KEK-PR0C--95-11 Proceedings JP9607178 of the Fifth Workshop on Japan Linear Collider (JLC)
Kavvatabi, Miyagi, Japan, February 16 - 17, 1995
Positrons
10 GeV 1.54 GeV S-band S-band Pre-Accelerator Injector
500 GeV 10-30 GeV X-band Linac S-band Linac for Positron Generation 1.54 GeV Positron Damping Ring
NATIONAL LABORATORY FOR HIGH ENERGY PHYSICS, KEK
KEK Proceedings 95-11 December 1995 A/H
Proceedings of the Fifth Workshop on Japan Linear Collider (JLC)
Kawatabi, Miyagi, Japan, February 16 - 17, 1995
Editor: Y. Kurihara National Laboratory for High Energy Physics, 1995
KEK Reports are available from: Technical Information & Library National Laboratory for High Energy Physics 1-1 Oho, Tsukuba-shi Ibaraki-ken, 305 JAPAN Phone: 0298-64-1171 Telex: 3652-534 (Domestic) (0)3652-534 (International) Fax: 0298-64-4604 Cable: KEK OHO E-mail: LIBRARY@JPNKEKVX (Bitnet Address) [email protected] (Internet Address) Contents
Physics at LEPII T. Tsukamoto 1
Present Status of 1.54 GeV ATF Linac S.Takeda 12
Button-type beam-position monitor for the ATF damping ring F. Hinode 28
Recent progress on cathode development and gun development at Nagoya and KEK
M.Tawada 31
Development of polarized e+ beams for future linear colliders M. Chiba 38
The LASER beams with very long focal depth for photon-photon collider
K. Matsukado 78
An interactive version of GRACE and catalogue of e+e- interactions as its application
S.Kawabata 92
The GRACE system for SUSY processes M. Jimbo 98
Search for dynamical symmetry breaking physics by using top quark T. Asaka 108
Status of R&D for the vertex detector Y. Sugimoto 121
Progress report of calorimeter subgroup Y. Fujii 123
I NEXT PAGE(S) | left BLANK JLC '95
2 J! 1 6 5 (*)
I. %$ 9:00-10:05 (KEK) MM W$- (&] EE5OO 7-^~>3 7 7*^ BB|*J (KEK)
2. WKH io:2o-ii;4O Vs=500GeV"C r ^ - # 7 - fi JL A & ^ ? e+e'->w+w" Kiifc £ £ &t & Vector Resonance KEK Search for Dynamical Symmetry Breaking by Using Top Quark yy ii 7 4 y—tr cOHiggs 7> ^ r-Y -
3. Wk 1:00-2:40 The GRACE System for SUSY Processes
GRACE tCATALOGUE KEK JUS Event Generator for Wpain Production KEK ^M Photon Photon Collider Plil^)^C'^, iS^^C IR Issues KEK
4. jDiltf (1)2:50-4:10 ^S JIIM KEK KEK Klystron Modulator 5. JMtf (2) 4:20-5:20 Damping Ring i: Vacuum Chamber O|r1^§ KEK Damping Ring ffl Beam Position Monitor CO fn\%fe KEK Alignment ^#i 60 ?l^t KEK ft 1*1 ISffloltT-^t'-f #> KEK Compton — in — 2 M 1 7 B i. iffilfeS 9:00-10:00 JM.& Baft *a#7-?->3 7yMR]tfT KEK }\\ffl VTX Progress Report KEK CAL Progress Report KEK 2. Wm 10:15-11:45 Jig BT£ The Measurment of Weak Boson Properties at Linear and Hadron Collider R. Szalapski Probing the Weak Boson Sector in eY -» zy and 7e-» ze S. Y. Choi Consequences of New Interactions in the EW Boson Sector ftHil&W Higgs Production and Decay in an Extended Supersymmetric Standard Model KEK ||T Precision SUSY Studies at Future Linear Collider KEK SJ'J^, S-quark correction to e+e- w+w-@JLC MJZ HJM = 3. Uffl8frWi ?-U 3. M 4. Physics at LEP II Toshio Tsukamoto Department of Physics, Saga University, Saga-shi, Saga, 840 Japan Abstract After the successful running as a Z factory, LEP machine is to be upgraded to the second phase with doubling the beam energy in order to operate beyond the W+W~ threshold. Here an overview is presented of the physics topics at LEP II. 1 Introduction It has passed 6 years since the first commissioning of LEP in 1989. The machine performance has been improved year by year and more than 15 millions of Z° events are now recorded. The Standard Model is checked out precisely and systematically through a variety of processes. One of the most fundamental parameter in the Standard Model Mz is measured down to 2 x 10~5 accuracy. The measurements of total width as well as partial decay widths of Z° enable one to determine the number of light neutrino species to be 3 with 5 per mil precision and put strong constraints for the invisible light particles at the same time. The couplings to Z° is derived with ever-improving accuracy using the forward-backward asymmetry of leptons and the r polarization measurements and the lepton universality is found to stand. So far, the Electroweak Standard Model is strongly supported by these LEP measurements. LEP II program will push it further to clarify the gauge structure in the Electroweak Stan- dard Model by studying W boson properties and gauge couplings. On the other hand, LEP has been the highest energy e+e~ collider and provided a unique opportunity in searching for new particles. Direct searches extensively done at LEP I resulted in no evidence at the moment. However, good news is that the properties of unknown particles are estimated indirectly based on the precision measurements of Z°. In fact, LEP I data is sensitive to the electroweak radiative correction and the mass of top has been evaluated assuming the Standard Model (Mtop ~ 176 ± 10t\l GeV). Recent observation of top quark at Fermilab [1] gave consistent values of Miop = 176 ± 8 ± 10 GeV (CDF) and Mtop = 199±" ± 22 GeV (D0) with similar accuracy to the indirect estimation from the Z° measurements. For the yet-unknown Higgs boson, one can also perform a global fit to MH using all the existing electroweak data. The situation is, however, quite different from Mtop case. Only very weak logarithmic dependence to the radiative correction is given for MJJ while Mtop has quadratic dependence. Although the minimum of x2 value indicates lower Higgs mass around 100 GeV [2], it is still premature to extract MH value with meaningful precision. Discussion on the grand unification of running coupling constants implies that the super- symmetric energy scale should be below 1 TeV assuming the minimal SUSY extension of the Standard Model [3]. If this minimal SUSY scenario is true, supersymmetric particles has to be found below 1 TeV. Especially, the lightest Higgs boson predicted by the minimal SUSY can exist below 130 GeV [4] which might be discovered in the near future. Now the LEP is going to increase its energy and one can expect a number of physics outputs from the new phase of LEP. Major physics issues at LEP II can be summarized as follows. • Study of W properties and gauge couplings through e+e~ —> W+W~ as a W pair pro- duction machine • Search for Higgs bosons and new particles as well as new exotic phenomena as a highest energy e+e~ machine 2 Machine upgrades for LEP II To double the beam energy, higher power of RF system has to be installed, because of the significant energy loss due to synchrotron radiation which goes as the 4th power of the energy. It is approved that 176 super conducting cavities are added to the existing 120 copper cavities in the next phase of LEP. With this configuration, a maximum energy of y^ = 175GeV can be accessible. The mass of W can be measured precisely in this first upgrade. Further increased energy is, however, apparently preferable in searching for new particles and new phenomena. By removing half of copper cavities and installing 70 more super conducting cavities, the energy can be increased up to 190 GeV. From the physics point of view, this second upgrades might be crucial since there could be a gap in the Higgs discovery range between LEP and LHC. The scenarios for further upgrades are now under discussion. It is said that 205 GeV can in principle be achieved without civil engineering [5] but is unlikely to be realized at the moment. Target integrated luminosity for LEP II is 500pb-1 within 3 years. Despite of great success of 8 bunch operation by the pretzel scheme started in 1992, it was found to be difficult to achieve the goal. Bunch train schemes were investigated instead. The result of the studies under LEP I operation looks very promising. For 1995 run, LEP will be operated at Z° energy with 4 trains in 4 bunches. Due to the total current limitation, 4 trains in 2 or 3 bunches will be adopted for LEP II. 3 Mw measurement Current available value for Mw from hadron colliders is 80.26 ± 0.16 GeV [6] which agrees quite well with the indirect estimation by LEP data: 80.32 ± 0.06 GeV. The masses of the W and Z are related by the formula: TO M2 L W V2GF(l - M& I Ml) (1 - Ar) where Ar is the radiative correction to the boson mass. Thus the precise measurement of gives another strict check of the radiative correction. The size of the radiative correction is changed by Mtop and MH- Unknown parameter MH can be evaluated indirectly, once Mtop is determined with a good precision by Tevatron. Figure 1 shows the relation between Mw and Mtop for different MJJ. Also shown in the figure are the Mw value from hadron colliders and Mtop from Tevatron measurements [1], Further improvement both on Mw and on Miop measurements could constrain the value of MH- Together with precise results from Z° measurements, the self- consistency of the Standard Model can be checked in a model independent way. Three independent methods are being investigated to measure the mass of W at LEP II [7]: • Direct mass reconstruction of the final state • Excitation curve near WW threshold • End point of lepton energy decayed from W As a result of the Monte Carlo studies, the first one is the most promising in terms of accuracy. Final states can be used in this analysis are 4 jets and 2 jets + a lepton. Statistical errors for both channels are similar and combining them, 55 MeV per SOOpb"1 per experiment can be achieved. Systematic errors are now under investigation. At present the error from experiment and theory is expected to be 10~35 MeV and that from beam energy uncertainty to be 15 MeV. Using excitation curve is straightforward method to extract Mw- It is, however, most sensitive just above the WW threshold, which means we have to sacrifice WW statistics for other physics analyses. Preliminary study indicates that we could obtain 100 MeV accuracy per 100 pb"1 per experiment. In addition, we have to pay attention to theoretical uncertainty of the prediction for excitation curve, which has to be less than 1% near the WW threshold. The third method examines the energy spectrum of leptons from W*. The lepton energy Ei is kinematically limited to \E( — jE&eotn| < \JEleam — Mjy/2, in which the cut-off of the distribution is used to determine Mw- The edge is substantially smeared when we include the width of W and initial state radiation. This method, in fact, requires excellent energy resolution for decayed leptons and accuracy obtained by this method is not so attractive. The error only in statistics would be 150~300 MeV per 500 pb"1 per experiment. It may be useful only for cross-checking. 3 80.7 oCD 80.6 - I i i i i I i i i i i i i i I i i i i I i i X/X//I I I I I I I 1 I I I I I I I I I 120 130 140 150 160 170 180 190 200 210 220 Mtop (GeV) Figure 1: Mw versus Mtop for various Higgs masses assuming the Standard Model. The cross bar indicates the current measurements of Mw and Mtop by Tevatron col- lider. Hatched area shows the expected accuracy in the future for Mw by LEP II (AMvr=35 MeV) and for Miop by hadron collider (AM4op=5 GeV). 4 Three Gauge Boson Couplings The couplings of fermions to the electroweak gauge boson Z° have been tested at LEP I and SLC and found to agree quite well with the Standard Model prediction. Together with the confirmed universality of the lepton couplings, strong support for the gauge theory description of electroweak interaction is now provided. Although these tests are achieved at per mill level in precision, the knowledge of the couplings between gauge bosons is still poor. Assuming C and P invariance, three boson couplings can be parameterized using 5 variables: Szww, KZi K7, ^z-> ^"i [8]- Any deviation from the Standard Model value, gzww = cotOw, K-Z — «-y = 1 and Xz = A7 = 0, suggests new physics. Present experimental limits come from associated W7 and Z7 production at the Tevatron [9]: —1.6 < A«7 < 1.8, |A7| < 0.6, -8.6 < A«2 < 9.0 and \XZ\ < 1.7. At LEP II, direct tests of the three gauge boson couplings (7"W+W*" and Z°W+W~) can be done through the s-channel diagrams of e+e~ —* W+W~. The W+W~ total cross section has very poor sensitivity to the anomalous coupling constants below y/s ~ 200 GeV. To maximize the sensitivity, one has to use all the possible angular information of the decayed fermions from W as well as the charge of the W. The most promising channel to be used is WW —> qqlv[i = /i,e). Relatively low background is expected to this channel and the charge of W can be determined from the decayed lepton. The result of the single parameter fit is shown in Figure 2. The sensitivity depends strongly on the centre of energy and the higher energy is crucial for this measurement. The expected sensitivity is 0.03-0.1 for and 0.06-0.1 for AA, depending on the model and the available beam energy [10, 11]. 0.21 - 0.19 - 0.17 - a 0.15 - 0.15 - 0.13 - 0.13 _ 0.11 - 0.11 - 0.09 - 0.09 - 0,07 - 0.07 - 0.O5 - O.OS - 0.O3 - 0.03 - 0,01 - 0.01 - 170 ISO 190 200 210 170 180 190 200 210 ECM IGeV] ECM [GeV] Figure 2: Expected sensitivities to the anomalous coupling constants based on Monte Carlo studies [10]. 5 Higgs searches After the discovery of the top quark at Tevatron [1], only Higgs boson is missing in the Standard Model. At LEP I, Higgs boson is produced via, e+e~ —> Z° —t H°Z°\ Since the production cross section falls off rapidly as MH increases, the current search domain covered up to ~60 GeV is actually close to the sensitivity limit at LEP I. The detection of Higgs at LEP II can be done using the process: e+e~ —> Z0* —> H°Z° where the produced Z° is on mass-shell while it is off mass-shell at LEP I. This enables us to put kinematical constraints on the final particles coming from Z° at LEP II. ! Z° 104r eV —> hadrons 3 CL 10 c 2 o 10 ==== •4-o4 r LEP LEF'200 ======CD CO If) 10 w + o e e" -> H°Z° o 4 o io-V i ' '""" 1 Iml -2 • .... i ... \ X /. 10 50 100 150 200 250 (GeV) Ecm Figure 3: Cross section for the Standard Model Higgs as a function of the centre of mass energy for several Higgs masses. Figure 3 shows the production cross section as a function of centre of mass energy. For > 60GeV, the cross section is actually higher at LEP II than at LEP I. Also seen in Figure 3, LEP II has the advantage in terms of signal to background ratio which is about 10~2 at LEP II compared with 10~5 at LEP I. The steep rise of the cross section above the threshold enables us to explore the mass range close to the kinematical limit. The signatures of the Higgs productions are; • Two jets * missing qq,Z° vv) • Two jets qq,Z° £+£-) t Four jets qq, Z° Several other known processes can have similar topologies at LEP II energy region. The cross sections of possible background processes are listed in Figure 4. Among them, W+W~, Z°Z° and e^W* are the most serious ones. To suppress the backgrounds, b-tagging plays a crucial role since the branching ratio of H° —* bb is very high (85%) while that of Z° -+ bb is — (5 — Vs=175GeV Vs=190GeV 10J e+e~—> .a a. 2 10 r- + + c e e"~>e e"qq o o e+e" en 10 r- 10 w CO eV->eVZ o L. + o e e"—> evW eV-»ZZ + 10 e e"->ZH 10 MM=60.70,BO MM= 60,70,80,90 e+e" —> -2 -2 10 10 Figure 4: Cross sections for background processes compared with that for Z°H° for = 175 GeV and 190 GeV. 600 50 70 80 90 100 ^600 "a. 500 J400 300 200 100 0 50 60 70 80 90 100 2 mH (GeV/c ) Figure 5: Discovery potential for the Standard Model Higgs for = 175 GeV and 190 GeV [12] only 15% and W* does not have decay mode to b-quark. All four experiments at LEP have installed high precision micro vertex detectors by which b-quark can be identified efficiently using its long life time information. Further kinemaiical requirement can be posed on Z° decay side to suppress backgrounds. The resulting efficiency ranges from 15% to 50% depending on the event topologies. Based on Monte Carlo studies [12], the luminosity to detect Higgs signal is estimated for different centre of energies, which is shown in Figure 5. With luminosity of 1, Higgs can be discovered up to M# ~ \/s — lOOGeV. Minimal SUSY extension of the Standard Model (MSSM) requires two Higgs doublet. In MSSM, there exist five physical Higgs bosons: one neutral CP-odd scalar (A0), two neutral CP-even scalars (h°, H°) and two charged scalars (H*). The mass relations predicted by this model requires the mass of the lightest Higgs h° should be smaller than M% at tree level. This upper bound has to be modified since the Higgs mass receives the large radiative correction from heavy top and scalar top. For top mass of ~175 GeV and stop mass of ~1 TeV, the upper bound is estimated to be less than 130-140 GeV [4] and LEP II could cover large part of it. This is one of the strong motivations to increase the beam energy of LEP II as high as possible. The reactions to produce the lightest Higgs h° are; e+e" -> h°Z° and e+e" -+ h°A°. The cross section of the former process compared with the Standard Model (e+e~ —> Z°H°) is suppressed by a factor sin2(/? — a) while that of the latter is proportional to cos2(/3 — a). So they are complimentary in terms of the production rates. Strategy for the detection [12, 13] is quite similar to that of the Standard Higgs boson except for the four jets channel in h°A° search. One can maximally use b-tagging for this channel because of the larger branching ratio of A0 -> bF. The accessible region of SUSY Higgs particles depends strongly on the available centre of mass energy. It also depends slightly on the choice of SUSY parameters. Figure 6 shows the covered region in (M/,, Mj\) plane [14]. One can see large area can be covered by LEP II. 6 Beyond the Standard Model SUSY is expected to be an attractive candidate beyond the Standard Model. Although its direct evidence is not found yet so far, it can naturally avoid the instability of the Higgs mass and could accoi'nonodate with gravity. To realize the grand unification of three gauge interaction, simple extrapolation based on the minimal Standard Model is now excluded by the precision measurements of the coupling constants at LEP while the Minimal Supersymmetric Standard Model can achieve it [3]. Among SUSY candidates, charginos (x*) are expected to be as light as Z° and its production cross section would be large (~ a few pb"1). They decay into the lightest neutralino and a pair of fermions where the neutralino goes through the detector without interaction, resulting in missing energy. General strategy with missing energy can also apply to other SUSY events such as sleptons and stops [15]. Background from W+W'~ events can be reduced by imposing further kinematical cuts. o dUU 1 1 1 i i i • i i i 1 1 ' H 111 1 1 i a | O :Vs = 175 GeV (0 150 (C) j en o ****** E 100 ~ ^ *•*** — ) •' \/v \,—' (D) : 50 r sensitivity • not allowed m m I by hZ + hA if by MSSM : - i i i | 1 • !••! i i t i i 0 i i ) i i i 1 _ T • n O : Vs = 190 GeV a 150 (c) : - • a E 100 J : . r »•** [_ (A) ^ (D) : 50 a - a • 1 f 1 - 0 Ml ' ' ' ^ ' 'i 'r^ 1 ~ r T « o :Vs = 210 GeV 150 0) ; — (B) y i W - o / E 100 - «^» _ I (A) ^ (D) j a 50 a a - a a ,; n III III III i , I , 0 20 40 60 80 100 120 140 h mass (GeV) Figure 6: Regions of detectability of h° at LEP II for y/s=l75, 190 and 210 GeV in the (Mh, MA) plane [14]. (A) 3a sensitivity region. (B) Sensitivity depends on other SUSY parameters. (C) Non-sensitivity region. (D) Theoretically disallowed region. — 9 — As was done at LEP I, a number of exotic particles such as heavy quarks, heavy leptons, heavy neutrino, excited leptons, leptoquarks etc. should be searched for and most of the search domains will be doubled at LEP II [16]. Besides direct evidence beyond the standard model, new phenomena could emerge as a tiny effect in the standard processes. In this context, LEP II has much higher sensitivity than LEP I since such an effect is obscured by the huge contribution of Z° pole at LEP I. For example, Z' can only be searched through Z-Z' interference at LEP I, while one can directly examine the Z' propagator as a tail effect up to 1 TeV range at LEP II [17]. Another example is four-lepton contact interaction which is motivated by composite model. By measuring total and differential cross sections of lepton pair final states (e+e~ and (i+fi~), one can look for up to ~10 TeV range of the effect on the cut off parameters. 7 Summary The energy upgrade of LEP enables us to study basic properties of W boson. The mass of W can be measured with a precision about 60 MeV per experiment if sufficient luminosity of 500pb~x is delivered. Three Gauge couplings 7W+W~ and Z°W+W~ will be directly checked with less than 10% precision. LEP II will also play a crucial role as a discovery machine. Many direct and unambiguous searches can be done there. The discovery potential of Standard Higgs ranges up to \/s — 100 GeV. One can obtain large coverage of parameter space for the Minimal Supersymmetric Standard Model. In the framework of MSSM, the lightest Higgs boson is expected to be smaller than 130 GeV and this signal of an extended Higgs sector could be within the LEP II search domain. If one of the sparticles is light enough to discover, LEP II could be the first place where we would find physics beyond the Standard Model. References [1] F. Abe et al. (CDF Collaboration), Phys. Rev. Lett. 74(1995) 2626; S. Adachi et al. (D0 collaboration), Phys. Rev. Lett. 74(1995) 2632. [2] LEP Electroweak Working Group, LEPEWWG/95-01 (internal note); The LEP Collaborations ALEPH, DELPHI, L3, OPAL, and the LEP Electroweak Working Group, CERN-PPE/94-187. [3] W. de Boer, G. Burkart, R. Ehret, W. Oberschulte-Beckmann, V. Bednyakov and S. Kavalenko, IEKP-KA/95-07; U. Amaldi, W. de Boer, P. Frampton H. Fiirstenau and J. Liu, Phys. Lett. B2S1 (1992) 374. [4] Y. Okada, M. Yamaguchi and T. Yanagida, Prog. Theor. Phys.Lett. 85 (1991) 1; J. Ellis, G. Rdidolfi and F. Zwimer, Phys. Lett. B257 (1991) 83 and Phys. Lett. 262B (1991) 477; -~ 10 H.E. Haber and R. Hempfling, Phys. Rev. Lett. 66 (1991) 1815; R. Barbieri, M. Frigeni and F. Caravaglios, Phys. Lett. B258 (1991) 167; R. Barbieri and M. Frigeni, Phys. Lett. B258 (1991) 395. [5] S. Myers and C. Wyss, First General Meeting of the LEP2 Workshop, CERN, 2-3 Feb 1995. [6] F. Abe et al. FERMLAB-PUB-95/033-E and FERMILAB-PUB-95/035-E; C.K. Jung, 27th International Conference on High Energy Physics, Glasgow, Scotland, 20-27 Jul 1994. [7] W.J. Stirling and N. Kjaer, First General Meeting of the LEP2 Workshop, CERN, 2-3 Feb 1995. [8] M. Bilenky, J.L. Kneur, F.M. Renard a.nd D. Schildknecht, Nucl. Phys. B409 (1993) 22; K. Hagiwara, R.D. Peccei, D. Zeppenfeld and K. Hikasa, Nucl. Phys. B282 (1987) 253. [9] H. Aihara, International Symposium on Vector Boson Self-Interactions, UCLA, 1-3 Feb 1995; F. Abe et al (CDF Collaboration), Phys. Rev. Lett. 74(1995) 1936; S. Errede, 27th International Conference on High Energy Physics, Glasgow, Scotland, 20-27 Jul 1994. [10] R. Eichler, LEP Committee, CERN, 3 Nov 1992. [11] R.L. Sekulin, Phys. Lett. B338 (1994) 369. [12] P. Janot, First General Meeting of the LEP2 Workshop, CERN, 2-3 Feb 1995. [13] A. Sopczak, CERN-PPE/94-73. [14] J. Rosiek and A. Sopczak, Phys. Lett. B341 (1995) 419. [15] G. Coignet, LEP Commitee, CERN, 3 Nov 1992; H. Baer, M. Brhlik, R.Munroe and X. Tata, FSU-HEP-950501 and UH-511-829-95; M.L. Mangano and S. Katsanevas, Second General Meeting of the LEP2 Workshop, CERN, 15-16 Jun 1995. [16] G. Giudice, Second General Meeting of the LEP2 Workshop, CERN, 15-16 Jun 1995. [17] C. Cerzegnassi, First General Meeting of the LEP2 Workshop, CERN, 2-3 Feb 1995. 11 Present Status of 1.54 GeV ATF Linac Seishi Takeda National Laboratory for High Energy Physics, KEK Oho 1-1, Tsukuba, Jbaraki 305 Abstract This paper describes the present status of 1.54 GeV S-band injector linac of the ATF (Accelerating Test Facility) for JLC. 1 INTRODUCTION The cross sections of electroweak interactions are inversely proportional to square of center-of-mass energy and then the luminosity should be increased with square of the center-of-mass energy. The luminosities higher than several 1033 cm~2 s~* are required ranging from 0.3 to 0.5 TeV in center-of-mass energy [1]. The luminosity is given by where N , Nb, /rep? cr*, a*, H& are the bunch population, number of the bunches in a bunch train, repetition rate of the linac, horizontal and vertical beam size at the interaction point and enhancement factor by the disruption, respectively. The beam power of two linacs is given by P6 = NNbEJTep = PacV, (2) where Ec is the center of mass energy. The beam power is limited by the wall-plug power Pac and the over-all efficiency rj from the wall-plug power to the beam. The total charge NNb accelerated in a RF pulse is also limited by the wall-plug power. Combining the above two equations, the luminosity is given by The higher bunch population ./V results in higher luminosity, but iV is limited by the beamstrahlung at the interaction point. The multi-bunch acceleration, Nb > 1, is required to increase both the over-all efficiency 7/ and the luminosity. To obtain the narrow spread of energies at the interaction point, a flat beam is preferable to reduce the energy loss by beamstrahlung [2]. — 12 — The small spot size at the interaction point is essential to obtain higher luminos- ity. The lower beam emittance results in smaller spot size. However the beam size is limited due to the synchrotron radiation losses in final focus quads [3]. The param- eters of JLC-I have been optimized by the computer simulation to obtain the lumi- nosity of several 1033cm~2s~l. In order to obtain the flat beam of 3 nm and 300 nm, the normalized vertical and horizontal beam emittance should be 3 x 10~8 m-rad and 3 X 10~6 m-rad, respectively. The generation of low-emittance multi-bunch beam from a damping ring is one of the essential key points to realize the high luminosity. The emittance preservation in main linacs is also essential key point to realize the high luminosity. The emittance growth of multi-bunch by long-range trans- verse wake fields can be eliminated by using damped structures [4, 5] and detuned structures [6, 7]. the precise alignment of main linac is required to reduce the single- bunch emittance-growth by the short-range transverse wake-fields. The alignment tolerance of the X-band main linac is 10 [im. The tolerance is inversely proportional to square of the rf frequency. The tolerance of C-band linacs is 40 fim and that of S-band linacs is 160 fim. As for the longitudinal wake-fields, the energy spread in a single-bunch produced by the short-range range wake-fields can be minimized by the off-crest acceleration. The energy spread of multi-bunch generated by the long-range wake-fields should be compensated to minimize the energy spread at the interaction point. The energy compensation among the multi-bunch will be settled down the <5f scheme at the ATF linac. The theoretical and technical difficulties of the JLC machine arise from the re- quirements for high luminosity and high accelerating gradient. These difficulties will be settled down by the R&D works of the Accelerator Test Facility. 2 ATF (Accelerator Test Facility) The ATF project was started in 1988 in order to stimulate the R&D work for the JLC project [8]. On the first stage, a high-gradient S-band linac unit was constructed in TRISTAN Nikko Experimental Hall [9, 10]. The project has been extended to construct an accelerator system as a proto-type machine of the real JLC machine [11]. At present the ATF is under construction in TRISTAN Assembly Hall. As shown in Figure 1, ATF consists of five major accelerator parts: an S-band in- jector linac [12], a beam transport-line [13], a damping ring [14], a bunch compressor and electron and positron sources. ATF will verify the multi-bunch scheme of linear colliders in all parts from the injector to the bunch compressor. The low-emittance multi-bunch acceleration is a key issue to increase the luminosity by increasing the acceleration efficiency of the linac. ATF generates, accelerates, damps, and com- presses a train of 20 bunches with 2 x 1O10 electrons/bunch and 2.8 ns bunch spacing. The amount of total number of electrons in a bunch train is approximately half that of the JLC-I machine. The multi-bunches are accelerated by the accelerating gra- dient of 33 MeV/m which is approximately same gradient of JLC-I machine at the initial stage. The goals of the vertical and horizontal beam emittance to be achieved 1 o are 3 X 10 8 m-rad and 5 X 10 6 m-rad, respectively. The goal of the bunch length is 100 //m with a one-stage compressor with a large (1/50) compression ratio. The amounts of the beam emittance and the bunch length are same those of the JLC-I machine. Bunch compressor Water cooling & Air condition facility Wiggler magnet \ Polarized eledr nsourc« Water cooling 8 Air condition facility \ RFgun s Control room 1.54G»VS-band LINAC Dampedcavity Wiggler magnet Poslt/on source Choks mods damped structure (RSD) I a DC pow«r supply (or modulator -120m. Figure 1: ATF (Accelerator Test Facility). 3 1.54 GeV ATF Linac The 1.54 GeV ATF injector linac is designed to accelerate multi-bunch electrons for injection to a low-emittance damping ring [15]. The multi-bunch beam at the ATF consists of a bunch train, each of which is a train of 20 bunches. The bunch population is 2xlO10 electrons and the bunch spacing is 2.8 ns. The maximum energy spread of 90 % of electrons in a multi-bunch is 1.0 % full width, and the normalized emittance at the end of the linac is less than 3xlO~4 m-rad (lcr). The linac is operated at a repetition rate of 25 pps to circulate five bunch trains in the damping ring and extract bunch train of low-emittance to the bunch compressor at 25 pps. The 1.54 GeV ATF injector linac consists of an 80 MeV pre-injector linac, eight units of regular accelerator section and two units of energy compensation accelerating structures. The parameters of the 1.54 GeV ATF injector linac are summarized in Table 1. 4 80 MeV Preinjector Linac The required specification of preinjector results from the energy acceptance and dynamic aperture of the damping ring [16]. Although the energy spread among Table 1: Parameters of 1.54 GeV ATF injector linac Beam Energy 1.54 GeV Bunch Population 2xl010 electrons/bunch Bunches/Train 20 Bunch Spacing 2.8 ns Repetition Rate 25 pps Energy Spread (Full Width <1.0% Beam Emittance <3xlO~4 m-rad (la) Total length 88 m Pre-injector 18 m Linac 70 m (active length: 48 m) 80 MeV Pre-Injector Beam Energy 80 [105 maximum] MeV Number of Bunches 20 Bunch Population 2xlO10 electrons Bunch Separation 2.8 ns Regular Accelerating Sections Accelerating Structure 27r/3 mode constant gradient Total length 3 m Total number 16 Accelerating Field Maximum Field 43 [52 maximum] MV/m with Beam-loading 33 [40 maximum] MeV/m RF Frequency 2.856 GHz Feed Peak Power 200 MW/Structure Klystron Klystron Peak Power 80 [85 maximum] MW Klystron Pulse Length 4.5 /is Number of Klystrons 8 RF Pulse Compression Dual-iris SLED Power Gain 5.0 at peak Klystron Modulator Total Number 8 Energy Compensation System Accelerating Structures RF Frequency 2.856 + 4.32727 GHz RF Frequency 2.856 - 4.32727 GHz Klystron Total Number 2 Klystron Peak Power 50 MW Klystron Pulse Length 1.0/is Klystron Modulator Total Number 2 — 15 — bunches will be compensated by a special accelerating structure in the 1.54 GeV ATF Linac, the energy spread within a bunch is determined by the bunch length at the exit of the preinjector. Also, the dynamic aperture of the damping ring determines the maximum emittance of the Linac beam. Assuming no emittance growth in the linac, the specification of the maximum emittance will be applied to the preinjector. Table 2 shows the requirement for the beam. Table 2: Required specification of the SO MeV preinjector. Beam energy 80 MeV Number of pulses 20 Bunch population 2.0xl010 Bunch separation 2.8 ns Bunch length (FWHM) <10 ps Normalized emittance <3xlO~4 rad m (rms) Energy spread (FWHM) <1 % (each bunch at 1.54 GeV) The numbers of the bunches in a bunch train and the bunch population required from the damping ring are 20 bunches and 2xlO10 electrons per bunch at the end of the linac respectively. The bunch length (FWHM) is less than 10 ps. As shown in Figure 2 the preinjector consists of a thermionic electron gun, two sub-harmonic bunchers, four single cell bunchers, an accelerating structure, a matching section of beam lattice, an energy analyzer and beam instrumentations. e-GUN OTRl B(Buncher) SHEf)(357MH2) lSHB2(357MHz) .Faraday Cup Figure 2: Schematic drawing of the 80 MeV preinjector. 4.1 Thermionic Electron Gun System The thermionic gun system consists of a triode with type EIMAC Y646-E or Y796 grid-cathode assembly [17, 18,19]. The HV pulser provides a pulse voltage of 240 kV — 1(5 — with 3 /is pulse duration. A grid pulser is installed on the HV station set up near the electron gun. Grid pulses with pulse separation of 2.8 ns are produced by gating the output of a 357 MHz signal generator synchronized with sub-harmonic bunchers. The multi-bunch from the gun has a bunch length of 1 ns FWHM. Each bunch contains 3xlO10 electrons. The bunch population is larger than the specification of the linac beam because of an about 70 efficiency in the buncher section. The parameters of the beam produced by the thermionic electron gun are shown in Table 3. Table 3: Parameters of the beam produced by the thermionic electron gun. Beam energy 200 keV Number of pulses 20 Bunch separation 2.8 ns Bunch population 3.OxlO10 Population tolerance < i 4.2 Buncher Section In order to realize the bunch length required for the preinjector, bunching by multi- bunch loaded-cavities is an issue to be solved. High current bunch beam produces a cumulative loading voltage in the bunching cavity, which distorts the cavity voltage. To avoid a phase shift of the bunching voltage due to the beam-loading, low R/Q cavities, which reduce the beam-loading voltage, are used. To obtain more flexibility for buncher tuning, four single-cell standing-wave cavities with low R/Q for the buncher as well as low R/Q Sub-harmonic bunchers. 4.3 Accelerator Section and Beam-loading The accelerator section of the 80 MeV preinjector is a 3 m-long constant-gradient accelerating structure. The geometrical dimension and rf specification are identical to the structures of the accelerator section of the 1.54 GeV linac. The energy loss due to the transient beam-beam loading at time t after beam injection is (t-tf) where i0, r0, r, tj are instantaneous current of the beam and shunt impedance, attenuation parameter and filling time of the accelerating structures, respectively. The energy spreads of the multi-bunch due to the beam-loading are evaluated to be 5 % and 1.25 % in full width at the bunch population of 2xlO10 and 0.5xlO10, respectively. Since transition radiation is produced by each electron in a bunch, it carries information concerning the bunch structure. The OTR (Optical Transition Radiation) monitor is used to measure the bunch profile and length bunch by bunch. The bunch by bunch energy spread can be measured by the OTR monitor and a bending magnet as shown in Figure 3. The energy spread is evaluated to be 1.24 % in full width at the bunch population of 0.5xl010 electrons [20]. 81.5 5 81.0 . Q. CD m m S? 80.5 - CD E 80.0 . CO CD CO 79.5 10 15 20 25 Bunch Number Figure 3: Bunch-by-bunch energy and spread measured by an OTR monitor. The bunch population is 0.5xlO10. A wire scanner is routinely used for bunch-by-bunch emittance measurement in the ATF preinjector. A MCP-PMT with gate technology detects gamma generated from the fire by colliding the beam electrons. 5 Accelerator Section of 1.54 GeV Linac The role of the accelerator section of the linac is to accelerate a multi-bunch from 80 MeV to 1.54 GeV with a minimum energy spread and a minimum emittance growth. The accelerator section of the 1.54 GeV ATF injector linac located down- stream of the 80 MeV pre-injector consists of eight rf units, two energy compensation units, a linac lattice, and beam monitors as shown in Figure 4. The beam-transport line is located between the exit of the 1.54 GeV ATF linac and the injection kicker of the damping ring. The rf unit of regular section consists of an E-3712 klystron, a modulator [21], a dual-iris SLED cavity [22], rf waveguides, two 3 m-long accelerating structures and rf dummy loads as shown in Figure 5 [23, 24]. The klystron produces an rf peak power of 80 MW with a pulse duration of 4.5 /is. The rf power is fed into a dual-iris SLED cavity, and the rf phase is reversed at 3.5 fxs. A peak power with a pulse duration of 1.0 fis is extracted from the SLED cavity. The rf power from the SLED is divided into two rf waveguides to feed a peak power of 200 MW into a 3 m-long accelerating structure. The accelerating field distributes from 52 MV/m at the downstream of the ac- celerating structure to 42 MV/m at the upstream of the structure. The energy gain 18 — U1UC EHHANCEO KWU. HIQHWH nMuuiiutti tr nwe LIUCFOH EWiur COWWUTION IMIIWr-llfttI ton |jh gl] ra I go gp] 142i UHi- To Enirgy Compenmallon #2 Figure 4: RF system for the regular section of the linac. 80 MW 4.5 (is 2-lrls SLED J 3: g C4 3 m Ace. j [ 3 m Ace. 52 MV/m 52 MV/m 42 MV/m ^J ^"w,» __>>. 80 <•a£ hr- 120 MeV 120 MeV Energy gain with beam loading Figure 5: RF unit of the regular section of the linac. — 19 of the first bunch among twenty bunches is 119.25 MeV in a 3 m-long structure. As shown in Figure 6, the energy gain of following bunches is reduced by the long-range longitudinal wake-fields, that is the beam-loading. The last bunch gains energy of 112.54 MeV. The energy spread is evaluated to be ±2.6 % at the bunch population of 2.OX1O10 and 20 bunches. 125.0 1 I" 122.5 without Beam Loading ( Ne = 0.0 ) r 120.0 120.05 MeV m < + • * 117.5 - 7.51 MeV (Beam Loading) I with Beam Loading CD (Ne = 2.0E+10) 112.54 MeV I 112.5 Energy Spread : ± 2.6 % w 110.0 . I . , . . I . , , 5 TO 15 "20 Bunch Number Figure 6: Energy gain of multi-bunch in a 3 m-long accelerating structure. 5.1 Active Alignment System In order to avoid any emittance growth in the linac, the accelerating structures should be aligned to less than 200 /xm r.m.s. of the vertical and horizontal directions. The support tables of the accelerator section of the linac have an active mover mechanism and wire-position sensors to align the linac components with a tolerance of less than 20 ^m r.m.s. of the vertical and horizontal directions. The 91 m-long wires are streched in both sides of the linac from the preinjector stage to the end of the linac. One end is fixed to the preinjector stage, shich does not have an active mover mechnism; the other end is streched by a tension weight of 33.5 kg. Assuming no kink on the wire, and uniformity of the wire, the sag of the wire is expressed as cosh - cosh ^(-- (5) where 1, T, p are total length of wire, tension and weight of the wire per unit length. Each position sensor consists of a pair of induction coils electrically connected in series, and mounted on a vertically movable offset stage fixed at a support stage. The center position of a pair of the induction coils is pre-caliblated on the calibration stand. The sensors are installed at four corners of the support table for Q-magnets and beam monitors, and a short support table for an accelerating structure. As for the long support table for the two accelerating structures, six sensors are installed at four corners of the support table, and both sides of the center of the table. --• 20 — The wire position is detected by a synchronous detection of the signal from the defferential coils using 60 kHz, 100 mA AC current on the wire. The frequency of 60 kHz is chosen in the frequency range suitable for the use of lock-in amplifier in order to obtain a stable measurement. The resolution of position sensor is 2.5/im. The dynamic range of the sensors is ±2.5 mm, which is determined by the gap length between two induction coils. The linac support tables are machined with an accuracy of less than ±10/um. The left side of the support table has a reference line parallel to the beam axis. The accelerating structures, Q-magnets and beam monitors are aligned to the reference line with an accuracy of less than ±10/«n. Each sensor installed on the left side of the linac has two pairs of induction coils to detect the vertical and horizontal positions from the wire, These sensors are aligned along the sag of the wire with a vertical offset. Therefore, the reference line is aligned in straight line. Each sensor installed on the right side of the linac stages has a pair of induction coils for measuring only the vertical position. As a result, the support tables are vertically and horizontally aligned with an accuracy of less than 20/fm r.m.s.. 5.2 Energy-Compensation System Klystron Modurator Klystron Modurator Klystron Klystron 50 MW 50 MW 1.0 j .1.0 MS 5 +4.32727 MHz 2856 - 4.32727 MHz MZZ Constant Gradient Constant Gradient Accelerating Structure Accelerating Structure designed at designed at 2856 +4.38727 MHz 2856 • 4.32727 MHz Constant Gradient Accelerating Structure Ls e3 m Pin-max = 50MW Eg-max = 26 MeV / m Ecomp-max = 80 MeV/unit Figure 7: Af energy compensation system. In the damping ring the variation of bunch spacing is not acceptable, the energy compensation system by using four dipole magnets is not applicable. The pro- posed Af energy-compensation system (ECS) is a new idea to compensate for the multi-bunch energy by keeping the bunch separation synchronized with the rf fre- quency. By passing the multi-bunch through an accelerating structure driven at an rf frequency which is slightly larger or smaller than the fundamental frequency, the multi-bunch would be obtained the different energy gain caused by the phase shift. — 21 — As a result, the energy spread of the multi-bunch is compressed to a small value, which is required from the damping ring. The compensation energy depends on the position of electrons in a bunch, since the bunch has a bunch length and the compen- sated field has a slope of the part of sinusoidal wave. If the bunch is compensated by both a negative slope and a positive slope: the efFect of the slope is canceled and the bunch would be accelerated or decelerated by a flat-top field. The system has high flexibility for bunch populations from zero to 4xlO10 electrons/bunch by adjusting the rf power of the klystrons. The accelerating and decelerating energy gain in the ECS structures would be optimized to obtain the minimum energy spread of the multi-bunch. The system consists of two klystrons and two 3 m-long accelerating structures designed at two rf frequencies: slightly higher and lower than the fundamental fre- quency as shown in Figure 7. In order to simplify the timing system, the rf frequency deviation was chosen to be 4.32727 MHz, which is just twice the damping ring rev- olution frequency, and is also the 660th sub-harmonic frequency of the 2856 MHz fundamental frequency [25]. Figure 8 shows the energy distribution of the multi- bunch after being compensated by the Af energy compensation system. 125.0 116.4 | 122.5 with ECS Energy Spread : ± 0.07% I 116.2 e 120.0 116.02 MaV 116.0 I 117.6 115.8 115.64 MeV | 115.0 » 112.5 115.6 10 15 10 15 20 110.0 Bunch Number Bunch Number Figure 8: Energy gain of multi-bunch in an accelerating structure after being com- pensated by the Af energy compensation system. 5.3 Lattice The lattice of the 1.54 GeV linac has been designed by using SAD simulation code. The beam acceptance of the linac is set to = 7 x 10 m, (6) which is ±4.8cr of the incoming beam, which has the following parameters at the downstream of the preinjector. E = 80 MeV (7) 4 = 3 x 10~ m (8) 22 — ax = ay = 0 (9) px = £„ = 1.93 m (10) | Ap/p | < ±1 % (11) a, = 1.5 mm (12) 1 1 i • I : Ji > i; r E 1= entrance of cavily exit of cavity •- 10 o s I* s \S.A] t: SO 17 II ft" 16 16 «!«"« 12 6 i: 3 't? } 3 IJ 3 ' LI I L2I L3I L4I CM L5 Lfl L7 L8I L9LI0 L1IL12IC2L13LI4.U5L16 S3 TDTDDDD s s s s Figure 9: Lattice of the 1.54 GeV ATF linac. Figure 10 shows the result of a simulation of emittance blow-up due to the wakes of cavities under misalignment of accelerator components with an orbit correction using the beam position monitors. The simulation is performed by assuming the injection error Az=Aj,=l mm, and skew rotation error of quads is 0.2 rnrad. 2 IOlo/bunch. 20 bunches, Al-2.8 ns, = En 3 10"^ in, A^,yini»l inm,A©"0.2 mrod .' ' ' ' 1 ' ' ' ' 1 ' ' '/ 1 ' ' ' '/ ~ miximarp of 12 ucnplti / y\ I AxB.yn mm 1 / / =: 3 - / / : *v $ 2 - J * ivtntc of 12 Umple» I i^£=. = =* ~- 1 1 , i i i i 1 I I i i • i i i i i 0.5 1.5 Figure 10: The lattice, fl function, and the beam size with an emittance. 23 Table 4: Schedule of ATF Construction Date Milestone 1993. 8 Beam acceleration up to SOMeV in the ATF Injector Linac by using one 3 m S-band structure. 1993.12 Generation of High Accelerating Gradient with SLED cavity (52 MeV/m maximum; 40 MeV/m average) 1993.12 Completion of the Shielding Hall for the Damping Ring. 1994. 7 Acceleration of the multi-bunch beam up to SOMeV includingg 0.5 m-lonS-band Test Structure for Choke-Mode Damped-Cavity. 1994. 8 Bunch by bunch measurements of beam emittance, length, energy and energy spread. 1995. 5 Test of Hot Model of Damped Cavity will be started. 1995. 6 1.2 GeV part of S-band Linac will be completed. 1995. 6 Beam Transport Line will be completed. 1995. 9 Beam Tuning of 1.2 GeV S-band Linac will be started. 1995.11 Alignment System including the supporting tables with active movers of DR ring will be completed. 1995.12 Vacuum System for the Damping Ring will be completed. 1996. 2 Control System for the ATF will be completed. 1996. 1 Magnet System for the Damping Ring will be completed. 1996. 9 RF System for the Damping Ring will be completed. 1996.10 Extraction Line, Emittance and Bunch Length Measurementm System will be completed. 1996.11 ATF Damping Ring will be completed. 1996.12 Beam Operation of ATF Damping Ring will be started. — 24 — The single bunch energy spread at the exit is at = 0.3 %, and | Ap/p| <0.75 %. These results show that a misalignment of less than 500 fim r.m.s. is not serious after a simple orbit correction. 6 Time Schedule The SOMeV preinjector linac is independently operated for the multibunch gen- eration and acceleration. The beam instrumentation is develped by using the multi-bunch beam from the preinjector. The regular units of the 1.54-GeV linac were almost completed by Autumn '94, except for the instrumentation and energy- compensation system. RF processing of the regular sections are to be under per- formed by July '95, and will be connected to the 80 MeV pre-injector in Aug. '95. The beam transport line will be completed in the region from the bending magnet near to the end of the bV'-- to a movable beam damp near to the injection septums of the damping ring. ATF project milestones are shown in Table. 4. The work for the damping ring has been scheduled, a new AC power sub-station will provide most of the power for the damping ring. The new sub-station will be completed before about the middle of 1995. The cooling facility for the damping ring will also be completed before about the middle of 1995. The beam operation of ATF damping ring will be started at the end of 1996. References [1] "JLC-P, ed. by JLC group, KEK Report 92-16 (1992). [2] K. Yokoya, "Electron Energy Spectrum and Maximum Disruption Angle under Multi-Photon Beamstrahlung", SLAC-PUB-4935, March 1989. [3] K. Oide, Phy. Rev. Letter, 61, 1988, pp 1713. [4] T. Shintake et al, "High Power Test of HO-Free Choke-Mode Damped Accel- erating Structure", Proc. of the 1994 International Linac Conference, Tsukuba, August 1994, pp 293-295. [5] T. Taniuchi, M. Yamamoto, K. kubo, T. Higo and T. Takata /'Damped Struc- tures for JLC Main Linac", in Proceedings of 8th Symposium on Accelerator Science and Technology, IPCR, Saitama, Japan, Nov. 1991, pp 155-157. [6] M. Yamamoto, T. Taniuchi, K. kubo, T. Higo and T. Takata, "Simulation of Long Range Wake Field in a Detuned Structure Based on Equivalent Circuit Model", in Proceedings of 8th Symposium on Accelerator Science and Technol- ogy, IPCR, Saitama, Japan, Nov. 1991, pp 170-172. [7] M. Yamamoto, T. Higo and K. Takata, "Analysis of Detuned Structure by Open Mode Expansion", Proc. of the 1994 International Linac Conference, Tsukuba, August 1994, pp 299-301. [8] Y. Kimura, "Electron-Positron Linear Collider R&D Program at KEK", Pro- ceedings of the 6th Symposium on Accelerator Science and Technology, Tokyo, Oct 1987. [9] K. Takata, "The Japan Linear Collider", in Proceedings of the 1990 Linear Accelerator Conference, Albuquerque, U.S.A., Sept. 1991, pp 18-20. [10] K. Takata, The JLC Project and ATF, Proc. of the 3rd Workshop on Japan Linear Collider (JLC) (held at KEK, Feb. 1992), KEK Proceedings 92-13, pp 1-6. [11] S. Takeda et al., "ATF (Accelerator Test Facility)", Conference Record of the 1991 IEEE Particle Accelerator Conference, San Francisco, May 1991, pp 2047. [12] S. Takeda et al., "1.54 GeV S-band Linac for Accelerator Test Facility", Proc. of 15th International Conference on High Energy Accelerators (held at Hamburg, July 1992) pp 839-841. [13] S. Kuroda et al., "ATF Beam Transport Line", Proc. of the 1994 International Linac Conference (held at Tsukuba, Aug. 1994), pp 83-85. [14] J. Urakawa, KEK Proceedings 91-10, 1991, pp 18-33. [15] H. Hayano et al., "1.54 GeV Injector Linac for ATF Damping Ring", Proc. of the 1994 International Linac Conference, Tsukuba, Aug. 1994, pp 50-52. [16] H. Hayano et al., "An 80 MeV Injector for ATF Linac", Proc. of the 1994 International Linac Conference, Tsukuba, Aug. 1994, pp 236-238. [17] T. Naito et al., "Direct Generation of Multi-Bunch with Thermionic Gun", Proc. of the 15th International Conference of High Energy Accelerators (held at Hamburg, July 1992) pp 164-166. [18] T. Naito et al., "Generation of Multi-bunch Beam with Thermionic Gun for the Japan Linear Collider", Proc. of the 1992 International Linac Conference (held at Ottawa, Aug. 1992) pp 88-90. [19] T. Naito et al., "Multi-bunch Beam with Thermionic Gun for ATF", Proc. of the 1994 International Linac Conference (held at Tsukuba, Aug. 1994), pp 375- 377. [20] T. Naito et al., "Bunch by Bunch Monitors for ATF Injector Linac", Proc. of the 1994 International Linac Conference (held at Tsukuba, Aug. 1994), pp 887- 889. [21] M. Akemoto et al., "Pulse Modulator for 85 MW Klystron in ATF Linac", Proc. of the 1994 International Linac Conference (held at Tsukuba, Aug. 1994), pp 415-418. [22] H. Matsumoto et al., "High Power Test of a SLED System with Dual Side- Wall Coupling Irises for Linear Collider", Nucl. Instr. and Meth. A330 (1993) pp 1-11. [23] H. Matsumoto et al., "High-Gradient S-band Test Linac for Japan Linear Col- lider", Proc. of the 1992 International Linear Collider (held at Ottawa, Aug. 1992) pp 296-298. [24] H. Matsumoto et al., "High Power Test of a High Gradient S-band Accelerator Unit for the Accelerator Test Facility", Proc. of the 1992 International Linac Conference (held at Ottawa, Aug. 1992) pp 62-64. [25] T. Korhonen et al., "R&D of the ATF Timing System", Proc. of the 1994 International Linac Conference (held at Tsukuba, Aug. 1994), pp 831-833. 27 — BUTTON-TYPE BEAM-POSITION MONITOR FOR THE ATF DAMPING RING F.Hinode KEK, National Laboratory for High Energy Physics,l-1 Oho, Tsukuba-shi, Ibaraki-ken, 305 Japan Button-type beam-position monitors (BPMs) were fabricated for the ATF damping ring. The BPM was designed to achieve a position resolution of less than 5 \lm; fabrication of the first 40 BPMs has been completed. For this BPM, a beam test was carried out at the 80-MeV injector part of the ATF LINAC. All of the bunch signals in the multi-bunch beam were clearly observed without any discharge. A calibration of the BPMs was also performed in order to check their offset from the electrical center to the mechanical center as well as their position detection sensitivity, The result shows good uniformity in position detection. Figure 1: Pickup electrode for the BPM. I. INTRODUCTION for mechanical strength and vacuum leakage. The threshold level of the brazing strength of the central conductor is 60kg The ATF [1] is under construction at KEK in order to against the tensile forces. No vacuum leak has occurred after study the feasibility for the linear-collider (JLC [2]). The ATF heat-cycle tests from liquid-nitrogen temperature up to 200 damping ring will be operated at 1.54 GeV with a multi-bunch °C. Furthermore, all of the electrodes have been checked electron beam, which will have a the vertical emittance of 5 x concerning their dielectric strength (> 1000 Volts). 10"11 m-rad. To achieve such a low-emittance beam we must correct for the dispersion of the orbit, which is small (T| < 2 B. BPM Block mm) in the long wiggler section of the damping ring. A precise measurement of the dispersion is indispensable. Each BPM is installed at a location near to every Usually, the actual dispersion is obtained by comparing each quadrupole magnet and sextupole magnet; there are 120 closed-orbit distortion under the conditions of different RF BPMs in the ring. We have already fabricated 40 BPMs in frequency (Af**7 -10 kHz). For this reason, the requirement for 1994. the resolution of the BPM is less than 5 |i.m [3], Furthermore, EBW, the impedance of the components, such as the vacuum chamber, can be the source of a single-bunch instability which could degrade the beam quality. To avoid such an instability, the total longitudinal impedance must be less than about 0,2 Q [1]. The results of wake-field calculations indicate that the 5 mm impedance is very small on the electrode of the button-type compared with the directional-coupler type [4]. We therefore selected the button-type electrode for the BPM [5], n. CONFIGURATION OF THE BEAM- POSITION MONITOR Figure 2: Cross-sectional view of the BPM. The BPM block was machined from a mass of aluminum A. Pickup Electrode alloy; four pickup electrodes were welded onto the block by The feedthrough consists of a central conductor with a electron beam welding (Figure 2). It has both horizontal and button electrode, an outer conductor with an SMA connector vertical reference planes, which are used for aligning the BPM and an insulator, as shown in Figure 1. The central conductor to the Q magnet. Both planes are also used as a reference for is made of Kovar and the button electrode is made of stainless the calibration process to obtain the electrical center of the steel. Taking into account the impedance, the curvature of the BPM block. After installation of the vacuum chamber with button shape is the same as the inner radius of the BPM block. BPMs to the beam line, the offset of the BPM to the Q magnet The insulator is ceramic AI2O3. The outer conductor with the will be measured with an accuracy within 50 (i.m by using the SMA connector is made of aluminum alloy with titanium reference planes. joined by an HIP transition, Some electrodes have been tested — 28 — III. BEAM TEST To study a low-emittance beam in detail, each bunch signal in the multi-bunch beam should be clearly distinguished. Furthermore, the BPMs must be operated stably against 3 x 10'° particles per bunch at maximum. Since the feedthrough of the BPM has a cavity structure, there is a possibility that a discharge would occur due to the multipactering effect caused by resonance fields excited by the wake-field, which are induced by the beam. At the Figure 4: Observed signal of a single-bunch beam. TRISTAN main ring, in fact, such a discharge effect, caused by the TEn coaxial mode, was observed on a button-type particles. There is a ringing shape at the signal tail. This is BPM placed in the RF cavity section [6]. The result of a caused by multiple reflections of the wall current at ihe edge resonance measurement shows that the lowest resonance of the BPM chamber due to the discontinuity. The difference frequency is 11 GHz, which is consistent with the result in the inner diameter between the vacuum chamber of the calculated by the MAFIA code. According to the calculation LINAC ( B. Beam Test using a Single-bunch Beam An observed signal under single-bunch operation is 10 iis/div shown in Figure 4. The peak pulse height is 5 V for 1.6 x 10'° Figure 5: Observed signal of a multi-bunch beam. sampling oscilloscope SHBl(357MHz) SHB2(357MHz) RG-213/u (~ 33m) BPM E-GUN WC3 \ \ . PB1 PB2 B(Buncher) PRM2 (Prc'bunchcr) WCM " Faraday Cup Figure 3: Setup of the beam test. almost the same shape as that of the single-bunch beam. The In this way we obtained a relation between the measured signal tail of each bunch is sufficiently small when the next position (x, y) and the set position (X, Y) of the wire position bunch signal arrives. for each monitor. The calibration was performed at 169 points in the central area, which is a 3.6 mm square region with a 0.3 mm step. As IV. Electrical Characteristics a result, there was no remarkable distortion in the mapping of all the BPMs. In the central region of ±0.5 mm the distortion was less than 10 Jim, while it was 100 |im at a point 1.8 mm A. Capacitance of Electrode away from the center. This distortion is acceptable for our Before and after electrodes were weldec onto blocks, the practical use. In the central region, however, it is only 2% and capacitance of each electrode was measured using a thus quite good. capacitance meter at a frequency of 800 Hz. The electrical- We obtained the distribution of the offsets of the capacitance values are distributed in the 2.5 ~ 4.7 pF region. electrical center to the mechanical one from the calibration Thus, four pickup electrodes with similar capacitance values data. The mean values of the offsets are X = -19 \im and Y = were grouped and welded onto the monitor block in order to 58 Jim, and the standard deviations are -90 jxm in both obtain uniform signals. directions. The main source of this deviation came from the errors in manufacturing the BPM block and pickup electrode. B. Mapping for the Calibration The second fabrication of 40 BPMs and the calibration Every BPM block was calibrated so as to obtain an offset are now progressing. between the mechanical and electrical centers of the BPM block [1,5]. The mechanical center is defined by the horizontal V. SUMMARY and vertical reference planes within the accuracy of machining. The electrical center was measured by the Button-type beam-position monitors (BPMs) were following procedure. fabricated for the ATF damping ring. A beam test of the BPM The BPM block was mounted on a fixed stage, and a 50 was carried out at the 80-MeV injector part of the ATF |J.m diameter tungsten wire was strung coaxially in the BPM LINAC. All of the bunch signals in a multi-bunch beam were block. Both ends of the wire were placed in a V-shaped clearly observed without any discharge up to 2 x 1010 groove made of ceramics, which was installed on both x and y particles per bunch. Calibration of the BPMs was also movable stages. Each stage has a high positioning accuracy of performed; the result shows good uniformity of position 0.1 Jim. One side of the wire was soldered to the SMA detection. connector. The other end was terminated so as to match into a 50 Q line; attached 100 g weights provide a proper constant tension. The wire was put in the base position of a gauge, VI. REFERENCES which has the same reference planes as that of the BPM, and observed directly by a microscope. Thus, the wire was [1] ATF Design and Study Report, to be published as a precisely aligned to the mechanical center of the BPM block KEK Internal Report. with an accuracy of 40 |im. The calibration was performed by [2] JLC Group, KEK-Report 92-16. using a pulse signal with a 5 ns rise time. Signal processing is [3] M. Tejima, Procs. of the SLAC/KEK Linear Collider based on the same principle as that for the SLAC/FFTB [8], Workshop on Damping Ring, KEK proceedings 92-6, 1992, The readout electronics, which are a combination of a pulse- p. 126-132. stretcher amplifier, a Track&Hold ADC and a pulse generator, have a resolution of 5 (im. We measured the four output [4] M. Takao et al., KEK-Report 91-14. [5] M. Tejima et al., Proceedings of the 1994 signals (V), V2, V3, V4) induced electrostatically on each electrode due to pulse signals transmitted on the wire. In order International Linac Conference, Vol. 2, pp. 914-916. to obtain the beam position, the following two calculation [6] M. Tejima et al., KEK Preprint 90-183. steps were performed. The first is a normalization procedure [7] A. J. Hatch, Nucl. Instr. and Meth. 41(1996) 261 -271. 1 (x , y'), given by [8] H. Hayano et al.; KEK Preprint 92-118, H. Hayano et al.; SLAC-PUB-5691. where k is a coefficient of the sensitivity on the position measurement, which depends on the geometry of the monitor chamber. It was obtained by a measurement for each BPM. Secondly, we converted the normalized results to the geometrical position (x,y) according to — 30 — Recent Progress on Cathode Development and Gun Development at Nagoya and KEK M. Tawada, T. Nakanishi, S. Okumi, H. Aoyagi \ S. Nakamura2 , M. Tsubata3 , K. Togawa, C. Takahashi, Y. Tanimoto 4, C. Suzuki Department of Physics, Nagoya University, Nagoya, 457, Japan M. Yoshioka, H.Matsumoto, Y. Takeuchi, T. Omori, Y. Kurihara KEK, National Laboratory for High Energy Physics, Oho 1-1, Tsukuba-city, Ibaraki-ken, 305, Japan T. Saka, T. Kato New Material Research lab., Daido Steel Co. Ltd, Nagoya, 457, Japan T. Baba and M. Mizuta Fundamental Research Laboratory, NEC Corporation, Miyukigaoka 34, Tsukuba-city Ibaraki-ken, 305, Japan K.Nishitani Accelerator Technology Corporation, Koyo Sen'l Bldg. 36-7 Namiki-cho Hachioji-shi, Tokyo 193 Japan Abstract We have improved the performance of high polarized electron photocathodes. The new generation photocahode for polarized electron source becomes to provide much higher degree of the electron polarization and also the quantum efficiency of more than 1%. The design of those cathodes are reviewed and the status of the polarized electron guns which have been constructed at Nagoya and KEK are reported. 1 Introduction The polarized electron beam will play an important role in the experiments by the future e+e~ Linear Colliders, such, as Japan Linear Collider(JLC) [1]. In order to prepare the polar- ized electron beam for JLC, we have been developing the high polarization photocathodes and forwarding a construction of the operational gun. The new generation of photocathodes for polarized electron source have been developed to realize a much higher degree of polarization than 50% and a high quantum efficiency, such as about 1%. The breakthroughs against a polarization limit of 50% by GaAs came in 1991, and 71% by AlGaAs-GaAs superlattice by KEK/Nagoya/NEC and 86% [2] by strained GaAs by Nagoya/Osaka Pref./Toyota Tech. Inst./Daido [3] were achieved. Further improved photocath- odes have been developed based on new ideas, such as the resonant absorption photocathodes for strained GaAs, modulated doping for superlattice and the strained layer for superlattice. Present address:Japan Synchrotron Radiation Research Institute (JASRI), Kamigori, Ako, Hyogo 678-12, Japan 2Present address:Physikalislies Institute der Universitiite Bonn, Bonn, Germany 3Present address:Nihon Koshuha Co. Ltd. ^Present address:Power Reactor and Nuclear Fuel Development Corporation, O-arai Engineering Center, Narita-cho 4002, O-arai-cho, Higashiibaraki-gun, Ibaraki-ken, 311-13, Japan 31 — 2 Cathode Improvement The photocathdes such as strained GaAs and superlattice had a big advantage in polarization, but the quantum efficiency is was relatively low compared with GaAs and seems not to be sufficient for JLC's polarized electron source. The SLAC group found that the maximum charge which could be extracted from the strained GaAs cathode was limited not by the space charge limit but by the electron trapping at the NEA (Negative Electron Affinity) surface, using the SLC polarized gun, which generated a 2.5 nsec single bunch beam with repetition rate 120 Hz. Tins effect, which is called "Cathode Charge Limit Effect" will give the more serious limitation on the extracted peak current for multi-bunch beam which required at JLC. So, it is very important to develop a photocahode with a good NEA surface (high quantum efficiency) and to investigate this phenomena using the high current multi-bunch beam. We have developed the two kinds of techniques to improve the quantum efficiency without lowering the polarization, i.e. resonant absorption method for Strained GaAs and modulated doing method for AlGaAs-GaAs supeiiattice. 2.1 Resonant absorption Strained GaAs Incident light \) Rita •-- •-. •-• •-- •-• j GaAs 1/4 DBR GaAsP GaAs Figure 1: The principle of the Fabry-Perot optical cavity with DBR. The absorption of the incident lights is enhanced between the DBR and the surface of GaAs. The structure of resonant absorption strained GaAs photocathode is shown Figure 1. Two parallel mirrors; A Fabry-Perot cavity is formed by a quarter wave DBR(Distributed Bragg Re- flector) and a surface boundary between GaAs and vacuum. Incident laser lights can be partially confined in a Fabry-Perot cavity. The real DBR was composed of 30 pairs of Alo.iGao.9As and Alo.6Gao.4As layers and had the reflectivity higher than 90% within the laser bandwidth of about 50nm. The enhancement of the quantum efficiency is expected at the wavelength which satisfies the resonant condition; 2nL = mA (m-.integer), where n, L are refractive index and total length •- 32 — -6 700 800 900 10 Wavelength (nm) Figure 2: Experimental results of the polarization and the quantum efficiency by the strained GaAs with DBR. of optical cavity, respectively. The obtained .result is shown in Figure 2 , where the quantum efficiency of the DBR cathode is indicated by solid circles, while that of normal strained GaAs is done by open circles. The quantum efficiency of about 1.3% was achieved at the wavelength of 860nm which gives the polarization of about 70%. In principle, this device does not bring the decrease of electron polarization because the same helicity electrons are produced by reflected lights as by incident lights, due to that the photon circular polarization is reversed on reflection and so is the the direction of electron emission. This expectation was also demonstrated by the experiment [4]. 2.2 Modulated Doping Superlattice In general, to realize NEA surface the semiconductor material of photocathode must be heavily p-doped . But the heavy doping may also cause the spin depolarization of electrons due to the electron-impurity scattering. Thus by higher doping only near the surface region than that of bulk of cathode, we can expect both of high quantum efficiency and high polarization. We have applied this modulated doping technique on AlGaAs-GaAs superlattice. The result is shown in Figure 3, where (a) is the result of the modulated doping superlattice and (b) is that uf the uniformly doping superlattice These cathodes were tested by a Nagoya cathode test system, whose bias voltage was -4kV and a base pressure was about lxlO~9 Torr. — 33 — 700 750 800 Wavelength (nm) Figure 3: The experimental results of the polarization by both modulated doping sample(a) which are indicated by solid circles and uniformly doping sample (b) are indicated by open circles. The arrows show the wavelength of the maximum polarization of each sample. Both samples have the same Be-doping density of about 5xlO17/cc inside superlattice, while a high density of about 4xlO19/cc is doped only for sample (a)to a surface GaAs layer with 50 A thickness. As the shorter wavelength photons can give the much higher quantum efficiency at the threshold of the band gap excitation, the wavelength shift of sample (a) brings a great advantage for the quantum efficiency improvement. The good quantum efficiency of 0.5% and the polarization of 70% at 757nm was obtained. Using the modulated doping superlattice, the total charge extracted by SLC gun with a bias volatge of -120kV was measured and a record performance of 2.3xlO11 electrons in 2.5 nsec pulse was achieved [5]. This was a nice data, because a so-called "Cathode Charge Limit Effect" was not observed in this experiment. We are preparing a test to extract a double pulse beam from AlGaAs-GaAs superlattice next experiment. 2.3 Strained Layer Superlattice The modulated doping AlGaAs-GaAs superlattice have a high quantum efficiency but the polarization is still lower than that of strained GaAs. This seems to be caused due to the higher order band mixing between a heavy hole band and a light hole band near the F-point. It is important to improve the polarization of a superlattice photocahode as much as possible, maintaining high quantum efficiency which assures a high charge current. We introduced a strain to a superlattice since the strain applied on the superlattice is expected to eliminate the higher order mixing of heavy hole and light hole bands [6]. Our choice of material for the first test is the InGaAs-GaAs strained-layer superlattice grown on a GaAs substrate. The new sample has a similar structure to sample of the two dimensional superlattice with modulation doping, but except in that the InGaAs layer are strained. The polarization of 89% was observed at the wavelength of 911nm, while the typical quantum efficiency was lxlO~3%at the same wavelength. unstrained- strained- unstrained- stralned- GaAs InGaAs GaAs InGaAs E0(GaAs)=1420 Eth=1357 Figure 4: The band structure of the Ino.isGao.ssAs-GaAs strained layer superlattice. Obviously the polarization of superlattice cathode is improved, but the quantum efficiency is not enough. We consider that the Fermi-bump caused by modulation doping may obstruct electron escape to the vacuum. Thus we believe that the quantum efficiency of InGaAs-GaAs strained layer superlattice can be improved by optimizing the modulation of acceptor concen- tration. 3 Gun Development The performances required for JLC polarized electron beam are high intensity, multi-bunch beam with a long cathode lifetime. As a first step, we are constructing the proto-type of DC -lOOkV polarized guns with measurement systems of the polarization both at Nagoya and KEK. The KEK's polarized electron source system is almost completed, and will be used for cathode development in near future. The schematic view of this system is shown in Figure 5 schematically, which consist of a DC gun, a Wien filter, a Mott polarimeter and a tunable CW Ti-sapphire laser. KEK Polarized Electron Source view port view mirror screen monito Figure 5: Schematic diagram of polarized electron source. The base pressure of the gun is 5xlO~nTorr and the dark current at an accelerating voltage of -lOOkV is applied is less than lOnA. In order to investigate the fundamental properties of the electric discharge mechanisms, we are constructing on High Gradient DC Gun at KEK, which is shown in Figure 6. An electrode is mounted on the ceramic insulators, and instrumentation to measure the dark current and the residual gases are installed. The top of electrode can be easily exchanged for the tests of the material search. The main pump system comprises a combination of ion pump and non-evaporation getter pump. All of the components were surface treated using both of new type photocathodes and assembled very carefully in clean room so as to avoid any contamination by dust. The base pressure of 8xlO~12 Torr was already attained and further performance test is scheduled in near future. We will try the high voltage test. — 36 — ROUGH PUMP ^ SYSTEM NEG VIEWING EXTRACTOR PORT GAUGE HVDC POWER SUPPLY ~- 300 kV SF6 NK dean Z SUS316L Figure 6: Schematic diagram of the experimental apparatus for the ultra-high-vacuum DC gun. 4 Conclusion We have demonstrated the usefulness of a resonant absorption strained GaAs and a modulated doping superlattice for the quantum efficiency improvement by a factor 10. The polarization of 90% of superlattice could be attained by introducing a strain to superlattice structure. The fundamental study to realize a gun with a extremely low dark current at high voltage was planned and the construction of the test gun is in progress. References [1] JLC Group, KEK Report 92-16 [2] T. Omori et al, Phys. Rev. Lett. 67 (1991) 3294 [3] T. Nakanishi et al., Phys. Rev. Lett. 66 (1991) 2376 [4] T. Saka et al., Jpn. J. Appl. Phys. 32 (1993) L1837-1840 (Part2, No. 12B) [5] Y. Kurihara et al., Jpn. J. Appl. Phys. 34 (1995) 355-358 [6] T. Omori et al., Jpn. J. Appl. Phys. 33 (1994) 5676-5680 Development of polarized e+ beams for future linear colliders M. Chiba1, A. Endo2, R. Hamatsu1, M. Hirose2, T. Hirose1 M. Irako1, N. Kawasaki1, T. Kumita1, Y. Kurihara3 T. Matsumoto1, H. Nakabushi2, T. Okugi1, T. Omori3 Y. Takeuchi3, M. Washio2, J. Yang1, M. Yoshioka3 1Department of Physics, Tokyo Metropolitan University, Hachioji, Tokyo 192-03, Japan 2Research and Development Center, Sumitomo Heavy Industries, Ltd. (SHI), 63-30, Yuuhigaoka, Hiratsuka, Kanagawa 254, Japan 3National Laboratory for High Energy Physics (KEK) Tsukuba, Ibaraki 305, Japan Abstract We have so far been carrying out systematic investigations to cre- ate polarized e+ on the basis of two new methods. One method is to use p+ decay of radioactive neuclei with short life-time produced with a proton cyclotron and the other method is to use e+e~ pair creation from polarized 7 beams made through backward Compton scattering of laser lights. Here we describe technical details on productions of polarized e+ and measurements of the polarization. The experiments of producing polarized e+ will soon start. Although the e+ intensity is not sufficiently high, we will acquire lots of know-how for further development of polarized e+ sources with high quality which will pos- sibly be applied to future linear colliders. This report is written based on the talks at the 5th JLC Workshop, and includes progress since then. — 38 — 1 Introduction Since development of polarized e~ source has been successfully done [1] [2] [3] [4], we can expect a highly polarized e~ beam in future linear collid- ers, for which we require higher performance on the e~ source than for the SLC. When we discuss physics at future linear colliders in this literature, we always assume that e~ beam is highly polarized. Questions that arise next are; • Can we also polarize an e+ beam? « If an answer is YES, what can we gain from the collider in which both of e~ and e+ beams are polarized? From the point of view of high energy physics, the polarized e+ beam should play important roles in next generation e+e~ colliders. First, we discuss cases that we study physics of the Standard Model in which e~ and e+ always in- teract in the helicity combination of e^ejT (RL) and e£e# (LR), except in the case of two photon interaction. Therefore if we choose the helicity of e~ (note helicity equals chiraity in massless limit), the helicity of e+ , which can in- teract with the e~ , is automatically selected. However, even in this case, the polarization of e+ plays a particular role. If e+ beam is not polarized, half of e+ in the beam cannot interact because half of collisions occurs in the "wrong" combination, such as e^e^ (RR) and e£e£ (LL). Then polarization of e+ beam gives us a factor two increase of an effective luminosity if both e~ and e+ beams have 100% polarization, because we can always choose the "correct" combination of helicities (RL and LR). In actual colliders, whose beams are not 100% polarized, it is convenient to define the quantity C cor- responding to the fraction of "correct" helicity combinations of e~ and e+ in + the beams: C = |(1 — P+P-) where P+ and P_ are polarizations of e and e~ beams, respectively (P± = 1 means the polarization of 100% right handed helicity beam). We can also increase an effective polarization Pejj — 1SP~_p as shown in the comparison of two cases, namely (i) 90% left-handed e~ beam and a non-polarized e+ beam and (ii) 90% left-handed e~ beam and 80% po- larization of a right-handed e+ beam. The latter gains a factor 1.7 increase of an effective luminosity than the former; C^p V- 09Pp ==o\ = l-^ and hence the latter has effective polarization of 99% while the former has 90%. 39 — In addition, the polarization of both e+ and e beams will suppress dras- tically the systematic error on measurement of the left-right asymmetry of a cross section; ^ a(LR) - a(RL) LR a(LR) + a{RL) ' When we measure ALR, most part of systematic errors arises from the polar- ization measurement of a beam in the case of single beam polarization. If we polarize both of e+ and e~ beams, propagation of errors from the polarization measurements to ALR can be drastically supressed[5]. In studying new physics beyond the Standard Model, the polarization of an e+ beam would play much essential roles, because some models al- low interactions in "wrong" combinations (RR and LL). The supersymmet- ric extension of the Standard Model, for example, allows reactions such as, e e + R fl ~^ &RX ve , and e£e£ —>• e^x"^, where e* and y^ stand for a sealer electron and a chargino, respectively. Thus, by choosing the helicities of e~ and e+ beams as RR or LL, we can study these interactions free of the backgrounds from the Standard Model interactions, e.g. e^e^ —> W+W~, eteR —* W^W"". Furthermore, if the mass of e^ are smaller than that of x^, the interactions mentioned above have lower threshold than e+e~ —> X+X~ to produce x±. According to these considerations, a future linear collider, e.g. JLC [6] will first be operated to achieve sufficiently large luminosity in the LR and RL modes which enables us to discuss in detail the physics of the Standard Model and to check small amount of discrepancy from the theory. Then, we will operate it in the LL and RR mode to search/study a special issue that would be suggested from analysis of data obtained in the LR and RL operation. There might be long period of time and one can also expect drastic progress of accelerator technology before polarized positron beams are ac- tually realized in future linear colliders. Therefore we attempt to create polarized e+ on the basis of different two ideas, one of which will be later decided to adopt for future linear colliders. It is well known that e+ emit- ted through /?+ decays are longitudinally polarized with the helicity v/c, v and c being the velocities of e+ and light. Until now, one gets polarized e+ only from j3+ decays of radioisotopes. Since numbers of e+ from isolated radioisotopes used in off-line manner is quite limited, it is necessary to de- — 40 — velop new methods to generate large amount of polarized e+ . The first method is to use fi+ decay from radioisotopes which are produced through (p,n) interaction by means of on-line usage of a compact proton cyclotron (Fig.l) resulting in possible utilization of radioisotopes with short life-time. This method provides continuous current of e+ , even if we use a pulse beam of protons, because the time structure of the proton beam is smeared by a decay constant of radioactive nuclei. Therefore, when we apply this method to linear colliders, we need an efficient and depolarization-free buncher. In chapter 2, details of the first method will be described. The second method is to use e+e~ pair creation of a polarized 7 beam, which is made by backward Compton scattering of laser lights on the e~ beam supplied from a linac. We will use 1.54GeV e~ beam from the linac of the Accelerator Test Facility (ATF) that is under construction at KEK for R&D works on future linear colliders (Fig.2). The e+ beam produced by this method has the same time structure as that of the initial laser/e~ beam. In the VLEPP project [7] and the DLC/TESLA [8] [9] project, it is also proposed [10] [11] that the polarized e+ beams are generated by e+e~ pair creation from polarized 7 beams. The difference between the methods in these proposals and our second method is how to generate the polarized 7 beam. In the VLEPP and the DLC/TESLA projects, a 7 beam is generated as polarized radiation from a few hundreds GeV e" beam which goes through a long helical undulater. Details of the second method will be described in chapter 4. Future linear colliders require large amount of e+ with complex time structure. For example, JLC requires intensity of 1014 e+ /sec, and the pulse structure shown in Fig.3. We set a first-stage goal of development, for each method, which is reasonably realistic from technical point of view. The first step of the /5+ decay method is to create continuous e+ beams of 105/sec with reasonably good magnitude of polarizations. For the polarized 7 method, we will create 104/sec polarized e+ pulses with 10 Hz, whose time structure is much simpler than that of JLC. Since there are long distance between any of first stage goals and JLC requirements, many years of further efforts and innovations will be required, even if the first stage will be achieved in near future. In Chapter 5, we briefly mention the expected technical problems for realization of the polarized e+ source of future linear colliders and some applications in the fields other than high energy e+e~ colliders. 41 — 2 Production of polarized slow e+ ?s using a compact cyclotron 2.1 Production system of slow e+ 's The group of Sumitomo Heavy Industries, Ltd.(SHI) could success- fully produce intense slow e+ for the first time using a compact cyclotron. Recently, many laboratories have developed slow e+ sources by means of ra- dioisotopes (RI)[12] and an electron linear accelerator (LINAC)[13] in order to research material surfaces. Generally the former is compact and inex- pensive, but the intensity of the slow e+ is not so high, while the latter method can produce intense slow e+ but its cost is very high and the system is normally very large. It is expected that our slow e+ production system using a compact cy- clotron [14] will be able to provide polarized e+ with high intensity emitted from j3+ decay radioisotopes which we produce in a nuclear reaction caused by proton irradiation of a target material. We have chosen aluminum (Al) as the target material so that the reaction 27Al(p,n)27Si with a threshold energy of 5.8 MeV takes places. The radioisotope 27Si thus produced has the (3+ decay branching of 100% and a half-life of 4.1s. The large maximum energy of 3.85MeV for the /3+ rays emitted from 27Si favors to achieve large helicity of e+ because of large magnitude of v/c. 27 Fig. 4 shows saturated radioactivity of Si for the proton current 1/J,A. Using this distribution, we calculate the energy distribution of e+ at the production point of e+ which has enough energy to go out of Al. The /3+ rays with wide energy spread are converted to slow e+ with good energy resolution at a moderator. Most of the high efficiency moderators have negative-work- function, which enable us to create slow e+ with narrow energy width (e.#.Ni(100)), or maximum yield (e.g. W(110)). At present, we use a 25 ^m thick polycrystal W foil as a moderator. However, using the Monte Carlo program SPG[15], we will make further examination to determine the most suitable thickness of W and the distance between Al and W for efficient creation of highly polarized e+ . The calculated intensity of the slow e+ is 4.3 x 105 slow e+/s when a proton beam with an energy of 18 MeV and a current of 1/JA irradiates an aluminum target with thickness of 1.8 mm and the slow positrons are — \2 — extracted from the moderator with the conversion efficiency of 10 4 from j3+ rays to slow e+ . Based on this value an intensity of 3.0 x 107 slow e+/s should be achieved using the compact cyclotron which Sumitomo Heavy Industries, Ltd. manufactures (commercial name: CYPRIS; maximum proton current: 70 //A). In addition, if the proton beam current and the conversion efficiency are improved up to 1 mA and 10~3 respectively in the future, it might be possible to produce a super intense e+ beam with intensity of the order of 109 e+/s. This number is the maximum of possible intensity which can be accomplished using a cyclotron and a conventional moderator. The slow e+ generated from the moderator are transported to the po- larimeter set 27m downstream using a 100G magnetic field produced by solenoid coils. The intensity of the slow e+ are determined by stopping all e+ and measuring the annihilation gamma-rays with a Ge semiconductor detector. Fig. 5 shows the e+ beam profile observed by a microchannel plate (MCP) with the diameter of lOmm^. The e+ intensities of 2 x 106 e+/s and 5 x 105 e+/s have been obtained after the transport of 10 m and 25 m, respectively, for the proton current of 30 fiA. These values are considerably lower than the theoretical maximum intensity of 1.1 X 107 e+/s estimated for the Al target thickness of 1.8 mm and thus the intensity should be improved in the future by refinement of various adjustments. 2.2 Helicities of slow e+ 's In order to estimate helicities of slow e+ emitted from the moderator, we simulate trajectories of e+ in the Al-target and the W-moderator using the Monte Carlo program, GEANT [16]. For e+ emitted from the Al target, we first calculate the velocity v and the angle 6 at the production point, 6 being the direction of e+ measured from the z- (beam) axis. Fig. 6 (a) shows en- ergy distributions of intial e+ produced inside Al. Among these e+, only high energy e+ can pass through Al as shown in Fig. 6(b). Until now, nobody has succeeded to derive an exact formulation for the spin precession of slow e+ in materials. However, the Michigan group clarified experimentally that no sig- nificant depolarization occurs when e+ is moving and loosing their energies in various materials [17]. We therefore assume for the calculations of polariza- tions that the depolarization of e+ does not take place during thermalization processes. If the spin directions are not changed inside the materials i.e. Al, + we can determine the z component of the e polarization Pz = (v/c)cos9 at — 43 the entrance of the moderator VV with the diameter 10mm placed at the dis- tance of 10mm from the Al target. As shown in Fig. 7, we obtain an average + e polarization, < Pz >= 0.68: There exist the negative Pz which are caused by backward scattered e+ 's. We further trace these e+ through the mod- erator and check how the polarization changes. The simulation clarifies as shown in Fig. 9 that the polarization increases as the energy becomes higher. Then we will be able to extract e+ with large polarizations if we utilize the thicker moderator which can absorb low energy e+ ( e.g. E < lMeV ) : This was actually observed by the Michigan group [17]. We are now investigating the most suitable thickness of the W moderator by means of a Monte Carlo Simulator SPG which includes thermarization processes. 2.3 Spin precession in electric and magnetic fields A schematic view of the experimental apparatus is shown in Fig.8. Slow e+ 's are bunched into pulses with the time spread of 150psec, accelerated electrostatically up to 2~20KeV and injected into the polarimeter. Use of a Wien filter (spin rotator) is considered as an option to reduce possible systematic errors. The initial polarization of e+ produced in the j3+ decay is subject to depolarization in the beam generation and transportation pro- cesses. Therefore, we have to study in detail possible depolarization at each section of the beam line. A spin precession during the beam transportation in electric and magnetic fields should be described by the relativistic the- ory of electromagnetism. We have developed a computer program POEM (POlarized beam simulator in Electric and Magnetic fields)[18] , based on GEANT, to simulate the trajectory and polarization of the slow positrons in static electric and magnetic fields. The relativistic equation of motion for charged particles moving with the velocity v in the electric field E and the magnetic field B is; —* 4(m7t0 = e(£ + t?x-), (1) at c and the spin of the particle obeys the Thomas' equation [19] d J JLJJX [(i + )5(i) it mc K2 i' 2 7 where s is the spin of the particle defined in the nonrotating rest system of the particle, g is the ^-factor of e+ , J3 = v/c is the ratio of the velocity v to the speed of light c, and 7 = 1/Vl - P2- All physical quantities (u, B, E, t) other than s are defined in the laboratory system. Eq.(2) includes contribution from the Thomas precession arising from rotation of the proper coordinate system of the particle with respect to the laboratory system. Hence the trajectory and polarization are calculated by solving the differential equations (1) and (2) using the Runge-Kutta method. A simulation for one of the simplest cases, i.e., the beam transportation along a straight beam pipe with an axially uniform magnetic field of 100G, is performed for e+ with the longitudinal and transverse kinetic energies of 10 eV and 1 eV, respectively. As shown in Fig.10 (a), the transverse component of the spin increases by ~ 10~5 after the transportation of 100cm suggesting that the depolarization effect in the 27-m beam line, guided by the solenoid coils, is negligible if the beam line is straight. However, the actual beam line has seven curved sections with 90 degree of arcs of about 70cm in radii to reduce backgrounds from the cyclotron. Fig. 10 (b) shows results of the simulation for the spin precession along one of the curved sections. The transverse component of the spin increases by 1.8 x 10~3 after the e+ trajectry is bent through 90 degree of the arc. This result is consistent with the naive calculation of the effect caused by the anomalous magnetic moment of e+ , 9~ • 7 • f • Therefore, we can conclude that no large depolarization occurs in our beam transportation system. 3 Measurement of the e+ polarization We need to design the apparatus of polarization measurement in the solenoidal beam transportating and focusing system. The mesurement of e~ polarization has been accomplished using either Mott(e~ -nucleus) or M0ller(e~ -e~ ) scattering. In principle, similar methods, i.e., Mott or Bhabha scattering could also be applied to determine the e+ polarization. However the asymmetry, i.e., the fractional change in counting rate when a completely polarized beam has its polarization flipped, in Mott scattering of e+ is very small, since e+ is repelled from the nucleus. The asymmetry in Bhabha scattering are also much reduced over M0ller asymmetry at energies below 1 MeV since the Pauli principle is not operative for e+-e~ collisions. + sz{e + e-) ^37(s2)/^27(sz=0) 2 1 *. = l(Tft) forbidden ^ ' A37 3 ' 372 1 1 sz = O(lfr) A27 3 ' 372 Table 1: | and [ mean that the e spin is up and down, respectively, "ft and + •Ij- the e spin. A27(Si) is the decay rate of 27 with total spin sz = 0. We thus discuss other methods of the polarization mesurement, applicable to slow e+ under keV. In order to establish the reliable method to measure e+ polarization, we attempt to construct measuring devices based on two different methods as described below. 3.1 Polarization measurement using a ferromagnetic target In the case of free annihilations of unpolarized e+ and e~ at rest, the branching ratio -S37/27 of 37 to 27 annihilation is represented as •*• •UHnt/i.'v537/2 7 -— •T= -— r=7j, where A27 and A37 are the decay rates of 27 and 37 annihilations for unpo- larized e+ and e~. For e+-e~ annihilation at rest, spin states of e+ and e" in parallel (which is called triplet state) annihilates into 37, while anti-parallel spin states (which is composed of the singlet(50%) and triplet(50%) states) annihilate into 27 and 37, respectively. The decay rates of the 27 and 37 annihilations for the total spin sz are shown in Table 1. The numbers of 37 and 27 annihilation events change by reversing the direction of e+ spin or e~ spin. The asymmetry A is represented as A anti—para ~ -P (for Pe+Pe- < 1), R e —•para. - 3 — '16 — + where RpaTa is the ratio of 37 to 27 annihilations of e and e~ with parallel spins and Ranti-Para that with the anti-parallel spins. Pe+ and Pe- stand for the polarization of e+ and e~, respectively. We utilize a ferromagnetic iron-foil as a polarized e~ target. To obtain full polarization of 3d-electrons under magnetic field of 100G, the target should be very thin (~ 10 fim). The foil is placed inclined by 20° from the beam direction to cover the full beam size (Fig. 11). The parallel and anti-pallarel spin states are realized by changing the direction of the external magnetic field. Since the ratio of numbers of polarized e~ to that of total e~ is expected to be~ 7% in iron[20], we estimate the analysing power of the target to be 9.3%. However, it is reported that surface electrons of ferromagnetic metals are not polarized due to oxidation[21] and thus the analysing power is decreased to \ when 0.5 monolayer oxygen is attatched at lxlO"11 Torr. In order to avoid this surface effect, we should inject e+ into inside of the target. As shown in Fig.12, 40-electrode plates are located in the vaccum chamber to accelerate e+ up to 20 keV before e+ is incident on the target. To suppress dispersions of the beam, Helmholz magnetic field is applied through the accelarating device. For measurements of 47 and 57 decaying from a positronium, we con- structed the multi-gamma ray spectrometer with 32 modules of Nal(Tl) scin- tillation counters[22]. On the basis of these experiences, we make a spectrom- eter to measure 37 events which consists of 10 Nal(Tl) scintillation counters (diameter of 3 inch and length of 4 inch), each located on a side of a decagon with lead shields and a photomultiplier tube (Fig.13). The front face of each Nal(Tl) crystal is located at a distance of 180 mm from the center of the spectrometer. We obtain the geometrical acceptance G ~ 10~3 for the case that the Nal(Tl) scintillators detect simultaneously three 7-rays. 3 + 5 + Assuming Pe+ = 0.5, G = 10~ , e beam intensity 10 e /s, and £3^/27 = 3^2, it is estimated that one week operation is necessary to determine the P within 10 % statical error. 3.2 Ortho-Positronium quenching in magnetic field The basic idea of the polarization measurement using ortho-positronium quenching in magnetic field is described in Ref.[17],[23]. In the external — -17 magnetic field B, the singlet state of a positronium (para-Ps state) I/J3 and a sub-state of the triplet state (ortho-Ps state) i^ti171 = 0) are perturbed to form two mixed states, while other triplet sub-states, ^(m = ±1) are not perturbed. Hence the mixed states are; (3) V-L -ty- and (4) where y = x/(l + Vl + x*), and x = 0.0276B (for B in kG). These states, $ and •?//, are called an ortho-like-Ps and a para-like-Ps, respectively. The time spectrum of Ps decays in the magnetic field B can be written as ^^^ ^ (5) Here, No is the total number of generated Ps, l/Xt, l/X't, and I/A!, are the life times of ortho-Ps (1/% = 140nsec), ortho-like-Ps, and para-like-Ps, respectively and e = x/y/l + x2. Using the field dependent parameter ?/, 2 2 2 2 we obtain X't = (Xt + y Xs)/(l + y ), and A'a = (As + y Xt)/(l + t/ ), where 1/AS = 0.125nsec is the life time of the non-perturbed pare-Ps. The quantity P in eq.(5) is the e+ polarization, and (j> is the angle between the e+ spin direction and B. Number of the generated ortho-like-Ps, North0-iike{^), can be determined from the shape of the time spectrum. Since Northo-Uke[Q) is related to the polarization P; ^^i£)=i(l-ePcos + the e polarization P can be evaluated by measuring N0Ttho-Hke{fy for the known strength of the magnetic field. The quantity e in eq.(6) is considered to be the analysing power of this method and is presented in Fig. 14 as a function of B. Consequently we obtain the life time of ortho-like-Ps as a function of B as seen in Fig. 15. A schematic view of the slow e+ polarimeter is shown in Fig. 16. An e+ beam with an energy of about 2keV enters along the magnetic field B through a hole in one of the pole pieces of the magnet and strikes the surface of a microchannel plate (MCP) target. The MCP output signal is used as the start pulse for the time measurement. A portion of the incident e+ captures e~ 's at the surface of the MCP and then forms Ps, which, leaving the surface, annihilates into 7 rays. Four modules of Nal(Tl) scintillation counters surrounding the MCP target supply stop signals when they detect 7 rays. Particular features of the magnet design is as follows; • Magnitude of the magnetic flux density can be changed in the range between 2kG and 8kG, which is necessary to reduce systematic errors in the polarization measurement. • Gradient of the magnetic field at the entrance to the MCP target is minimized to supress the depolarization of incident e+ 's. • Inhomogeneity of the magnetic field at the neighbour of the MCP target is less than 1%. The reason why this magnet has holes in both sides of the pole pieces is to achieve better uniformity of the magnetic field and to facilitate to install the MCP. The tapered shapes of the pole pieces are to minimize the gradi- ent of the magnetic field. The behavior of the magnetic field calculated by POISCR [24] is shown in Fig.17, where the origin of the coordinates is the center of the magnet, the z-axis is the direction of the beam line, and the r- axis points radially outword from the center of the beam lines. Fig. 18 shows the systematic error expected in the polarization measurement, caused by the inhomogeneity of the magnetic field at the neighbour of the MCP target. The e+ trajectry and the spin direction in the polarimeter are simulated, as shown in Fig. 19 , by POEM for the magnetic field calculated by POISCR. The simulation reveals that e+ can reach the target with reasonablly low depolarization as seen in Fig. 19 resulting in the possible determination of the polarization with small systematic errors. •19 — 4 Polarized e+ beam using Compton scatter- ing 4.1 Conceptual design of the polarized e+ beam line As the second method to produce a highly polarized e+ beam, we propose a relatively compact facility. Fig. 2 demonstrates the basic idea to employ Compton scattering of an unpolarized e~" beam and circularly polarized laser lights as a 7 source for pair-creation. We are now preparing to produce polarized e+ beam at the Accelerator Test Facility (ATF) for linear colliders in KEK. The pulse-laser is required to keep enough yield of 7-rays. We will use Nd:YAG laser, Continuum , NY81C-10. Its power is 550 mJ in 6ns pulse at a wave length of 532 nm for 2 times harmonic doubler wave, and the maximum repetition rate is lOIIz. The laser light is used as lmm diameter beam. The e" beam of about 2 x 1010 per bunch used for our experiment is supplied from the 1.54 GeV electron linac at ATF (see Fig. 20). Though the maximum repetition rate of ATF electron linac is 25 Hz, we are going to drive it with 10 Hz as a result of limitation for the laser repetition rate. Fig. 21 shows the differential cross section for Compton scattering of unpolarized e~ with an energy of 1.54 GeV and circularly polarized lasers with 532 nm wave length. The energy distribution for the initial laser light with the helicity of —1 and the scattered 7-ray with the helicity of ±1 are respectively shown in the figure with the dashed- and dotted-lines. The polarization of the scattered 7-ray is shown as a function of the 7 energies in Fig.22, where we can see that 7-rays produced at the high-end of the energy spectrum are highly polarized. These polarized 7-rays are utilized to create polarized e+ through pair creation on a thin target. To simplify the problem, we consider a case of incident 7-rays with a fixed energy and a fixed helicity. The differential cross section for the 7 beam with an energy of 40 MeV and a helicity of +1 is shown as a function of the energy of produced e+ in Fig.23. The energy distribution of e+ produced with the helicity of +1/2 and —1/2 are respectively shown by the dashed- and dotted-lines in the figure. Since the cross section is proportional to a square of a charge number Z2 of a nucleus, we present the cross section normalized by Z2 in the figure. The calculation is performed numerically based on the helicity amplitude in the laboratory 50 — flame with Born approximation using HELAS[25] and BASES/SPRING[26], in which the recoiling of the nucleus and the shielding effect of e~ are not taken into account. The pair created e+ has high polarization around the high-end of the energy distribution as shown in Fig.24. When we required for calculation of energy and helicity distributions of e+ created in our experiment, we have to consider that the scattered 7-ray do not have monochromatic energies. We therefore have to average over the energy and the polarization distributions of the incident ( Compton scattered ) 7 beam. Fig. 25 shows the energy distribution of pair created e+ weighted by the energy distributions of incident 7-rays for each helicity state with the helicity of = ± 1/2. The polarization obtained in the similar manner is shown in Fig.26. 4.2 The optimization of target thickness If we choose a thicker foil target for pair-creation, the production rate for e+ per incident photon becomes greater, but it is necessary to consider e+ decays and depolarizations through spin flip as a result of bremsstrahlungs in the target. For this reason, we have to determine the target thickness for pair-creation to optimize the polarization and the intensity of e+ . The calculation is performed based on the probability of various collisions and stopping forces using EGS4 and that of the spin flip using HELAS. In this simulation the e+ converter is assumed to be Tungsten because of the large atomic number Z=74. The production rate of e+ with various polarization values per incident photon is shown in Fig.27 indicating that the production rate of 80% polarized e+ is saturated at 3mm, which should therefore be the best target thickness. When we want to produce higher polarization, it is necessary to take out the e+ 's with higher energy. The lower limitation of the energy for 80% polarized positron is 43MeV as shown in Fig.27. In the experiment designed for ATF, we will produce about 1.4 x 104 e+ per bunch, when we take into account the power and the spot size of the laser, number of e~ per bunch , beam profile and optics of e~ . 51 — 5 Discussion and applications We have been developing high-quality polarized e+ beams on the basis of two new ideas which might promise us lots of interesting applications to wide areas of physics. Concerning the method of (3+ decay, it has been dis- cussed that the beam transport system using the magnetic field might give rise to depolarization because of spiral motion of e+. We therefore examined precisely possible process of the spin precessions in elecrtric and magnetic fields by using the simulation program POEM which we have developed ac- cording to the relativistic formalism of motion and spin precession of charged particles. In particular, we checked carefully magnitudes of depolarization in various parts of our apparatus where the magnetic field changes and con- sequently attributes to the spin precession. Indeed we found by POEM to be able to keep the depolarization below few percent level by adjusting the shape of magnetic fields and by accelerating e+ to increase the ratio of the longitudinal energy to the transverse energy. The second idea on the basis of the laser-Compton scattering leads us to realize relatively simple processes to create polarized e+ as compared with the other methods in which the polarized 7 beam are generated from a few hendred GeV e~ going through a long helical undulator. Our present attempt is certainly the first step to forecast further developments along the lines of our new idea. One of our technical interests is to achieve polarized e+ beams for future linear colliders which will start with the center of mass enegy 300~500 GeV. If we want to apply the present two methods to future linear colliders, we have to surmount severe difficulties in related with the beam intensity and the bunch structure. Certainty, progress of technology in near future will be much faster than what we can now imagine and hence some of present technical difficulties will be gradually solved. Here we propose one of ideas to achieve the polarized e+ beams for a future linear collider such as JLC based on present levels of technology. For the first method using the j8+ decay, 3.0 x 107 slow e+/sec is expected for a moderator with the conversion efficiency of 10~4 as described in Sec. 2.1. By naively extending the calculation, the production of 2.3 x 1015 e+/sec is possible before thermalization in the moderator using a thicker (6mm) target of liquid aluminum and a high current (100mA) proton accelerator with an energy of 40MeV. Such a high current of protons may be achieved by a — 52 proton linac. It is reported that a moderator with the conversion efficiency of about 20% will possibly be available using a foil of silicon (or diamond) crystal sandwiched by electrodes which supply high electric field to extract thermalized e+ [27]. Therefore, if we can collect 50% of e+ emitted from the Al target in 4?r solid angle and create the complex pulse structure, as shown in Fig.3, with about 50% efficiency, intensity of 1014 e+/sec required in JLC can be achieved. An idea of the buncher we propose is illustrated in Fig.2S. The contin- uous beam of polarized e+ is transported to the buncher with the energy of leV. Firstly, the e+ beam is led to a potential gap of V(t) = -0.04 + 0.122 + 0.0036£2, repeated every 6.7msec corresponding to 150Hz (Fig. 29a). Here, the time dependent potential V is in volt and the time t is in msec. Then, the beam is bunched into pulses with the time spread of about 37/^s after transportation of 7620m. Secondly, the pre-bunched beam is accelerated up to 6keV to reduce space charge effect and led to the second potential gap of V{t') = -80 + 64.9*' -f 0.243/'2, where t' is the time in //sec. The cycle of the potential (—18.5/us < t' < 18.5 fis) is synchronized to the e+ pulses of 37,us time spread coming into the second potential gap, repeated with the period of 6.7msec (Fig. 29b). The beam structure of 560ns train is obtained at 7600m downstream from the second potential gap. Then the beam is accel- erated up to 200KeV which is sufficiently large as compared with the energy spread of ±1200V before the acceleration. Finally, the accelerated beam is transported to an RF cavity with the frequency of 179MHz to produce 100 bunch structure in each of the 560ns trains. The efficiency of this buncher is possibly higher than 50%, which is able to satisfy the requirement for the polarized e+ source described above. For the second method using the laser Compton-scattering, hundred CO2 laser modules would be able to create the pulse structure of JLC. As shown in Fig. 3, the JLC beam consists of 150 trains per second, each of which contains 100 bunches in the time interval of 5.6 nsec. Each bunch with a good beam quality to be forcused to the incoming e~ beam within diffrac- tion limit has 10J and 500psec effective pulse specification. Fig. 30 shows a basic idea on the system architecture that parallel lines of pulesed CO2 laser oscillator-amplifiers are operated sequencially at 150Hz and each line is operated with 5.6nsec. time inteval. The pulse width and timing are con- trolled through semiconductors switched by a module of NdrYAG lasers of 500psec.,89MHz specificatin. The proposed laser system is well within the 53 reach by existing technologies of high power laser architecture, although we will have to challenge and solve lots of technical problems i.e. managements of the large system, a damage threshold of the output window and optics etc. The technical details will be described in another publication[2S]. For the efficient e+ production, we need to consider the energy of the e~ beam. If we use the C02 laser described above, the production rate of e+ increases as the e~ beam energy increases and saturates above 5GeV. Adopting the CO2 laser and the 5-GeV e~ beam, we can obtain the e+ po- larization higher than 80% if we select the e+'s with the energy greater than 25MeV. If we assume a same capture section as that of an unpolar- ized e+ source for JLC (efficiency = 40%)[29] and the transfer efficiency of the beam from the pre-damping ring to the interaction points is 50%, in- tensity of 1.5 x 108e~'s per bunch is expected at the interaction point for e~ beam with 5 x 10loe~'s per bunch. This is about two orders of magnitude smaller than the required value for future linear colliders. To realize the polarized e+ beam with the designed intensity, e"~ beam with (9(1012) e~'s per bunch is needed. Our approach for achieving the required intensity is, therefore, to improve or optimize each of the system component. Use of a multiple converter target is one of candidates to be considered. In the field other than high energy physics performed with high energy e+e~ colliders, a polarized e+ beam of low energy will be highly useful, too. In elementary particle physics, it is interesting to produce a large amount of polarized spin-triplet Ps by a polarized e+ beam to study CP invariance in the lepton sector [30], being complementary to studies in high energy e+e~ interactions. In solid state physics, a slow e+ beam injected on a sample will play unique role in study of spin dependence of surface magnetism by measuring the 27/37 decay rate of Ps [31]. In biophysics, there has been one of the biggest questions in long years that amino acids (and ribose), on which terrestrial life is based, show maximal optical activity that all amino acid (ribose) of life are L-type (D-type). Before the parity violation in weak interaction was established, the only possible explanation was that a random fluctuation in chemical and biological evolutions caused the asymmetry. On the other hand, an attractive possibility of causal mechanisms is that the parity non-conservation in the week interaction necessarily leads us to expect the selection of the L-amino acid (the D-ribose) [32]. Here, polarized e+ beam will also be powerful tool to study this possibility [33] [34]. Finally we will briefly mention our schedule of experiments to produce the polarized e+ beam. The various components of the apparatus for the /?+ decay method are now being constructed and tested at Tokyo Metropolitan University. Before summer, the test of all components will be finished and then the devices will be carried to the cyclotron site at SHI Examination k, Inspection, LTD. Since the method of using Lhe ferromagnetic target is, as described in sec. 3.1, subject to backgrounds arising from the decay of oPs created at the surface of the target, we might better start the polarization measurement by the quenting methods in which we can also expect large statistics. Concerning the laser-Compton scatering, the ATF linac will be planned to operate in this autumn. We are now constructing the optics system and the detectors placed at the end of the licnac and will observe the polarization of Compton scattered 7-rays from the end of this year. In the next year, we will produce polarized e+ and successively perform to measure the polarization of e+ by Bhabha scattering using the ferromagnetic target. 6 Acknowledgements We would like to acknowledge Dr. K. Fujii and Dr. K. Hagiwara of KEK, Dr. M. Jimbo of Tokyo Keiei University, and Dr. T. Kon of Seikei University for their suggestions on physics study in future linear colliders with a polarized e+ beam and for their continuous encouragement. We would also express thanks to Dr. S. Okada of Japan Atomic Energy Research Institute to give us helpful information on the simulation program SPG. We thank Dr. S. Terada and Dr. K.H. Tanaka of KEK to give us many suggestions on the design of the ferromagnetic target and on the magnet for oPs quenching. This research is partially supported by a grand in Aid for Scientific Re- search from the Ministry of Education of Japan and a research fund of KEK for cooperative developments. References [1] T. Omori, Y. Kurihara, K. Itoga, Y. Takauchi, M. Yoshioka, N. Nakan- ishi, H. Aoyagi, M. Tsubata, S. Nakamura, T. Baba, M. Mizuta, Proc. of High Energy Accelerator Conference, Humburg 1992, (World Scientific, Singapore, 1992), Vol-I, p.157. [2] H. Aoyagi, H. Horinaka, Y. Kamiya, T. Kato, T. Kosugoh, S. Nakamura, T. Nakanishi, S. Okumi, T. Saka, M. Tawada, M. Tsubata, Phys. Lett. A16T (1992), 415. [3] T. Maruyama, E. L. Garwin, R. Prepost, G. H. Zapalac, J. S. Smith, J. D. Walker, Phys. Rev. Lett. 66 (1991), 2351. [4] Y. Kurihara, T. Omori, Y. Takeuchi, M. Yoshioka, T. Nakanishi, S. Okumi, M. Tawada, K. Togawa, M. Tsubata, T. Baba, M. Mizuta, R.K. Alley, H. Aoyagi, J.E. Clendenin, J.C. Frisch, G.A. Mulhollan, P.J. Saez, D.C. Schultz, H.Tang, K.H. Witte, Jpn. J. Appl. Phys. 34 (1995), 355. [5] K. Flottmann, DESY 95-064. [6] JLC-I, JLC Group, KEK-Report 92-16(1992). [7] V.E. Balakin, Proceedings of ECFA workshop on e+e~ linear collders (LC92), MPI-PhE/93-14 ECFA 93-154 Vol. I, (1993) pp.243, edt. R. Set- tles. [8] T. Weiland, Proceedings of ECFA workshop on e+e~ linear collders (LC92), MPI-PhE/93-14 ECFA 93-154 Vol. I, (1993) pp.121, edt. R. Set- tles. [9] M. Tigner Proceedings of ECFA workshop on e+e~ linear collders (LC92), MPI-PhE/93-14 ECFA 93-154 Vol. I, (1993) pp.227, edt. R. Set- tles. [10] V.E. Balakin, A.A. Mikhailichenko, INP 79-85 (1979). [11] K. Flottmann, DESY 93-161(1993). [12] T.Aruga, S.Takamura, M.Hirose and Y.Itoh, Phys. Rev. B 46 (1992) 14411, G.R.Brandes, K.F.Canter and A.P.Mills, Jr., Phys. Rev. Lett. 61 (1988) 492, W.E.Frieze, D.W.Gidley and K.G.Lynn, Phys. Rev. B 31 (1985) 5628. [13] Y.Ito, M.Hirose, S.Takamura, O.Sueoka, I.Kanazawa, K.Mashiko, A.Ichimiya, Y.Murata, S.Okada, M.Hasegawa and T.Hyodo, Nucl. In- strum. Methods A 305 (1991) 269, T.Akahane, T.Chiba, N.Shiotani, S.Tanigawa, T.Mikado, R.Suzuki, M.Chiwaki, T.Yamazaki and T.Tomimasu, Appl. Phys. A 51 (1990) 146, L.D.Hulett, Jr., T.A.Lewis, D.L.Donohue and S.Pendyala, in: Positron Annihilation, Eds. L.Dorikens-Vanpraet, M.Dorikens and D.Segers (World Scientific, Singapore, 1989) p.586. [14] M.Hirose, M.Washio and K.Takahashi, Appl. Surf. Sci. 85 (1995) 111. [15] S. Okada and H. Kaneko, Applied Surface Science 85, 149 (1995). [16] GEANT, Detector Description and Simulation Tool, CERN Program Library Long Writeup W5013. [17] J.Van House and P.W. Zitzewitz, Phys. Rev. A 29, 96 (1984). [18] T. Kumita, Preprint of TMU HEP/EXP 95-3 [19] V. Bargmann, L. Michel, and V.L. Telegdi, Phys. Rev. Lett. 2, 435 (1959), Classical Electrodynamics, J.D. Jackson, John Wiley & Sons. [20] P.S.Cooper et al, Phys.Rev.Lett. 34 (1975) 1589. [21] D.W.Gidley et a/., Phys.Rev.Lett. 49 (1982) 1779. [22] S. Adachi et al., Phys. Rev. Lett. 65, 2634 (1990), S. Adachi et al, Phys. Rev. A 49, 3201 (1994). [23] G. Gerber et al, Phys. Rev. D 15, 1189(1977), P.W. Zitzewitz et al., Phys. Rev. Lett. 43, 1281(1979). [24] THE CERN-POISSON PROGRAM PACKAGE (POISCR), CERN Program Library Long Writeup T604. [25] M. Murayama, I. Watanabe, K. Hagiwara, KEK Report 91-11. [26] S. Kawabata, Compt. Phys. Commun., 41 (1986) 127. [27] S. Tanigawa et a/., RADIOISOTOPES, 41 (1992) 535 (in Japanese). [28] Proposed method to make a highly polarized e+ beam for future linear colliders Y. Kurihara et aL, to be submitted to J.J.P.S. [29] 'ATF Design and Study Report', KEK-Internal 95-4, ed. F. Hinode et al. (1995). [30] W. Bernreuther, U. Low, J.P. Ma, 0. Nachtmann, Z. Phys. C41, (1988) 143. [31] D.W. Gidley, A.R. Koymen, and T.W. Capehart, Phys. Rev. Lett. 49, 1779 (1982). [32] F. Vester, T.L.V. Ulbricht, and Ii. Krauss, Naturwissenschaften 46, 68 (1959). [33] R.A. Hegstorm, Nature 297, 643 (1982), and references therein. [34] G.W. Gidley, A. Rich, J.V. House, and P.W. Zitzewitz, Nature 297, 639 (1982) , and references therein. — 58 — Proton Polarized e+ (18 MeV) Al Target Moderator (2 mm) (W, 25 fim) Figure 1: Polarized e+ source using f3+ decay ** -w^" ""••«-•"-•«•••••«« -I... ..mJC...m. Polarized Photon *•>„ (MeV order ) Pair Creation^ . ^ Laser (circular polarize) e+ Polarized Positron Figure 2: Polarized e+ source using backward Compton scattering 6.7msec 560nsec 10 +t 10 e s/bunch i 100 bunches/train Figure 3: Time structure of the e* beams of JLC E Al target 27AI(p,n)27Si 18 MeV proton 1 fiA 8.5 G Bq •r> 0 0.5 1 1.5 Depth (mm) Figure 4: Saturated radioactivity of 27Si — 60 — T E E o i Figure 5: Positron profile - 61 — 14000 '- 12000 E- 10000 8000 (a) 6000 4000 2000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Energy Figure 6: Energy distribution of e+ at (a) the production points and (b) the surface of the Al target. 500 - h 400 - 300 - /I 200 - 100 ft IIII i_-i j-t-L i-r-r-P-tii i i 1 i i i i 1 i i 'IIIMIMMI,,,, I ' ' -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 (v/c)*cos(sita) Figure 7: Polarization of e+ along the beam direction at the entrance of the W moderator. 62 — Solenoid Transport Moderator Target CYPRIS Proton Cyclotron Figure 8: A schematic view of the experimental apparatus. — 63 — « 3500 r in O •4-t '55 o C§ -1 -0.75 -0.5 -0.25 0.25 0.5 0.75 (v/c)*cos(sita) Figure 9: Correlation between energy and polarization along the beam di- rection. 1.2 r 1 z - i i i i I i i i i I i i i i I ' i ' i I i i i ' I ' ' i i I i i i i I ' i ' ' I ' i ' 0 10 20 30 40 50 60 70 80 90 100 -4 x 10 Spin (Longitudinal) 0.2 0.1 0 -0.1 1 i i i I i i i i I i i i i i i I i i i _L 0 10 20 30 40 50 60 70 80 90 100 Spin (Transverse) 1.2 1 0.8 0.6 I I 1 I I I I I I I I I I I I I I I I III I I 0 50 100 150 200 250 Spin (Longitudinal) I li d 1 11 0.002 iililiiiSlllllin 0.001 0 -0.001 : , , . . 1 i i , i , I I 1 1 I 1 . 1 , I , . 1 , I , I 0 50 100 150 200 250 Spin (Transverse) Figure 10: Spin motion calculated by POEM in the straight beam line (a) and the bent beam line (b). Longitudinal and transverse components of the spin are shown as functions of the path length in cm. Here, magnitude of the spin is normalized to unity. The curved section is between 100cm and 210cm in the path length. — 65 Solenoid Coil 40 mm y / positron beam ~ 10 mm. 25 mm (a) (b) Figure 11: (a). A schematic view of the ferromagnetic iron-foil target. (b).The foil is inclined at 20° to the positron beam. Figure 12: The vaccum chamber to accelarete the positrons up to 20 keV. Lead shield Ferromagnetic target Figure 13: A cross-sectional view of the spectrometer. Figure 14: Analysing power e as a function of the magnetic flux density B. 67 — 150 o o w c 100 £ 50 0 Figure 15: The relation between the magnetic flux density (B) and the life time of ortho-like-Ps (1/A{). — 68 stop signal TDC PMT- y-ray- solenoide coil- positron beam MCP 700 mm 460 mm Weight: 450 kg Figure 16: A schematic design of the polarimeter. — 69 0.6 R = 3 mm r 0.4 ••- \ 0 0 0.2 - \ i i i i i i i i taaosaeccs m 0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2 0.225 BZ VS Z 0.6 -7 0 mm 0.4 •"•tf""i•"• 0 • . ; 0.2 • i i i i ,,II ,I BZ VS R R Figure 17: Magnetic flux density calculated by POISCR. 0.8 - w w 0.6 —• '- < > 0.4 "-• a 0.2 - , i».... • f -. "ill ' t . . i i i i i i i 10 B(kG) Figure 18: Expected systematic error, due to inhomogeneity of the magnetic field, as a function of the magnetic flux density B. — 70 — (a) Positron trajectory Q) 0.5 0 0.4 3 = 1 °' I"; ' £-0.2 '~ _j; : .9- 0.1 ~ / CO "7" " 1 1 1 1 i i . . 1 I 1 | i i i . • i . , 0 U-.-T i 0 2.5 7.5 10 12.5 15 17.5 20 22.5 25 _. . _ • Z (cm) SpinT ^ ' 0) a> - \ cO.98 ll D - 'Q0.96 - 1 1 1 1 .... , , , , 1111 .... , , , , , , , , , . , , 0 2.5 7.5 10 12.5 15 17.5 20 22.5 25 (c) SpinZ Z(> Figure 19: (a).Trajectory, (b).transverse componet and (c).longitudinal com- ponents of the spin simulated by POEM in the polarimeter. Here, the MCP target is placed at the right end (z=25cm). Longitudinal and transverse components of the e+ kinetic energy are 2keV and 5eV respectively, at the entrance of the polarimeter. — 71 ACCELERATOR TEST FACILITY FOR JLC (KEK JLC) Bunch compressor Water cooling & Air condition facility Wiggler magnet '" •**>mi'" !••••'•« ^^ K=^ Polarized electron source Water cooling & Air condition facility Control room | Damped cavity Wiggler magnet Positron source Choke mode damped structure (RSD) DC power supply lor modulator . 120m. Figure 20: The Accelerator Test Facility (ATP) 1 . , . , 1 . . . . _ :...... 1 .... 14 : Total h = +l h = -l 12 I 10 XI a s ... ^v^ *> >- '—^ .fur1 m - to *»i • • j •a # * • i • . i - " *_ • 1 1 1 1 .... , . . . i . ,-^.H ^ . . • 10 20 30 40 50 60 70 EY [ MeV ] Figure 21: Differential cross section for the Compton scattering of unpolar- ized e" and circularly polarized laser. — 72 — 1 * j ' ' 1 1 j • • * 1 / 0.5 " 7 ; / C ..„.., 3 • • / J/£,, -0.5 — r "z* i: • i ; -1 i..,. 1 ,. , , 10 20 30 50 60 70 80 Br [ MeV ] Figure 22: The polarization of the scattered 7 " h=+l/2 h=-l/2 Total 700 ! • _—— 600 500 - / \ : 400 / S 300 [ "?""* \ • 200 / • # >O 100 i • - . • 1 \ . . 1 . 1 1 ' 1 i 5 10 15 20 25 30 35 40 Ep [MeV] Figure 23: da/dE of 40MeV photons with h = +1. — 73 — ...... • 0.5 — y o •a a • •0.5 — • • : : ; 1 i : • : : ! i : .... -1 | . , . . 10 15 20 25 30 35 40 Ep [MeV] Figure 24: The polarizaation of pair-created e+ 0.012 1 1 1 1 .... ' ' ' ' ! h=+l/2 : h=-l/2 0.010 Total \ 0.008 : \ - B • > \ (2 0.006 : / * :t. \ - g \ 4 "^ u 0.004 r • - • -/" 1 • Js 0.002 |\ 10 20 30 40 50 60 70 80 Ep [MeV] Figure 25: Energy distribution of e+ (h = ±1) weighted by the energy dis- tribution of the incident 7. 1 * ' ' 1 ..., i.... \ \ ^li i 0.5 -" 1 : j : / g ;: : : """:" »•»•••• : : j " : : i i i : : : : | : : : : 1 -0.5 — ! ' : • i .... I,.,.!.... -1 10 20 30 40 50 60 70 80 Ep [MeV] Figure 26: The resultant polarization of 0.07 O 40% Polarity 0.06 • 60% Polarity I i_U X 80% Polarity a o o 0.05 - •g. c o o 0.04 ••CO- ..in.. T3 • n 1 0.03 Pi 0.02 ..o a X | 0.01 •x- 3 4 5 Target Thickness [mm] Figure 27: Production rate of e+ per photon. — 75 — 6.7ms 37//s 560ns 5.6nsX100 DC beam RF Cavity 1st Potential Gap 2nd Potential Gap „ , . , 1 st Accelerator 2nd Accelerator ~6keV -200keV Figure 28: A schematic view of the highly efficient buncher we propose. The path length along the beam line between the first potential gap and the first accelerator is 7620m, while that between the second potential gap and the second accelerator is 1300m. 6.7msec V(t') 6.7msec +1200V OV -0.4V -1200V 37 fi sec 37 fx sec (a) (b) Figure 29: Time dependence of (a) the first potential gap and (b) the second potential gap. — 76 — Nd: YAG 500 ps 89 MHz Pulse Width Control Laser CO 2 100 ns 150 Hz Beam Oscillator Amplifier No. 1 —* —>. Delivery Beam No. 2 Oscillator Amplifier \ I * - >- Delivery Beam Oscillator Amplifier No. 3 - Delivery Beam Oscillator Amplifier No. 4 Delivery Interaction I Point I I I Beam Oscillator Amplifier No.100 —•- I- Delivery Figure 30: A schematic view of the laser system we propose. — 77 — The LASER beams with very long focal depth for Photon-Photon Collider Koji MATSUKADO Department of Physics Hiroshima University Higashi-Hiroshima, 724, Japan Abstract We investigated characteristic features of a quasi Bessel beam for the purpose of the application to Photon-Photon Collider. In our Experiment, the generated Ek'ssel beam kept 21 fim spot size for a distance of greater than 800 mm. 1 Introduction Interaction point LASER Beam High Energy Photon Conversion point Electron Beam oooo Figure 1: Photon-Photon Collider. In Photon-Photon Collider, two real photon beams with a few hundreds GeV energy collide each other. These high energy photons are produced from LASER photons by means of Compton scattering with high energy electron beam from linac, say, JLC [1]. This report describes a status of our investigation of the LASER focusing system for Photon-Photon Collider. In particular we consider the diffraction free beam called Bessel Beam. — 78 — First of all, we review a LASER focusing system based on conventional optical tech- niques [2] to clarify the reason why the diffraction free beam is needed. For the purpose of effective e — 7 conversion with Compton scattering, the condition, * Tlphoton 1. (1) must be satisfied. Where oc is the Compton total cross section that is about 2x10 25cm2 1S in our case. ?ip/i0ion the number of LASER photons per square centimeter defined as, E nphoton - (2) where E is a LASER energy per pulse, A is a cross section of the LASER beam. This condition is satisfied with a parameter, for example, • A = Ifivn. • E=10J/pulse. • Beam spot radius ~ 10//m. In addition, we must consider a focal depth problem: If e — 7 conversion region has a smaller volume, i.e.,a shorter focal depth, the conversion efficiency becomes less. Unfortu- nately, a small beam waist radius inhibits a deep focal depth. Fig.2 shows this situation. Furthermore, making a small beam spot radius requires a short focusing distance(f) or large spot size of incident beam(D), since the beam spot radius is proportional to f/D. This constraint makes the design of optical system for photon-photon collider difficult. All these troubles are caused by a beam divergence, i.e., the diffraction. Can we avoid the diffraction? Figure 2: Focusing. Left: Tight focusing and short focal depth. Right: Soft focusing and deep focal depth. — 79 — 2 Bessel Beam 2.1 The Diffraction free beams. In 1987, J.Durnin published some papers concerning with diffraction free beams [3][4]. Here, we derive the fundamental equation. We solve the Helmholtz equation, in the form, if> (x,y,z;t) = After some calculation, \ dy2 a J\ (f>(x,y;a)= F (u) exp [i (a cos u • x + a sin u • y)] du, (7) Jo where F (u) is an arbitrary function. It is easy to show that eq.(7) is equivalent to (f>(x,y,a) = 27r£anJn(ar), (8) where tit! 2 2 2 an = / F(u)exp(inu)du, r = x + y . (9) JO As our diffraction free beams are expressed by the sum of Bessel functions Jn, they are called the Bessel beams. 2.2 Intuitive interpretation Next, we shall discuss why Bessel Beams are diffraction free or whether they disagree with the uncertainty principle or not. The intensity distribution of Bessel beam is shown in fig.3, and written as 2 I(x,y) = J n(ar). (10) Since Bessel functions approach to the form — 80 — gradually, eq.(10) can be replaced by (12) irar As shown in fig.3, I(x,y) has many ring shaped maxima. Since the circumference of each ring is approximately 2-irr, the integrated intensity over each ring is; / • 1-KT « constant. (13) This fact means that all rings have almost equal number of photons, i.e., Bessel beams have infinite expanse. This fact is the answer to our questions mentioned above. A "real" Bessel beam can not be generated, because it requires the infinite size of optical system and the infinite power, in other words the Bessel function is not square-integrable. However we can make a "quasi Bessel beam" with an optical system with a finite size. Figure 3: The intensity distribution of (Jo) Bessel Beam. 2.3 Making of (Quasi) Bessel beam Hereafter only the Oth order Bessel beam will be considered (Jo beam), because all of Bessel functions (Jn(r), n > 0) except for the Oth order one take value 0 at r=0. In the case of Jo beam, the function F(u) is reduced to an arbitrary constant value. We start from the equation (see eq.(7)), /•2T7 i{;(x,y,z;t) — / duexp [i(acosu • x + asinu • y + f3 • z)]exp(—iut). (14) Jo We define a vector, k1 = (a cos u, a sin u, /?) (15) 81 — as shown in fig.4. Then, eq.(14) is rewritten as ij}(x,y,z;t) = / duexp [i (fc' • r — ojt)]. (16) JO This equation, means the situation that the electro-magnetic fields with wave number vector k' are superimposed from n = 0 to u = 2K on each points as shown in fig.4. If we realize such an electro-magnetic field on infinite space, we get a real Bessel beam. However we can prepare only a finite size optical devices, for example showing in fig.5. Figure 4: The wave number vector of Bessel beam. In this case, a quasi Bessel beam is generated. Quasi Bessel beam has two remarkable features: • There is a maximum distance called the maximum range zmax> beyond which the Bessel like nature disappears. t Within the range of 0 < z < zmax, the amplitude is proportional to z. These features are easily seen from fig.5. 3 Experiments Following the theory mentioned above, we made two kinds of optical devices which gener- ate the Jo beam, and performed some measurements and analysis concerning with them. In this chapter, the results of these experiments will be shown. 3.1 Experimental Setup We show the experimental setup in fig.6. We used the LASER (He-Ne; 633 nm, 10 mW) as a light source. The LASER spot size was fitted to the "Bessel beam generator" through a lens system. Generated Bessel beam was detected by a beam profiler (Photon Inc. model 2180) which adopts a slit-measurement as shown in fig.6. While it has merits: • High S/N value, • Alignment is very easy, - 82 — K Figure 5: The behavior of Quasi Bessel beam. we must take into account of the fact that the measured profile with the detector {M(x)) is different from real one (/(x,t/)), because it is transformed as, M (x) = / / dxdyf(x,y). (17) J J (slit at x) We can not directly solve eq.(17) to obtain f(x,y), we assumed two types of f(x,y), e the theoretical value based on the specification of our Bessel beam generator, and substituted them into eq.(17). The resultant M(x) were compared with the measured M(x). Our theory was examined by using the former function. The parameters as a Bessel function were determined by using the latter. — 83 — LASER He-Ne Beam 633 ran lOmW Expander Ge-Detector Beam Profiler Photon Inc. Aperture Model 2180 Figure 6: Experimental setup. Upper: Setup. Lower: The structure of Beam Scan. 0.135 ^Central , Spot i Radius Figure 7: The definition of central spot radius. — 84 — 3.2 Devices and measurements 3.2.1 Slit with concentric circle pattern or hologram We made a device which has some concentric circles with 0 or 1 transmissivity [5]. Fig.8 helps us to understand the action of this device: the incident plane wave from left side generates various order of diffraction lights as a result of passing the slits. Bessel beam is generated by ±lst order diffracted lights. In the region close to the device (z ~ zmax/2), higher order (\n\ > 2) diffracted lights disturb the field. 1.53 mm 0,36 mm Figure 8: Bessel beam generator: Slit with concentric circles. It was designed to have specifics, • Zmax = 830 mm. • Central spot radius = 108 fim. The "central spot radius" was defined as the radius where j£ (ar) takes the value 0.135. See fig.7. Fig.9 ~ fig. 12 are results of our measurements. 1 ' ' ' ' 1 • i' • • • i • • • • i • ' • •. 20 — i Bessel c A Gaussian Beam S$^%, liBeam 10 — 5 — L - •'••,- 1 , , , ," 200( 400 600 800 1000 h z(mm) Figure 9: The z-dependence of beam intensity (calculated values). Comparing the Bessel beam and Gaussian beam with the same spot radius a = 108 /xm. I 12J •Measured values 7.5 600 800 z(mm) Figure 10: The z-dependence of measured intensity. Comparing the measured values and the theoretical ones based on eq.(17). — 86 — 1 - ' " " ' I ' ' ' * 1 " " 1 ' 1 ' ' ' ' 1 ' '; en 125 xi 11? fi - 100 _r__ -U^ -_ 108 (im s 1 spo t i 75 — ntra l 50 — 25 |- : n = .... 1 .... 1 .. . . 1 . , . . 1 , , soo z(mm) Figure 11: The z-dependence of the central spot radius.In this figure, the value 108 shows the designed value. 0.8 x (mm) Figure 12: The x-dependence of beam intensity measured by using a pin-hole shown in the figure. This plot can be considered as an approximately real profile. — 87 — 3.2.2 Cone Lens Index: n=1.46 (Quartz) 10 mm Figure 13: The cone lens and the action. The action of this device will be clear from fig. 13 [6]. The lens was designed to have the specifics: • Zmax = 2490 mm. • Central spot radius = 21 [ITCI. Fig.14 ~ fig.17 are results of our measurements for the cone lens. 1000 Gaussian beam I • ... 1 ...... I ..- z(mm) Figure 14: The z-dependence of beam intensity. Comparing the Bessel beam and Gaussian beam(calculated) with the same spot radius or = 21 /xm. The intensity values of Bessel beam were deduced from the AJ$(ar) fitting method. — 88 — .... I .....,., z(mm) Figure 15: The z-dependence of beam spot radius. The value 21 fim is designed spot size.The reason why the spot size reduces at sniall z may be due to the imperfection of our lens shape. 0.15 0.2 x (mm) Figure 16: The x-dependence of calculated beam intensity. The difference between the calculated value and measured one in fig.17 in x-direction can be understood by accuracy of cone angle. — 89 — l50 £> ° • • • • 1 • • • • 1 • • • • i • • • •. fc Original 1u C 1250 \ Scanning f| Slit—*; Aperture — • 1 Additional \- '• '\ Slit \ 750 —r\ r \ Fitted I i 1 ^Measured 230 — 0.05 0.1 0.15 x(mm) Figure 17: The x-dependence of measured beam intensity. The results of pin-hole like measurement described in the caption of fig.12. 4 Conclusion We succeeded in generating the quasi Bessel beams. Two types of Bessel beam generators were made and examined. 1. Slit. The results are as follows; • The measured zmax value was coincident with the designed value 830 mm. • Through the region 0 < z < 830 mm, the beam kept its beam spot radius 108 iim. • The profile of generated beam was somewhat off from the ideal Bessel function. 2. Cone lens. The results were as follows; • The central spot radius was kept less than 21 \im through the distance more than 800mm. Though the designed zmax value was 2490mm, the measurements were limited to z < 800 mm by the size of our optical bench. • The Generated beam profiles were well following to the Bessel function. • The behavior of beam profile and central spot radius made differences from the theoretical values. These facts were owing to the imperfection of lens shape. Much still remains to be done; • The study how to apply the Bessel beam to Photon-Photon Collider must be done. 90 • In the present work, Bessel beams were made of the conventional LASER (Gaussian beam) by means of Bessel beam generators. Especially in the Slit (Hologram), the conversion was not much effective. We have to develop more effective device, for example a refraction hologram, etc. References [1] I. F. Ginzburg et al., Nucl. Inst. Meth. 205, 47 (1983). [2] D. L. Borden, D. A. Bauer, D. 0. Caldwell, preprint SLAC-PUB-5715. [3] J. Durnin, J. Opt. Soc. Am. A4, 651 (1987), [4] J. Durnin, J. J. Miceli, Jr. H. Eberly Phys. Rev. Lett. 58 1499 (1987). [5] A. Vasara, J. Turunen, and A. Friberg J. Opt. Soc. Am. A6, 1748 (1989). [6] N. E. Andreev, Y. A. Aristov, L. Y. Polonskii, and L. N. Pyatniskii, Sov. Phys. JETP 73 969 (1991). 91 — © World Scientific Publishing Company AN INTERACTIVE VERSION OF GRACE AND CATALOGUE OF e+e~ INTERACTIONS AS ITS APPLICATION T. Ishilcawa, S. Kawabata and Y. Kurihara National Laboratory for High Energy Physics, Tsukuba-rdty, Ibarahi, SOS Japan GRACE system is an excellent tool for calculating the cross section and for generating event of the elementary process automatically, However it is not always easy for beginners to use. An interactive version of GRACE is being developed so as to be a user friendly system. Since it works exactly in the same environment as PAW, all functions of PAW are availablefor handling any histogram information produced by GRACE. As its application the cross sections of all elementary processes with up to 5-body final states induced by e+e" interaction are going to be calculated and to be summarized as a catalogue. 1. Introduction In order to calculate the cross section of the elementary processes automatically GRACE system was developed in 1987 and is being implemented gradually.1 Basic feature of the most recent version is as follows; (i) Preparation of input file To use GRACE we have to prepare the input file, which specifies the elemen- tary process, namely, by giving the names of particles in the initial and final states, the order of perturbation and the code number of kinematics in the kinematics library used for the numerical calculation. (ii) Generation of Feynman graphs for the process All Feynman graphs for the process are generated by the command gengraph and their information is saved in a file. (iii) Draw the Feynman graphs The procedure gracefig draws all generated Feynman graphs on the X-terminal and allows us to select some graphs and also to make a EPS file to print them on a piece of paper. (iv) Generation of FORTRAN source code and Makefile The numerical value of the differential cross section is calculated by the he- licity amplitude formalism2, whose FORTRAN source code is generated in a suited form for the numerical integration and event generation program f)O 2 Interactive Version of GRACE and its Application BASES/SPRING3 by the command genfort. In addition to this, a Makefile is also generated for the gauge invariance test, the numerical integration, and the event generation, (v) Edit FORTRAN source code (a) To change physical constants from default ones we must edit the source file setmas.f. (b) When there is no suited kinamatics routine for the process in-the kinematics library, we must code it by ourselves. (c) To apply some kinematical cuts for calculating the cross section, we must add them in the subprogram KIMEM. (d) To change the integration parameters of BASES3 the subprogram KINIT must be altered. (e) Either to change the default set of histograms or to add some his- tograms to them, we must modify the subprograms KINIT and KFILL. GRACE system supposes to make the histograms of several fundamental variables as the default, e.g. the integral variables etc. (vi) Compile and link the source program The executables gauge, integ and spring are made by the command make. (vii) Run the gauge invariance test program In GRACE we can calculate the cross section either in unitary gauge or in covariant gauge. By sampling one point in the phase volume, by calculating the numerical value of the differential cross section at the point in both gauges and by testing their mutual consistency, we can make a minimal gauge invari- ance test. Before calculating the cross section it is recommended to run the gauge invariance test invoked by the command gauge. (viii) Numerical integration The command integ invokes the numerical integration of the differential cross section over the phase volume. The result will be printed on the termi- nal. By the redirection like "integ > file name", you can keep it in a file. The probability information is automatically produced and saved in the file bases.data, according to which the event generation is done. A special care should be taken in this step is to watch the numerical stability of the integral, (ix) Event generation Once the numerical integration converges with a good numerical stability, we can generate events by the command spring. The program spring samples a hypercubes in the phase volume of the integral variables according to the probalility information made in the previous step, and sample and test a point in the hypercube. Each time when the program spring accepts a point in the phase volume as an event, the program comtrol returns to the main program mainsp, where the four vectors of final particles are calculated and saved in a file. For this purpose users must modify the main program mainsp. f before the compilation. — 93 — Interactive Version of GRACE and Us Application 3 This works already interactively as you see, but works in the command mode. Some experts like the command mode, but it may be a trouble maker or an unfriendly system for the beginner. It is much better to support both the command and the menu modes. Furthermore, each time to change some parameters users must edit the relevant source program directly. This stutation should be also improved. 2. A New Version of GRACE grc++ In a new version of GRACE grc++ both command mode and menu mode are supported. To construct this system we have decided to utilize the commamd interpreter KUIP4 developed at GERN and to install grc++ in the CERN utilities to realize the same environment as PAW++5. grc++ opens three different windows, i.e. "Executive Window", "Main Browser" and "Graphics 1" as PAW++ does. Giving a command in the input pad of the ex- ecutive window, we can use GRACE in the command mode as described in the previous section. In the object window of the main browser are there many direc- tories corresponding to CERN uitilities, among which GRACE has been installed. Under the directory GRACE, there are two directories eeto2 and General. When we select General, a menu of the procedures of make_diagram, diagram, grc- fort, grcmake, gauge, integration, spring, kinem and setmas are displayed in the object window. (i) make_diagram The procedure make_diagram creates a new menu window, where the input parameters for the elementary process are to be given. When "ok" button in the menu is pressed after giving the parameters, the input file it; created and then all Feynman graphs are generated automatically. The information like the number of Feynman graphs • • • are printed on the transcript pad of the executive window. (ii) diagram A Feynman graph drawer gracefig is invoked. (iii) grcfort The FORTRAN source code- and the Makefile are generated. (iv) Edit FORTRAN source code We are improving the system so that we may have as less chance to edit files directly as possible, but items (c) and (e) are still remained at present. They will be improved in near future. (a) To change the physical parameters like masses, widths, etc., the procedure setmas is prepared, by which we can change these parameters in a menu window. (b) When no suited kinematics routine was found in the library, we have to write it by ourselves. In grc++ all possible kinematics routines for the final states with up to five bodies are going to be prepared. When it is acomplished, there will be less chance for users to write the kinematics 4 Interactive Version of GRACE and its Application routine by themselves. (c) To apply the kinematical cuts for calculating the cross section, we have to modify the kinematics routine KINEM. (d) To change the integration parameters of BASES and to make numer- ical integration, the procedure integration is prepared. (e) To make different histograms from the default ones, we have to change the subprograms KINIT and KFILL. (v) grcmake The procedure grcmake is to compile and link programs for the gauge in- variance test, the numerical integration, and the event generation. (vi) gauge The gauge invariance test is invoked, (vii) integration The procedure integration opens a menu window, where we can change the integration parameters for BASES. By pressing "ok" button, we can start the numerical integration and can see its result on the transcript pad of the executive window, (viii) spring The procedure spring opens a menu window, where the number of generat- ing events is to be given. By pressing "ok" button, we can start the event generation. (ix) setmas In the setmas menu window, the default values for the physical parameters are shown. Those parameters changed in the window are replaced by the new ones. (x) kinem When kinem is selected, three vi windows are created for modifying subpro- grams KINIT, KINEM and KFILL. (xi) eeto2 This is a special procedure, by which we can calculate the center of mass energy dependency of the total cross section of the elementary processes e+e~ —+ xy. We have applied this procedure to obtain the result for the catalogue described in the next section. As far as we use the default setting of grc++, we can calculate the cross section of the elementary processes automatically. Only when we make some diffrerent histograms from the default ones or apply some kinematical cuts to the phase volume, we have to modify the relevant subprograms. Another new feature of grc++ is utilization of PAW++ function to display the histograms and scatter plots accumulated during the numerical integration and the event generation. Since BASES/SPRING has a proper histogram package, accumulation of histograms and scatter plots is done by the proper package. After the termination of the integration or the event generation, the contents of histograms and scatter plots are copied into the HBOOK format. Thus we can display them - 95 - Interactive Version of GRACE and Us Application 5 in the graphics 1 window in terms of PAW++ functions. If we had the NTUPLE function5 in grc++, we could make hitograms and scatter plots of any variables even after the termination of the integration and the event generation. It will be supported in near future. 3. Catalogue of e+e~ interaction As an application of grc++ we are planning to calculate the cross sections of all elementary processes with at most five final state particles in the e+e~ interaction. The values of masses and widths for particles are taken from PDG6. For example, Mz = 91.187 GeV r = 2.49 GeV Mw = 80.22 GeV rzv = 1.942 GeV, where as the width of W boson we took the sum of widths for all known decay channels of W instead of PDG value. In order to see the effect of Higgs mass, the calculation was done for the following three cases: (1) MH = 91.187 GeV Tj•i = 2.49 GeV (2) = 80.22 GeV Tjtj = 1.942 GeV (3) without Higgs. The following calculations have been carried out for each process for each case of Higgs mass: (i) The total cross sections at the center of mass energies from the threshold of final state ( or 180 GeV if the threshold is less than 180 GeV ) to 1 TeV with the energy step of 20 GeV. (ii) The differential cross sections at the center of mass energies of the threshold ( or 180 GeV if the threshold is less than 180 GeV ), 500 GeV, and 1 TeV. The result of each process is described in the following feature: (i) Feynman diagrams in the unitary gauge drawn by gracefig. (ii) Remarks The numbers of graphs both in unitary and covariant gauges, the code number of kinematics routine, the kinematical cuts if necessary, the integration pa- rameters both for the yfs dependence of the cross section and the differential cross sections, and the effect of Higgs mass are tabulated as remarks. (iii) Total cross section The yfs dependence of the total cross section is drawn for each case of Higgs mass. If the Higgs mass effect is negligible small, only for the case (1) the numerical values of the cross section at CMS energy of every 20 GeV are listed in a table. Otherwise, three tables are made. (iv) Differential cross sections The momentum and the polar angle distributions for each final particle are displayed as well as the correlations among them are shown both by the scatter plots and by the lego plots. — 96 — 6 Interactive Version of GRACE and its Application We have already calculated all e+e~ processes with two or three body final states, of which lists are given in table 1. Table 1. Two or three body fined states tt bi uu dd e+e- e+e- -* jj+ fi~ fepe 77 iZ ZZ W+W~ tH ZH HH 667 uu-y dd-y e+e-7 j/ei/e7 777 7Z7 ft {j*^ W+W--/ 7H7 ZHi HH-i ttz bbZ uuZ ddZ e+e-Z e+e- -H. u+ uTZ Ve&eZ v n 2 zzz w+w-z ZHZ HHZ tbW+ udW+ e"PeW+ fl-j/liW+ UH bbH uuH ddH e+e-H H+H~H V&VeH UpVfxH W+W-H HHH These results are being summarized as the catalogue of e+e interaction part I and will be reported in very near future. In addition to this, we will open all source codes, used in the calculation of catalogue, for anyone to use them like a data base or a program library. A cknowledgement s On the way to develop grc++ we have received much encouragement and support from many people. We would like to thank our colleagues in MINAMITATEYA group, particularly, Y. Shimizu, K. Kaneko, K. Kato, H. Tanaka and J. Fujimoto for helpful discussions for getting the system better. The support from CERN CN division was indispensable for this development. We are also grateful to express our sincere gratitude to Professors H. Sugawara, S. Iwata for the encouragement. We are indebted to companies Fujitsu limited, KASUMI Co, Ltd. and SECOM Co Ltd. for their kind supports and understanding our work. This work was supported in part by Ministry of Education, Science and Culture, Japan under Grant-in-Aid for International Scientific Research Program No.03041087 and No.04044158. References 1. T. Ishikawa, T. Kaneko, K.' Kato, S. Kawabata, Y. Shimizu and H. Tanaka, "GRACE manual" KEK Report 92-19, February 1993. 2. H. Tanaka, Comput. Phys. Commun. 58, 153 (1990). 3. S. Kawabata, Comput. Phys. Commun. 41, 127 (1986). S. Kawabata, "A New Version of the Multi-dimensional Integration and Event Gener- ation Package BASES/SPRING" KEK preprint 94-197, February 1995. 4. CERN CN Division, KUIP, CERN Program Library Long Writeup 1102, October 1994. 5. CERN CN Division, PAW++, CERN Program Library Long Writeup Q121, September 1993. 6. Particle Data Group, Phys. Rev., D50 1173 (1994). — 97 — The GRACE system for SUSY processes Masato JlMBO Computer Science Laboratory, Tokyo Management College, Ichikawa, Chiba 272, Japan (e-mail: [email protected]) MlNAMI-TATEYA collaboration Abstract We introduce a new method to treat Majorana fermions on the GRACE system which has been developed for the automatic computation of the matrix elements for the processes of the standard model. In the GRACE system, we already have such particles as Dirac fermions, gauge bosons and scalar bosons within the standard model. On the other hand, in the SUSY models there are Majorana fermions. In the first instance, we have constructed a system for the automatic computation of cross-sections for the processes of the SUSY QED. The system has also been applied to another model including Majorana fermions, the minimal SUSY standard model (MSSM), by the re-definition of the model file. 1 Introduction It has been a promising hypothesis that there exists a symmetry called supersymmetry (SUSY) between bosons and fermions at the unification-energy scale. It, however, is a broken symmetry at the electroweak-energy scale. The relic of SUSY is expected to remain as a rich spectrum of SUSY particles, partners of usual matter fermions, gauge bosons and Higgs scalars, named sfermions, gauginos and higgsinos, respectively [1]. The quest of these SUSY particles has already been one of the most important pursuits to the present high-energy physics [2]. Although such particles have not yet been discov- ered, masses of them are expected to be O(102) GeV [3]. In order to obtain signatures of the SUSY-particle production, electron-positron colliding experiments are preferable because the electroweak interactions are clean and well-known. Thus we hope SUSY particles will be found out at future TeV-region (sub-TeV region) e~e+-colliders such as CLIC, NLC and JLC [4]. For the simulations of the experiments, we have to calculate the cross-sections for the processes with the final 3-body or more. We have already known within the standard model that the calculation of the helicity amplitudes is more advantageous to such a case than that of the traces for the gamma matrices with REDUCE [5, 6]. The program — 98 package CHANEL [7] is one of the utilities for the numerical calculation of the helicity amplitudes. It, however, is also hard work to construct a program with many subroutine calls of CHANEL by hand. Thus we need a more convenient way to carry out such a work. The GRACE system [8], which automatically generates the source code for CHANEL, is one of the solutions. The system also includes the interface and the library of CHANEL, and the multi-dimensional integration and event-generation package BASES/SPRING v5.1 [9]. In the SUSY models, there exist Majorana fermions as the neutral gauginos and hig- gsinos, which become the mixed states called neutralinos. Since anti-particles of Majorana fermions are themselves, there exists so-called 'Majorana-flip', the transition between par- ticle and anti-particle. This has been the most important problem which we should solve when we realize the automatic system for computation of the SUSY processes. In a recent work [10, 11], we developed an algorithm to treat Majorana fermions in CHANEL. In the standard model, we already have such particles as Dirac fermions, gauge bosons and scalar bosons in the GRACE system. Thus we can construct an automatic system for the computation of the SUSY processes by the algorithm above in the GRACE system. In this work, we present the check list of the system at this time, and one of the results. 2 Majorana fermions into new GRACE In Fig. 1, we present the system flow of GRACE [12]. The GRACE system has become more flexible for the extension in the new version called 'grc' [13], which includes a new graph-generation package. With this package, every graphs can be generated based on a user-defined model. It is necessary for us to make the interface and the library of CHANEL and the model file for including the SUSY particles. The method of computation in the program package CHANEL is as follows: 1. To divide a helicity amplitude into vertex amplitudes. 2. To calculate each vertex amplitude numerically as a complex number. 3. To reconstruct of them with the polarization sum, and calculate the helicity ampli- tudes numerically. The merit of this method is that the extension of the package is easy, and that each vertex can be defined only by the type of concerned particles. Here we propose an algorithm [10, 11] for the implementation of the embedding Ma- jorana fermions in CHANEL as follows: 1. To calculate a helicity amplitude numerically. 2. To replace each propagator by wave functions or polarization vectors, and calculate vertex amplitudes. 3. Not to move charge-conjugation matrices. — 99 User input Theory (Lagr angian) Process(particle, order) Particles and interactions Diagram generator Diagram description Drawer Kinematics database ! Feynman diagrams 1 Matrix element generator Function for matrix element (FORTRAN) kinematics generated CHANEL code code library convergence BASES(Monte-Carlo integral) information ... Cross section Distributions SPRING (event generator) Distribution Parameters Specified number of j generated events » Fig. 1. GRACE system flow 100 — • method 1. To choose a direction on a fermion line. 2. To put wave functions, vertices and propagators along the direction in such a way: i) To take the transpose for the reverse direction of fermions ii) To use the propagator with the charge-conjugation matrix for the Majorana-flipped one. As a result, the kinds of the Dirac-Majorana-scalar vertices are limited to four types: (1) UTU (2) t/TrZ7T (3) Z7cTrT!7T (4) U^^C-W where t/'s denote wave functions symbolically without their indices, and C is the charge- conjugation matrix. The symbol T stands for the scalar vertex such as The vertices (2)~(4) are related to the vertex (1) which is the same as the Dirac-Dirac- scalar vertex in the subroutine of CHANEL. Thus we can build three new subroutines for the added vertices. We have performed the installation of the subroutines above with the interface on the new GRACE system. 3 Numerical results At the start for the check of our system, we have written the model file of the SUSY •QED. In this case, there is only one Majorana fermion, photino. Next we have extended the model file and the definition file of couplings for the MSSM. The tests have been performed by the exact calculations with the two methods, our system and REDUCE. In Table I, the tested processes are shown as a list. The references in the table (without [11]) are not the results of the tests, but for help. Here we present the results for the single-selectron production within the SUSY QED process at the energy \/s = 190 GeV. The masses of the concerned particles are M^ = 50 GeV, M-eR = 100 GeV and Mih = 130 GeV in the calculation. This is the case that the pair-production processes occur for both selectrons at the JLC-I energy, but they do not at the LEP-II energy. The Feynman diagrams for this process, which are drawn by the program package 'gracefig' [18] in the new GRACE, are shown in Fig. 2. In Fig. 3, we show the angular distribution of the outgoing positron in the process e~e+ —» %7e+. Here we use BASES for the calculation from the REDUCE output. The result is in beautiful agreement with the value that is obtained by GRACE at each bin — 101 — Graph 3 + + Fig. 2. Feynman diagrams for e e —> eR7e . to-1 c- 10-2 _ lO*-5 — O u to 10-4 _ cos9e Fig. 3. Angular distribution of the positron in e~e+ — The symbol 0 stands for the result from REDUCE. The histogram indicates the result from GRACE. 102- Process Number of diagrams Comment Check Reference SUSY QED e e -> eReR 2 Majorana-flip OK 2 in internal lines OK [11] 2 OK — -f. ~— ~+ Including pair OK [14] c —t "R^R t o annihilation OK [14] 1 Values are OK [14] ^R^L 1 equal OK [14] + e~e —> 77 4 F-B symmetric OK [11] e-e+ _,, -y^ 12 Final 3-body OK [15] e~e+ —> e^7e+ 12 Including every OK [16] element for tests [17] MSSM 8 4 Majorana fermions OK /v X /V1 3 OK Table I. The list of the tested processes. of the histogram. Since the two diagrams with the one-photon exchange dominate in this case, there is a steep peak in the direction of the initial positron. In such a case, the equivalent-photon approximation (EPA) works well [17]. In Fig. 4, we show the 2s distribution of the selectron in the process e~e+ — The quantity Ts is denned in Ref. [17] as where PT denotes the transverse momentum of the selectron. We show also the PT distribution of the selectron in the process e~e+ —> e^je+ in Fig. 5. Here we calculate the two dominant diagrams for comparison with the result from EPA. 4 Summary We introduce a new method to treat Majorana fermions on the GRACE system for the automatic computation of the matrix elements for the processes of the SUSY models. In the first instance, we have constructed the system for the processes of the SUSY QED because we should test our algorithm for the simplest case. The numerical results convince us that our algorithm is correct. It is remarkable that our system is also applicable to another model including Ma- jorana fermions (e.g. the MSSM) once the definition of the model file is given. We have calculated the processes e~e+ —> 777 and e~e+ —> e^7e+ within the SUSY QED. We should calculate the single-photon event from e~e+ —• XiXi7 t^l' an(^ ^e resultant — 103 — 1 ' ' ' I 10" 10-2 i"b •a 10-- 1.2 1.3 1.4 Fig. 4. Ts distribution of the selectron in e e+ - The symbol 0 stands for the result from REDUCE. The histogram indicates the result from GRACE. The curve represents the result from EPA. Iff1-4 _ i i t i I t i * i I « i i i 1 i t 10 30 40 so 60 PT [GeV] Fig. 5. PT distribution of the selectron in e~e+ — The symbol <> stands for the result from REDUCE. The histogram indicates the result from GRACE. The curve represents the result from EPA. — 104 — single-electron (positron) event from the single-selectron production e~e+ —> e^x^e^ [16] as soon as possible. It should be emphasized that the GRACE system including SUSY particles is the powerful tool for this purpose. 5 Acknowledgements This work was supported in part by the Ministry of Education, Science and Culture, Japan under Grant-in-Aid for International Scientific Research Program No.04044158. One of the author (M.J.) and Dr. H. Tanaka have been also indebted to the above-mentioned Ministry under Grant-in-Aid No.06640411. References [1] H.P. Nilles, Phys. Rep. 110 (1984), 1. H.E. Haber and G.L. Kane, Phys. Rep. 117 (1985), 75. M. Chen, C. Dionisi, M. Martinez and X. Tata, Phys. Rep. 159 (1988), 201. R. Barbieri, Riv. Nuovo Cimento 11 (1988). R. Barbieri et al, Z PHYSICS AT LEP 1, CERN Report CERN 89-08 (1989) Vol. 2, p.121. [2] C Dionisi, in Proceedings of XVII International Meeting on Fundamental Physics, PHYSICS AT LEP, Lekeitio, April 23-29, 1989, edited by M.A.-Bemtez and M. Cerrada, (World Scientific, Singapore, 1990), p.71. ALEPH Collaboration, D. Decamp et cil, Phys. Lett. 244B (1990), 541. DELPHI Collaboration, P. Abreu et al, Phys. Lett. 247B (1990), 157. Proceedings of the Joint International Lepton-Photon Symposium & Europhysics Conference on High Energy Physics, Geneva, Switzerland, July 25-August 1, 1991, edited by S. Hegarty, K. Potter and E. Quercigh, (World Scientific, Singapore, 1992). [3] R. Barbieri and G. Giudice, Nucl. Phys. B296 (1988), 75. T. Kon and M. Jimbo, in Proceedings of the First Workshop on Japan Linear Collider (JLC), KEK, October 24-25, 1989, edited by S. Kawabata, KEK Report 90-2 (1990), p.280. M. Jimbo, in Proceedings of the Second Workshop on Japan Linear Collider (JLC), KEK, November 6-8, 1990,. edited by S. Kawabata, KEK Proceedings 91-10 (1991), p.185. [4] Proceedings of the workshop on Physics at Future Accelerators , La Thuile and CERN, January 1987, edited by J.H. Mulvey, CERN Report CERN 87-07 (19S7). C. Ahn et ai, SLAC-Report-329 (1988). Proceedings of the Third Workshop on Japan Linear Collider (JLC), KEK, February 18-20, 1992, edited by A. Miyamoto, KEK Proceedings 92-13 (1992). [5] H. Tanaka, T. Kaneko and Y. Shimizu, Comput. Phys. Commun. 64 (1991), 149. — 105- [6] I. Watanabe, H. Murayama and K. Hagiwara, in Proceedings of the Third Workshop on Japan Linear Collider (JLC), KEK, February 18-20, 1992, edited by A. Miyamoto, KEK Proceedings 92-13 (1992), p.265. [7] H. Tanaka, Comput. Phys. Commun. 58 (1990), 153. [8] T. Kaneko, in New Computing Techniques in Physics Research, edited by D. Perret- Gallix and W. Wojcik, Edition du CNRS, Paris, 1990, p.555. T. Kaneko and H. Tanaka, in Proceedings of the Second Workshop on Japan Linear Collider (JLC), KEK, November 6-8, 1990, edited by S. Kawabata, KEK Proceedings 91-10 (1991), p.250. T. Kaneko, in New Computing Techniques in Physics Research II, edited by D. Perret-Gallix, World Scientific, Singapore, 1992, p.659. T. Ishikawa et ah, Minami-Tateya group, in GRACE manual, KEK Report 92-19 (1993). and References therein. [9] S. Kawabata, Comput Phys. Commun. 41 (1986), 127. S. Kawabata, KEK preprint 94-197 (1995) [To be published Comput. Phys. Com- mun.). [10] M. Jimbo and H. Tanaka, Talk presented at JPS meeting, Fukuoka, March 28-31, 1994. [11] M. Jimbo, H. Tanaka, T. Kaneko and T. Kon, preprint TMCP-95-1 (1995) [To be published in the Proceedings of INS Workshop "Physics ofe+e~, e~7 and 77 collisions at linear accelerators", INS, December 20-22, 1994]. [12] Minami-Tateya collaboration, The document file for GRACE version 1.1, kek/minami/grace/grace.tar.Z at ftp.kek.jp (130.87.34.28), (1994). [13] T. Kaneko, preprint KEK-CP-020 (KEK Preprint 94-83 / MGU-CS/94-01), (1994). [14] M. Jimbo, in Proceedings of the Second Workshop on Japan Linear Collider (JLC), KEK, November 6-8, 1990, edited by S. Kawabata, KEK Proceedings 91-10 (1991), p.185. M. Jimbo, Memoirs of Tokyo Management College I (1993), 101. and References therein. [15] T. Kon, Prog. Theor. Phys. 79 (1988), 1006. and References therein. [16] M. Jimbo, T. Kon and T.Ochiai, Rikkyo University preprint RUP-87-1 (1987). M. Jimbo, Prog. Theor. Phys. 79 (1988), 899. and References therein. [17] M. Jimbo and M. Katuya, Europhys. Lett. 16 (1991), 243. M. Jimbo and M. Katuya, in Proceedings of the KEK Summer Institute on High — IOC- Energy Phenomenology, KEK, August 21-25, 1990, edited by K. Hikasa, KEK Pro- ceedings 91-8 (1991), p.84. and References therein. [18] S, Kawabata, Talk given at this workshop. 107 April 1995 Search for Dynamical Symmetry Breaking Physics by Using Top Quark T.Asaka, N.Maekawa, T.Moroi, Y.Shobuda, and Y.Sumino Department of Physics, Tohoku University Sendai, 980 Japan Abstract We report first results from the investigation on. probing dynamical mech- anism of the electroweak symmetry breaking using top quark. We consider the case where the top quark mass originates from a fermion-antifermion pair condensate, which necessitates (i)strong interaction to cause condensation, and (ii)4-fermi interaction to give top mass. From the observed top mass and the unitarity constraint, we obtain, for the 4-fermi interaction (ii), lower bounds for its strength and upper bounds for its intrinsic new mass scale as we vary the type of the strong interaction (i). Bethe-Salpeter and Schwinger-Dyson equations are solved numerically to study the dynamical symmetry breaking effect semi-quantitatively. — 108 1. Introduction The SU(2) x U(l) gauge theory for describing elcctroweak interactions are very suc- cessful both theoretically and experimentally. However all the experimental tests has been done only for its gauge part. We know that weak bosons and various fermions have their own masses. Because these mass terms are forbidden by gauge symmetry, a mechanism of symmetry breaking is needed, but we have few knowledge about this mechanism so far. Two possibilities have been considered for a symmetry breaking mechanism. In the so- called standard model (SM), symmetry breaking sector consists of an SU(2) gauge doublet of elementary scalar field (Higgs field). By the non-zero vacuum expectation value of this field, gauge symmetry is broken. On the other hand, in models such as technicolor (TC) model [1], new fermions are introduced for its symmetry breaking sector and their condensate breaks symmetry dynamically. Here we consider the latter possibility and intend to search for dynamical symmetry breaking physics by using top quark. Why do we use top quark? Top quark mass is much heavier than other fermions and is of the same order value as the electroweak symmetry breaking scale [2]. When gauge symmetry is dynamically broken by a fermion-antifermion pair, two things are needed to explain fermion masses. First, we need a strong attractive force to form condensate. Second, 4-fermi interaction, which corresponds to the fermion-Higgs Yukawa interaction in the standard model, are needed. Top quark, for example, acquires its mass through a 4-fermi interaction, C = -^HmLtR + h.c. , (1) where M is the new physics scale and \& denotes a new fermion which is introduced in the symmetry breaking sector. From the fermion condensate (XP\I/), top gets mass as This condensate (ty$\ also gives mass to the gauge bosons, because the condensate has a non-zero SU(2) charge. Therefore in the TC-like model, we expect a following inequality: 1/3 < v ~ 0(250 GeV), (3) where v represents the electroweak symmetry breaking scale. The reason we have inequality rather than equality is because there may exist a condensate, which does not give mass to the top quark while giving masses to the gauge bosons, other than this /\&\I/\. From the 109 — observed value of mt a 175 GeV [2], the lower bound for the strength of 4-fermi interaction (~ 1/M2 ) in eq.(l) is obtained via eqs.(2) and (3). We also get the lower bound for the non-standard effect to e.g. Ztt vertex as shown in fig.l. Similar argument holds for other fermions, but the most stringent lower bound tr 4-fermi interaction Figure 1: The Feynman diagram for the non-standard correction from the dynamical sym- metry breaking to the Ztt vertex. for the non-standard gauge-fermion coupling is posed in the case of top quark. This is the reason why we use top quark in searching for a mechanism of dynamical symmetry breaking. To study the dynamical symmetry breaking effect, we solve numerically the Schwinger- Dyson (SD) and Bethe-Salpeter (BS) equations. For the QCD case, taking /„. as the only input parameter, AQCDX^^), Mp,Mai,Ma0,fp, and fai have been calculated by this approach, which meets the experimental value within 20 ~ 30 % accuracy; see refs.[4, 5]. Therefore we expect to calculate the oblique corrections, the Ztt vertex, etc., semi- quantitatively. In this note, we report, as our first results, the numerical estimation of 4-fermi in- teraction which gives rise to the top mass. From the experimental value of mt and the unitarity constraint, we obtain lower bounds for the strength of the 4-fermi interaction and upper bounds for its intrinsic new mass scale by considering various types of the strong interaction which is needed to cause a fermion-antifermion condensate. 2. Numerical Estimation of the Four-Fermi Interaction Now we consider that top quark mass comes dynamically from a fermion-antifermion pair condensate. In this case, a strong interaction which causes the condensate and a 4-fermi interaction which gives a top mass are required. Here we show the theoretical — 110 — estimation of this 4-fermi interaction using the observed top quark mass and the unitarity constraint. Model First, we explain the model we consider here for the symmetry breaking sector. We consider the one-doublet TC-like model in which the SU(NTC) gauge interaction cause fermion condensation. New fermions which are fundamental representation of SU(NTC) are introduced as QL = I „ I , UR, DR, (4) with the weak hypercharge, Y(QL) = 0, Y(UR) = \, Y(DR) = -\ , (5) where these quantum numbers are chosen to cancel the anomaly. JVd sets of above new fermions are introduced in the breaking sector. We also introduce the kernel K which characterizes the strong interaction between \& and *, where ty denotes the above fermion U or D. Now in the case where the SU(NTc) gauge interaction cause the condensation, we can approximately write down the explicit form of the kernel K. In the improved ladder approximation in Landau gauge [3], the kernel K in the momentum space is written as [4] (the momentum configuration is defined as fig.2 ) where C2 is the second Casimir invariant of the SU(N) fundamental representation and the running coupling constant g(p, k) is used. Figure 2: The Feynman diagram for the kernel K(p, k). The helix denotes the techni-gluon propagator. Ill — Once the kernel is fixed as in eq.(6), we can write down the SD eq. and BS eq. which are the self-consistent equations for the full 2-point and 4-point Green functions of $, respectively. In fig.3, these equations are shown diagrammatically. By solving these equa- tions numerically following ref.[4], we obtain the mass function of \&, S(p), and the decay constant, F^. SDeq. BSeq. Figure 3: The graphical representations of the SD eq. and the BS eq. The line with a blob denotes the full propagetor of ty. Top Quark Mass In order to give top quark mass, we also introduce a 4-fermi interaction j^ h.c., (7) where qi denotes the ordinary quark weak doublet and G is the dimension-less coupling. In this 4-fermi interaction, M denotes a new physics scale. Because 4-fermi interaction cannot be a fundamental interaction, there should be a scale M above which this interaction will resolve. Although above this scale there will exist some fundamental theory (such as extended technicolor model[6]) which induce the 4-fermi interaction, we will leave the whole theory and consider only this 4-fermi interaction below the scale M. Now top quark acquires mass through the 4-fermi interaction (7) as shown in fig.4. Using the full propagator of "[/" which is a solution to the SD equation W, top mass can be calculated as [*] In fact, the SD eq. shown in fig.3 is incomplete. Because of the 4-fermi interaction (7), we should consider the mixing of SD eqs. of top quark and "[/", but here we neglect this effect. — 112 — Figure 4: The diagram for the top quark mass. where with x — p% = — p2. Here the new physics scale M is defined as the upper bound of the integral. In order to set the mass scale of the theory, we use the decay constant F* which is obtained by solving the BS eq. of "£/" and "Z?". From the gauge bosons masses, the Fv is normalized as, Mw > ^, (11) where g is the SU(2)L gauge couping constant. Here we have inequality rather than equality, because there may be fermion condensates other than \UU\ and \DD\ , which do not contribute to the top mass while giving masses to the gauge bosons. By substituting the observed top mass mt — 175 GeV [2], we obtain the coupling G for a given M. We consider the case where the strong interaction is given by the SU(2) and SU(3) gauge interaction, and show the G~M contours in figs.5 and 6, respectively. In each figure, the upper region of the lines are allowed because of the inequality in eq.(ll). The behavior of the coupling G can be understood by the momentum dependence of the mass function Y,(x). The E(x) is almost flat for x < A^c, while decreasing rapidly for x 3> A^c as S(z) ~ -(lnx)^"1 , (12) x where B represents the coupling flow and, using the lowest order coefficient of the /? function (/?o), B = /?o/12C2. Now Arc is defined as the scale where the leading-logarithmic running — 113 — coupling constant diverges. Prom this momentum dependence, G behaves as, 2 2 G ~ ~ (for M «A TC), (13) M2 G "onus (fMM2»A->- (14» 2 For M «C Aj-C, the coupling G becomes larger as M —> 0, because the upper bound of the integral is getting smaller. We neglect this region of M for considering that the r 2 new physics scale is at least larger than Arc. F° M > AyC, from the equation (14) we obtain the lower bound of the strength of the 4-fermi interaction in eq.(7) ( ~ G/M2 ). This result agrees with the naive discussion given in the introduction. Moreover, we consider the fixed coupling constant case ( QED like theory or walking technicolor theory [7] ) and show the G-M contour in fig.7. In this case, the asymptotic behavior of £(x) is different from the former case, and is E(z) ~ 4= (x "> oo), (15) and the coupling G increase linearly in the large M region. Therefore the lower bound of the strength is weaker than the former case. Unitarity Constraint Next, we consider the unitarity constraint and put the upper bound for G. First of all, we explain the unitarity bounds which are used here. Let us consider a two-body to two-body scattering (1+2—» 3 + 4 ). We take the helicity of each particle as Ai,A2, A3, and A4. In the CM. frame the partial-wave expansion for the helicity amplitude is, M J where A; = Ai - A2, A/ = A3 - A4, # = 2|pi|/-N/s, /?/ = 2|pf|/v/s and d XuXf(8) is the Wigner's d function, pi(pf) is the common CM. momentum for the initial (final) state and y/s is the CM. energy. The angle (6, |RerJ| < 1 (for an elastic channel) (17) \TJ\ < 1 (for an inelastic channel) (18) —114- Using these unitarity bounds, we put the upper bound of G. We consider the two-body to two-body scattering of fermions through the 4-fermi inter- action (7) at the energy scale E ^> A^c where the effect of confinement can be neglected. Matrix elements of these processes grow as ~ E2 as the scattering energy is increased and would break the above unitarity bounds. Before reaching the energy scale M, the 4-fermi interaction should resolve to retain unitarity. Thus the upper bound for G is obtained by assuming that the matrix elements do not break the unitarity below the scale M. The most stringent bound comes from the scattering of it —> UU process. From the 4-fermi interaction (7), this process occurs in a scalar channel (J=0) only. The helicity amplitude for this process is, ?.., , M\ _!._,_! ^s (19) 2'2'2'2 2' 2' 2' -2 Ml in the massless limit mt — my = 0 and the unitarity bound eq.(18) gives By the above assumption, the upper bound for the coupling G is obtained as G < 8TT. (21) This bound is shown in figs. 5, 6, and 7. From these figures the typical value of the coupling G is found to be of order 10 (G/4TT is of order 1) in all cases, which is a rather strong coupling constant. Note that this unitarity constraint does not hold for M < Arc and is necessary and not sufficient condition for unitarity. Combining the previous estimation of the top mass and this unitarity constraint, we obtain the upper bounds for the intrinsic new mass scale M in eq.(7). For the SU(2) and SU(3) gauge interaction case, the upper bound is around 5 TeV. For the fixed coupling constant case, this bound is much weaker than the former case and is around 15 TeV. SU(2) gauge interaction case G Figure 5: Allowed region in the G-M plane for mt = 175 GeV. The 517(2) gauge interaction is considered as the attractive force to cause the condensation. The lower bounds for G from the top quark mass is shown with curved lines in the case that the number of the set of "£/" and "D", Nd is taken as Nd = 1, 2, and 3. In this case, ATc = 1-7, 1.1, and 0.7 TeV for Nd — 1,2, and 3, respectively. The upper bound for G from the unitarity constraint is also given. From this figure, we can obtain the upper bounds for the scale M. — 116- SU(3) gauge interaction case Figure 6: Allowed region in the G-M plane for mt = 175 GeV. The SU(2>) gauge interaction is considered as the attractive force to cause the condensation. The lower bounds for G from the top quark mass is shown with curved lines in the case that the number of the set of "[/" and "£>", Nd is taken as Nd = 1, 2, and 3. In this case, ATC = 1.3, 0.9, and 0.7 TeV for A^d = 1,2, and 3, respectively. The upper bound for G from the unitarity constraint is also given. From this figure, we can obtain the upper bounds for the scale M. 117 — Fixed coupling constant case Nd = 6 Nd= 9 10 15 M (TeV) Figure 7: Allowed region in the G-M plane for mt = 175 GeV. The interaction of a fixed coupling constant considered as the attractive force to cause the condensation. The lower bounds for G from the top quark mass is shown with curved lines in the case that the number of the set of "[/" and "D", Nd is taken as Nd = 3, 6, and 9. The upper bound for G from the unitarity constraint is also given. From this figure, we can obtain the upper bounds for the scale M. 3. Conclusion and Discussion We have considered the case where the top quark acquires mass from a fermion- antifermion condensate. In this case, (i) a strong interaction to cause condensation and (ii) a 4-fermi interaction to give top mass are needed. We consider two possibilities for the strong interaction. One is the non-Abelian gauge interaction ( SU(2) and SU(3) ) case. The other is the fixed coupling constant case. From the observed top quark mass and the unitarity constraint, we obtain the following results for a 4-fermi interaction in each case. c Lower bound for its strength is obtained. • Coupling G/4TT is typically of order one. • Upper bound for the new physics scale M is obtained: • M < 5 TeV for the non-Abelian gauge interaction case. • M < 15 TeV for the fixed coupling case. The above results will be the bases in studying the dynamical symmetry breaking mech- anism by using top quark. From these results, we intend to calculate the non-standard corrections for the oblique parameters, the Ztt vertex, the Wtb vertex and so on. Solv- ing SD eq. and BS eq. numerically, we can analyze these non-perturbative effects semi- quantitatively. Our goal will be to discuss the detectability of these corrections at NLC and LHC, and probe the dynamical symmetry breaking mechanism by using top quark. 119 — References [1] S. Weinberg : Phys. Rev. D13 (1976) U7;Phys. Rev. D19 (1979) 1277 L. Susskind : Phys. Rev. D20 (1979) 2619 For review, E.Farhi and L. Susskind : Phys. Report 74 (1981) 277 [2] F. Abe et al. ( CDF Collaboration ) : Phys. Rev. Lett. 74 (1995) 2626 S. Abachi et al. ( DO Collaboration ) : Phys. Rev. Lett. 74 (1995) 2632 [3] V.A. Miransky : Sov. J. Nucl. Phys. 38 (1984) 280 K. Higashijima : Phys. Rev. D29 (1984) 1228 [4] K.-I. Aoki, M. Bando, K Hasebe, T.Kugo and H. Nakatani : Prog. Theor. Phys. 82 (1989) 1151 K.-I. Aoki, M. Bando, T. Kugo, M.G. Mitchard and H. Nakatani: Prog. Theor. Phys. 84 (1990) 683 [5] K.-I. Aoki, T. Kugo and M.G. Mitchard : Phys. Lett. 266B (1991) 467 [6] S. Dimopoulos and L. Susskind : Nucl. Phys. B155 (1979) 237 E. Eichten and K. Lane : Phys. Lett. 9OB (1980) 125 [7] B. Holdom : Phys. Lett. 15OB (1985) 301 K. Yamawaki, M. Bando and K. Matumoto : Phys. Rev. Lett. 56 (1986) 1335 T. Akiba and T. Yanagida : Phys. Lett. 169B (1986) 432 T. Appelquist, D. Karabali and L.C.R. Wijewardhana : Phys. Rev. Lett. 57 (1986) 957 — 120 — Status of R&D for the Vertex Detector Y. Sugimoto KEK, National Laboratory for High Energy Physics Tsukuba, Ibaraki, 305 Japan In the experiment at JLC, precise measurement of primary, secondary, and tertiary vertices is very important. For example, 6-quark jet tagging by detecting the decay vertex of B-meson is essential for the search for light higgs particles which decays predominantly into bb pair. For this purpose, the vertex detector will be put very close to the inter- caction point. As a. candidate for the vertex detector, study of charge cupled devices (CCDs) is going on. Because of their pixel structure, CCDs have an advantage of 3-dimensional measurement of tracks of charged particles. A disadvantage of CCDs is the smallness of their signal. At present, they" are used at -90°Cr to get high S/N ratio in SLD experiment at SLAC. To over- come this disadvantage, an R.fcD is going on. The goal of the present R&D is to develop CCD detectors which can be operated at room temperature (i.e. > 0°C) with sufficiently high S/N ratio. The noise of CCDs due to dark current becomes serious as the operating temperature goes up. There are two trivial ways to solve this problem; one is to increase the signal, and the other is to decrease the noise. In order to increase signal, we will use CCD detectors which have thicker depletion thickness than CCDs usually used for video cameras. The noise due to dark current is proportional to \Zhft, where id is the dark current per unit area and t is the integration period. Therefore, the noise can be reduced by adopting quick readout and by using high quality silicon wafers with small id. The small id can also be achieved by adopting a special biasing scheme. So far, we have obtained two types of CCD samples; one from Hamamatsu and one from EEV. The sample from Hamamatsu, type S5466, is originally designed for astronomy. The sensitive area is 12.3 mm square with 512 x 512 121 — I •3 f O 0 • i i • I i i • i I • i . • I . • • i I i • i i I • i • t -20 -10 0 10 20 30 40 Temperature (degree) Figure 1: Output pulse height due to dark current as a function of temper- ature. The integration period is 4 sec. pixels. The temperature dependence of the CCD output due to dark current was measured. The result, is shown in fig. 1. If the gain of 6 j-iV/e given in the data, sheet is assumed, the dark current is ~ 1700 electrons at 4°C. Since the shot noise is the fluctuation of the dark current, this CCD sample gives the noise of \/l700 ~ 40 electrons at 4°C. If the readout period is reduced to few msec (the interval of the JLC beam crossing) instead of 4 sec, the noise will be much smaller. With 20 /mi thick depletion layer, we can expect ~ 1600 electron-hole pair for a MIP track. So, the noise level of this CCD sample is satisfactory to get sufficiently high S/N ratio. The sample from EEV is provided as a CCD camera. The notable char- acteristic of this CCD is its capability of quick readout. Th maximum rate is 5 msec/frame. As far as the S/N ratio is concerned, it looks feasible to operate CCD detectors at room temperature. However, there are still maity items to be considered; radiation hardness, power consumption (cooling), optimization of detector configuration, etc. We formed a CCD R.&D subgroup to carry out these R&D programs. Radiation hardness of CCD detectors will be the main issue of our study in this yea.r. 122 Progress Report of Calorimeter Subgroup JLC calorimeter subgroup KEK, Kobe University, Konan University, and Shinshuu University Presented by Yoshiaki Fujii (KEK) Preface Due to the tragic earthquake which attacked Kobe area just one month ago, I could not have a discussion with collaborators at Kobe/Konan Universities. Some of their works, espacially simulations and silicon pads, are thus not included in this talk. I hope we can report them at a next oppotunity. 1. Introduction Our task comprises two issues: one is a simulation study to set our design goal, and the other is a detector R&D to achieve the design goal. Construction of a full detector simulator, named JIM, was initiated last year. A proto- type was released last August, which includes major detectors and structural components, as well as an event generator. Integration of calorimeter part has been done by I.Nakamura, a Kobe student, and thus the results can not be reported here. For a detector R&D, we set two major targets. One is whether we can achieve re- quested energy resolution or not, and the other is whether we can detect photons inside 2 Tesla magnetic field. After the 4th JLC Workshop in March 1993, we carried out a beam test of calorimeter test modules to investigate energy resolution and other performances. The preliminary results were already reported at the 2nd JLC Detector Workshop [1] in February '94, and at JPS meeting in March '94. The results are briefly reviewed in the following section. In the last summer we carried out tests of photon detectors in magnetic fields up to 2.5 Tesla. The results were reported at JPS meeting in Sep. '94, and were recently submitted for publication. The results are also briefly presented here. Based on these results, R&D plans in JFY '95 will be discussed. 2. Brief Review of Beam Test of Calorimeter Test Modules. One tower of lead/scintillation fiber calorimeter test modules (quote SciFi hereafter), four towers of lead/plastic scintilator sandwich test modules (quote SW hereafter), and a set of silicon pad detector and a preshower detector were constructed in 1993 and in 1994. Their beam test was carried out at KEK in November '94, to study their energy resolutions for electrons and for pions, e/7r separation, and e/7r pulse height ratio (compensation). Measured energy resolution for electrons was in good agreement with EGS simulation. The constant term was ~ 1.2% and was reasonably small. The energy resolution for pions, however, had a large constant term as shown in Fig. 1. This constant term can be interpreted to be due to shower leakage. 123 - - 80 * I i r i i i LINE = 41.9%/VE + 7.7% J 60 b 40 20 0 i i i i 1 i t i i I i i 1 . i 0 0.2 0.4 0.6 0.8 1 1/2 1/VEBEAM [GeV" ] Figure 1: Energy resolution of a SW module for pions. o > o 0 O Electron data O Pion data + Pion data (Leak Corrected by GEANT) 0 , , , , 0 1 BEAM ENERGY [GeV] Figure 2: Beam energy vs. measured energy for electrons and for pions. — 124 — 250 i—r i—i—r Sum of 2 PMTs for Electrons > 200 - O 150 n g o •§ 100 o *J | Module 1 o $ Module 2 § Module 3 £ 50 $ Module 4 0 i I t i i I t 1 I 0 12 3 4 5 BEAM ENERGY [GeV] Figure 3: Number of photoelectrons measured with a SW module for pions. In Fig. 2, measured energy is shown for various beam energies. The measured pion energy is almost equal to the measured electron energy after shower leak correction with GEANT. However data shown in Fig. 2 are not corrected for WLS attenuation, and the results can not be directly interpreted as an indication of compensation. Final results should come soon. Number of photoelectrons detected with PMT's of SW modules are shown in Fig. 3. 160 photoelectrons/GeV does not significantly affect the aimed hadron energy resolution of 40%/Vf. Number of photoelectrons of the preshower module is shown in Fig. 4 for pions. The preshower test module has 6 layers of Imm-thick plastic scintillator plates. Thus for events which did not initiate shower in the preshower detector, average number of photoelectrons are 6 p.e./layer for MIPs. Since the preshower is expected to have 3.4 times better light collection efficiency than the presently postulated EM calorimeter, these values are inter- preted to be 1.8 p.e./layer for MIPs, or 140 p.e./GeV assuming e/mip=0.67, in the case of presently postulated EM calorimeter. Statistical contribution to the aimed EM energy resolution of 15%/y/E is thus not negligible, and improvements of light collection efficiency should be studied. 3. Tests of Photon Detectors in Strong Magnetic Field. In the JLC detector, the calorimeter is presently desinged to be located in 2 Tesla magnetic field. It is popular to locate crystal calorimeters in magnetic fields, since they generate a large number of photons and PIN silicon photodiodes can be used to detect — 125 — 60 - Sum of 4 PMTs for Pions 40 C) O o <> o 20 $ All Pions $ with Pion Cut 0 , , , , I , , • • I , 0 12 3 4 5 BEAM ENERGY [GeV] Figure 4: Number of photoelectrons measured with a preshower detector for pions. them. Sampling calorimeters, however, generate only a small number of photons as shown in the previous section, and high-sensitivity photon detectors operational in magnetic fields are necessary. We have tested three types pf photon detectors in strong magnetic fields: a Fine- mesh Photomultiplier Tube (FMPMT), a Hybrid Photomultiplier Tube (HPMT), and a Vacuum Avalanche Photodiode (VAPD). The tests were carried out at KEK-PS. A normal conducting magnet 8D320 and a superconducting magnet SKS were used for the tests. The setup is schematically shown in Fig. 5. It is well known that FMPMT's are operational in magnetic fields of strength up to about 1 Tesla [2, 3, 4]. It is, however, expected to apply a 2 Tesla magnetic field at JLC, and their operationability must be studied. We tested a Hamarnatsu H261lSXA(24). This is one of the test pieces of equipment aimed at very high gain, with the number of dynode stages being increased to 24. The results are shown in Fig. 6. Though the gain has steep angular dependence, the gain is large enough even at 2.5 Tesla. With careful axis alignment, FMPMT's can be used at JLC environment. HPMT is a new photon-detection device developed by DEP in collaboration with the Canberra, INFN and CERN-LAA projects [5, 6]. It comprises a photocathode and a large- area PIN silicon photpdiode, facing very closely to each other in a vacuum. A high voltage is applied between the photocathode and the photodiode so that photoelectrons are electro- statically accelerated for high gain. We tested a commercially available HPMT (PP350B) [7]. The results are shown in Fig. 7. It has gain of ~1100 and is quite insensitive to the angle. As described in the previous section, number of photoelectrons from calorimeter modules are expected to be ~S0 p.e./GeV/PMT. In the case of HPMT, preamp noise of 126- iMAGNET LED (blue) Frost Glass Rotate B Optical Filters SLIT FMPMT/HPMT/VAPD ,v-k-\ " Figure 5: Schematical view of the test setup. 8 10 1111 i i i 1 El ! 1 ', . , ! . . . . 1 , 1 " ' " = : HV=-2.7kV B=1.5T i 111 1 D 107 _ • D —_ : D B=2.0T \ x "• x x 106 ^ X - o°oO°o X 0 105 _ B=2.5T 0 0 X E o - , , , 1 , , , , 1 , 10' • 1 , . . .9. . ." 0 10 20 30 40 50 Angle [degree] Figure 6: Gain of FMPMT vs. angle between FMPMT axis and magnetic fields. — 127 1250 \ l i HV=-6.0kV 1000 750 500 o 2.5T(SKS) 250 !• 2.0T(SKS) X 2.0T(8D320) 0 1 0 20 40 60 Angle [degree] Figure 7: Gain of HPMT vs. angle between HPMT axis and magnetic fields. ~1000 electrons is expected. This introduces ~12MeV readout noise in the case of ~1100 gain. For precision measurement, noise of a few MeV is preferable, and thus gain of ~4000 is necessary. R&D's for high-gain HPMT's should be done. VAPD is a hybrid photon-detection device similar to HPMT, which uses a large-area avalanche photodiode instead of a PIN silicon photodiode as the photoelectron detector [8, 9]. VAPD is therefore expected to have a much higher gain than that of the HPMT. We tested a commercially available VAPD [10] in magnetic fields up to 1.0 Tesla. For the VAPD, we could not apply a magnetic field greater than 1.0 Tesla, due to some magnetic material inside the device. The results are shown in Fig. 8. 4. Plan in JFY'95. We have made an improved version of a silicon pad detector last year, based on the results of the beam test. Beam tests of this detector will be carried out this summer. As described in the previous section, high-gain HPMT's will be the best solution. R&D's will continue to achieve gain >4000. Number of photons from the EM calorimeter is not very large and may affect the energy resolution. Therefore we should experimentally confirm that we can really achieve the aimed EM energy resolution. Construction of a fine-sampling EM module will be done, and a beam test of it will be carried out with new silicon pad detector. Simulation works are one of the most important issues. In JFY'94, however, we could not work on this so much. In collaboration with JIM group, we plan to complete a calorime- ter part of the detector simulator. 128 — 200000 HV=-13.0kV Bias=2.25kV 150000 B=0.5T 100000 o O 50000 _r i I . , 9 0 20 40 60 Angle [degree] Figure 8: Gain of VAPD vs. angle between VAPD axis and magnetic fields. References [1] K.Ishii et.al., Proceedings of the 2nd JLC Detector Workshop, KEK Proceedings 94-2, April 1994. I.Nakamura et.al., ibid. Y.Fujii et.al., ibid. [2] F. Takasaki, K. Ogawa and K. Tobimatsu, Nucl. Instr. and Meth A228 (1985) 369. [3] H. Kichimi et al., Nucl. Instr. and Meth. A325 (1993) 451. [4] J. Janoth et al., Nucl. Instr. and Meth. A350 (1994) 221. [5] S. Basa et al., Nucl. Instr. and Meth. A330 (1993) 93. [6] H. Arnaudon et al., Nucl. Instr. and Meth. A342 (1994) 558. [7] PP350B technical data, Delft Electronische Producten, Dwazziewegen 2, Roden, Postbus 60, 9300 AB Roden, the Netherlands. [8] P. Cushman and R. Rusak, Nucl. Instr. and Meth. A333 (1993) 381. [9] S. J. Fagen, "Vacuum avalanche photodiodes can count single photons", Laser Focus World, Nov. 1993 (PennWell Pub. Co.). [10] 748-73-75-631 technical data, Advanced Photonix Inc., 1240 Avenida, Acaso, Camar- illo, CA93012, U.S.A. — 129 Electrons 1.54 GeV 10 GeV I S-band S-band Injector Pre-Accelerator 500 GeV X-band Linac 1.54 GeV Electron Damping Ring