from axisd lattice constantLindhard angle ( critical angle )Transverse energy of particles < Potential depth at a 5.4 3

Beam size, divergencemmmm

Extraction lineQM3 magnetinjection lineextraction lineExperimental setup

Data Acquisition systemThe procedure of grabbingpictures and movement of two goniometerswas synchronized with the beam gate.

Pictures were taken only when electron beam hit the FOS plate.

Simulation (1)Beam divergence: 5.2 mradBeam divergence: 3.0 mradLarger beam divergence Smaller deflectioncrystal angle f, (mrad)deflection angle, (mrad)deflection angle, (mrad)crystal angle f, (mrad)

Application to collimation - ILC (International Linear Collider) - Creation of system to remove beam tails Spoiler copper 8.6 mm thick (0.6X0) X0 : Radiation lengthAbsorber copper 4.3 m thick (30X0) proposed by A. I. Drozhdin

Bent crystal silicon 2 mm thick(0.02X0) Deflection efficiency of 250-GeV positrons with the 2 mm Si crystal bent at 0.1 mrad is 80%Deflected beam tail can be localized and is not scattered anywhere.

For practical realizationMany issues on electron beam deflection by crystals are unclear.For example How is an efficiency of deflection? How is a maximum angle for being trapped in the potential ? How is a length of crystal ? ( How is dechanneling length ? )Necessity of basic studies on electron beam deflection by crystal

Electron beam deflection

Lindhard critical angle for axis of Si crystal: 0.7 mradBeam divergence > Lindhard angleverticalhorizontalBeam divergence at crystal positionBeam divergence as a function of QM3 currentcurrent of QM3 magnet, (A)current of QM3 magnet, (A)beam divergence, (mrad)beam divergence, (mrad)

Vertical beam divergence: 3.0 mrad ( QM3: 2.0 A )projectionBeam center Weighted mean in 2s regionAnalysis of imagesvertical projection, (mm)Fitting with double Gaussian

Lindhard string continuous potentialConditions for simulation Simulation ( axial channeling )a Thomas-Fermi radiusr distance from axisd lattice constant, it is 5.43 A for SiZ1e charge of incident particleZ2 atomic number, 14 for SiC Lindhard constant ( ) 4th order of Runge-Kutta method primitive simulation - without consideration of single and multiple scattering, channeling radiation and crystal imperfection Incident angles of particles limited to the twice of Energy of electrons: 150 MeV Thickness of the crystal: 16 m

Searching of c2 minimumBeam divergence found to be 0.6 mrad, and the normalization for deflected beam intensity 1/0.93beam divergence, (mrad)n number of datap number of adjustable parameters (=2)yiexp i-th experimental vaueyisim data from the simulationwi =1/si2 weight of each experimental point, where si is a standard deviationnormalization factor for d.b. intensity

Physics process electron beamscattering with uniform transverse momentum in initial condition deflection anglecrystal axis ftransverse momentumin initial conditionWidth of profile broadening due to a helical motionDeflection angle smaller angle than crystal angleNo effects for a bent crystal ?straight crystal

Crystal and proton beamMaterial: SiliconSize: 3 x 0.3 x 10 mmBending angle: ~ 32.6 mradPlane: (111)Lindhard angle:0.051 mradParameters of crystalbendingangle, 32.6 mradEnergy: 12 GeVIntensity: 1012 protons/spillSize: 15 x 12 mmDivergence: < 5 mrad15mm12mmParameters of the proton beam

First of all, I express my thanks to organizing committee for giving me an opportunity of my talk here. Id like to talk about experiments on beam deflection by crystals. The collaborators are from Hiroshima, KEK, and IHEP-Protovino. IHEP Collaborators provided a bent Silicon crystal and came to KEK for joining an experiment. Recently, we performed two kinds of experiments. One is a beam separation of a 12 GeV proton beam at KEK-PS. We used a bent Silicon crystal with the length of 10 mm and the bent angle of 32.6 mrad. A goal of the study is an application to J-PARC. One of the applications would be a beam separation in the slow extraction line at J-PARC. We are also interested in a beam collimation in J-PARC in the future. We would like to learn practical know-hows from Tevatron and LHC for these application at J-PARC. Another experiment is a beam deflection of 150 MeV electron beam at the electron beam facility in Hiroshima University. We used a straight Silicon crystal with thickness of 16 micro-meter. A crystal axis of was tilted to the beam axis. Our scope is understanding of basic properties on the electron beam deflection. We are also interested in application to a beam collimation in ILC. In my talk, first, I will introduce the proton and the electron beam experiments. After these, I will give our future prospect and conclusion.

First, Im talking about the experiment on proton beam separation of the 12 GeV proton beam at KEK-PS. This is 12 GeV proton synchrotron ring and there are two slow extraction lines and two experimental halls. This line is the neutrino beam line for K2K experiment. Our experimental setup was here. It was constructed on EP2 beam line in the East counter hall here. This is the experimental configuration for the beam separation. 12 GeV primary beam hit the bent crystal, which was mounted on the goniometer. Proton beam went out and main part of the beam hit a fluorescence plate in downstream. A deflected beam hit a CsI plate on the fluorescence plate. The profile of main beam and the spot of deflected beam were observed as a image at the position of these plates. We took images as changing theta. Theta is an angle between the crystal and the beam axis on the horizontal direction. This is a typical raw image at the position of the fluorescience plate. We can easily see a spot of the primary beam, here. However, the spot of the deflected beam is not seen in this image. After a background subtraction, we can see the spot of the deflected beam. From the background-subtracted image, we can extract the intensity of the deflected beam. This figure shows theta dependence of the deflected beam intensity. The horizontal axis is theta. The vertical axis is the intensity of the deflected beam. The intensity of the deflected beam is changed as theta changing. In this experiment, a beam divergence was unknown. So we made simulations by using the CATCH code with different beam divergence and compared with the experimental data. These are results. The colored circles represent the simulated values. This is 5mrad case, 1mad case, 0.5mrad case, 0.3mrad case. The behavior of the simulation was changed as the beam divergence changing. A chi-square analysis leads to the extraction of the beam divergence, which is 0.6 mrad. The simulated values with the beam divergence of 0.6 mrad were superimposed in the plot of the experimental data. The open circles are simulated values, while the closed circles are experimental data. From the data and obtained beam divergence, we can extract the crystal efficiency. This is a definition of crystal efficiency. Number of deflected protons is expressed as a product of angle efficiency, number of incident protons upon the crystal and the crystal efficiency. The angle efficiency is expressed as a ratio of number of incident protons within Lindhard angle to number of total incident protons. This value depends on the beam divergence. Number of incident protons upon the crystal was obtained from the cross-sectional area of the crystal and the beam profile, which is calculated to be 3 times 10 to the 9. From these values, we obtained the crystal efficiency of 23 %. Summary on proton experiment. We demonstrated the beam separation of 12 GeV proton beam with the bent crystal. Experimental results are qualitatively understood. The chi-square analysis gives the beam divergence of 0.6 mrad. By using the value and data, the crystal efficiency of 23 % was obtained. Next, I will report on experiment of electron beam deflection at the 150 MeV electron beam facility in Hiroshima University. This is 150 MeV electron circulation ring, which is called as REFER ring. It means Relativistic Electron Facility for Education and Research. The electron beam deflection experiment was carried out using the REFER ring. This is layout of REFER. A 150-MeV electron beam from a microtron is injected into the ring and circulating. A part of the beam is extracted in the extraction line. The typical beam intensity is about 10 to the 4 per second. Experimental setup was constructed in this area. The vertical beam divergence at the crystal was controlled by this quadrupole magnet. This is experimental configuration for beam deflection. The electron beam hit a straight Silicon crystal with thickness of 16 micrometer and also hit Fiber Optic plate with a scintillator in downstream. A beam profile was observed as a image at this plate. The crystal axis was initially aligned parallel to the beam axis. We observed the beam profile in each combination of theta and phi. Theta is angle of horizontal direction, phi is angle of vertical direction. We can extract the deflection angle from the beam profile. Lindhard angle of 150 MeV electron for axis is calculated to be 0.7 mrad. This shows obtained deflection angle as a function of the crystal angle, phi. The vertical beam divergence was 3.0 mrad. Theta was 0 mrad. We saw the characteristic curve, which said that the deflection by channeling occurred. In the minus side of phi far from zero, we did not see any deflection. As phi changing to the plus direction, the deflection angle starts increasing at certain angle. The deflection angle reaches the maximum and starts decreasing. After reaching the minimum, the deflection angle starts increasing again and reaches zero. In the plus side of phi far from zero, we did not see any deflection again. The direction of the deflection was consistent with the direction of axis. It is noticed that the beam divergence is larger than the Lindhard angle, 0.7 mrad. I show you two typical results with other beam divergence, 3.8 mrad and 5.2 mrad. In 3.8 mrad case, the behavior due to deflection effect remained, but the maximum deflection angle became small. In the case of 5.2 mrad, we did not see similar behavior due to deflection effects. From these results, the maximum deflection angle depends on the beam divergence.

We defined deflection magnitude, delta, which was derived by fitting the plot with 1st derivative Gaussian function. This figure shows the normalized deflection magnitude as a function of beam divergence. It is clear that the deflection magnitude depends on the beam divergence. The beam divergence becomes larger, the deflection magnitude becomes smaller. We made a simple simulation and compared with the experimental data. These figures show the results. This is in the beam divergence of 3.0 mrad, and other is in 5.2 mrad. On rough behavior, we can see that the simulation agrees with experimental results. However, on details, there is discrepancies between simulation and experimental data. One reason is that the simulation code includes only electron dynamics under the Lindhard string potential.Summary on electron experiments. The results showed clear evidence of ability to use crystals for handling 150 MeV electrons under the following condition; the beam divergence is larger than 0.7 mrad, which is Lindhard angle. The deflection magnitude depends on the beam divergence. The beam divergence becomes large, the deflection becomes small. On rough behavior, the simple simulation agrees with the data. On details, there are descrepancies between the simulation and the data. Our future prospect is here. Studies for the beam separation in J-PARC are our next target. For studying the crystal collimation in future, we would like to learn practical know-how from Tevatron and LHC. Another direction is to get understanding of electron beam deflection aiming at the future application of a beam collimation in ILC. We will make an investigation on basic properties under the condition; beam divergence is smaller than Lindhard angle. For this purpose, we are planning the next deflection experiments at KEK-ATF. KEK-ATF is the electron beam facility providing a high quality electron beam for ILC study and it is quite attractive. The electron beam has the characteristic of ultra-low emittance, small beam size, etc.. First, I introduce the beam separation in J-PARC. This is layout. The 50 GeV synchrotron ring is here and has a slow extraction line. In this machine, the intensity is quite strong, which is 10 to the 14 per second. Separation point is here and this is a candidate for inserting the crystal. Compared with the conventional separation system, amount of material in the beam line is smaller, because only a part of the crystal would be inserted in the main beam. We expect smaller beam loss in this beam line. In addition, the crystal beam separation provides the smaller beam profile and smaller emittance in this separation line. There are benefits on providing higher quality proton beam for experiments with 50 GeV primary proton beam. On the electron beam handling, we are planning the deflection experiment at the KEK-ATF aiming at the future application of the beam collimation in ILC In this experiment, the following condition is satisfied; beam divergence is smaller than Lindhard angle. Furthermore, the effect of multiple scattering is smaller than that with 150 MeV beam. In the case of 16 micro-meter straight Si crystal, the multiple scattering is 0.14 mrad, Lindhard angle for axis is 0.24 mrad, and beam divergence is 0.1 mrad on x direction and 0.001 mrad on y direction, even if the beam is focus to 10 micro-meter. Thus, we can see the deflection effect more clearly. We can investigate precise measurements compared with 150 MeV electron beam experiment. I show you the layout of KEK-ATF. This is the ATF dumping ring, the extraction line. ATF2 beam line is the test beam line for final focus in ILC. We are planning that the setup is constructed in the space in front of the beam dump, here. There is a possibility of using the ATF2 beam line for the deflection experiment in the future. This is a proposed setup. The target chamber and 2nd chamber are set in front of the beam dump. The vacuum level is kept for the pressure less than 10 to the -7 Torr. In this facility, users have to take account of radiation safety issues, because the dumping ring is on the ground. In this setup, only the crystal and fluorescence plate are mounted on the beam line. This situation is similar to that in ATF2 beam line. Conclusion. We performed two kinds of experiments. One is the beam separation with the bent crystal for 12 GeV beam. Other is the beam deflection with the straight crystal for 150 MeV electron beam under the following condition; the beam divergence is larger than Lindhard angle. On the proton beam, our next target is study for the beam separation in J-PARC. The crystal collimation in J-PARC is also interesting, but it is in future. We are also planning the deflection experiment at KEK-ATF for understanding the basic properties of beam deflection. These results can be linked to a feasibility test of the crystal collimation in ILC. Thank you for attention.The is results of simulation on the trajectory of electrons. Each points represent crystal axis of . This is initial position. The behavior is like a random walk. Right figure shows the detail orbit around axis. As you see, electron orbit is not closed. The electron is trapped with the string potential and circulating for a certain time, then get out from the trapping potential. This is characteristic of electron channeling. The reason of getting out is not multiple scattering or other incoherent scattering. These are examples of the measured beam profile at FOS plate position. You can see that this case, the vertical size of the beam is smaller than this case. This is our setup including the extraction line. All setup is inside of vacuum system. The 150 MeV electrons are extracted from the REFER ring and lead to this vacuum system. The silicon crystal is located here. The beam size at the crystal position is controled by the quadrupole magnet, which is called QM3. The electrons passing through the crystal irradiate FOS plate. Generated lights are reflected by a mirror, and then detected by a system of an image intensifier and a CCD camera. This is a photograph of the extraction line. This small quadrupole magnet is the QM3, which controls beam divergence at the crystal.One of motivations of electron beam handling is beam collimation in ILC. Conventional design of beam collimation system is proposed and composed of two parts. Beam tail is scattered by spoiler, then it absorbed by absorber. It is assumed that the length of spoiler is 8.6 mm, which corresponds to 0.6 X0. One application is to replace the spoiler to bent crystal. For example, the deflection efficiency for the 2 mm Si bent crystal with 0.1 mrad and 250GeV positrons is 80 %. The length of 2 mm corresponds to 0.02 X0. In addition, deflected beam tail can be localized and is not scattered anywhere. It seems that the reduction efficiency becomes high. For practical realization of electron beam collimation, there are many hurdles in present. We have to clarify many issues. For example, How is an efficiency of deflection ? How is a maximum angle for being trapped in the potential ? How is a length of crystal ? Anyway, if we will develop the collimation system, we have to study the basic properties on electron beam deflection as a first step.

150 MeV is too low energy for bent crystal, because in actual cases, bent crystal intends to become thick. Low energy electron beam has disadvantage due to large influence of multiple scattering. Then, we use a thin and straight crystal.If the crystal is tilted to electron beam direction, the beam is expected to deflect like this picture. In this case, it is possible to observe electron beam deflection by crystal.Beam divergence at the crystal position was estimated from measurements of beam size and calculation of the beam line optics, as a function of the current of QM3.The horizontal divergence is almost constant. The vertical divergence is from 3.0 mr to 5.2 mr. Please note that the Lindhard angle for axis of silicon crystal is 0.7 mr, and the beam divergence is larger than the Lindhard angle for our case.I show analytical process of images. The 2-dimensional profile was projected to this direction. The projected profile is a distribution on the vertical position. This distribution was fitted with a double Gaussian shape, and then a weighted mean within +- sigma was calculated. We define a beam center as a weighted mean.In order to understand our experimental results, we made Monte-Carlo simulations. In this calculation, we used the Lindhard string potential. The formula is here. The dynamics of electron can be calculated under this string potential. Track of each electron is calculated with the 4th order Runge-Kutta method. No consideration has been made for multiple scattering and channeling radiation. The electron energy is 150 MeV and the thickness of the silicon crystal is 16 micrometer. Using a straight crystal, it is said that the asymmetrical scattering occurs due to a bias ob the initial condition in transverse momentum. It is called doughnuts scattering. In the scattering, the deflection angle has the tendency of becoming smaller than the crystal angle. It seems that the bent crystal has no effect.