Transcript
The MICE Beamline instrumentation for a precise emittance measurement M. Bonesini (on behalf of the MICE Collaboration)
Abstract—The international Muon Ionization Cooling Experiment (MICE) will perform a systematic investigation of ionization cooling of a muon beam. The demonstration comprises one cell of the US Study II neutrino factory cooling channel. As the emittance measurement will be done on a particle-by-particle basis, sophisticated beam instrumentation is needed to measure particle coordinates and timing vs RF. A PID system has been constructed and installed at RAL, in order to keep beam contamination (e, π) well below 1%. The muon beamline has been characterized, obtaining µ+ rates up to ∼ 30 good muons per ISIS spill (for a 1 V · ms beam loss). A preliminary measure of the beam emittance, using a particle-by-particle method with only the TOF detector system, has been developed. Index Terms—Ionization cooling, neutrino beams, neutrino factory, muon collider.
I. I NTRODUCTION The proposed neutrino factory [1], [2] is a muon storage ring with long straight sections, where decaying muons produce collimated neutrino beams of high intensity and well defined composition, with no uncertainties in the spectrum and flux from hadronic production. The physics performance of a neutrino factory depends not only on its clean beam composition (50% νe , 50% ν µ for the µ+ → ν µ νe e+ case), but also on the available beam intensity [3]. The cooling of muons (accounting for ∼ 20% of the final costs) will increase the neutrino factory performance and reduce the muon beam emittance up to a factor 2.4 (as described in reference [4] with a cooling section 75 m long). A neutrino factory will be the most efficient tool to probe the neutrino sector and observe CP violation in leptons. In addition, at the high energy frontier, colliding muon beams may be a valuable option, benefitting from the use of point-like particles and the much higher mass of a muon with respect to that of the electron [5]. The relatively modest muon cooling needs of a neutrino factory might be traded off by using a larger aperture machine, such as a Fixed-Field Alternating Gradient accelerator [6], [7], but no practical muon collider [8] is conceivable without cooling of at least three orders of magnitude. The MICE experiment [9] at RAL aims at a systematic study of a section of the cooling channel of the proposed US Study 2 [10], attaining a 10% effect for a 6π·mm rad beam. The 5.5 m long cooling section consists of three liquid hydrogen absorbers and eight 201 MHz RF cavities encircled by lattice solenoids. As conventional emittance measurement techniques, M. Bonesini is with Sezione INFN Milano Bicocca, Dipartimento di Fisica G. Occhialini, Milano
Fig. 1. View of the MICE experiment at RAL. The muon beam from ISIS enters from the left. The cooling channel is put between two magnetic spectrometers and two TOF stations (TOF1 and TOF2) to measure particle parameters.
based on profile monitors, reach barely a ∼ 10% precision, a novel method based on single particle measurements has been used. Particles are measured before and after the cooling section by two magnetic spectrometers complemented by TOF detectors. For each particle x,y,t,px,py , E coordinates are measured. In this way, for an ensemble of N (∼ 106 ) particles, the input and output emittances may be determined with a precision up to 0.1%, that allows a sensible extrapolation of the results to the full cooling channel. The experiment will be done in several steps, of which the first one (STEP I) is the characterization of the beamline. II. T HE MICE STEPI
BEAMLINE DETECTOR SYSTEM
The secondary muon beam from ISIS (140-240 MeV/c central momentum, tunable between 3 − 10π· mm rad input emittance) enters the MICE cooling section after a Pb diffuser of adjustable thickness (see figure 1 for details). Muons originate from π decay inside a 5 m long SC solenoid upstream of the first PID detectors. Pions are produced dipping a hollow titanium cylindrical target into the ISIS proton beam. The dip depth and timing with respect to the ISIS beam cycle determine the production rate. The MICE target was used during the STEPI data taking in 2009 and 2010 and operated for over 550,000 actuations until summer 2011. The produced secondary particle rate is derived from the ISIS beam loss, as measured by argon filled ionization chambers [11] (BLM) nearby the synchrotron main ring. The integrated signals of
the chambers in ISIS super period SP7 (in V · ms units) is proportional to the secondary particle rate. A sketch of the present MICE beamline, with installed detectors for STEP I, is shown in figure 2.
Fig. 2. Sketch of the present MICE beamline, with installed detectors for STEPI.
The MICE beamline is based on the interplay of the two dipole magnes D1 and D2. A π 7→ µ beam is selected when pD1 ' 2 × pD2 . This beam will be used to demonstrate muon cooling and has a minimal contamination from pions. Alternatively, by putting pD1 ' pD2 all particle species (electrons, muons and pions) are transported and the beam (“calibration beam”) may be used for detector calibration. The driving design criteria for MICE beamline detectors are robustness, in particular of the trackers, to sustain the severe background conditions near the RF cavities and redundancy in PID in order to keep beam contaminations (e, π) well below 1% and reduce systematics on the emittance measurements. PID is obtained upstream of the first tracking solenoid by two TOF stations (TOF0/TOF1) [12] and two threshold Cerenkov counters (CKOVa/CKOVb) [13], that will provide π/µ separation up to 365 MeV/c. Downstream the PID is obtained via an additional TOF station (TOF2) and a calorimeter (EMCAL), to separate muons from decay electrons and undecayed pions. All TOF detectors are used to determine the time coordinate (t) in the measurement of the emittance. To determine the timing with respect to the RF phase to a precision of 50 a detector resolution ∼ 50 ps is needed for TOF0. To allow a better than 99% rejection of pions in most of the incoming muon beam, a resolution ∼ 100 ps for the TOF measurement between TOF0 and TOF1 is needed. All these requirements imply a conservative request of ∼ 50 − 60 ps for single TOF station resolution. All the TOF stations share a common design based on fast 1” scintillator counters (made of Bicron BC404 or BC420) along the x/y directions (to increase measurement redundancy) read at both edges by fast conventional R4998 Hamamatsu photomultipliers 1 . TOF0 planes cover a 40 × 40 cm2 active area and TOF1 and TOF2 cover respectively a 42×42 cm2 and 60 × 60 cm2 active area. The counter width is 4 cm in TOF0 1 1” linear focussed PMTs, typical gain G ∼ 5.7 × 106 at B=0 Gauss, risetime 0.7 ns, TTS ∼ 160ps
and 6 cm in the following ones. The TOF stations must sustain a high instantaneous incoming particle rate (up to 1.5 MHz for TOF0). R4998 PMT rate capabilities were tested in the laboratory with a dedicated setup based on a fast laser 2 . The rate capability was increased by the use of an active divider for PMTs. The PMT signals, after a splitter, are sent to a CAEN V1290 TDC, following a Lecroy 4415 leading edge discriminator, for time measurements and are digitized by a CAEN V1724 FADC 3 to give the pulse height for the time-walk correction. After time-walk corrections and the calibration procedure with impinging beam particles (see reference [12] for details), the TOF detector timing resolution can be measured by using the time difference ∆txy between the vertical and horizontal slabs in the same station (see figure 3). The obtained resolution on the difference is σxy ∼ 100 ps for TOF0 and TOF2, σxy ∼ 120 ps for TOF2 4 . The obtained performances are compatible with requirements. Resolutions are compatible in the TOF0 detector (4 cm wide slabs) and the TOF2 detector (6 cm wide slabs), showing that path length fluctuations effects are negligible. A hint on the intrinsic stability of TOF detectors is shown in figure 4. The beam line Cherenkov detectors will help to provide a cleaner muon beam for the MICE experiment by reducing backgrounds in the time-of-flight (TOF) detector particle identification. The detector consists of two aerogel Cherenkov counters [13], named CKOVa and CKOVb. The aerogels are of high density with indices of refraction na = 1.07 and nb =1.12 respectively. These correspond to densities ρa,b = 0.225 and 0.370 g cm−3 . In figure 5 an exploded view of one Cherenkov detector is shown. The aerogel tiles are two layers thick (2.3 cm) and cover a 46×46 cm2 area. Four 8 inch EMI 9356KB PMTs collect the Cherenkov light in each counter. The expected particle rate in the CKOV detector is so high, that the digitisation of the pulse profile is done using a very high frequency sampling digitiser CAEN V1731 (1 GS/s maximum sampling rate) connected directly to the PMTs through a coaxial cable. The downstream calorimeters (EMCAL), made of two separate calorimeters: KL and EMR, is not intended to be used for energy measurement: its main goal is to provide separation between muons and decay positrons. In addition it should be able to separate muons from undecayed pions. In the MICE calorimeter, EMR determines precisely the muon momentum by range meadurement, while KL acts as an active pre-shower to tag electrons from muon decay. The EMCAL detector consists of a Pb-scintillating fiber calorimeter (KL), of the KLOE type [14], with 1-mm diameter blue scintillating fibers glued between 0.3 mm thick grooved lead plates (see figure 6 for more details) followed by an electron-muon ranger (EMR), made of a ∼ 1m3 fully sensitive segmented scintillator 2 A home-made system based on a Nichia NDHV310APC violet laser diode and an AvetchPulse fast pulser (model AVO-9A-C laser diode driver) was used. This system gave laser pulses at ∼ 409 nm, with a FWHM between ∼ 120 ps and ∼ 3 ns and a max repetition rate of 1 MHz 3 the same ADC is used also for the KL calorimeter readout 4 This translates into ∼ 50(60)ps resolution for the full TOF0/TOF2 (TOF1) detector with crossed horizontal and vertical slabs. The worse resolution of TOF1 is probably due to the poorer quality of the PMTs used.
time resolution
Fig. 3.
Time difference ∆txy between vertical and horizontal slabs in TOF0,TOF1 and TOF2.Trigger is on TOF1.
75
TOF Resolution
TOF0 TOF1 TOF2
70 65 60 55 50 45 40
2500
2550
2600
2650
2700
2750
2800
2850
run number
Fig. 4. Stability of the time resolution of the TOF stations vs run number (the elapsed time is about one month). Trigger is on TOF1.
Fig. 5. Aerogel Cherenkov counter blowup: a) entrance window, b) mirror, c) aerogel mosaic, d) acetate window, e) GORE-TEX reflector panel, f) exit window and g) 8 inch PMT in iron shield.
block. The “spaghetti” design for KL offers the possibility of fine sampling and optimal lateral uniformity. The overall detector dimensions, including magnetic shielding and housing of photomultiplier and voltage dividers, are approximately 120 × 4 × 160 cm3 . The active volume of 93 × 4 × 93 cm3 is divided into 7 modules, which are supported by an iron frame. The iron frame shields the PMTs from magnetic fields. KL has a thickness of 2.5 X0 and ∼ 0.15λint . KL has 21 cells and 42 readout channels (one for each end of a cell). In Figure 7 a schematic view of one exploded KL module is given. The light is collected by Hamamatsu R1355
Fig. 6. Left: schematic layout of KL extruded lead layer and composite with fibers; right: a photograph of a three cells module of KL with Winston cone lightguides.
D
C
B
A
Fig. 7. Exploded view of one KL module, showing the assembly of the fibre lead module, light guides (A), mu-metal shieldings (B), PMTs (C) and voltage dividers (D).
PMTs with voltage dividers E2624-11, providing differential output pulses on twisted pair cables with 120 Ω impedance at 50 MHz. The signal from the PMTs is sent to a shaper module, which shapes and stretches the signal in time in order to match the sampling rate of the flash ADCs. The flash ADC modules are the same 14 bit CAEN V1724 with 8 channels used for the TOF stations. Figure 8 shows the global layout of the KL
assembly.
Muons 80 MeV/c at KL entrance
0.1
150 MeV/c at KL entrance 220 MeV/c at KL entrance 320 MeV/c at KL entrance
0.08 0.06 0.04 0.02 0
500
1000 1500 2000 2500 3000 3500 4000 4500 5000 ADC product
Electrons
Pions
Fig. 8. Global layout of KL assembly (the exploded view shows the various custom made components for support and magnetic shielding).
The expected resolution of KL σE ' 5%/E is fully dominated by sampling fluctuations and is linear for electrons or photons in the range 70-300 MeV. In the upper panel of figure 9 the KL response to muons at different momenta is shown. The deposited energy reaches its minimum (∼ 30 MeV) at 300 MeV/c (minimum ionization in lead). The bottom panel shows instead the KL response to pions. In the plots, the x-axis represents the sum of ADC products from all slabs in KL above a given threshold5. The middle panel of figure 9 shows the typical response of KL to electrons. The fraction of 80 MeV/c electrons which pass through KL is ∼ 70%. In the present setup the KL calorimeter is followed by three 1 inch thick 10 × 100 cm2 scintillator bars, put vertically side by side behind the middle of the detector, to tag any surviving particles (Tag counters) and measure the KL “transparency” for muons. They will be removed, after the final installation of the EMR calorimeter. The fraction of electrons, muons and pions passing through KL is shown in figure 10. KL must act as a pre-sampler for EMR, introducing a minimal perturbation to incoming muons and pions. From figure 10 one can see that the threshold for this is around 140 MeV/c. A raw time estimate may be inferred for incoming particles also for KL, using a simple linear interpolation of the rising edge of the sampling FADC signal. This provides some 5 The ADC product is defined as the product of the digitised signals from the left and right sides of one slab divided by their sum: ADCprod = 2 × ADCleft × ADCright /(ADCleft + ADCright ). The factor of 2 is present for normalisation. The product of the two sides compensates for the effect of light attenuation.
0.09
135 MeV/c at KL entrance
0.08
165 MeV/c at KL entrance
0.07
320 MeV/c at KL entrance
240 MeV/c at KL entrance
0.06 0.05 0.04 0.03 0.02 0.01 0
500
1000 1500 2000 2500 3000 3500 4000 4500 5000 ADC product
Fig. 9. KL response (normalised to the number of events) to muons for various incident momenta (top panel), to pions (bottom panel), with various momenta, and to 80 MeV/c electrons at the entrance of KL (middle panel).
redundancy in the global time measurement. Laboratory tests have shown that a ∼ 210 ps preliminary resolution may be obtained by selecting cosmic muons impinging at the center of a KL cell. The KL response as compared to the TOF one for imcoming particles is shown in figure 11. Clear bands for e, µ, π may be seen. The PID detectors have been installed in steps in the MICE Hall at RAL ( see figure 12) in 2008 and 2009. For the next steps of MICE the presently installed PID detectors will be complemented by the downstream electron muon ranger (EMR) and two trackers inside the spectrometer solenoids, before and after the cooling section. Each tracker [16] is composed of five stations, each consisting of three layers of 350 µm diameter scintillating fibre doublets, read out by Visible Light Photon Counters (VLPCs). The EMR detector (see figure 13) improves the downstream rejection of the µ → e background and is a fully active tracker calorimeter made of 48 planes of extruded triangular scintillator bars (59
Fig. 10. The fraction of electrons, muons and pions passing through KL (“KL transparency”).
N (mm)
Fig. 13. Exploded of the EMR detector. PMTs used for the readout are also shown.
3 6 10
140 39(80) 98(207) 95(183)
pz (MeV/c) 200 240 58(171) 57(237) 112(527) 85(198) 78(200) 89(174)
TABLE I C OLLECTED DATA IN SUMMER 2010. E VENTS ARE BEAM DATA IN PARENTHESIS .
Fig. 11. Time of flight, as measured by TOF detectors, versus KL response as ADC product.
each plane), with Bicron BCF-92 wavelength shifting (WLS) fibre readout. Each plane is read on one side by a single Phillips XP2972 PMT and on the other side by a 64 channel Hamamatsu H7546B multianode PMT. The detector granularity allows individual tracks reconstruction and to measure energy deposition in each bar. III. B EAMLINE C HARATERIZATION The beamline has been characterized by the use of the TOF system, with data taken mainly in summer 2010. Table III shows a summary of the collected data.
P OSITIVE
Figure 14 shows, as an example, the distribution of the timeof-flight between TOF0 and TOF1 or TOF0 and TOF2 for a high emittance muon (π → µ) beam and a low emittance calibration beam. The first peak which is present in both distributions is considered as the time-of-flight of the positrons and is used to determine the absolute value of the time in TOF2. A natural interpretation of the other two peaks is that they are due to forward flying muons from pion decay and pions themselves. Using TOF identification, it was possible to determine the muon rate for the π → µ beam as a function of target dip into the ISIS beam (measured as beam loss in V · ms). This is shown in figure 15. Up to 30 (6) good muons were obtained for the positive (negative) beam for a 1 V ·ms beam loss with a pion background of ∼ 3 − 5% (∼ 1%) as preliminarily estimated by MC simulations. The pion contamination in the beam (especially the positive one) may be reduced with suitable cuts on the time-offlight between TOF0 and TOF1, at the expense of a reduced muon tagging efficiency, or by introducing in the analysis the informations coming from the Cherenkov counters. IV. P RELIMINARY
Fig. 12. Fish-eye view of the present status of the MICE installation. After the last MICE beamline quadruplets Q7-Q8-Q9, in the middle, TOF2 and KL may be seen.
IN KEVT.
MEASURE OF BEAM EMITTANCE WITH TOF DETECTORS
The final measure of emittance will be done in the MICE beam on a particle-by-particle basis with the trackers (to measure x, y, x0 = Px /pz , y 0 = py /pz .E for each particle) and the TOF stations (to measure t). Due to a delay in the delivery of the tracking solenoids, in MICE STEPI, the emittance was
TOF0 -> TOF1
Entries 8599 Mean 27.89 RMS 0.5426
800
TOF0 -> TOF1
8404 29.63 1.782
600
700 600
500
500
400
400
300
300
200
200
100
100 0 20
Entries Mean RMS
22
24
26
28
30
TOF0 -> TOF2
32
34
36
38
40
time of flight [ns] Entries Mean RMS
500
7371 36.59 0.631
0 20
22
24
26
28
30
TOF0 -> TOF2
32
34
36
38
40
time of flight [ns] Entries Mean RMS
450 400
5833 38.7 2.158
350
400
300 300
250 200
200
150 100
100
50 0 30
Fig. 14.
32
34
36
38
40
42
44
46
48
50
time of flight [ns]
Average Muon Rate (tracks per 3.2ms spill)
32
34
36
38
40
42
44
46
48
50
time of flight [ns]
Time of flight between TOF0 and TOF1 (top) and TOF0 and TOF2 (bottom) for the muon (left) and calibration (right) beams.
Muon TOF Track Rate Vs Beam Loss with Cuts 26.2ns < dt < 32ns for 15th June 2010
16
14
12
10
8
6
4
2 0 0
0.5
1
1.5
2
2.5
Sector 7 Integral Beam Loss (V.ms) Muon TOF Track Rate Vs Beam Loss with Cuts 26.2ns < dt < 32ns for 16th June 2010
Average Muon Rate (tracks per 1ms spill)
0 30
50
40
30
20
10
0 0
0.5
1
1.5
2
2.5
3
3.5
Sector 7 Integral Beam Loss (V.ms)
Fig. 15. Average muon TOF track rate per spill as a function of induced ISIS beam loss for a negative π → µ beam, with a 3.2 ms spill gate (top), and for a positive π → µ beam, with a 1 ms spill gate (bottom).
preliminary measured with the TOF system only [17], deriving from them also the x, y, x0 , y 0 information for each particle and measuring pz from the time-of-flight between TOF0 and TOF1. From the precise timing informations from PMTs at the edges of each scintillation counter a transverse impact coordinate (x or y) is reconstructed with a resolution better than 1 cm. With suitable cuts on the time-of-flight between TOF0 and TOF1, a sample of muon is selected with a pion contamination of the order of some per-cent. Once the initial and final x, y particle coordinates are measured, the muon track through the present MICE channel (including the two Cherenkov counters, drift spaces and the last Q7,Q8,Q9 quadruplet of the STEPI MICE beamline) is estimated as (u1 , u01 ) = M(pz ) · (u0 , u00 ), with M(pz ) transfer matrix and u = x or y. The muon momentum pz is initially estimated via the formula: pz s = E ∆t with s track length and ∆t time-of-flight between TOF0 and TOF1. With a separation of ∼ 7.7 m between TOF0 and TOF1, pz may be measured with a resolution σpz = (E 2 /m2 )σt /t. For the the MICE baseline beam, with = 6π mm and pz = 200 MeV/c, this corresponds to about 1%, giving a comparable resolution on the transfer matrix M. An iterative procedure, based on the transfer matrix, is then used to recompute s and pz . This procedure correct a track length bias (∼ 5 MeV/c) on pz reducing it to less than 1 MeV/c. At this point x0 , y 0 are evaluated from the initial and final muon positions. The horizontal and vertical trace space distri-
Fig. 16.
Reconstructed (top) and simulated (bottom) data for the trace plots for the baseline MICE µ − beam = 6π mm and pz = 200 MeV/c.
butions (x, x0 ) and (y, y 0 ) are then plotted and the transverse emittances x and y may be computed, giving an estimation of the normalized emittance via the formula: pz √ N ' x · y m Figure 16 shows the trace plots for the MICE µ− baseline beam, with = 6π mm and pz = 200 MeV/c, for both experimental data and MC simulation. Even if the agreement is not perfect, the beam occupies the desired regions in the trace space. All beams show an RMS beam size of the order of 5-7 cm. V. C ONCLUSIONS The fist step of MICE, corresponding to characterize the incoming muon beam, has been mainly accomplished, by the construction of the muon beamline and the PID detectors. Obtained detector performances are compatible with requirements and a preliminary measure of emittance, with TOF only, has been realized. R EFERENCES [1] Koshkarev, D. G., “Proposal for a decay ring to produce intense secondary particle beams at the SPS”, CERN/ISR-DI/74-62,1974. [2] Geer S., Phys.Rev. D57 (1998) 6989. [3] Bonesini, M. and Guglielmi, A., Phys.Rept. 433 (2006),65. [4] Choubey S. et al, International Design Study for the Neutrino Factory, Interim Design Report, IDS-NF-20,2011. [5] Ankenbrandt et al., Phys.Rev.ST Accel.Beams 2(1999) .081001 [6] Rees, G.H. and Kelliher, D.J., “New High Power Scaling FFAG Driver Ring Designs”, Proceedings 46th ICFA Advanced Beam Dynamics Workshop HB2010,2010, p. 54-56.
[7] Edgecock R., Int.J.Mod.Phys.A26 (2011) 1736-1743 [8] Geer S., “Muon colliders and neutrino factories”, LINAC10 Conference, Tsukuba, Japan, 2010. [9] G. Gregoire et al., MICE Proposal to RAL, 2003. [10] S.Ozaki et al., BNL-52623, June 2001 ; M.M. Alsharo’a et al., Phys. ReV. ST. Accel. Beams 6,081001 (2003); R. Palmer et al., arXiv:0711.4275. [11] Clarke-Gayther, M.A. and Borden, A.I. and Allen, G.M., “Global Beam Loss Monitoring Using Long Ionisation Chambers at ISIS”, Proceedings EPAC 94, 1994, London, p. 1634. [12] R. Bertoni et al., Nucl. Instr. and Meth. A615 (2010) 14. [13] L. Cremaldi et al., IEEE Trans. Nucl. Sci. 56 (2009) 1475. [14] A. Aloisio et al., KLOE coll., Nucl. Inst. and Meth. A 494 (2002) 326. [15] D. Lietti et al., Nucl.Instrum.Meth. A604 (2009) 314. [16] M. Ellis et al., “The design, construction and performance of the MICE scintillating fibre trackers”, to appear on Nucl. Instr. Meth”, arXiv:1005.3491. [17] M.A.Rayner, PhD thesis, Oxford University, 2011.
