The partial truncated icosahedron phoswich array for detection of low energy charged pions and light charged particles

Accepted Manuscript The partial truncated icosahedron phoswich array for detection of low energy charged pions and light charged particles A. Zarrella, E. Churchman, J. Gauthier, K. Hagel, L. Heilborn, A. Jedele, Y.-W. Lui, A.B. McIntosh, A. Rodriguez Manso, A. Wakhle, M.D. Youngs, S.J. Yennello

PII: DOI: Reference:

S0168-9002(18)31476-1 https://doi.org/10.1016/j.nima.2018.10.161 NIMA 61483

To appear in:

Nuclear Inst. and Methods in Physics Research, A

Received date : 18 June 2018 Revised date : 6 September 2018 Accepted date : 21 October 2018 Please cite this article as: A. Zarrella, et al., The partial truncated icosahedron phoswich array for detection of low energy charged pions and light charged particles, Nuclear Inst. and Methods in Physics Research, A (2018), https://doi.org/10.1016/j.nima.2018.10.161 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

*Manuscript Click here to view linked References

1

2

3

4 5 6

7 8 9

The Partial Truncated Icosahedron Phoswich Array for Detection of Low Energy Charged Pions and Light Charged Particles A. Zarrellaa,b,∗, E. Churchmanb,c , J. Gauthierb , K. Hagelb , L. Heilborna,b , A, Jedelea,b , Y.-W. Luib , A. B. McIntoshb , A. Rodriguez Mansob , A. Wakhleb , M.D. Youngsb , S.J. Yennelloa,b a

Chemistry Department, Texas A&M University, College Station, Texas 77843, USA b Cycotron Institute, Texas A&M University, College Station, Texas 77843, USA c Texas Lutheran University, Seguin, Texas 78155, ,USA

10

Abstract

11

The Partial Truncated Icosahedron (ParTI) phoswich array has been de-

12

signed for the detection of low energy charged pions and other light charged

13

particles for the study of pionic fusion reactions. The array consists of 15

14

plastic/CsI(Tl) phoswich detector units arranged in approximately one hemi-

15

sphere of a truncated icosahedron geometry. The phoswich detectors’ particle

16

identification capabilities have been characterized. A pulse shape discrimi-

17

nation technique has demonstrated isotopic identification for Z = 1 and Z

18

= 2 particles and elemental identification up to at least Z = 3. Utilization

19

of digital electronics allows for increased sensitivity in event triggering and

20

independent verification of pion detection. A calibration technique for the

21

array has also been developed.

22

Keywords:

23

Phoswich, Pion Detection, Pulse Shape Discrimination

∗

Corresponding author

Preprint submitted to Nuclear Instruments and Methods A

September 6, 2018

24

25

1. Introduction The ParTI phoswich array was designed for use in an experiment to mea-

26

sure the pionic fusion rection 4 He +

12

27

identifying charged pions. Pionic fusion is the process by which two nuclei

28

fuse during a collision and then deexcite through the exclusive emission of

29

a pion [1]. The final state consists of a fusion residue which has the same

30

number of nucleons as the reacting system and a pion. The species of the

31

fusion residue and the charge of the emitted pion satisfy charge conservation.

32

The process has been observed in many reacting systems [2, 3, 4, 5, 6, 7, 8, 9]

33

with reported reaction cross sections ranging from the order of 100’s of nb

34

to 100’s of pb for the smallest to largest systems. This extremely coherent

35

process requires that more than 140 MeV of excitation energy from collisions

36

of complex ions is focused into only 2 degrees of freedom - the rest mass

37

and kinetic energy of the emitted pion. These reactions represent the most

38

extreme limit of subthreshold pion production.

C →

16

N + π + for the purpose of

39

Pionic fusion events can be identified through either the detection of the

40

characteristic fusion residue with mass equal to the colliding system and ki-

41

netic energy within a well-defined range around the center of mass velocity or

42

through the detection of the emitted pions. An experiment has been designed

43

at the Cyclotron Institute at Texas A&M University which is the first experi-

44

ment to be sensitive to both the fusion residue and the emitted charged pion.

45

This manuscript will describe the ParTI array which is responsible for the

46

pion detection aspect of the experiment and the complete characterization

47

of its particle identification (PID) and selective triggering techniques.

2

48

2. The Partial Truncated Icosahedron (ParTI) Phoswich Array

49

In its current configuration, the ParTI phoswich array is comprised of 15

50

plastic/CsI(Tl) phoswiches arranged on the faces of a truncated icosahedron

51

geometry and covers 19.7% of the solid angle. This results in a detection

52

efficiency of 21.6% for pions produced in the pionic fusion reaction of inter-

53

est discussed in the previous section. The target is positioned in the center

54

of the truncated icosahedron shape and the distance from the target to the

55

faces of the phoswich detectors is 12.1 cm for pentagonal phoswich geome-

56

tries and 11.8 cm for hexagonal phoswich geometries (phoswich geometries

57

will be discussed more in Section 2.1). Each face of the array is an individual

58

aluminum frame which holds a single phoswich detector. Those frames are

59

held into the truncated icosahedron shape using aluminum tabs machined to

60

orient the frames at the proper relative angles. A photograph of the popu-

61

lated ParTI array is shown in Figure 1. Because each detector frame is an

62

independent part, the array geometry can be changed for different experi-

63

mental needs. The ParTI phoswiches can be used in various configurations

64

including a wall or ring by machining appropriate tabs and different portions

65

of the truncated icosahedron geometry can be accessed with existing parts.

66

2.1. The ParTI Phoswiches

67

The ParTI phoswiches are composed of a layer of EJ-212 fast-scintillating

68

plastic and a layer of CsI(Tl) that are optically coupled together and to

69

a single photomultiplier tube (PMT). The array uses 3 different phoswich

70

geometries. Of the 15 total phoswiches in the current construction, 12 have

71

geometries of regular pentagons or hexagons which correspond to the faces

3

Figure 1: A photograph of the populated ParTI phoswich array viewed from the upstream side. (color online)

72

of the truncated icosahedron shape. Each side of the regular polygons is 3.8

73

cm. The 3 other phoswiches are modified hexagonal geometries such that

74

the three detectors fit in a special hexagonal frame with a hole in the center

75

for the beam to pass through. This can be seen at the center of the array in

76

Figure 1.

77

Regardless of the geometry, each of the phoswiches has the same compo-

78

nent construction. The front face of the detectors is a 3 mm thick layer of 4

79

EJ-212 which was chosen for its fast rise and decay times (0.9 ns and 2.4 ns,

80

respectively). The next component is a 1.5 cm thick layer of CsI(Tl). The

81

CsI(Tl) scintillation has two prominent components (600 ns and 1 µs), both

82

of which are much slower than the fast plastic which results in a response

83

that can be separated from the plastic response. Following the CsI(Tl) crys-

84

tal is a 2.54 cm thick Lucite light guide which couples the polygonal shape

85

of the scintillating components to the circular face of a 1924a Hamamatsu

86

PMT. The plastic/CsI(Tl) and CsI(Tl)/light guide couplings are made with

87

a thin layer of BC-630 optical grease. The light guide/PMT coupling is made

88

using RTV-615 optical adhesive. The front face of each phoswich is covered

89

with a layer of 420 µg/cm2 aluminized Mylar® to provide a uniformly thin

90

and reflective barrier. All other faces of the phoswiches are wrapped with

91

white teflon tape and no Mylar. Figure 2 is a photograph of two completed

92

phoswiches, one pentagonal geometry and one hexagonal.

Figure 2: A photograph of two ParTI phoswiches outside of the array. (color online)

5

93

3. Fast vs. Slow Particle Identification and Calibration Method

94

The selection of two scintillating components with very different decay

95

characteristics allows for a fast vs. slow pulse shape discrimination (PSD)

96

method in order to achieve PID. The ParTI phoswiches were simulated us-

97

ing Geant4 [10] in order to predict their PID capabilities. In each Geant4

98

phoswich event a single particle is fired into the front face of the simulated

99

detector. The identity of the particle was selected at random from 10 possi-

100

ble species: neutron, γ, proton, deuteron, triton, 3 He, α, 6 Li, π + , π − . The

101

range of simulated particle energies varies by particle type. For protons,

102

deuterons, and tritons the energy range is from Emin = 5 MeV to Emax = 70

103

MeV, 90 MeV and 110 MeV, respectively. For tritons, 3 He, α, and 6 Li events

104

the energy range is from Emin = 5 MeV to Emax = 2 ∗ Z ∗ 60 MeV where

105

Z is the particle’s proton number. The charged pion energies range from

106

Emin = 5 MeV to Emax = EP∗ F , the energy above the pionic fusion threshold

107

for the pionic fusion experiment mentioned in Section 1. Lastly, the angle

108

of trajectory with respect to the face of the detector was randomized on an

109

event-by-event basis with a maximum simulated angle corresponding to the

110

distance from the center of the phoswich to the corners. This accounts for

111

the range of possible energy losses due to the different effective thicknesses

112

of the phoswich components as a function of incident angle. An example of

113

a Geant4 simulated phoswich response to an incident alpha particle is shown

114

in Figure 3. The vertical axis is the count of photons reaching the PMT face

115

and the horizontal axis is the time since the arrival of the first photon at the

116

PMT in ns.

117

The PSD is accomplished by integrating the waveform response inside two 6

1600 1400

Photon Count

1200 1000 800 600 400 200 0

0

100

200

300

400 time (ns)

500

600

700

800

Figure 3: A Geant4-simulated phoswich response to an incident alpha particle. The vertical axis is the number of collected photons and the horizontal axis is time since first photon arrival in ns. The red shaded region corresponds to the fast integration window and the green shaded region corresponds to the slow integration region. (color online)

118

time regions, one corresponding to the scintillation response of the fast plastic

119

and the other corresponding to the scintillation of the CsI(Tl) component.

120

The fast integration region starts at the beginning of the simulated response

121

and spans 28 ns. The fast integration window is shown in Figure 3 as the red

122

shaded region and encompasses the sharp peak produced by the fast plastic

123

response. The slow integration region begins 48 ns after the start of the

124

phoswich response and spans 400 ns. The slow integration window is shown 7

125

as the green shaded region in Figure 3. The widths and locations of these

126

integration regions were determined by investigating many configurations of

127

the gates in the Geant4 simulation and comparing the PID quality at each

128

iteration.

129

The two integrations can be used to produce the fast vs. slow PID plot

130

shown in Figure 4. The horizontal axis is the photon sum inside the slow

131

integration gate and the vertical axis is the photon sum inside the fast in-

132

tegration gate. The different color points correspond to different particle

133

types - blue for neutrons, green for γs, red for charged pions, and black for

134

the other light charged particles which are labeled next to their respective

135

lines. The neutron and γ responses largely lie on top of each other to form

136

the neutron/γ line. The charged pion PID line contains events from both

137

π + and π − primary events. These events are clearly separated from the

138

neutron/γ line below and from the proton line above. Thus, Geant4 predicts

139

that the ParTI phoswiches will be sensitive to charged pions and capable

140

of efficiently identifying them. Furthermore, the simulation indicates that if

141

one records the entire phoswich response waveform, there is a second charged

142

pion identification method that can be utilized for π + primary events.

143

After implantation in the detector material (or possibly in flight on the

144

way to the detector), the π + particle with a mean lifetime of 26 ns will decay

145

into a µ+ + νµ . This 2-body decay results in 4.12 MeV of kinetic energy

146

for the µ+ . The phoswich response to this decay is difficult to separate

147

from the implantation response, however, since it follows so shortly after the

148

implantation. The µ+ is likely to have also stopped in the detector material

149

(unless it was produced close to an outside wall of the detector) and will

8

×103 Nuclei Events

140

Z=3

Gamma Events Neutron Events

Fast Integration (photons)

120

Pion Events

100 80

Z=2

60

Z=1

40 20 0 0

100

200

300 400 500 600 700 Slow Integration (photons)

800

900

×103 1000

Figure 4: The fast vs. slow PID figure produced by the Geant4 simulation of phoswich responses to neutrons, γs, and light charged particles. The vertical axis is the integrated photon count inside the fast gate and the horizontal axis is the integrated photon count inside the slow gate. The phoswich responses to different particle types are shown in the various colors (online) and marker styles - blue triangles for neutrons, green squares for γs, red open circles for charged pions, and black closed circles for other light charged particles whose proton numbers are noted next to each line.

150

subsequently decay into e+ + νe + ν¯µ with a mean lifetime of 2.2 µs. This

151

will produce a scintillation response from the stopping of the e+ . Since it is a

152

3-body decay, the positron will have a distribution of possible kinetic energies

153

which is peaked around 50 MeV [11]. These positrons have enough energy

9

154

that they are likely to leave the detector before completely stopping. The

155

energy deposited on their way out and their longer mean lifetime compared

156

to the scintillation components of the phoswich, though, make the resulting

157

phoswich response identifiable.

158

Figure 5 shows an example of one of these π + primary events simulated

159

in Geant4. Unlike in the waveform in Figure 3, there is a second peak from

160

the µ+ decay which is clearly visible. A second method of identification for

161

π + events can be established using this characteristic shape of the phoswich

162

response. This method cannot be efficiently used to identify π − primary

163

events as the capture lifetime of the µ− on the nuclei in the detector material

164

is much shorter than the decay lifetime [12].

165

Experimentally, each phoswich response is digitized and recorded using

166

the Struck Innovative Systemme (SIS) 3316 VME digitizer with 250 MHz

167

sampling rate. With access to the experimental waveform responses, the

168

same analysis as described above can be performed on the experimental data.

169

3.1. PID and Calibration using Secondary Beams

170

For the purposes of characerizing and calibrating the ParTI phoswiches, a

171

beam of 16 O was accelerated to 3.9 MeV/nucleon by the K500 superconduct-

172

ing cyclotron at Texas A&M University and impinged on a 9 Be production

173

target. The Momentum Achromat Recoil Spectrometer (MARS) [13] was

174

used to separate and focus secondary beams of known species and energy

175

onto ParTI phoswiches which were mounted at the spectrometer focal plane.

176

The data can then be used to create a calibration template for the phoswiches.

177

In experiments for which calibration cannot be done using radioactive sources

178

or calibration beams, these calibration templates can be used to perform a 10

1200

Photon Count

1000

800

600

Muon Decay Pulse ↓

400

200

0

0

500

1000 time (ns)

1500

2000

Figure 5: An example Geant phoswich response waveform created by an incident π + particle. The vertical axis is the number of photons collected and the horizontal axis is the time since first photon arrival in ns. The second pulse is the detector response to the µ+ daughter’s decay and its presence in the waveform can be used to identify the π + primary.

179

180

successful calibration. Figure 6a is the fast vs.

slow PID figure for a hexagonal geometry

181

phoswich produced using a compliment of secondary beams from MARS.

182

The vertical axis is the integration of digitized ADC channels inside the fast

183

gate and the horizontal axis is the integration inside the slow gate. The

184

spectrometer was focused on the N = Z particle line for 5 magnetic rigidity 11

185

settings. From bottom to top, one can see the neutron/γ line extending out

186

from approximately (0,0). Next is a band corresponding to protons which

187

are uncorrelated with the spectrometer settings as they are primarily pro-

188

duced in reactions on the MARS slits located directly in front of the focal

189

plane chamber. The proton band is followed by a series of 5 spots which

190

are the phoswich responses to the 5 energies of deuteron secondary beams

191

with magnetic rigidities (Bρ) = 1.0 Tm, 1.2 Tm, 1.3 Tm, 1.4 Tm, and 1.5

192

Tm. Above the deuerons are 4 spots which are a result of alpha secondary

193

beams at the 4 highest Bρ settings mentioned above (the lowest energy al-

194

pha beam does not punch through the plastic layer). Lastly, there are two

195

spots above the alphas which correspond to two energies of 6 Li secondary

196

beams at the two highest Bρ settings (6 Li at the lower rigidity settings stop

197

in the plastic layer). Spanning the entire vertical range to the left of the data

198

is the ”punch-in” line which is populated by charged particle events in the

199

phoswiches that stop in the fast plastic and do not punch into the CsI(Tl)

200

layer.

201

The process for creating the ParTI phoswich calibration templates gener-

202

ally follows the technique discussed in ref. [14]. The first step of the method

203

is correcting for the shapes of the punch-in and neutron/γ lines. Incident

204

neutrons and γs which populate the line are only depositing energy in the

205

CsI(Tl) layer of the phoswich. Because no energy is deposited in the fast plas-

206

tic, the neutron/γ line should lie on the horizontal axis in ∆E-E PID space.

207

In fast vs. slow space, though, it is clearly sloped. This is due to the fact

208

that, despite its relatively slow rise time, the CsI(Tl) scintillation begins at

209

the moment of the charged particle interaction. Thus, there is some non-zero

12

×103

90

100

Punch-In

Fast Integration

70 60

6

Li

50 40

102

α

30

d

0 0

50

p

100 150 Slow Integration

70 60

102

50 6

Li

40 30

10

10

103

80

103

20

×103

90

80

Fast Integration

100

α

10

20

nγ

d

10 200

1 3

250 ×10

(a)

0 0

50

p

100 150 Slow Integration

200

1

250 ×103

(b)

Figure 6: (a) The fast vs. slow PID figure from a hexagonal geometry phoswich exposed to secondary beams at the focal plane of MARS. The vertical axis is the integration of waveform response inside the fast gate in ADC channels and the horizontal axis is the integration inside of the slow gate. The red curves are fits of the punch-in and neutron/γ lines and each PID band is labeled with the corresponding PID. (b) The same fast vs. slow phoswich data from panel (a) after the correction for the punch-in and neutron/γ lines has been made. (color online)

210

integration of the CsI(Tl) response inside the fast integration gate. Similarly,

211

the punch-in line corresponds to particles stopping inside the fast plastic and

212

depositing no energy inside the CsI(Tl) layer and yet the responses do have

213

some slow integration due to the fast scintillation “leaking” into the slow

214

gate in the same way. In order to move forward with the energy calibration,

215

these contributions must be corrected for. Also, the intersection of these two

216

lines represent (0,0) in energy space but are shifted in fast vs. slow space

217

due to statistical fluctuations in the waveform’s baseline event-by-event.

218

The two red lines in Figure 6a represent polynomial fits of the punch-in

219

(second order) and neutron/γ (first order) lines. These two fits are used to 13

220

correct the fast and slow integrations of the waveforms event-by-event using

221

Equations 1 and 2 where F and S refer to the fast and slow integrations,

222

(Sint , Fint ) is the intersection point of the two fits, and Mn/γ is the slope of

223

the neutron/γ line fit. The values of a1 and a2 are the coefficients from the

224

second order fit of the punch-in line after the fast vs. slow space has been

225

inverted to produce a fit of slow integration as a function of fast integration. F 0 = (F − Fint ) − (S − Sint )Mnγ

(1)

S 0 = (S − Sint ) − (F − Fint )2 a2 − (F − Fint )a1

(2)

226

The corrected fast vs. slow PID space is shown in Figure 6b where the

227

neutron/γ line is now on the horizontal axis and the punch-in line is on

228

the vertical axis. This corrected fast vs. slow PID figure (and a figure pro-

229

duced through the same process for a pentagonal geometry) is the calibration

230

template that can be used to calibrate the ParTI phoswiches in future ex-

231

periments. In those cases, the raw fast vs. slows should be corrected for

232

the shapes of the punch-in and neutron/γ lines and the location of the in-

233

tersection in the same way that was done above. Then, the fast and slow

234

integrations should be scaled such that they match the appropriate template.

235

The scaling process accounts for gain differences between experiments arising

236

from different PMT biases or intrinsic gain differences between PMTs.

237

Energy calibrations are produced for both the energy deposited in the fast

238

plastic and the CsI(Tl) crystal from the fast vs. slow template in Figure 6b.

239

These energy calibrations can be used for any future phoswich data which

240

has been corrected and properly scaled to match the template. 14

241

4. Characterization of the ParTI Phoswich Response to π +

242

Three ParTI phoswiches (2 hexagonal and 1 pentagonal) were transported

243

to the Paul Scherrer Institute (PSI) in Switzerland where the detector re-

244

sponses to π + implantations were studied. The experiment was performed

245

in the πM1 beam line and copper degraders were used to degrade the pion

246

beams to access the energy range of interest for pionic fusion reactions, ap-

247

proximately 5 MeV to 30 MeV. Figure 7a shows the fast vs. slow PID figure

248

for the π + data collected in a representative phoswich at PSI along with a

249

set of proton/deuteron reference beams also taken at PSI. The charged pion,

250

proton, and deuteron PID lines are labeled. The pion PID line is clearly

251

separated from the proton and neutron/γ lines. The pion line is also popu-

252

lated with the µ+ daughters of pions that have decayed in flight on the way

253

to the phoswiches. The resolution of the detectors is not sufficient, though,

254

to separate the two particles as their masses only differ by approximately 35

255

MeV and their charges are equal. This ambiguity does not affect the any of

256

the results of the following analyses, however, and this PID line will continue

257

to be referred to as the pion PID line.

258

Like before, the vertical axis in Figure 7a is the integration of the phoswich

259

response waveform inside the fast gate in ADC channels and the horizontal

260

axis is the integration inside the slow gate. Panel (b) of Figure 7 shows

261

a closer look at the pion PID line from panel (a) with the software PID

262

gate for pions (solid) and a gated background region (dashed) overlaid. The

263

characteristics of the events inside the two gated regions can help further

264

characterize the phoswich responses to π + implantations.

265

While the π + beam data in both panels of Figure 7 were being collected, 15

16

×103

10

×103

Pion Cut 14 2

Fast Integration

Fast Integration

10

d

8

10

p

6

Background Cut

8

10

12

103

6 102

4

4

0 0

10

20

10

2

π

2

n/γ 30

40 50 60 70 Slow Integration

104

1

80

90

100 ×103

0 0

(a)

10

20 30 Slow Integration

40

1

50 ×103

(b)

Figure 7: (a) The fast vs. slow PID figure from a representative ParTI phoswich during the PSI experiment. The vertical axis is the integration of the waveform response inside the fast gate in ADC channels and the horizontal axis is the integration inside of the slow gate. The PID lines corresponding to the neutron/γ line and the phoswich responses to π + , proton, and deuteron implantations are labeled in the figure. (b) A closer look at the π + PID line from panel (a). The software PID gates for pions (solid) and a gated background region (dashed) are overlaid. This figure does not include any data collected during the proton/deuteron beam runs. (color online)

266

the data acquisition was employing a “muon decay trigger.” When imple-

267

mented, a trigger is produced by the SIS3316 digitizer upon the second in-

268

stance of the signal passing threshold within a digitization window. This

269

trigger was developed in order to facilitate the measurement of very low

270

probability pion production reactions (in this case, pionic fusion) by selec-

271

tively triggering on pion primary events. Phoswich responses to π + events

272

are likely to pass the trigger conditions due to the second pulse produced by

273

the muon daughter’s decay while all other particle interactions do not have

274

this characteristic waveform shape.

16

275

4.1. π + Waveform Analysis

276

As mentioned above, the characteristic phoswich response due to pion

277

implantations that are followed by decays of the muon daughters can be

278

used to distinguish pion events from other charged particles. Figure 8 shows

279

one of the digitized phoswich responses for an event that falls inside the pion

280

PID gate in Figure 7b. The vertical axis is the ADC value in channels and

281

on the horizontal axis is time since the start of the digitization window in

282

ns. The sharp peak followed by the slow decay is the phoswich response to

283

the charged particle implantation and the second pulse with the slower rise

284

around 8000 ns is due to the daughter muon’s decay inside the CsI(Tl) layer.

285

The time between the initial implantation of the pion into the detector

286

and the muon-like decay pulse is defined as ∆t. Because ∆t varies from decay

287

to decay, the time locations of the two pulses have to be determined event-

288

by-event. Because the charged particle implantation always begins with a

289

sharp rise from a clean baseline, the time associated with the implantation is

290

considered to be the first bin of the digitized response that is above a software

291

threshold. The decay pulse, however, is more difficult to locate because the

292

rise is typically more gradual and it is often riding along the fall from the

293

implantation response so there is not a good baseline. Fortunately, because

294

the muon decay trigger is selecting that second pulse, the time associated

295

with the decay pulse can be approximated to be the location of the trigger

296

within the digitization window which is constant for all events.

297

The distributions of ∆t for events inside the pion PID gate are shown

298

for each of the three phoswiches in the PSI experiment in the left column

299

of Figure 9. The right column shows the ∆t distribution for events inside 17

1100

ADC Channels

1000 900 800

∆t

700 600 500 400 7000

7500

8000

8500 Time (ns)

9000

9500

10000

Figure 8: An experimental waveform for an event whose fast and slow integrations fall inside of the pion PID gate in Figure 7b. The vertical axis is the ADC value in channels and the horizontal axis is the time since the beginning of the digitization window in ns. The sharp peak followed by the slow decay toward baseline is the phoswich response to the charged pion implantation and the second, slower rising pulse is the response to the daughter muon’s decay inside the CsI(Tl) crystal. The time between the implantation and the decay is defined as ∆t. (color online)

300

the gated background regions for the three phoswiches. Each one of the

301

histograms is a measurement of the decay curve associated with the incident

302

particle species (or, possibly, multiple particle species) inside of the PID line

303

or background area. Beyond the requirement that all analyzed events must

304

have produced detector responses that meet the muon decay trigger criteria,

305

all events in this analysis have a ∆t of at least 1000 ns. By not considering the

18

306

lower region of the ∆t distributions, we can be confident that the measured

307

fast and slow integrations of the implantation pulses are not contaminated by

308

the decay-like pulses. Also, this eliminates cases where statistical fluctuations

309

in the implantation response satisfy the muon decay trigger prematurely.

310

For each of the decay curves in Figure 9, an exponential fit of the form of

311

Equation 3 can be implemented. Here, the exponential term, λ, is the decay

312

constant of the particles inside the gated region and the constant offset, a0 , is

313

allowing for a flat background. The red curves overlaid in Figure 9 represent

314

these fits for the pion-gated data in each of the ParTI phoswiches. In each

315

pion-gated panel, next to each fit line is the value of τ extracted from the fit

316

of the decay curve where τ = 1/λ. N = N0 e−λdt + a0

(3)

317

It is clear from the comparison of the pion-gated data to the gated back-

318

ground data that the particles populating the pion PID line are decaying

319

after implantation and that those particles populating the background re-

320

gion are not showing any obvious decays which is the expectation if the

321

events populating this region are from random coincidences. The mean life-

322

times calculated from the decay constants in Figure 9 are all within 1% of

323

the known value for τµ+ , 2197 ns.

324

4.2. The Effect of the Muon Decay Trigger

325

As previously mentioned, the muon decay trigger was developed in order

326

to increase the ParTI array’s sensitivity to pion-like events. During the

327

pionic fusion experiment, the array is located inside the target chamber and,

19

1100

4000

τ = 2223 ns

3000

1000 2000 900 1000

900

τ = 2211 ns

Counts

Counts

3000

2000

1000

800

700

180

3000

τ = 2205 ns 160

2000

140 1000 120 1000

2000

3000

4000

5000

6000

7000

8000

1000

2000

3000

4000

5000

6000

7000

8000

Time between Pulses (ns)

Figure 9: (Left column) Decay curves for events inside of the pion PID gates for each of the three phoswiches with the exponential fits of the form of Equation 3 overlaid in red. The extracted mean lifetime, τ , from each of the fits is given in all 3 cases. (Right column) The same ∆t analysis for events inside the gated background regions for the 3 ParTI phoswiches. (color online)

328

consequently, the phoswiches are subjected to a high rate of background.

329

With a standard triggering scheme, the acquisition would be swamped by

330

these uninteresting events resulting in an unacceptable dead time for the

331

measurement of low cross section reactions. The effect of the muon decay

332

trigger on pion event sensitivity is shown in Figure 10.

333

In both panels of Figure 10 the vertical axis is the integration of the 20

7

×103

7 Pion region

6

6

Background region

5 102

4 3

10

2

Fast Integration

Fast Integration

5

102

4 3 10

2

1 0 0

×103

103

1 10

20

30 40 Slow Integration

50

1

60 ×103

(a)

0 0

1

10

20

30 40 Slow Integration

50

60 ×103

(b)

Figure 10: (a) A representative fast vs. slow PID figure for data collected with a π + beam while the SIS3316 digitizer’s standard trigger was implemented. (b) The fast vs. slow PID figure for the same phoswich and same beam as panel (a) but with the muon decay trigger implemented. A gated region inside the pion PID area (solid) and a gated region in a background region (dashed) are overlaid in both figures. (color online)

334

phoswich response inside the fast gate and the horizontal axis is the inte-

335

gration inside the slow gate. The data shown was collected by the same

336

phoswich in the PSI experiment and for the same π + beam. In panel (a) the

337

acquisition is being triggered using the SIS3316’s standard CFD triggering

338

functionality and in panel (b) the muon decay trigger is implemented.

339

Immediately, it is clear that the muon decay trigger has significantly

340

increased the pion sensitivity as the ratio of the population in the pion line

341

to all other background events has increased. In order to quantify the effect,

342

two software gates are created and shown overlaid in the two figures. The

343

size of the solid-lined gate was chosen as to encompass the primary region

344

corresponding to the pion beam spot in Figure 10b. The dashed-lined gate

345

surrounds a region populated by background events. Both gates are the same 21

346

size and their size and positions do not change between panels. A ratio of the

347

number of events inside the two gates shows that the muon decay trigger has

348

increased the pion selectivity by approximately a factor of 10. This result

349

should be considered a proof of the muon decay trigger’s usefulness, not an

350

expectation to be applied to future experiments. The increased selectivity

351

that can be obtained using the muon decay trigger is dependent upon the type

352

and rate of background particle implantations as well as hardware thresholds

353

and the amount of noise in the electronics.

354

5. Conclusion

355

The ParTI array has been designed for the detection of charged pions and

356

other light charged particles resulting from pionic fusion experiments. The

357

plastic/CsI(Tl) phoswiches have been constructed and have been shown to

358

have good PID capabilities for pions and light charged particles using fast vs.

359

slow PSD. The SIS3316 digitizers have been used to record phoswich response

360

waveforms which have been shown to be able to aid in the detection of pion

361

events using the characteristic shape produced by the decay of daughter

362

muons. The muon decay trigger which was developed for the pionic fusion

363

experiment is working as intended with the ability to increase sensitivity to

364

pion-like events by at least an order of magnitude. The results reported

365

here indicate that the ParTI array meets the necessary criteria for use in the

366

pionic fusion experiments for which it was designed. It has the potential to

367

also be useful for a range of experiments which require a flexible array with

368

large angular coverage for the detection of light charged particles.

22

369

370

6. Acknowledgments This work was supported by the Robert A. Welch Foundation [Grant A-

371

1266] and the U.S. Department of Energy DOE [Grant DE-FG02-93ER40773].

372

We would like to thank the excellent staff at the Texas A&M University Cy-

373

clotron Institute for providing the experimental beams and technical support.

374

We would also like to thank the Paul Scherrer Institute for providing us with

375

beam time and, especially, the PSI πM1 beam line group for making the

376

experiment possible.

377

References

378

[1] P. Braun-Munzinger and J. Stachel, Pion Production in Heavy-Ion Col-

379

lisions, Annual Review of Nuclear and Particle Science, 37 (1987) 97 -

380

131.

381

382

383

384

385

386

387

[2] D. Horn, et al., Pionic Fusion of Heavy Ions, Physical Review Letters, 77 (1996) 2408 - 2411. [3] Y.Le Bornec, et al., Coherent Pion Production near Threshold with a 3

He Projectile, Physical Review Letters, 47 (1981) 26 - 30.

[4] L. Joulaeizadeh, et al., Pionic fusion in light-ion systems, Physics Letters B, 694 (2011) 310 - 315. [5] W. Schott, et al., Observation of the

12

C(3 He,π + )15 N reaction near

388

threshold using recoil detection, Physical Review C, 34 (1986) 1406 -

389

1410.

23

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

[6] M. Andersson, et al., Pionic fusion study of the 6 He halo, Nuclear Physics A, 779 (2006) 47 - 62. [7] L. Bimbot, et al., Experimental study of coherent pion production in 4

He(3 He,π + )7 Li and a test of the reaction mechanism, Physics Letters

B, 114 (1982) 311 - 314. [8] J. Homolka, et al., Measurement of the

12

C(3 He,π + )15 N reaction cross

section near threshold, Physical Review C, 38 (1988) 2686 - 2691. [9] N. Willis, et al., Coherent pion production in 10

6

Li(3 He,π + )9 B and

B(3 He,π + )13 C close to threshold, Physics Letters, 136 (1984) 334 -

336. [10] S. Agostinelli, et al., GEANT4 - a simulation toolkit,Nuclear Instruments and Methods A, 506 (2003) 250 - 303. [11] J. Barlow, et al., The momentum spectrum of electrons from muon decay, Proceedings of the Physical Society, 84 (1964) 239 - 244. [12] A. Grossheim, et al., Decay of negative muons bound in

27

Al, Physical

Review D, 80 (2009) 053001.

406

[13] R.E. Tribble, R.H. Burch and C.A. Gagliardi, MARS: A momentum

407

achromat recoil spectrometer, Nuclear Instruments and Methods A, 285

408

(1989) 441 - 446.

409

[14] D.A. Cebra, et al., The light response of plastic scintillator and the

410

calibration of large arrays, Nuclear Instruments and Methods A, 313

411

(1992) 367 - 372. 24