- Email: [email protected]
Scintillating fiber trackers with optoelectronic readout for the CHORUS neutrino experiment
NUCLEAR INSTRUMENTS &METHODS IN PHYSICS RESEARCH
Nuclear Instruments and Methods in Physics Research A 344 (1994) 143-148 North-Holland
Section A
Scintillating fiber trackers with optoelectronic readout for the CHORUS neutrino experiment S. Aoki e, J. Dupont d, J. Dupraz d, J.-P. Fabre d, W. Flegel d, D . Frekers g, B. Friend G. Gregoire f, M. Gruwé c, H. Heynitz g, K. Hiller ', K. Höpfner b, J. Konijn V. Lemaitre f, P. Lendermann b, *, R. Meijer Drees d, L. Michel f, C. Mommaert °, M. Nakamura ', P. Nappey d, K. Niwa ', J. Panman d, R. Pintus d, S. Reynaud d, F. Riccardi h, G. Stefanini d , G. Van Beek `, P. Vilain `, J.L. Visschers G. Wilquet K. Winter d, H.T. Wong d
d
d,
d,
NIKHEF, Amsterdam, The Netherlands Humboldt-Universitdt, Berlin, Germany IIHE ULB-VUB, Brussels, Belgium CERN, Geneva, Switzerland e Kobe University, Kobe, Japan f Universiti Catholique de Louvain, Louvain-la-Neuve, Belgium x Westfdhsche Wilhelms-Universadt, Miinster, Germany INFN, Naples, Italy `Nagoya University, Nagoya, Japan Institut für Hochenenergiephysik, Zeuthen, Germany b
d
h
J
The CERN WA95/CHORUS collaboration has been constructing a hybrid detector to search for (v,,-v,)-oscillations . An essential component of the detector is a scintillating fiber tracking system for precise track reconstruction of particles. The physics motivation is explained, and the detector layout, the tracker design, its optoelectronic readout and test beam results are presented.
1. Physics motivation and concept The question whether neutrino flavours mix at some level and the related question whether neutrinos have non-zero mass is one of the remaining great challenges in experimental particle physics [1]. The objective of the CHORUS experiment at CERN is to search for neutrino oscillations in the v, - v. appearance channel in the CERN SPS wideband neutrino beam [2]. The construction and first tests of the detector in the neutrino beam will be completed by the end of 1993, data taking will start in 1994. The search for (vL-v,)-oscillations is based on the detection of the charged current reaction v,N -> -r-X in a background of v,,-induced interactions . The neutrinos from the subsequent decay of the -r - create a missing transverse momentum. Because of the small lifetime of the 3 X 10 -13 s) the use of nuclear T-
(-
* Corresponding author . Elsevier Science B.V. SSDI0168-9002(93)E1062-3
emulsion as a target is essential in order to obtain the spatial resolution necessary to detect the decay vertex . For an event candidate with missing transverse momentum the region to be scanned in the emulsion is determined with a system of scintillating fiber trackers . Changeable emulsion sheets between the fiber arrays and the bulk emulsion blocks serve as a tracker/emulsion interface, they will be replaced every three weeks to keep the beam muon background in the regions to be scanned at an acceptable level . Using computer assisted microscopes one can then find the primary vertex in the bulk emulsion from where the search for the decay kink is finally performed (Fig . 1) . The sensitivity of the CHORUS Experiment is displayed in Fig. 2. In the case (v9-vT)-oscillations would occur close to the present limit for the oscillation parameters, it should be possible to observe around 70 events (and 0.4 background events) in two years beamtime . However, if there are no oscillations, the present limit for sine 20, at large AM2 will be improved by a factor of 16 . 111. TRACKING AND VERTEX DETECTORS
144
S. Aoki et al. /Nucl. Instr. and Meth . in Phys. Res. A 344 (1994) 143-148 Tracking emulsion plate
Scintillation liber tracking array
v
Fig. 1 . Tracking in the target region of the CHORUS detector, the thin emulsion sheets SS, CSI and CS2 serve as a tracker/emulsion interface. 2. The layout of the CHORUS detector The layout of the CHORUS detector is shown in Fig. 3 and the upstream part in Fig. 4. The target region consists of two identical modules each containing two blocks of bulk emulsion and four fiber trackers . The magnet tracker is used for the charge determination of muons and hadrons, which is essential for background reduction. The tracks of the particles traversing the hexagonal-shaped magnet and their momentum are measured with three hexagonal fiber trackers. The emulsion target, the tracker system and ." --vT
90% c I
limits
û
âa
sin' 20,.,
Fig. 2. 90% confidence limits and sensitivity of the CHORUS experiment to (v,,-v,)-oscillations, if there are no oscillations the present limit for sine 20N, at large Am t will be improved by a factor of 16 .
the magnet will be placed in a "cold box" at 5°C, to reduce fading of tracks in the emulsion as well as ageing of the scintillating fibers. Together with the calorimeter and the muon spectrometer, this setup allows an efficient kinematical selection of v,N event candidates and a significant reduction of the number of events to be scanned in the emulsion by one order of magnitude.
3. The scintillating fiber trackers The bottom part of a target module is displayed in Fig. 5. Each of the four trackers consists of two pairs of projection planes, separated by 10 mm of honeycomb support structure. The two projections of each pair are turned by 90 degrees with respect to each other, the second pair is turned with respect to the first one by _+ 120 mrad in order to be able to resolve track ambiguities. Each projection plane consists of seven layers of staggered scintillating plastic fibers (Fig . 6) [3]. The fibers have a length of 220 cm, and a diameter of 500 Wm, the thickness of one projection plane is about 3.2 mm . Each fiber layer is painted with a Ti0 2 based white paint which acts both as a glue and an extramural absorber to avoid crosstalk between fibers. The far end of the ribbons is equipped with an aluminiumsputtered mirror . The design of the optoelectronic readout of the magnet tracker is identical to that of the target tracker, the fibers have the same diameter and they form projection planes with the same thickness . Each of the three hexagonal-shaped trackers consists of six paddles forming two projection planes .
S. Aoki et al. I Nucl. Instr. and Meth . in Phys. Res. A 344 (1994) 143-148
ILTNE I
VETO ST
EHOL40N TARGETS a---) NG---
1
HIGH RESOLTITTON CAIDBDEETER
ST z STREAMER TUBES TI-T2 d TRIGGER SCINTTUATORS Tk - TORODIAL Fe 1EAGNET DRIFT CHAMBERS
HEXAGONAL
F188R TAROSr
xAGNET
145
DC ST
TU
cc DC
TW
DC
TH
DC
TN
DC
TH
DC
TN
DC ST
BEAM
ANTI
COUNTER
DLWONII FDERE TRACRERS SHIE
SPECTROMETER
C
T13WERATURE l STABH.TLED CHA1®ERJ
15rn neutrino detector, consisting of emulsion target and tracker, hexagonal magnet and tracker, calorimeter and muon spectrometer. ti
Fig. 3. Side view of the
CHORUS
4. The optoelectronic readout The optoelectronic readout consists of chains of four image intensifiers [4] followed by a CCD [5] as shown in Fig. 7. Except for the microchannel plate all image intensifiers are of the electrostatically focussing type . The first image intensifier provides a demagnification of the corresponding part of the track picture by a factor of 100/25 . The quantum efficiency of the
Fig. 4. Upstream part of the
photocathode is - 12% . The phosphor crystal PH allows a photon gain of - 10y/-y for the first stage. However, the light emission is quite slow (90% light emitted after - 450 lts), so the number of photons gained within 10 ws is too small to collect enough photons for a full usage of the MCP-efficiency . Therefore a second image intensifier with a gain of - 10y/-y is required between the first stage and the MCP. For a quick transmission of the signal to the MCP it is
CHORUS
neutrino detector . 111 . TRACKING AND VERTEX DETECTORS
146
S . Aoki et al. / Nucl. Instr. and Meth. t o Phys . Res . A 344 (1994) 143-148
Y L
11 w~r T_ Z'
R'Y'
SCIFI TRACKER
SCIFI TRACKER
lA
1H
Y: Vr vv~i
SC[FI TRACKER 2A
Fig. 5. Layout of one of the two identical target modules. 11CY
ELECTROSTATIC
ELECTROSTATIC
TIBER BUNDLE
EIBCCROSTATLC
rr
sT
SCIFI TRACKER 2H
CCD
10025
25
25
25
GATING
Fig. 6. Staggered arrangement of the fibers in a projection plane. equipped with a fast phosphor P46. The MCP itself provides a gain of - 4000y/y before the picture is projected onto the CCD by a fourth image intensifier with a demagnification of 25/11 . The CCD itself con-
.
r `I r r Y r r V V r r r r r r r r 500
y
I
TI V
AM
Fig. 7. Design of the four-stage optoelectronic readout chain .
tains 228 X 550 pixels . Taking the total demagnification into account, the pixel size of 16 wm X 23 wm corresponds to a ratio of - 4 pixels per fiber. The CCD sensor contains an image zone and a memory zone with an antiblooming gate in the image zone to provide a 1 ws fast clear. The transfer time from the image to the memory zone is - 120 [Ls, while the read out, performed during the interburst interval of the neutrino beam, takes about - 20 ms . Therefore the system can read out two events per burst. For the readout of all hodoscopes 58 optoelectronic chains are needed . The coupling of the fiber ribbons to the chains is displayed in Fig. 8 for all planes of one projection including the corresponding planes turned by ± 120 mrad . The endface of each projection plane is subdivided into 20 subsequent substrips, 16 strips are coupled to one optoelectronic chain. The distance of
147
S. Aoki et al. I Nucl. Instr. and Meth . in Phys. Res. A 344 (1994) 143-148 BEAM
z z+ zZZ z+ z
two of these strips on the CCD surface is in the order of the pixel size, enough to overcome picture distortions caused by the optoelectronics and to relate each hit on the CCD surface unambiguously to the corresponding strip.
5. Test results Many tests have been performed in order to optimize the properties of the fibers and the optoelectronic chain. Fig. 9a shows a track picture of a 5 GeV-pion
MM E 28 E _ 26 24 22 2018 -d 16
45
50
55
60
65
70 mm Fig. 9. Track picture: (a) Raw data before application of any cuts, one box corresponds to one pixel, the box size indicates the pixel pulse height; (b) Subcluster coordinates and fitted track after noise subtraction, before the fit a correction shift transverse to the beam direction had to be done for each ribbon separately in order to correct some ribbon misalignment . 111. TRACKING AND VERTEX DETECTORS
148
S . Aoki et al. / Nucl. Instr. and Meth. in Phys . Res. A 344 (1994) 143-148
0.35
transverse direction
0 .3 0.25 0.2 0.15
FWHM = 0 34mm
0 .1 0.05 mm Fig. 10. Spot size of multiplicity-1-clusters . taken with four fiber ribbons equipped with mirror and one optoelectronic chain. Fig. 9b is the same event after noise subtraction with the ribbons aligned and the hits located by the analysis software . The best-fitted trajectory is superimposed. The number of detectable hits per projection plane for tracks traversing the ribbons at the far end (220 cm distance to the readout) is (5 .3 ± 0.3) hits per ribbon . The attenuation length of the light is 750 cm . Comparisons with measurements without mirror give a mirror reflectivity greater than 80%. There are two aspects which define the spatial resolution of the tracking system . The "spot size" is the size of a single-multiplicity-cluster (i .e . a spot on 0.1 0.08 -
mm
the CCD caused by one photoelectron in the first image intensifier stage), it is due to the intrinsic resolution of the optoelectronic readout chain. The measured FWHM is 415 P.m in the readout direction of the CCD and 340 wm in the transverse direction (Fig . 10). The "track residual" is the deviation of the hit from the best-fitted trajectory and the dominant contribution is from the finite size (460 wm inner diameter) of the fiber. The distribution of the track residual is depicted in Fig. 11, showing a FWHM of 350 wm . Adding the full width half maxima for the track residual and the spot size in the readout direction (which will be perpendicular to the dominant track direction) in quadrature gives an estimate of the "two-track resolution" of the CHORUS hodoscopes which is about 540 lLm. Together with the value for the hit density, this is sufficient for a precise determination of the impact point of a certain fraction of tracks on the emulsion surface and consequently a minimum scanning time for one event. References
0.06 FWHM = 0.35mm
0.04 0.02 -
-1 .5
-1
-05
0
05
1
1 .5
mm coordinates with respect to Fig. 11 . Residual of the subduster the fitted track.
[1] K. Winter (ed .), Neutrino Physics (Cambridge University Press, 1991). [2] N. Armenise et al ., CHORUS Proposal, CERNSPSLC/90-42 (1990) . [31 Kuraray SCSF38, Kuraray Co . Ltd., Japan. [4] Image intensifier suppliers: (100/25) electrostatic Hamamatsu Photonics, Japan; (25/25) and (25/11) electrostatic - Delft Electronische Producten (DEP), The Netherlands ; (25/25) MCP - Proxitronic Funk GmbH & Co ., Germany. [5] Thomson TH7864 CCD sensor, Thomson, France .
Nuclear Instruments and Methods in Physics Research A 344 (1994) 143-148 North-Holland
Section A
Scintillating fiber trackers with optoelectronic readout for the CHORUS neutrino experiment S. Aoki e, J. Dupont d, J. Dupraz d, J.-P. Fabre d, W. Flegel d, D . Frekers g, B. Friend G. Gregoire f, M. Gruwé c, H. Heynitz g, K. Hiller ', K. Höpfner b, J. Konijn V. Lemaitre f, P. Lendermann b, *, R. Meijer Drees d, L. Michel f, C. Mommaert °, M. Nakamura ', P. Nappey d, K. Niwa ', J. Panman d, R. Pintus d, S. Reynaud d, F. Riccardi h, G. Stefanini d , G. Van Beek `, P. Vilain `, J.L. Visschers G. Wilquet K. Winter d, H.T. Wong d
d
d,
d,
NIKHEF, Amsterdam, The Netherlands Humboldt-Universitdt, Berlin, Germany IIHE ULB-VUB, Brussels, Belgium CERN, Geneva, Switzerland e Kobe University, Kobe, Japan f Universiti Catholique de Louvain, Louvain-la-Neuve, Belgium x Westfdhsche Wilhelms-Universadt, Miinster, Germany INFN, Naples, Italy `Nagoya University, Nagoya, Japan Institut für Hochenenergiephysik, Zeuthen, Germany b
d
h
J
The CERN WA95/CHORUS collaboration has been constructing a hybrid detector to search for (v,,-v,)-oscillations . An essential component of the detector is a scintillating fiber tracking system for precise track reconstruction of particles. The physics motivation is explained, and the detector layout, the tracker design, its optoelectronic readout and test beam results are presented.
1. Physics motivation and concept The question whether neutrino flavours mix at some level and the related question whether neutrinos have non-zero mass is one of the remaining great challenges in experimental particle physics [1]. The objective of the CHORUS experiment at CERN is to search for neutrino oscillations in the v, - v. appearance channel in the CERN SPS wideband neutrino beam [2]. The construction and first tests of the detector in the neutrino beam will be completed by the end of 1993, data taking will start in 1994. The search for (vL-v,)-oscillations is based on the detection of the charged current reaction v,N -> -r-X in a background of v,,-induced interactions . The neutrinos from the subsequent decay of the -r - create a missing transverse momentum. Because of the small lifetime of the 3 X 10 -13 s) the use of nuclear T-
(-
* Corresponding author . Elsevier Science B.V. SSDI0168-9002(93)E1062-3
emulsion as a target is essential in order to obtain the spatial resolution necessary to detect the decay vertex . For an event candidate with missing transverse momentum the region to be scanned in the emulsion is determined with a system of scintillating fiber trackers . Changeable emulsion sheets between the fiber arrays and the bulk emulsion blocks serve as a tracker/emulsion interface, they will be replaced every three weeks to keep the beam muon background in the regions to be scanned at an acceptable level . Using computer assisted microscopes one can then find the primary vertex in the bulk emulsion from where the search for the decay kink is finally performed (Fig . 1) . The sensitivity of the CHORUS Experiment is displayed in Fig. 2. In the case (v9-vT)-oscillations would occur close to the present limit for the oscillation parameters, it should be possible to observe around 70 events (and 0.4 background events) in two years beamtime . However, if there are no oscillations, the present limit for sine 20, at large AM2 will be improved by a factor of 16 . 111. TRACKING AND VERTEX DETECTORS
144
S. Aoki et al. /Nucl. Instr. and Meth . in Phys. Res. A 344 (1994) 143-148 Tracking emulsion plate
Scintillation liber tracking array
v
Fig. 1 . Tracking in the target region of the CHORUS detector, the thin emulsion sheets SS, CSI and CS2 serve as a tracker/emulsion interface. 2. The layout of the CHORUS detector The layout of the CHORUS detector is shown in Fig. 3 and the upstream part in Fig. 4. The target region consists of two identical modules each containing two blocks of bulk emulsion and four fiber trackers . The magnet tracker is used for the charge determination of muons and hadrons, which is essential for background reduction. The tracks of the particles traversing the hexagonal-shaped magnet and their momentum are measured with three hexagonal fiber trackers. The emulsion target, the tracker system and ." --vT
90% c I
limits
û
âa
sin' 20,.,
Fig. 2. 90% confidence limits and sensitivity of the CHORUS experiment to (v,,-v,)-oscillations, if there are no oscillations the present limit for sine 20N, at large Am t will be improved by a factor of 16 .
the magnet will be placed in a "cold box" at 5°C, to reduce fading of tracks in the emulsion as well as ageing of the scintillating fibers. Together with the calorimeter and the muon spectrometer, this setup allows an efficient kinematical selection of v,N event candidates and a significant reduction of the number of events to be scanned in the emulsion by one order of magnitude.
3. The scintillating fiber trackers The bottom part of a target module is displayed in Fig. 5. Each of the four trackers consists of two pairs of projection planes, separated by 10 mm of honeycomb support structure. The two projections of each pair are turned by 90 degrees with respect to each other, the second pair is turned with respect to the first one by _+ 120 mrad in order to be able to resolve track ambiguities. Each projection plane consists of seven layers of staggered scintillating plastic fibers (Fig . 6) [3]. The fibers have a length of 220 cm, and a diameter of 500 Wm, the thickness of one projection plane is about 3.2 mm . Each fiber layer is painted with a Ti0 2 based white paint which acts both as a glue and an extramural absorber to avoid crosstalk between fibers. The far end of the ribbons is equipped with an aluminiumsputtered mirror . The design of the optoelectronic readout of the magnet tracker is identical to that of the target tracker, the fibers have the same diameter and they form projection planes with the same thickness . Each of the three hexagonal-shaped trackers consists of six paddles forming two projection planes .
S. Aoki et al. I Nucl. Instr. and Meth . in Phys. Res. A 344 (1994) 143-148
ILTNE I
VETO ST
EHOL40N TARGETS a---) NG---
1
HIGH RESOLTITTON CAIDBDEETER
ST z STREAMER TUBES TI-T2 d TRIGGER SCINTTUATORS Tk - TORODIAL Fe 1EAGNET DRIFT CHAMBERS
HEXAGONAL
F188R TAROSr
xAGNET
145
DC ST
TU
cc DC
TW
DC
TH
DC
TN
DC
TH
DC
TN
DC ST
BEAM
ANTI
COUNTER
DLWONII FDERE TRACRERS SHIE
SPECTROMETER
C
T13WERATURE l STABH.TLED CHA1®ERJ
15rn neutrino detector, consisting of emulsion target and tracker, hexagonal magnet and tracker, calorimeter and muon spectrometer. ti
Fig. 3. Side view of the
CHORUS
4. The optoelectronic readout The optoelectronic readout consists of chains of four image intensifiers [4] followed by a CCD [5] as shown in Fig. 7. Except for the microchannel plate all image intensifiers are of the electrostatically focussing type . The first image intensifier provides a demagnification of the corresponding part of the track picture by a factor of 100/25 . The quantum efficiency of the
Fig. 4. Upstream part of the
photocathode is - 12% . The phosphor crystal PH allows a photon gain of - 10y/-y for the first stage. However, the light emission is quite slow (90% light emitted after - 450 lts), so the number of photons gained within 10 ws is too small to collect enough photons for a full usage of the MCP-efficiency . Therefore a second image intensifier with a gain of - 10y/-y is required between the first stage and the MCP. For a quick transmission of the signal to the MCP it is
CHORUS
neutrino detector . 111 . TRACKING AND VERTEX DETECTORS
146
S . Aoki et al. / Nucl. Instr. and Meth. t o Phys . Res . A 344 (1994) 143-148
Y L
11 w~r T_ Z'
R'Y'
SCIFI TRACKER
SCIFI TRACKER
lA
1H
Y: Vr vv~i
SC[FI TRACKER 2A
Fig. 5. Layout of one of the two identical target modules. 11CY
ELECTROSTATIC
ELECTROSTATIC
TIBER BUNDLE
EIBCCROSTATLC
rr
sT
SCIFI TRACKER 2H
CCD
10025
25
25
25
GATING
Fig. 6. Staggered arrangement of the fibers in a projection plane. equipped with a fast phosphor P46. The MCP itself provides a gain of - 4000y/y before the picture is projected onto the CCD by a fourth image intensifier with a demagnification of 25/11 . The CCD itself con-
.
r `I r r Y r r V V r r r r r r r r 500
y
I
TI V
AM
Fig. 7. Design of the four-stage optoelectronic readout chain .
tains 228 X 550 pixels . Taking the total demagnification into account, the pixel size of 16 wm X 23 wm corresponds to a ratio of - 4 pixels per fiber. The CCD sensor contains an image zone and a memory zone with an antiblooming gate in the image zone to provide a 1 ws fast clear. The transfer time from the image to the memory zone is - 120 [Ls, while the read out, performed during the interburst interval of the neutrino beam, takes about - 20 ms . Therefore the system can read out two events per burst. For the readout of all hodoscopes 58 optoelectronic chains are needed . The coupling of the fiber ribbons to the chains is displayed in Fig. 8 for all planes of one projection including the corresponding planes turned by ± 120 mrad . The endface of each projection plane is subdivided into 20 subsequent substrips, 16 strips are coupled to one optoelectronic chain. The distance of
147
S. Aoki et al. I Nucl. Instr. and Meth . in Phys. Res. A 344 (1994) 143-148 BEAM
z z+ zZZ z+ z
two of these strips on the CCD surface is in the order of the pixel size, enough to overcome picture distortions caused by the optoelectronics and to relate each hit on the CCD surface unambiguously to the corresponding strip.
5. Test results Many tests have been performed in order to optimize the properties of the fibers and the optoelectronic chain. Fig. 9a shows a track picture of a 5 GeV-pion
MM E 28 E _ 26 24 22 2018 -d 16
45
50
55
60
65
70 mm Fig. 9. Track picture: (a) Raw data before application of any cuts, one box corresponds to one pixel, the box size indicates the pixel pulse height; (b) Subcluster coordinates and fitted track after noise subtraction, before the fit a correction shift transverse to the beam direction had to be done for each ribbon separately in order to correct some ribbon misalignment . 111. TRACKING AND VERTEX DETECTORS
148
S . Aoki et al. / Nucl. Instr. and Meth. in Phys . Res. A 344 (1994) 143-148
0.35
transverse direction
0 .3 0.25 0.2 0.15
FWHM = 0 34mm
0 .1 0.05 mm Fig. 10. Spot size of multiplicity-1-clusters . taken with four fiber ribbons equipped with mirror and one optoelectronic chain. Fig. 9b is the same event after noise subtraction with the ribbons aligned and the hits located by the analysis software . The best-fitted trajectory is superimposed. The number of detectable hits per projection plane for tracks traversing the ribbons at the far end (220 cm distance to the readout) is (5 .3 ± 0.3) hits per ribbon . The attenuation length of the light is 750 cm . Comparisons with measurements without mirror give a mirror reflectivity greater than 80%. There are two aspects which define the spatial resolution of the tracking system . The "spot size" is the size of a single-multiplicity-cluster (i .e . a spot on 0.1 0.08 -
mm
the CCD caused by one photoelectron in the first image intensifier stage), it is due to the intrinsic resolution of the optoelectronic readout chain. The measured FWHM is 415 P.m in the readout direction of the CCD and 340 wm in the transverse direction (Fig . 10). The "track residual" is the deviation of the hit from the best-fitted trajectory and the dominant contribution is from the finite size (460 wm inner diameter) of the fiber. The distribution of the track residual is depicted in Fig. 11, showing a FWHM of 350 wm . Adding the full width half maxima for the track residual and the spot size in the readout direction (which will be perpendicular to the dominant track direction) in quadrature gives an estimate of the "two-track resolution" of the CHORUS hodoscopes which is about 540 lLm. Together with the value for the hit density, this is sufficient for a precise determination of the impact point of a certain fraction of tracks on the emulsion surface and consequently a minimum scanning time for one event. References
0.06 FWHM = 0.35mm
0.04 0.02 -
-1 .5
-1
-05
0
05
1
1 .5
mm coordinates with respect to Fig. 11 . Residual of the subduster the fitted track.
[1] K. Winter (ed .), Neutrino Physics (Cambridge University Press, 1991). [2] N. Armenise et al ., CHORUS Proposal, CERNSPSLC/90-42 (1990) . [31 Kuraray SCSF38, Kuraray Co . Ltd., Japan. [4] Image intensifier suppliers: (100/25) electrostatic Hamamatsu Photonics, Japan; (25/25) and (25/11) electrostatic - Delft Electronische Producten (DEP), The Netherlands ; (25/25) MCP - Proxitronic Funk GmbH & Co ., Germany. [5] Thomson TH7864 CCD sensor, Thomson, France .