26 Aug 1982 - Typical Single Spark Chamber Gap Termination . . . . . . . Operation of ..... the data processing electronics, the trigger electronics, the PDP ll/T55.
SLAC-255 UC-34D (T/EI
CHARMONIUM SPECTROSCDPY FROM R A D I A T I V E DECAYS OF
T H E J/$ AND
$'
*
J o h nE r t h a G l aiser S t a n f o r dL i n e a rA c c e l e r a t o rC e n t e r S t a n f o r dU n i v e r s i t y S t a n f o r dC, a l i f o r n i a 94305
A u g u s t 1982
P r e p a r e df o rt h eD e p a r t m e n to fE n e r g y u n d e rc o n t r a c tn u m b e r
DE-AC03-76SF00515
P r i n t e d i n t h eU n i t e dS t a t e so fA m e r i c a .A v a i l a b l ef r o mt h eN a t i o n a l T e c h n i c a lI n f o r m a t i o nS e r v i c e , U.S. D e p a r t m e n o tf Commerce, 5 2 5 8 P o r t Royal Road, S p r i n g f i e l dV, i r g i n i a 22161. P r i c eP: r i n t e d Copy A10, M i c r o f i c h e A01.
*
Ph.D.
dissertation.
ABSTRACT The
CrystalBall
NaI(T1)
detectorwasused
to
studythephoton
spectra from the following inclusive charmonium decays: e+e-
+
J/J,
L
y
+ hadrons
DatawerecollectedattheStanfordLinearAccelerator
Center‘s
electron-positron colliding beam machine, SPEAR, intermittently over a tuo and half year period, beginning at the end detector consists of
a highly segmented array
( 1 6 radiation lengths) covering
-98% of
of 1978.
The Cystal Ball
732 NaI(T1)
of
4n steradians, and
crystals is used to
measure the photon energies, directions, and lateral shower distributions.Centrallylocatedsparkandproportionalchambersare is
used for charged particle recognition. As a result, this detector excellently suited for studying the radiative transitions among the charmoniumfamily
i)
of states:
from
$’
to
~ ( 3 4 1 5 1 , ~ ( 3 5 1 0 1 , and ~ ( 3 5 5 0 ) ; ii) from 4”
state n c ( 2 9 8 4 ) ; and i i i ) from
+’to
thetriplet
P states,
and J / J , to the singlet
the radially excited singlet S state
~~‘(3592).
Theanalysis
of
following spectrum and 8,‘(3592), ratios are measured
1.8.106
$’
each with
and
a +4 MeV
to
2.2-106
J/$
x o ( 3 4 1 8 1x,1 [ 3 5 1 2 ) ,
of states:
be:
S
error on
decaysyieldsthe ~ ~ ( 3 5 5 8 1 ,n,(2984),
its mass. The branching
B(+’+YXo,l,z)
=
(9.9?0.5?0.8)%,
(9.020.520.7)%,
(8.0Yl.5?0.7)%,
and
(0.28+0.06)% and B ( J / W v C )
at the
respectively;
(1.27+0.36)%;
--
B($”+Yvc)
= (0.5-1.2)%
and B($’-vvC‘)
90% confidence level. Values for the natural line widths are
obtained:
=
r(xo,1,2)
respectively (90% C.L.); (90% C.L.).
(13.5-20.4)
T ( v C )=
MeV,
from the parent particle’s
as
impact point. Each bump with an
1 0 MeV became a final state particle candidate.
The identification of charged particles with the energy bumps was facilitated by a package of
cylindrical magnetostrictive spark chambers
and multiwire proportional chambers surrounding the beam pipe. Information from the central tracking chambers was used
in the current
analysis to separate neutral candidates from charged for the inclusive photon spectra.
To achieve the basic design philosophy of the detector,
that the apparatus be as closeto a 100% active calorimeter as possible, the materials and construction of the central chambers were selected to minimize their converter thickness. For a more detailed description the experiment than what follows, the reader while detailed a description
is referred to the list o f of
magnetostrictive spark chambers may be found in AppendixB.
-
16
-
of
the central
2.2
EXPERIMENTALLAYOUT TheSPEAR9e+e-collidingbeamfacilityattheStanfordLinear
Accelerator Center is a machine for circulating a single bunch each of electrons and positrons and et bunches are made
in parallel but opposite oval orbits. The to
collide at the two interaction regions,
where a detector can observe the final state and study the process. The SPEAR ring major diameter. The bunch size
e-
the
physics of
has *30 meter minor diameter and e40 meter
is typically 2.6 cm long (uz)
and less
than one millimeter in transverse dimension. The Crystal Ball located in
detector was enclosed
in an environmental dry room
(IR).
the SPEAR east interaction region
adjacent to the east
Located directly
IR was the experiment control room, which housed
the data processing electronics, the trigger electronics, the PDP ll/T55 data acquisition computer, and operator's
console. The detector proper
consisted o f the following elements, as shown in Figures 3 and 4 . 1) 2) 3) 4)
5)
The Central Ball with 672 NaI(T1) crystals. The Central Tracking Chambers Package. The End Caps: tracking chambers and 60 NaI(T1) The Luminosity Monitor. The Outer Hadron-Muon Separator (OHMS).
its optical properties would be
is hygroscopic and
Since the NaI irreparablydamaged
crystals.
if exposedtoevenasmallamount
of
hydration, the central NaI was hermetically sealed
surface
in two hemispherical
containers.Theendcapcrystalswereindividuallysealed
in
walled stainless steel enclosures. As an added precaution the dry room maintainedaconstanttemperature (dew-point
of
-42OC)
of (20?l)*C
andalowhumidity
environment for the detector. Aside from
- 17
-
thin
safeguarding the integrity of the NaI crystals, the dry room added to thestabilityofthecombinedNaIandphotomultipliertube(PMT) temperature sensitive output signals. The data from the
OHMS
part of
thedetectorwerenotusedintheanalysisforthisstudy.Located outside the dry ruom were the
500 KeV proton Van de Graaff accelerator
used for energy calibration of the NaICTl), and the prepulser, pulser, and high voltage power supplies used for the spark chambers. Dataatthetworesonanceswerecollectedover November, 1978 to April, 1981. summarized in Figure 5.
a
periodfrom
The histories of the J/+ and 9’ r u n s a r e
During this period
a wiggler magnet system was
installedatSPEAR.Thissystemwasdesigned
to
current ceiling and thereby the luminosity,
increasethebeam
addition to increasing
in
the synchrotron light available to the Stanford Synchrotron Radiation Laboratory, which was also conducting experiments at SPEAR. Data were taken uith the uiggler magnet both on and
of the
+’resonance yield as
a
off.
A
quantitative analysis
function of the wiggler magnet field
strengthhasalreadybeengivenby
M.
Oreglia.zTable
1
liststhe
typical beam currents and luminosities observed shortly after injection andtheaverageresonanceyieldsforeachrun.Basedonshortterm tests the J/$ yield was found to be best with the wiggler magnet off. qualitative comparison of the condition, as shown
4 ‘’
in T a b l e 1,
yield with the wiggler magnet
indicates an improved yield with the
magnet on.
-
18
-
A
--
I .6 0.8
-0.2q r-
0
600
/ 768 321
400
0 I
I .75
I .6
r-
-
0
I-
0.83
0.79
0.8
- - - - - - - - - - I
0.451 0 -. I7)O. 23 I
I
-3452 2000 I
0 I980
I979
+’
11-82
4416A4
F i g u r e 5 : J/JI and JI’ D a t aA c q u i s i t i o nH i s t o r i e s . The i n t e g r a t e dl u m i n o s i t i e sa n do b s e r v e d numbers o f h a d r o n i cd e c a y sa t t h e J/9 and JI‘ r e s o n a n c e sa r e shown v e r s u st i m e .
-
19
-
TABLE 1 T y p i c a l SPEAR Beam C u r r e n t s a n d L u m i n o s i t i e s a f t e r I n j e c t i o n . A l s os h o u n i s t h e SPEAR w i g g l e rm a g n e ts t a t u s , the t r i g g e rr a t e ,a n dt h e c o r r e c t e dr e s o n a n c ep r o d u c t i o ny i e l d . WIGGLER
RUN
BEAM CURRENT (ma)
LUMINOSITY (cm-zsec-l) ( 1030)
TRIGGER RATE
RESONANCE YIELD R A T E
(Hz)
(Hz)
ofi off off
4.8 5.1
0.48 0.52 1.01 0.31
2.6 2.4
1- 0 8
2.4
0.74
off on on on off on 12.2 (85%) o f f (15%)
8.4 11.8 11.3 11.4
1.2 1.9 1.8 1.8 1 1.4 0.7
2.1
0.52 0.72
CONDITION
1
J/$
J/3 2 3
J/$
+'' 1 3'
2
$'
4
3' 9' 9'
6 6
+'3
2.3
5
4.9
9
7.6
-
3.5 3.2 0.70
0.66 0.66
3.2 2.5 4.0
0.44
2.8
0.45
J H J CENTRAL BALL As shown i n F i g u r e 3,
t h ec e n t r a lb a l l t oc o n s i s t s
o f t w oh e m i s p h e r e s
o f 336 N a I ( T 1 )c r y s t a l se a c h ,p o s i t i o n e do n ea b o v et h eo t h e ra r o u n dt h e SPEAR beam
p i p eD.u r i nngo r m a d la t a k i nt ght ewhoe m i s p h e r easr e
c 1osed t o w i t h i n a b o u t
t h e b a l 1,
2 mm.
Two t u n n erle g i o n s
on
where 48 c r y s t a l sh a v eb e e nr e m o v e da, l l o w
o f t h e beam p i p e .
beam p i peand
f o r t h ei n s e r t i o n
A c e n t r a l ~ 5 0cm d i a m e t ehr o l l o w
dome s u r r o u n d st h e
c e n t r a lt r a c k i n gc h a m b e r s .
To u n d e r s t a n d t h e s p h e r eF . igure
o p p o s i t es i d e so f
b a l l a n dc r y s t agl e o m e t r i e s , *b e g i n
6 shows how t h es u r f a c e
i s d i v i d e di n t oa ni c o s a h e d r o n .
Each o f t h e 20 t r i a n g u l a rf a c e s( m a j o rt r i a n g l e s ) t r i a n g l e( m s i n torri a n g l e s ) .
i n turn
These
i s s u b d i v i d e di n t o a sr eu b d i v i d ei ndt o
c r y s t atl r i a n g l e s T. h e i vr e r t i c e sa r ep r o j e c t e db a c ko n t ot h es u r f a c e o ft h es p h e r et oo b t a i nt h ef i n a lc r y s t a lg e o m e t r y .
-
20
-
with a h o l l o w
4 9
GEOMETRY AND JARGON
20 -
"MAJOR TRIANGLE"
.H
80 -
INDIVIDUAL "MODULES"
720
"TUNNEL REGION"
11-82 4416A55
Figure 6: Construction Geometry for the Central Ball. Note the single layer o f crystals surrounding the tunnel opening.
-
21
-
7,
As shouninFigure
eachcrystalhasthreelongitudinaledges
constructed from radii extending from the interaction point. The crystals are towardsthe cone angle
16
IR.
inches long with triangular cross sections tapering Aninscribed(circumscribed)circlesubtends ( ~ 1 2 ~ with ) respect to the
o f "5.5O
IR.
a full
To ensure proper
internallightpropagationandinter-crystalopticalisolation,each crystal was individually wrapped in reflector paper and aluminized foil (aluminum foil for the bottom hemisphere and aluminized Mylar foil for the top hemisphere). The scintillator light output traveled through e l inchairgap,glasswindow,andanother
a
2 inchairgapbefore
entering a tuo inch diameter P M T . " After the crystals were fabricated, they were individually compensatedforlongitudinallinearity
by sanding the crystal sides.
This was done to control the scintillator light propagation inside the crystal.Thegoalwastoachieve attached
PMT a s
a
I3'Cs
a uniformpulseheightfroman sourcewasmovedlongitudinallyalongthe
crystalside.Thedepthofmaximumshowerdevelopmentforphotons increases with the log of the photon energy. Although this compensation removed a logarithmic bias in the photon energy,
it introduced random
inter-crystal energy dependent calibration errors. Based
o n the known
compensation curves, signal output versus longitudinal position of the I3'Cs
source(afewtypicalexamplesareshoun
inFigure
81,
these
errors are known to be t(2-3)%. The resulting configuration provides 94%
of 4~ solid angle coverage.
Thosecrystalsimmediatelysurroundingthetunnelswereshavedto smallerdimensionsthantheremainingstandardcrystals,toallow
-
22
-
SINGLE CRYSTAL SCHEMATIC
x
INTERSECTION REGION
Figure 7: Individual Crystal and Photomultiplier Tube. Each crystal is individually wrapped in reflector paper and aluminized foil (not shown). The small triangular end is typically e 2 i n c h e s on a side, and the large end is typically e 5 i n c h e s on a side. Due to the construction technique there were 1 1 different sizes of crystals.
-
23
-
PROBABLE SHOWER MAXIMUM(MeV)
175 400 1600 IO4 I
c3 -
I
I
1
AI
1. I
r
i
IO5
0.9
W
I -J 3
a W
> -
I .o
c
0
-l
a -I 0.9
0
.
I
W [r
c
0.9 41 41 1- 86 2A 3
-I
J 15 5 IO
0
D ISTANCE FROM S M A L L END (inches)
Figure 8: A Few Typical Crystal Compensation Curves. The signal output from a 137Cs source positioned along the side of crystal i s plotted versus the longitudinal position o f the source. The s i g n a l s were normalized by their averge value.
-
24
-
a
adequatecableaccesstothecentralchambers.Takingthisinto account,plusthefinitesizeoftheinteractionregionandthe requirement that there be an additional layer of crystals surrounding the central crystal struck by a photon, leads to an acceptance of 85% of 471
solid angle for clean photon detection and the best energy
measurement.
IHE
2.4
2.4.1
CENTRAL TRACKING
CHAMBERS
MaqnetostrictiveSparkChambers
Two cylindrical magnetostrictive spark chambers (MSSC), e a c h w i t h t w o
sparkgaps,provideddataforaccuratechargedparticletrajectory B g i v e s a detailed description of the central
reconstruction. Appendix
MSSC. Basically, the construction details for the proportionalchambers
(MWPC)
MSSC and multiwire
cylindricalshellswerethesame.The
conductive surfaces were formed from etched copper (MIJPC)
coated Mylar. For the
was used, while for the
MWPC
(MSSC)
or aluminum
MSSC 1.5 and 2 mil copper on
5 mil Mylar
0.5
mil aluminum on
2.5 mil Mylar was
chosen. Laminated cylindrical shells made from the metallized Mylar, epoxy, and Styrofoam were used as the basic chamber building blocks. Different diameter shells were concentrically connected using spacers at theirendstoformthechambers.Thefinalstructurewasextremely rigid, even though the shells were only
21.5
mm thick and up to
e15 c m
in radius, as shown Figure 9. Both MSSC had
a similar design except for dimension and cross plane
angles. The following structures are noted when moving from the inner to the outer surfaces:
-
25
-
ETCHED OUTER OUTER CHAMBER WIRES PLAIN CATHODES FOR BOTH CHAMBERS
/
ETCHED INNER CATHODE
.
I
1-83
44 16853
-
26
-
1)
innershell. shield ground plane ( 1 . 5 mil CUI. b) H . V . return ground etched cross plane gap 1 ( 1 . 5 mil CUI. spark gap 1 (90% Ne and 10% He). middle shell a) H . V . etched straight plane, gap 1 ( 1 . 5 mil CUI. b) tuo non-etched H . V . return current planes ( 2 mil C u each). c) H . V . etched straight plane, gap 2 ( 1 . 5 mil CUI. spark gap 2 . outer she1 1 . a1 H . V . return ground etched cross plane gap 2 ( 1 . 5 mil CUI. b) shield ground plane (1.5 mil CUI. a)
2) 3)
4)
5)
The etched copper tracings were
0,012
inches wide with a
1.0 mm center
to center spacing. In general, the two cross planes Mere helicity to
o f opposite
reduceambiguitiesintrackreconstruction.Thechamber
cross planes were inclined
30°
45O
for the inner chamber and
outer chamber. Both chambers, including the end cap gasmixture.Theinnerchambercovered
94%
MSSC,
of 4n,
for the
used the same
whiletheouter
chamber covered 71% o f 4n solid angle. Since the spark chambers had to be pulsed, their data were not used by the trigger logic. Information from the central
and MWPC was
ball
used for trigger determination. Once a trigger was asserted, voltagepulse
to thesparkchambersMasinitiated.
track was present
in one
a high If
anionization
o f the MSSC gaps due to a charged particle,
then a spark in that gap might occur. The ensuing current pulse along thechamber’setchedwires effect.
The
detectedusingthemagnetostrictive
Was
resultingacousticpulseswereconvertedtoelectrical
signals, then amplified, and finally digitized in the control room. The raw data indicated the point on the chamber circumference, projected along the etched wire, where a charged particle had passed.
-
27
-
+
Following a trigger, was applied to
a
by 211 ms,
the chambers for
350 volt ion clearing field pulse
duration of e 1 5 ms.
a
In addition a
constant field of 10 volts was applied at all other times. The combined effectwas
to
removetheuncombinedionsresultingfromthespark
discharge.
2.4.2 A
MultiNireProportionalChambers
single cylindrical multiwire proportional chambers (MWPC) with two
gaps was sandwiched between the two
4n sr.
Moving from the inner
to
MSSC’s.
The MWPC subtended
83% of
outer surfaces the structure was as
fol lows: 1)
inner shell a) shield ground plane b) etched negative H . V . cathode cross pi ane ( 6 2 O ) for gap 1 2) inner gap 1 ( 9 0 % A r and 10% C O z ) a ) 5 mm gap b) 1 4 4 signal anode wires at H . V . ground c) 5 mm gap 31 middle shell a) plain negative H . V . cathode plane for gap 1 b ) plain negative H . V . cathode plane for gap 2 4) outer gap 2 a) 5 mm gap b) 1 4 4 signal anode uires at H . V . ground c) 5 mm gap 5) outer shell a) etched negative H . V . cathode cross plane ( 9 0 ° ) for gap 2 b) shield ground plane The anode wires were 0.02 mm gold plated tungsten with a spacing o f 24.5 mm.
The MWPC data consisted
of a
list o f wires with signals above the
noise level, and the analog signals from the etched cathode strips. In genera1theMWPCspatialresolutionwascoarserthanfortheMSSC, although the MWPC anode wires had a substantially better efficiency (see
-
28
-
Table 21.
For these reasons the MWPC information was only used for
identifying (or "tagging")
charged particles and not for reconstructing
charged particle trajectories.
TABLE 2 Central Tracking Chamber Performance. The values given are for a single MSSC plane, MWPC anode wire plane, or MWPC cathode strip plane and pertain to inclusive hadronic decays of J1.L and 9 ' . The resolution for a track crossing N gaps would be =N-"* better. Also, the polar angle resolution contains a contribution due to z vertex ( c Z ~ 0 . 8cm ; the uncertainty in the measured interaction ranges from 1 0 - 1 0 0 mr at the chambers). AZIMUTHAL CHAMEER EFFICIENCY FWHM POLAR FWHM RESOLUTION (mr)
~
150520 115215 137
2555 2525 35k2 3522
Inner MSSC Outer MSSI: 95-gap 1 MWPC MWPC75-gap 2
2.5
(%I
RESOLUTION (mr> ~
~
~
~~~~~~~~~
~
~~~~
~~
70-75 70-75 90-98 (anode only) 90-98 (anode only)
180
E N D CAPS
The end caps were intended
to
increase the solid angle coverage
of
the detector, providing both energy deposition information and charged particle tracking. Fifteen standard Harshaw manufactured hexagonal L,,d
NaI(T1)
Figures 3
crystals w2re
and 4 ) .
Two gaps
used in each
20
of the end cap quadrants (see
o f magnetostrictive spark chambert2 were
placed directly in front of the crystals. This brought the solid angle coverage to 98% of 4n sr. End cap data Mere used in the selection of hadronic decay events, and as an additional source of candidate photons for
end
capparticleinformation
includedintheinclusivephoton
was
spectra.
-
n o reconstruction.
29
-
No
2.6
ACQUISITION
The PMT analog pulses were fed to the counting house, where they were integratedandheldoncapacitorsfordigiti2ati0n.l~UsingCMOS switching, the integrate and hold pulses were sequentially connected to a 200 M H z Tracor Northern
1 2 1 3 rundown 13 bit ADC
( 8 1 9 2 channels) for
digital processing. To accomodate the desired dynamic range in photon energies and not permit the
to limit the
ADC
required each PMT signal to be digitized at two corresponds to a low
gains.
resolution,
The low channel
energy deposition and is fed from the higher gain is best suited for high energy deposits and
circuit. ,The high channel
is fed from the lower gain circuit. This both channels, increasing high gain was
detector‘s
way a single
can handle
ADC
its dynamic range. For this experiment the
set for 0-200 MeV and the low gain was sei for
Fast-out analog sums
of 9
crystals (corresponding
to
0-4 GeV.
minor triangles)
were available for trigger purposes. T u o independent hardware trigger systems, both using the fast-out analog signals from the central ball
in coincidence with the beam cross
andMWPCinformation,evaluatedthetriggerassertion.Thetower trigger system,7*1h which was
a compact
T T L logic system, performed
analog summations corresponding to various geometries with in the ball; e.g., i )
ii)
the total ball energy excluding the tunnel modules
(Etot)
and
a count of the number o f major triangles with energy above “ 1 1 0 MeV
(NHT),
thenearminimumenergydepositionduetominimumionizing
particlestraversing
16 L,,d
of NaI.Inadditionthetowertrigger
hardware interrogated the MWPC readout hardware
to obtain the number of
groups of 8 contiguous wires, with at least cne hit
-
30
-
in coincidence from
each gap
Table 3
(NLp,,).
gives a list
of tower triggers and their
thresholdsusedforthedatainthisstudy.Theprimarytrigger consisted o f
an "OR"
of these two triggers,
in
addition to one and
this study. sometimes two other triggers which were not used in
TABLE 3 Primary Trigger Scheme A small part of the NMT, NL,,,, and Eto+ are explained in the text. early running for both J/$ and $' had the total energy threshold set at 1.7 and 2.0 GeV, respectively. This trigger had an active solid angle of 84% or 92% o f 45r sr, depending on whether or not the tunnel boarder crystals Mere included in the trigger hardware energy summations. C
LOG1TRIGGER
THRESHOLDS (GeV)
tot
+'
J/+
Energy:
Total Multiplicity:
> (NL,,,
(NL,,,
> >
The other trigger systemz
1.1 0.14
1.1 0.14 0.14 0.14
O).(NMT = 2)-Etot > 0) *(NMT = 3 ) -Eto+ > [NMT 2 4).Etot >
0.14
0.14
was constructed from modular
N I M logic.
Thissystem,whichwassomewhatredundanttothetowertriggerin function, made hardware analog sums for the top, bottom, and 140
The top and bottom were each required to pass a and the
total
2
energy had to be
650
MeV.
full ball.
MeV discriminator Constant fraction
discriminators were used to provide timing information. To satisfy the t o be present within
timing requirement, the signals had beam cross. The
8 ns of the
NIM trigger, with an active solid angle of
94% of 4n
sr, was combined as an t*OR*' with the primary tower elements to make the highlyredundantand
thus dependable final
trigger.Studies
o f the
trigger requirements using Monte Carlo simulated events (see Appendix C)
-
31
-
show the trigger efficiency to be the J/$ and
~ 9 9 %for inclusive hadronic decays of
4".
Following a trigger assertion a general event hold pulse
is
issued b y
the primary trigger logic. This signals the pending acquisition of the processeddata
by
ll/T55 computer.
the Data Acquisition Programs running on the PDP B y way of a CAMAC interface15 the information from
each element o f the detector is read into the computer memory. The data are stored both
i)
in internal buffers to be written to magnetic tape
for offline analysis, and
ii)
in common block areas for online program
analysis. Two Digital Corporation tape drives handled the data
2.7
I/O.
CALIBRATION
Periodic calibrations6 using and i i )
i>
137Cs as a 0 , 6 6 1 MeV
a Van de Graaff to generate
lgF(p,a)iSO*,
resulting
gamma ray source
0 . 3 4 MeV protons for the reaction
in a 6 . 3 1 MeV
gamma ray from the
l60* decay,
were made on approximately two week intervals during the runs. The pedestals for each crystal were sampled at the same of the calibration
time.
The results
led to a determination of the high and low channel
I t was convenient to parameterize
pedestals and the low channel slope. the high channel slope
as a ratio of the high channel slope to the low
channel sl ope. The ratios were set
by
comparing the ADC values for
crystal energy depositions within the range of both the high and low channels simultaneously ( 1 2 0 - 2 0 0 MeV). Using Bhabha events, YY events, and direct e+e- decayso f J/-# and 4", collected over the two week period corresponding to the 137Cs and Van de Graaff data,
it was possible to correct the initial calibration and
-
32
-
a
substantiallyimprovetheresolution.Theseeventsyielded sample of monoenergetic pairs at the known beam energies, the J/# and 1 . 8 4 2 0 GeV at the
clean
1 . 5 4 7 5 GeV at
+’.
Each major run was preceeded by a careful scan of the resonance peak as a function of the SPEAR beam energy. The peak set point was easily held to within sigma of
1.2
20.2
MeV.
The chromatic dispersion
in the beams has
This will broaden the Bhabha contribution
MeV.
a
to the
peak, but not the direct resonance decay contribution. However this statistical broadening
i s negligible compared to the NaI(T1)
resolution
I t was found that about two weeks
which i s 2 3 0 MeV at these energies. runningonresonanceproducedenoughe+e-and
yy
to
events
make
statistically significant calibrations of each crystals slope value. 1 . 8 4 GeV by 234% over
The Bhabha calibration improved the resolution at the I3’Cs and Van de Graaff calibration alone. There are several important consequences
to
this calibration
procedure.First,thecalibrationsareparameterizedin
a
waythat
assumesalinearrelationshipbetweenthedigitizedvalueforthe crystal signal and the energy deposition
in the crystal. This may not
be the case. For example the masses
of slow
from the invariant mass of their
and n’s,
RO‘S
tor0 photon decay products, are both (CY/, for the n o and
slightly but significantly below their known masses 22%
f o r the
3).
this
However,
correlations between the calibration variations with
Y
calculated
result i s
energies
and/or
complicated by
possible
directions. Second,
a time constant of less than two weeks but
longerthanafewhours,wouldbeaveragedover
in
calibrationmeasurement,andentirelymissed
-
33
-
by
theBhabha
theshortduration
137Cs/Van
de
G r a a f f runs.
A
discussion
r e s o l u t i o n may be found i n A p p e n d i x F.
-
34
-
o f the
detector's
energy
REFERENCES 1.
F.
KGningsmannand
K.
Bulos,Crystal
Ball memo CB-NOTE 154 or 251
(1980).
2.
thesis, Stanford Universi ty Report No.
M. Oreglia, Ph.D.
SLAC-236,
1980, (unpublished). 3.
M. Oreglia et ai., Phys. Rev. 925, 2259 (1982).
4.
Y.
Chan
5.
J.
Gaiser et al., IEEE Trans. on Nucl. Sci. NS-26,
6.
I. Kirkbride et a1 . , IEEE Trans. on Nucl. Sci. NS-26, 1535 (1979).
7. E.
D.
d., IEEE Trans. on
Data at Spear,"
at Spear," in
173 (1979).
in Proc. of the XIVth Rencontre de
Moriond, Les Arcs-Savoie, France, March 111-23, August, 1980;
333 (1978).
"Initial Studies of the Charmonium System Using the
Bloom,
Crystal Ball
Nucl. Sci. NS-25,
E.
D.
Bloom, "Results from the Crystal
Proc. of the
and Photon Interactions
1979, SLAC-PUB-2598, Ball Detector
1979 International Symposium on Lepton
at High Energy, Batavia, Illinois, August
23-29, 1979, SLAC-PUB-2425, November,
1979.
8.
R, Chestnut et al., IEEE Trans. on Nucl. Sci., NS-26, 4395 (1979).
9.
M. Sands, Stanford University Report
10.
No. SLAC-121, 1970.
Fabrication o f the hemispheres was performed under a joint effort by the Crystal Ball Collaboration and the Harshau Chemical Company, Solon, Ohio.
11.
C. Peck and F. Porter, Crystal Ball memo CB-NOTE Dl0 (1976).
12.
J. Tompkins, Crystal
13.
G. Godf rey, Crystal Ball memo CB-NOTE 121
(1977).
Ball memo CB-NOTE 232 (
1976).
14. 6. Godfrey, Crystal Ball memo CB-NOTE 131 (1978). 15.
G. Oxoby and J. Bernstein, Crystal
-
35
-
Ball memo CB-NOTE
113 (1977).
Chapter I I I DATA SELECTION 3.1
INTRODUCTION
1 . 8 - 1 0 6 are hadronic decays o f J/JI and
JI', respectively. The remaining
triggers resulted from a combination of background processes: Removable Backgrounds 1)
Cosmic rays intersecting the (termed cosmic ray events).
ball within the trigger timing window
2)
Degraded e+/e' from the beams which shower in the NaI(T11, or e+/ein the beams interacting with residual gas atoms in the vacuum pipe (termed beam gas events).
3)
Q E D reactions, e'e- + e'e-, IJ.'I.L-, or YY, including single photon radiative additions to these reactions (termed QED events). Direct resonance decays to lepton pairs are also included in this category (ie. J/JI or $' + R'R-I. Non-removable Backgrounds
4)
Non-resonance hadron production, ete(termed non-resonance events).
+
hadrons not via
J/$ or 4 ' '
hadronic decays while minimizing the inclusion of the background events. It
was
found that each of the above backgrounds, except the
non-resonance events, had a characteristic signature allowing
it to be 3h
effectively removed. The residual contamination due to backgrounds
-
36
-
1)
through 3) above amounted to 0.8% for the J/JI sample and 1.5% for the 9' sample. The number of non-resonance background events was calculated to be 1 7 ~ 1 0at~ J / J I and 53-103 at $',
and no attempt was made to remove
themfromthesample.Thecharacteristicsoftheinclusivehadronic decays were sufficiently different from the backgrounds to allow for 94% efficiency in the data selection.
a
A description of the parameters
characterizing the classes of events and the details of the selection criterion will be discussed below. During this discussion, the use of the term "background" shall refer to the first three classes of events only.
3.2 INTERPRETATION
O F THE CRYSTAL ENERGIES
To minimize the event selection bias and to simplify the Monte Carlo simulation
of theinclusivehadronicdecaysof
J/JI
and
JI',
it
was
decided to use only the crystal energy data and some properties derived
No information was required from the
from this in the event selection.
central or end cap tracking chambers. Cosmic ray and beam gas events were removed solely on the basis of their spatial energy deposition patterninthe
NaI(T1).
ThePEDeventselectionrequiredthemore
sophisticateddeterminationofparticletrackinformation
(i.e.p
number of tracks and their energies). Note that the use of the uord "track" in this study refers to both neutral and charged tracks. Asidefromtheanalysiswhichleadstotheisolationofthe individual final state particles, to be described below, each event may be analyzed as
a
unique pattern of energy. This leads to two
interpretations of an event:
i)
-
a mixture of isolated and sometimes
37
-
the
overlapping energy depositions produced ii)
interactions in the detector, and characteristic of a used
to
by
a pattern
of energy deposition
quantify the patterns. These were the
event's
total energyIandasymmetry,as
Section 3 . 4 .
Following is
a
will
Q E D event subtraction and to
bediscussedin
is pertinentto
Thisinformation
in
theinclusivephotonanalysis
IV.
Thecrystalmap
has
a
proventobe
usefulmode
of
JIJI hadronic decay,
a
cosmic ray event, a beam gas event, and
radiative PED event. The map projections are made along it's is
a J/JI
by dividing the ball
it.
majortriangleboundaries,andthenunfolding projected
graphically
10 and 1 1 show a
displaying the energy deposited in each event. Figures
crystal
total
descr iption of the particle identification
and track reconstruction techniques .
Chapter
final state particle
particular class of event. Three parameters were
multiplicity,
the
the
Each
it's energy inscribed.
as a small triangle, uith
The tunnel regions appear as hexagonal holes. Any energy deposited in the end caps is indicated in these holes. Readily visible are clusters of
crystals with non-zero energies, corresponding to points of impact by
thereactionproducts.
Also
apparenttotheeyearethewidely
different patterns of energy distribution in these typical events.
I t is clear from Figures
and 1 1
10
that most clusters appear to is one central
result from a single particle. This means that there crystal with the maximum energy, surrounded
by crystals with lesser
energies which can be attributed to the shower of
a
single particle
striking the central crystal. Some examples of energy clusters (connectedregions)areshowninFigure
12.
-
38
-
Anisolatedminimum
1217
RUN
EVENT
#
2485 ET0T= 27
ECM=
3095
ECM=
3684
ECR ETRKTYP 1 105CIR 57
L
2
950 162CIR
74N 3383 4
102 1018N
5
162 381 N
6
92
7
281
8
105 332
9
197
RUN s 1 2
102CG1 l02N N
197CG2
2305
EVENT
*
58
ET0T=
406
TRK T 217 218 N 215 215 CR
(b) Cosmic
11-82 4416832
Figure 10: Crystal Energy Maps-I. The top figure (a) indicates a typical J/1) hadronic decay. It is n o t a 1)’ decay. The bottom figure possible to distinguish this event from (b) illustrates a cosmic ray traversing tu0 lines of crystals separated bytheinnerhollowdomeofthedetector.Tracksidentifiedbythe standard production analysis are printed at the left,
-
39
-
RUN t~ 2 305 #
11
CR
TRK 17
12
27
1
2
EVENT
#
66
ET0T=
721
ECM=
3684
EVENT
i~
3002
ET0T=
3516
ECM=
3583
606 683
3
(a) Beam Gas
RUN I
CR
1091 TRK T
1 123 1708 C 2 1734 140 N 3 1525 1912 N
11-82 4416833
Figure 1 1 : Crystal Energy Maps-11. typically asymmetric beam gas event i s shown in the top figure (a), while a radiative Q E D e v e n t a t J/+ is shown i n t h e b o t t o m f i g u r e (b). The energy o f the crystal indicated b y tl*K*w i s shown above the map.
A
-
40
-
ionizing particle sometimes provides an extreme example of the single bump connected region, where
all the energy of the cluster is contained
in one crystal. Electromagnetic showers are examples
of very regular
energy deposits, both in the longitudinal and lateral development of the shower.Theirfluctuationsarerelativelysmall,thoughoccasional large fluctuations do occur.
As a contrast, interacting hadrons produce
extremely varied energy patterns in the NaI(T1).
Their fluctuations are
large and irregular when compared to electromagnetic showers. Inadditiontherearemuitiplebumpclusters,wheretheenergy 12).
depositions of several particles may overlap (see Figure cases the individual bumps had using the bump discriminator
In these
to be unfolded from the main clusters algorithm.'
The analysis began by finding
connDcted regions of energy defined as follows: A connected region i s a cluster of contiguous crystals all with an energy greater than 10 MeV. Contiguous means
crystals touching at either their vertices or their sides. The bump discriminator algorithm was then applied region to determine the number Initially there
of
to each connected
particles within the region.
is only a single identified particle
(bump),
which is
associatedwiththecrystalcontainingthemaximumenergyinthe connected region. Each crystal in the connected region was tested to see if it's
energy could be explained
of this bump. on the
as a
probable shower fluctuation
An empirically derived envelope function
was fine tuned
data to most effectively discriminate true particle bumps from
shower f 1 uctuat ions by hand scanning events. This necessarily required walking a fine
line between, on the one hand, finding too many fake
-
41
-
a large
bumps due to fluctuations, and on the other hand, introducing detectioninefficiency
by
radicallysuppressingfakebumpsand
real is
bumps. In the search for new bumps the energy of the ith crystal
compared to the empirically estimated maximum energy for the ith crystal resulting from a
fluctuation of a showering particle striking the bump is parameterized by
crystal.Theenvelopefunction
theintercrystal
6 , as follows: opening angle, 4
where Eb is the energy of the particle striking the bump module (the 14 was used for estimating Eb, see another bump if E ;
> f(46i).
below).
The
ith
crystal
is
considered
This process i s repeated until each
o f the
a bump or a modules in the connected region has been explained as either fluctuation o f an existing bump. Themultiplicity
of
theeventreferstothenumber
of
a particle
discrete energy bumps. Each bump became an impact site for candidate and was associated with
a track
observed
in the event track bank. On
occasion, the charged particle trajectory reconstruction programs, which used only the central spark chamber data, would find a track which was not associated with any energy bump in the detector.
A
hand scan
of
some of these events revealed the cause: charged particle trajectories were f i t individually
t o roads of spark chamber hits.
Only after all
the tracks were fitted was the common interaction vertex determined. The trajectories were then displaced to intersect the
vertex.
Sometimes
a poor initial vertex determination would move the trajectory far enough
-
42
-
(d)
4416A34
Figure 12: Connected Regions. Figure ( a ) illustrates two minimum ionizing energy clusters, while an electromagneticshower is showninFigure (b). Figure ( c ) s h o w sa n A two bump complex interacting charged hadron connected region. connected region i s shown in Figure ( d l .
-
43
-
it's trueenergybumptodestroythecorrespondence.The
awayfrom
result was a charged track with zero energy. charged particle energy deposit became overcounting b i a s
As a cotkequence the true
a fake "photon."
To remove the
of b u m p s
the multiplicity was defined as the number
rather than the number of tracks. Two generically different methods for measuring the particle's deposition were employed
in thestandardeventanalysis.Theenergy
sort algorithm2
was the more sophisticated method in that it
(ESORTI
attempted t o unscramble overlapping showers. It sorted each
energy
crystal's
energy, determining what fraction belonged to each particle. In most cases each crystal's
entire energy was attributed to
With regard to multiple bump energy clusters, the the maximum possible spatial resolution which of
a single particle.
ESORT
method achieved
was set by the resolution
the bump discriminator. The algorithm was fine tuned on
electromagnetic showers, since their energies were the
most reliably
measured. In contrast, the much simpler energy
as
the
sum
of
a
1 1 3 algorithm estimated the
geometricallyfixedpatternofcrystals
13 crystalswereinvolved(see
surrounding the bump module; usually Figure 131,
particle's
although bumps at the major triangle vertices involved only
12 crystals. In practice, the distribution of the number of crystals
with an energy crystals with a
>
0.5 MeV used in the 113 energy calculation, peaked at 8 standard deviation
= 3 crystals. This method made
no
attempt to unscramble overlapping energy deposits. On the one hand, it systematically underestimated the energy of electromagnetic showers, by not including the average ~ 2 . 2 %of energy deposited outside the c13.
-
44
-
On
theotherhand,
it sufferedfrom
a systematicoverestimateofthe
particle’s energy,
if another particle impacted sufficiently close by.
In this case there
was a
double counting o f the deposited energy since
113’s overlapped.Thiswasremediedbyrequiringthatthe
thetwo
interparticle opening angle, the 113
ESORT
Oij, satisfy C O S e i j
energy estimate yielded
a
0 . 5 is at least algorithm. The rise in the distribution for cos8 partially due to clustering of hadron-nuclear fragments, identified as neutrals (split-offs), neartheinteractingchargedparticle.The < 0.0 maybedue to a phasespacebias gentleincreaseforcos9 associated with the charged particle direction.
-
72
-
The radiative transitions to the nc(2984) candidate state from 9’ and J/JI
were studied with three photon selection criteria to evaluate the
sensitivity of
v C natural line
the measured branching ratios and the
uidth to the cuts and the background shape. The same set of Cuts, and
C),
as described above for 9‘ were used here.
Also
B)
the pattern cut
Dl was divided into two parts,.only one o f which was used here. Figures 23(a> and (b) correspond to Figures (dl
20(b)
and (c).
Figures 23(c)
and
show the J/9 and 9’ spectra resulting from a partial application of
the cut Dl.
The selection i s as follows:
6)
Same as above.
C)
Same as above.
E)
The pattern cut described
in 0)
above, removes
lateral energy distribution both too narrow to be
a
particle with
a
a
photon shower
andtoowidetobeasinglephotonshower.Thenarrowenergy deposition is
typical of minimum ionizing charged particles uhich
deposit their energy
in only a few crystals. That part
was not applied here.
o f the cut
Rather, in addtion to criteria A I - C ) ,
Figures
23(c) and (dl show the result of requiring that the particle showers merely not be too broad
(see
removing energetic
(Elr
TO’S
Appendix E ) .
>
600 MeV)
This has the effect of
where the showers from the
decay Y’S have merged, and is responsible for the improved signal to noise seen at the
J/3, spectrum end
point.
More important to the
present application, interacting charged particles also tend to have broad energy patterns and are suppressed b y this cut.
-
73
-
The
n,
signal Mas also s e e n in
spectra obtained b y applying the cuts A )
and Dl.
-
74
-
25 20 15 IO
5
15 IO
5 0
IO’
I o3
IO’ (MeV)
I o2
I o3 4416816
Figure 2 3 : B ) , C), and E ) J/3 and 3’ I n c l u s i v e Y S p e c t r a . and ( b ) r e s u l t from the same c u tasus s e d for The J/+ s p e c t r (aa ) F i g u r e s 20(b) and ( c )F. i g t i r e(sc ) and ( d l s h o w J/3, and 4“ s p e c t r a t h pe a t t e r n cut E l designed t o remove e n e r g e t i c obtained by applying no’s as crell as i n t e r a c t i n gc h a r g e dp a r t i c l e s . The removal o f f a s t no’s e s p e c i a l l ye n h a n c e st h e end p o i n t s t r u c t u r e as s e e ni nt h e J/+ spectrum.
-
75
-
4.2.3
2 ~'(3592)
Slightly different photon selection criteria were used for studying candidate state observed in the transition 4"
the n,'(3592)
photon energy of "92 MeV. 6)
Same as above.
GI
In addition to 6)
3
rnc' at a
The cuts are summarized here:
the particles were required to pass
a pattern cut
which rejected energy depositions too broad to be consistent with
E),
singlephotonshower.Thecutwassimilarto stringent.Intheregion
of
EY
a
butless
thiscutpredominantly
92 MeV,
removedinteractingchargedparticles,misidentified
as
photons.
Neutral tracks too close to interacting charged particles were also
HI
Instead o f criterion G I , and in addition to 61, an alternate pattern cut was applied. Using
a different parameterization for the lateral
energy profile, a selection criterion similar to but milder thanDl,
was used to suppress both interacting and minimum ionizing charged particlgs (see Appendix E).
Neutral tracks too close to any charged
particle were removed i f cosEitrack.charse I)
In addition to
HI,
.
The numerical
results are tabulated in Table 10. The problem of determining the
?I, natural
the statistical uncertainty in the signal J/3 EY
2
636 MeV in the transition 3'
-f
width. The error introduced due
YQ,
line width i s dominated by +
rqc, since the signal at
is fairly insensitive to the
to the uncertainty
nC
in the detector's
intrinsic resolution for the J/3 transition is small, since the observed
n c width is about equal this regard, the measurement. Figure fitted
nc
n,
t o the detector's
resolution at EY ~ 1 0 8MeV. In
natural width measurement
is
30 gives a plot of the variation
width, for each
fit shown in Figure 29.
-
90
-
similar to the
X0
in x 2 versus the The single dot
in
I I
I
I
I
I
I
300 200 IO0 7.5
c
Background , Subtracted
15
-
IO
5 0
-5 -10
200 I50 IO0
-
I
Backaround
I
IO
0 -2 -4
2800
8 0 . 1
-
-J
2900 3000 1
I
1
I
3 100
1
I
n
2920 2960
3200
3000
3040
I 20
1
IO0
80
X
.E
0
-F
N
v
5
50 40
60 4
2 0
-2 -4
2750 12-82
2850
2950 3050
3150
2940 2980 3020
RECOl L MASS
(MeV )
4416C39
Figure 2 9 : !3),C),E) Simultaneously Fitted J l 3 and 4" Spectra: qC. The results o f a simultaneous fit to the q c mass in 9' + rnC and J/$ -+ yqC. The fits (b) and ( d l correspond to J/J, spectra shown in Figures and the fits ( a ) and (c) to the 3' spectra shoun in 23(a) and (b), Figures 20(b) and ( c ) . The third fit, ( f ) and ( e ) , is to the J/3 and 3' spectra in Figure 23(c) and (dl, respectively. The preferred resolution value of u o = 2.7% was used here.
-
91
-
Figure
30
representsthe
f i t resultsforaresolutiongiven
by
Each half parabola shows the variation in x 2 obtained from
2.7%/Ey1/”.
the fit b y manually varying
r(qC)for
the two extremes in the resolution
uncertainty interval. The parabola on the left (right) corresponds to the uorst (best) estimated resolution.
A symmetric parabola would have
resulted if the resolution had been held fixed. Both the 3’ the
+
yqc and the J/$
and 1979
1978
acquisition, the
resonance
3
signals were first measured5 in
yqC
data.
Withthe
1980 and
1981
3’ and J/3 data samples were doubled, yielding the
results shown here.Severalcross-checksweremadeon
nC
signals, in all
verifying that they were present in subsets of the data, and regions
of
transitions
the to
data
NaI(T1).
Inaddition,bothsignalscorrespondto
a
astateyielding
consistentmassat
298424
confirmation^^^^ of the state in exclusive decays of the 3‘ and a l s o been reported.
-
92
-
MeV.
J/$ have
I
I04 m
0 h
z
I
I
I
I
I
.
12 -
X L
I
I
IO
-
8 I
78 2 I?
4
6
8
- 82
IO
!?
12 14 (MeV)
16 18 20 441b013
F i g u r e 30: x 2 D i s t r i b u t i o nV e r s u st h e q c NaturaL l i n eW i d t h . The v a r i a t i o n i n x 2 i s shown a s a f u n c t i o o ntfh e v C width i n the and J/+ y s p e c t r a shown i n F i g u r e 29. The s i m u l t a n e o u s f i t t ot h e ( v e r t i c l e d a s h e dl i n e s )i n c l u d e : i ) the c o n t r i b u t i o n st ot h ee r r o ri n i n x 2 and ii) s t a t i s t i c a lu n c e r t a i n t yt r a c e db yt h ep a r a b o l i cv a r i a t i o n s t h es y s t e m a t i cu n c e r t a i n t y i n t h er e s o l u t i o ni n d i c a t e db yt h es e p a r a t i o n i n t h e v e r t i c e sf otrh et w oh a lpf a r a b o l a . The h a lpf a r a b o l a on the r i g h (t l e f t )c o r r e s p o n d st ot h eb e s (t w o r s t ) limits i nt h er e s o l u t i o n . The p o i n t i n b e t w e e n r e s u l t s f r o m a r e s o l u t i o n o f 2.7%/EY1/@.
JI'
r
-
93
-
TABLE 10 Fit Results for then C State. The results from the simultaneous fitsto the three q Y and J/3r inclusive photon spectra shown in Figure 29. SPECTRUM
MASS
r
Ny(J/++Y?lc)
Ny(3'y+Yqc)
(MeV 1
(MeV) ~
2983.3k1.3 2986.521.2 2982.151.5
8)
C>
E)
~~~~
~
2595+423 13500'2642-1887 25382373 10300+1553-1476 22275277 4600'1220-943
~~
9.2+4.4-3.2 13.323.7 12.0+5.2-4.7
3' 2 ~ ~ ~ ' ( 3 5 9 2 )
4.3.3
The four
3rY
spectra shown
in Figure 2 4 were fit in the region of
MeVwithananalysisidenticaltothatusedforthe transition.
J/J,
The results of the fits are tabulated
-+
in Table 1 1 ,
92
r?lC while
Figure 31 shows four views o f the fitted signal using the central value for the detector's
resolution. The signal at
=92 MeV was first seen in
the inclusive photon spectrum from the 1978-1979
.I' data.
When combined
with the 1981 V data (doubling the data sample), the highly significant signal shown in the fits was identified' as the 8 c y candidate state at a mass of 359254 MeV (hereafter referred to as the ?lc' for brevity). The uncertainty in dominated
by
the determined q c ' natural line width
the statistical uncertainty in the signal and the
systematic uncertainty in the detector's ?lc' x1
is equally
resolution. In this regard the
width measurement has similar problems when compared to the narrow and
x2
states. Figure
correlated signal amplitude
32 shows the variation in
X*
a s a function of the
width.
qc'
and the highly The data
was fixed at it's were plotted from fits made while the resolution uncertainty value, in order to obtain an upper limit on the
-
94
-
lower
vC' width
I
I
I
I
I
I
I
I
I
I
I
I
60000 50000 -
-
40000 30000
I
I
I
I
I
20000
I
I
I
I
I
I
I
-
I
h
u)
.-c n $?
cu
v
-
- IO00
I
I
I
I
I
I
IO00
I
40000 30000
25000
-
20000 -
-
15000 -
l0000
20000 Background I500 - Subtracted
-
-
-
I
I
I
I
I
-
( d)
-
Background
1000 - Subtracted
1000 -
60
80
I 20
I 00
60
80
IO0
I 20
F i g u r e 31: B1,G)-I) F i t s t ot h e $‘ + yqc’ t r a n s i t i o n . The f i t t e ds i g n a l , JI’ + m c ‘ , f r o mt h ef o u r JI’ s p e c t r ai nF i g u r e 2 4 i s shown. These f i t s H e r e made u s i n g t h pe referre v ad l u ftoh ere from d e t e c t o r ’ rse s o l u t i o n . The d a t a shown i n T a b l e 1 1 w e r oe b t a i n e d f i t s u s i n% g h be e svt a l u feotrh ree s o l u t i o n within i t s u n c e r t a i n t y interval.
-
95
-
at
and signal strength. Upper limits were identified
by a change in
x 2 of 2 . 7 .
signal amplitude were calculated the amplitude obtained by
the 90% confidence level (C.L.1
fixing
The lower limits for the
by subtracting
r =
0.0 MeV,
1 . 3 - r ~ ~ t ~ t from i ~ t i ~ ~ l
and using the narrowest
detector resolution. The
follolfing crosschecks,
in addition
to
some
of thechecks
mentioned for the nc(2984) candidate state, were made on the analysis to verifythisinterestingresult.MonteCarlostudiesruledoutthe possibility of
a fake signal resulting from the photon background in
decays like $'
+
n0RoJ/q and $'
+
7)J/r),
where the no's
and the
n are
produced Cl05e to threshold. In addition, the charged particle spectrum was
examined for structure near 92 MeV, and none was found.
-
96
-
Li
23 W (u
22
(u
X
21
i
X v
A
z
8
6 4
2
23
28
22
27
21
26
20
25
l i
co
cu
\ lu
x
24
19
0
2
4
6
8
1
0
r
11-82
0 (MeV)
2
4
6
8
1
0
441 6 A l o
Figure 32: x 2 Distribution for the qc’ Fit. The x z distribution and the variation in the signal amplitude for 4“ + yqC’ are shown as a function of the q G ’ natural line width for the fits tothe inclusive Y spectrashoun in Figure 2 4 . Thedashedlines nc’ indicate the statistical contribution to the upper limit for the natural line width for the case with the best detector resolution.
-
97
-
TABLE 1 1 Fit Results for the nc’ State. The results from the four fitted inclusive photon Figure 2 4 are given below. SPECTRUM
E.y
(MeV)
JI’ spectra
( 9 0 % C.L. Upper Limit)
< < <
. The comparison to theory i s explained in the text. DATUM
XI
x0
(%I Inclusive r C . 8. Exclusive Hark-I I Mark-I DESY-Heidelberg B(xs)rJ/+) ( % I Inclusive Y
X2
B2tXJ)
r(xJ+'YJ/$') ( K e V )
Inclusive Y (90% C . L . ) Model Ref, 2
5.3.3
2.56-20.12'0.20
not seen 0.059-20.017 < 0.56 0.220.2 0.14t0.009
2.3820.40 2.420.6
0 . 6 0 2 0 . 1 712.421.5
28.422.1
51-167 117
2.420.8
2.520.4
0.9920.10+0.08 1.2620.22 1.120.3 1.2620.22 1.020.2
89-703 < 1183 240
30
5
Radiative Decays io the r ~ cand
Table 17
summarizes the pseudoscalar branching ratios obtained
averagingtheresultsfromeachspectrum.
by
All contributionstothe
uncertainty are given in the single error which includes a contribution due to the uncertainty
in the intrinsic detector resolution and a
normalizationerror.
The dominanterror
statistics of
their signals over a
-
is attributable to
large background which
122
-
28%
thelow
is further
complicated
by thecorrelation
of thefittedsignalamplitudesand
widths (except for the transition 9‘
-+
~ 7 ) ~ ) .
When the naive model prediction for the magnetic dipole transition J/*1
+ yvc
Kang,”
is properly corrected,
a rate
as indicated by Sucher, Feinberg and
i s predicted in close agreement with the experimentally
observed value,
if the confining part of the potential
17 includesthiscorrectedmodel
a scalarLorentzstructure.Table
estimate given by Value.
is taken to have
Martin,18 which is a
factor o f 2 below the naive 9’
Since a similar correction for the transition M1 theoryprediction6
lacking,thenaive
prediction for the hindered magnetic dipole transition using a Q C O
sum rule,l9 is found
4”
qc’
is
table.
A
+
is given in the
Ml
yqc
obtained
to be in agreement with the observed
value.
TABLE 17 Radiative Decays to the Pseudoscalar Candidates. obtained from The branching ratios B ( $ ” + ~ q ~ ’ l , B($‘-+yqc), and B(J/+yqc) t h e fitted inclusive photon spectra from J/3 and decays are given. The errors reflect the total uncertainty in the branching ratio values. A description of the reference t o theory may be found in the text.
+‘
DATUM
n
71C
B(4“-+Y1So) ( % I Incl usi ve Y Q C D Sum Rule Ref 19 Naive M 1 Theory Ref. 6 B(J/+-+Y~S~) (%I Inclusive Y Corr. Ml Theory Ref. 18 I
0.2830.06 0.35
(0.5-1.2)
90% C.L.
0.650.1 1-2720.36 (1.0-1.3)
-
123
-
REFERENCES
1.
W. Buchmiiller, Phys. Lett. 1128, 479 (1982).
2.
R. McClary and N. Byers, UCLA/82/TEP/l2,
3.
A.
4.
M.
1982, to be published.
B. Henriques et a l . , Phys. Lett. m , . 8 5 (1976). Aguilar-Benitz
Lett.
m, April
et
aJ.,
"Review of Particle Properties," Phys.
(1982).
5.
H.
Schnitrer, Phys. Rev. Lett. 3 5 , 1540 (1975).
6.
E.
Eichten et al., Phys. Rev.
@2l, 203
(1980).
The naive M I theory
predictionobtainedfromthisreferencewascalculatedwith
a
charmed quark mass of 1.6 GeV.
7. For example see:
T.
Appelquist, R.
Barnett and
M.
28, 387 (1978); W.
Rev. Nucl. Part. Sci.
Lane, Ann.
K.
Buchmiiller and S . - H .
H.
Tye, Phys. Rev. 023, 2625 (1981). 8.
The theoretical gluonic (hadronic) decay rates for the X J states can be converted into total widths using the data in Table 15.
Assuming
that the gluonic and radiative widths dominate, BhJ(Xphadrons)
=
(0.99420.002),
(0.71650.021),
respectively f o r
and (0.876t0.015),
To get the total widths the absolute QCO predicted gluonic
j=O,1,2.
widthsshouldbemultipliedby relative to the multiplied
x2
1/BhJ.
width, the
by Bh;?/Bho
x0
and
=(0.88+0.02)
Tocorrectthewidths x1
and
QCO estimates should be Bht/Bhl
=
(1.2220.04),
respect i ve 1 y . 9.
R. Barbieri, R. Gatto, E. Remiddi, Phys.
m,
Caffo, R.
Gatto, E.
Remiddi,Phys.
R. Barbieri, R.
Gatto, E.
Remiddi, Phys.
Barbieri, M. (1980);
Lett.
465 (1976); R.
Lett. Lett.
9 5 8 , 93 1068, 497
(1981).
10.
The energy dependent error
-
in
the branching ratio
124
-
is mainly
a
combination of
f i t and the photon
the statistical error from the
(as
energydependenterrorinthephotondetectionefficiency opposed t o the overall normalization error).
11.
C.
12.
The width of the
3 8 , 1324 ( 1 9 7 7 ) .
J . Riddick et al., Phys. Rev. Lett. $"
is taken to be (215+40) KeV.
This error shows
up in some o f the model predictions, when calculating
a branching
ratio. 13.
T. M. Himel et al., Phys. Rev. Lett. 4 4, 9 2 (01 9 8 0 ) .
14.
J. Whitaker et a l . , Phys. Rev. Lett.
15.
GI. Bartel et al., Phys. Lett. 738, 4 9 (21 9 7 8 ) .
16.
M. Oreglia et a l . , Phys. Rev. 025, 2 2 5 9( 1 9 8 2 ) .
17.
G.
37,
1596 ( 1 9 7 6 ) .
Feinberg and J. Sucher, Phys. Rev. Lett.
3 5 , 1740 ( 1 9 7 5 ) ; K.S.
Kang and J. Sucher, Phys. Rev. 018, 2698 ( 1 9 7 8 ) . 18.
A.
Martin,"QuarkoniumPhenomenology,"intheProc.
o f the X X I
International Conference on High Energy Physics, Paris, J u l y 26-31, 1982, Ref .TH.3397-CERN, August 26, 1982. 19.
M.
A.
Shifman, 2. Phys.
a,345
-
(1980).
125
-
Appendix A CHARMONIUM THEORY
The model
for a strongly bound system
a quark and antiquark
of
known as quarkonium. The quarkonium model applicable to the
is
J/ll family
is called charmonium, with reference to the flavor quantum number charm, carried b y
the quark antiquark pair (cE>. The members
have a net charm pair.
of
zero,
as
of
it cancels in the particle antiparticle
The states are distinguishable by other quantum numbers, such as
spin ( ' S I , orbital angular momentum (21, space parity (PI, (C),
the family
and energy
(E).
Because o f the symmetry
charge parity
in fermion-antifermion
bound states, not all the quantum numbers are independent.
If
3
i s the
total angular momentum, due to 2 and 3 coupling, then it is known' that:
c = ( - l ) L + s and P = - ( - ? I where
3 =t
+
3, 'S
sc + s ~ J, =
spectroscopic notation,
S
= 0 or
(A- 1 )
L,
L +
Sz,
L = 0,1,2>
1,
and S z = 0 , t l .
...
or S , P , D ,
-
126
-
in
These results are
independent o f the interaction, and predict the following two sets states:
...
of
I
singlet states
triplet states
s=o 'L
s=1 3L J
J
where J = L
where J = L-l,L,L+l, if L
>
0
(A-2)
o r J = l i f L = O
c = P
c = -(-llL
(-1)L
= -c
for L = 0,1,2,..
P = c
.,
is usedfor
wherethespectroscopicnotation,
a state with spin = S,
orbital angular momentum = L,
and total angular
momentum = J. Thecharmoniumenergy
(mass)
spectrumwithinthepurelyquantum
mechanical constraints given by (A-21,
is determined by the details of
the full strong interaction dynamics between the CE pair. In states with different quantum numbers will
this way,
have different masses to the
extentthatthequantumnumbersenterintotheinteraction.Foran interaction which depends only on the separation
of
the constituents,
the quantized radial excitations constrain the possible values for the angularmomentum.Fortheradialquantumnumber,
n =
1,2,3,.
. .,
indicating n-1 nodes in the radial wave function, the angular momentum is limited to L
= n,n-l,n-2,...,0.
-
The following states are possible:
127
-
(A-3)
n
2
1
L = O
The ordering of the states
...
3
in terms o f mass
will again depend on the
details of the interaction. Understanding spectroscopy viewed
as an observable property
o f the
underlying quark structure is essentially a bound state field theoretic a
problem in the strong interaction. Currently the best candidate for theory o f
is quantumchromodynamics
the strongforce
hoped, but not presently realizable, that
in
(QCD).
It
is
QCO the meson
spectroscopies can be calculated precisely. Perturbative calculations in QCD fail
I 1 Fermi, Nhere the
to converge for separations roughly
strongcouplingconstant,
1.
growslargerthan
as,
Theinteraction
energy between quarks separated a large distance increases linearly with distance, resulting in a constant attractive force o f 2 1 6 tons. confinement domain, perturbation theory breaks of
S 1
thescaleforseparationsroughly
interaction to be weak, i.e., a s
a -I
(3 U
- -2 I
W
>
T
.
-3
1 -
: 1 "b
-4
mC I
I
Y XY'
-5
0.0I
0.1
0.05
r
3-82
F i g u r e 40:
0.5 Urn)
I 4282A 1
Charmonium Potentials.
A c o m p a r i s o n s o f some o f t h e p o t e n t i a l s w h i c h h a v e b e e n u s e d t o f i t t h e A + B - R X p o t e n t i a l g w i t h x 2 0.1; b) Q C D inspired charmonium system: a)
p ~ t e n t i a l ; c~ C) o u l o m pb l u ls i n e a pr l u ls o g a r i t h m i ic n t e r p o l a t i o n C o u l o m bp l u sl i n e a rp o t e n t i a 1 ; Se )s a m e potential;4 d ) including running coupling constant dependence,? a,(R). indicate the R independent normalization uncertainty.
-
132
-
as c ) b u t The error bars
El ( n o spin f l i p , parity f l i p , Si =
Sf)
Ref. 1 1
is defined interms of the 6j symbol. M1 (spin flip, no parity flip, L i = Lf) Ref. 12 For transitions between 3S1 and ' S O only.
(A-5)
where, M, is the quark mass, Mi (Mf) is the initial (final) state mass, and I = radial matrix element = I 1 + I Z + 13 + It,
12
= is the spin-independent relativistic correction,
13
= (+)
it
multiple
ii) photon conversion
in the beam pipe and central chambers, i i i ) imperfect light transmission through the crystals and into the phototubes,
iv)
shower losses due to
the paper and aluminized Mylar between the crystals, as
well as to the
material o f the hemisphere cans, v ) phototube multiplication statistics,
-
151
-
vi)
vii)
electronic signal processing and digitization, and
the
interplay with the calibration process. In addition HETC treated kaons a later
by scaling to pions, with the same velocity as the kaon. In
version approximate corrections for the pion-kaon mass difference were made to the energy released in nuclear interactions. Regarding the omission
of
the end caps, this did not affect the of the trigger
trigger efficiency calculation since they Were not part
~ 4 % of the total energy ( ~ 4 %loss in
criterion. However, the absence of
solid angle) would lead to a slight underestimate
as part of the
hadronic event selection efficiency. This was included 25%
the inclusive
for
overall uncertainty in the event selection efficiency.
The central magnetostrictive spark chambers and multiwire proportionalchambersweresimulatedingreatdetail.Therawdata output
was of
thesameform
as forthe
real experiment.Although
simulated central chamber data Were not used in this study, they are included in the discussion for completeness. The Monte Carlo technique actually used for calculating the photon detection efficiency was designed
to remove as much bias
as possible
resulting from the simulation process. The method used combined single 4 ~ rsolid angle, with
photons generated isotropically over the entire real J.4
eventsobtainedusingthehadronselectionalgorithm.The
number o f Y’S generated was used to normalize the number detected at the endtoyieldanefficiency.Samples
of
prepared at Ey = 90, 1 4 5 , 2 1 0 , 320,
~ 5 0 . 1 0 ~gammaseachwere
and 500 MeV.
An immediate loss of
e7% resulted from applying the EGS code, since the end caps were not in
the simulated geometry package. This did not bias the final result,
-
152
-
as
all candidates for the inclusive photon spectra were required to pass a solid angle cut of I c o s S ~ ~< ~0.85. ~ ~ . The ~ ~ array of crystal energies for each Monte Carlo photon was summed uith the existing crystal energy
J/4 event with an
data of a single J/4 event. The result was a typical extra monochromatic photon. With the assumption that on the average
all
the bound charmonium
states have similar multibody hadronic decays, these mixed events were
4 '
usedtosimulateeither
or
YX
+
charmonium-like state. The number
of
J/+
+
YX,
where
X
is any
Monte Carlo gammas surviving the
loss mechanisms mentioned in Chapter I V , divided by the number generated is the photon detection efficiency (after including a correction for the photon conversion probability
in the beam pipe and inner chambers).
This technique contains a slight underestimating bias due to neglecting thedecayrecoilagainsttheradiativegamma,Thebiasshouldbe largest for the highest Monte Carlo photon energies. The uncertainty this effect is included
due to
in the 25% overall normalization error
for the photon detection efficiency.
C.3
MONTECARLOPERFORMANCE About 42% of the energy deposited
by HETC charged hadronic particles
came from the primary particle ionization on the average. An additional 20% (9%)
was generated by
secondary proton (charged pion) ionization.
The two photon decay of no's electron decays yielded
1.9%.
contributed 14%.
Secondary muons and their
The remaining *13% of the average energy
deposition was attributed to three processes known to produce split-offs in theMonteCarlo:
i) ionization by
-
153
-
heavynuclearfragmentswere
2 . 5 % o ft h et o t a le n e r g y ,
r e s p o n s i b l ef o r
ii) p r o m p t Y‘S
f r o mt h ed e c a y
o fe x c i t e dn u c l e ig a v e
3.5%) and iii) l o we n e r g yn e u t r o n sf r o mt h ed e c a y
oef x c i t e d n u c l egi a v e
7.1%.
I n t h e H E T C c o d et h , hee a v i e sp,h o t o n s ,
a n dn e u t r o n sf r o mt h el a s tt h r e em e n t i o n e ds o u r c e sw e r en o tt r a n s p o r t e d .
A l l t h e i kr i n e t i ce n e r g y generated.
was d e p o s i t e a dt th p e o i nwt h e r teh e w y ere
This i s o n l yc o r r e c tf o rt h en u c l e a rf r a g m e n t sw h e r e
a short
I t uas
p a tl he n g twh o u lbdeex p e c t efdr o m t h e si rt r o ni go n i z a t i o n . o b s e r v e dt h a t data.
HETC p r o d u c e df a m r o r es p l i t - o f f st h a nw e r es e e n
B y s u p p r e s s i n gt h el a s t h r e ep r o c e s s e s
i n the
HETC was made
t oc l o s e l y
r e p r o d u c e the r e a l s p l i t - o f f d i s t r i b u t i o n s . The q u a l i t yo tf h eM o n t eC a r l od a t a
was c h e c k e bdm y a k i n sge v e r a
47-50
c o m p a r i s o n s t o t h ree ar le s o n a n c de a t aF. i g u r e s
show t h et o t a
a n dc h a r g e dp a r t i c l es p e c t r a ,
m u l t i p l i c i t y a n de n e r g yd i s t r i b u t i o n s ,
Y
n e u t r a la n dc h a r g em u l t i p l i c i t i e s ,a n d
Y-Y
mass p l o t sf o r
4” andMonte
C a r l od a t ar e s u l t i n gf r o mt h es t a n d a r dp r o d u c t i o na n a l y s i sp r o g r a m s . As
can seen be
in
Figure
47,
tM h eo nCt ea r l ot o ‘ steanl e r g y
i n t h ed a t a .
d i s t r i b u t i o n i s s h i f t e ds l i g h t l yh i g h e rt h a ns e e n
i) c h a r g e dp a r t i c l e s
b e l i e v e dt or e s u l tf r o mt w om a i ns o u r c e s : d e p o s ist l i g h t l m y o re n e r g y e v e n t s ( - 2 0 4 MeV) exactT . his
This i s
( ~ 2 6 0 MeV)
i n HETC
otnhaev e r a gteh a n
in real
and ii) t h em o d eul s e dt og e n e r a t et h ed e c a y s
i s not
i s e x p e c t e dt ol e a dt o
a s l i g h to v e r e s t i m a t ef o rt h ee v e n t
within
s e l e c t i oenf f i c i e n c by w u , te l l p a r t i c l es p e c t r u mp r o d u c e db y
t h e 55%
HETC d e v i a t e s
error.
The
f r o m t h ed a t am a i n l y
charged
i n the
minimum i o n i z i npge aaknid tnhceo n t i n u u m below
2 0 0 MeV
m o s t l yb yc h a r g e dp a r t i c l e sw h i c hh a v er a n g e do u t ) .
The c o n t i n u u ma b o v e
200 MeV c o n t a i n sm o s t l yi n t e r a c t i n gh a d r o n s ,a n dg o o da g r e e m e n t there.
-
154
-
(populated
i s seen
r
I
I,
I
800 600 v, IZ 3
400
0 0
200 0
0
0.2
0.6 0.8 Eta+ /( 3.6 8 4 GeV
0.4
1.0
2000 I500 v,
I-
=3
1000
0 0
500
0 0
(
12-82 4416A42
F i g u r e 47:
T o t aE l nergy
IO 20 TOTALMULTIPLICITY
and M u l t i p l i c i t y f o r W Monte C a r l o and D a t a .
-
155
-
4000
3000
2000
IO00 A
15000 l u
10.000
5000
0
2
I
3
loglo ECHARGE
12-82
4 4416A71
Figure 4 8 : Y andCharged P a r t i c l eS p e c t r af o r Monte Carlo andData. Figure ( a ) compares p a r t i c l e si d e n t i f i e da sn e u t r a l s from t h ed a t a and from t h e Monte Carlo. The dominant c o n t r i b u t i o n sa r e from r e a l p h o t o n s i tnh ed. a t a and EGS generated showers i n t h e Monte Carlo. Charged hadron cascades generated b y H E T C a r e compared with real charged p a r t i c l e si nf i g u r e (b). The bump a t 3.2 GeV i s due t o r e a l and EGS showered e t and e‘from$‘+XJ/+Xete-.
-
156
-
2000
I500 Monte Carlo
1000
1 '. 500 c
I
0 0
10 20 NEUTRALMULTIPLICITY
3000
2000
1000
0 10
0 12- 8 2 441 6A43
Figure 49:
20
CHARGED MULTIPLICITY
Neutral and Charge Multiplicities f o r Monte Carlo and Data.
-
157
-
4000 3000 v,
I-
7 3
0
0
2000 1000 0
0
200
12-82
Figure 50:
Y-Y
400
600
My7
(MeV)
800
1000 44 16A70
Mass P l o t f o r Monte Carlo and Data.
-
158
-
Since Monte Carlo photons were used to measure the photon detection efficiencyunder
a varietyofselectioncriteria,includingpattern
cuts, the EGS generated shower profiles were compared with real photons. The source of real
Y’S
was from the fitted data sample
9’
-+
rrJ/13r
+
rrR+R-, although in the comparison, unfitted quantities were used. The following shower quantities were checked: the
>
1 1 3 geometry with an energy
51(a)),
ii)
1%
i ) the number of crystals in o f the photon energy (Figure
the spectrum o f energies deposited in each crystal (Figure
51(b>), i i i ) the x1114 energy ratio (Figure 52(a)), ratio (Figure
52(b)),
and v )
i v ) the 12/c4 energy
the 14/113 energy ratio (Figure
Very good agreement is seen between the the real gamma showers.
-
159
-
52CcI).
EGS produced photon showers and
X 0
I
250
-
11
2
0
1
I
200 (a)-
200
P Z
MONTE CARLO yls REAL y's
I50
150 IO0
0
IO0 50
50
0 12-82
k
0
0
0
IO CRYSTALS
2
4 4 4 16A64
F i g u r e 51: C r y s t aEl n e r g i e sf o M r o n t eC a r l oa n dR e aPl h o t o n s . Monte C a r l o a n dr e apl h o t o ns h o w e r sa r ec o m p a r e dF. i g u r e( as) h o w st h e d i s t r i b u t i o no tf h e number o f c r y s t a l w s i t h i tnh 1 e1w 3 hichava en 1 7 5 MeV < E t r a c k < 250 MeV. The e n e r g y > 1.0% o tf h et r a c ke n e r g yf o r i n e a c hc r y s t a lw i t h i nt h e n a t u r a l log o f the i n c l u s i v ee n e r g yd e p o s i t o n 1 1 3 i s p l o t t e d i n f i g u r e ( b ) f o r t h e same e n e r g yp h o t o n s .
-
160
-
x
MONTE CARLO y ' s R E A L y's
200
50
I-
6.50
1.00
0.75 R2/4
60
40
20
0
0.4
0.8
0.6 l/4
12-82
I .O 441 6A65
F i g u r e 5 2 : M o n t eC a r l oa n dR e aPl h o t o nS h o w eP r atterns. R2/!,, RIIh, Monte Carlo and r e apl h o t o ns h o w epr a t t e r n s compared.
-
161
-
and
Rt,/13
are
Appendix D SUBTRACTION
Neutral pions were all
combination of
identified3' fitted
of
Y'S
selecting the best
in the events by
1
mass pairs with the lowest globa
'Y-'Y
resultingfrom
Photon candidates from the entire
mo
decay
detector's
in
X2
1 ater
o f this best combination was stored for
statistic. The result subtraction
RECONSTRUCTED no's
OF
thephotonspectra.
acceptance (98%
of 4n sr)
were included in the combinatorial search. Neutral tracks were identified on the basis of the end cap and central tracking chamber data alone (no pattern cuts were used). to find the class of energetic v o thetwo
Y
A l s o the algorithm was not intended
decays (En roughly
>
600 MeV),
where
showersoverlapduetotheirsmallopeningangle.The
discussion which follows excludes this class ofm o decay. First the routine TGFIT operated on each Y-'Y pair. In the f photon pair opening angle
(eyy)
and energies (Eyl and Ey2)
it3; the
were varried
to minimize the x 2 statistic subject to the constraint that MZyy(fitted)
=
MZTo, where
x Z is defined as:
-
162
-
The uncertainties in the measurements were taken to be: and
IT($)
= 37 mr.
The
constraintfunction,
m i n i m i ~ a t i o nwas ~ ~ obtained by linearizing the
Mzyy
=
2Ey~Er~-(1-cos9yy),anditeratingthe
solution. The process would converge contribution from the last term For each combination
of
u e = 2.4WEy'''
in less than 5 iterations and the
in Equation
D-1
would be neglegable.
x 2 and the fitted
neutrals the
Y-Y momentum y-r
mass
for each pair was calculated from the
X2
4-vector were saved. Figure
53(a)
shows the unfitted
distribution input to TGFIT. The confidence level for one degree
(CL)
of freedom and those pairs Nith
a CL
-600 MeV).
Three parameters were used to Characterize single photon showers.
where the x 1 2,
4,
12maxD
o r 13
14,
and 1 1 3 are energy sums based
crystals as shoun in Figure
energy deposited
in
a
group of
module (111, t o a larger group are bound
in the interval
This h a p p e n s outside the 2 4
13.
on
the sum of 1,
The ratios compare
the
crystals, always including the bump
of
[0,1lD
crystals. while R z / +
In general,
Rl/q
and
Rb/13
may be larger than
1.0-
when the shower fluctuates across the bump module vertex (such that the
12max > 14).
Values of
the parameters
near 1.0 indicate a narrow lateral energy profile, while smaller values
-
156
-
indicate a
broaderprofile.Thedistributions
of these ratios were
stud.ied for well-defined charged particles and photons. A
variety
of
pattern cuts were used in the inclusive photon analysis, to suppress the charged particle contamination
described in Chapter I V ,
in the neutral particle sample. The details
of the photon selection
will be discussed below, with
criteria based on the ratio parameters reference to the cuts, A ) - I ) , AI-C)
0)
outlined in Chapter I V .
No pattern cut was applied.
This pattern cut was composed depositions of
R2,+
-
(RQ,~J
versus
The narrow energy
minimum ionizing charged particles were removed
rejectingparticleswith 0.9975
i)
of two parts:
-
Rb/13
Rt,/13
>
Figure 55(a)
0.95)1.23.
> 0.9975,
R2,b
0.9875,
by
>
Rz/t,
shows a scatter plot of
for minimum ionizing charged particles. The curve
indicates the cut used
to remove this background. Photons
populate the cut region (Figure
55(b)),
a1 so
though their main
distribution is centered outside the cut. Referring
to Figure 13,
this cut selects those photons which shower across the central crystal's
side (Rz,Q
the vertex (RZfQ
>
