Performance evaluation of a depth-of-interaction detector by useof position-sensitive PMT with a super-bialkali photocathode

Yoshiyuki Hirano • Munetaka Nitta •

Naoko Inadama • Fumihiko Nishikido •

Eiji Yoshida • Hideo Murayama • Taiga Yamaya

Received: 28 January 2013 / Revised: 2 August 2013 / Accepted: 4 August 2013 / Published online: 21 August 2013

� Japanese Society of Radiological Technology and Japan Society of Medical Physics 2013

Abstract Our purpose in this work was to evaluate the

performance of a 4-layer depth-of-interaction (DOI)

detector composed of GSO crystals by use of a position-

sensitive photomultiplier tube (PMT) with a super-bialkali

photocathode (SBA) by comparing it with a standard bi-

alkali photocathode (BA) regarding the ability to identify

the scintillating crystals, energy resolution, and timing

resolution. The 4-layer DOI detector was composed of a

16 9 16 array of 2.9 9 2.9 9 7.5 mm3 GSO crystals for

each layer and an 8 9 8 multi-anode array type position-

sensitive PMT. The DOI was achieved by a reflector con-

trol method, and the Anger method was used for calcu-

lating interacting points. The energy resolution in full

width at half-maximum (FWHM) at 511 keV energy for

the top layer (the farthest from the PMT) was improved and

was 12.0 % for the SBA compared with the energy reso-

lution of 12.7 % for the BA. As indicators of crystal

identification ability, the peak-to-valley ratio and distance-

to-width ratio were calculated; the latter was defined as the

average of the distance between peaks per the average of

the peak width. For both metrics, improvement of several

percent was obtained; for example, the peak-to-valley ratio

was increased from 1.78 (BA) to 1.86 (SBA), and the

distance-to-width ratio was increased from 1.47 (BA) to

1.57 (SBA). The timing resolution (FWHM) in the bottom

layer was improved slightly and was 2.4 ns (SBA) com-

pared with 2.5 ns (BA). Better performance of the DOI

detector is expected by use of a super bialkali

photocathode.

Keywords DOI � Super bialkali photocathode �PET � GSO � In-beam PET

1 Introduction

Position-sensitive detectors composed of a scintillator

array and photomultiplier tubes (PMTs) are commonly

used for positron emission tomography (PET). In addition

to the X–Y plane, the interacting points between a scintil-

lator and gamma-rays in the Z direction can be identified by

depth-of-interaction (DOI) techniques. Many DOI methods

have been proposed based on the stagger arrangement [1–

3], pulse-shape discrimination [4–9], reflector control [10–

13], dual-end readout [14–19], individual readout [20], and

monolithic scintillators [21–26]. We developed a DOI

detector which consists of the LGSO scintillator array and

a multi-anode type, position-sensitive PMT (PS-PMT)

(H8500, Hamamatsu Photonics K.K., Hamamatsu, Japan)

[11]. This detector was designed for the small OpenPET

prototype. Yamaya et al. [27] proposed the concept of

OpenPET, which is able to visualize a physically opened

space between two detector rings and has in-beam PET as

one of its applications. We demonstrated the proof of

concept [28], and Yoshida et al. [29] have reported the

system performance.

In-beam PET is one technique for dose verification in

particle therapy. In in-beam PET, a target positron distri-

bution is generated by nuclear interactions between pri-

mary particles and irradiated tissue, and the activity of the

Y. Hirano (&) � N. Inadama � F. Nishikido � E. Yoshida �H. Murayama � T. Yamaya

Department of Biophysics Molecular Imaging Center,

National Institute of Radiological Sciences, 4-9-1 Anagawa,

Inage-ku, Chiba 263-8555, Japan

e-mail: yhirano@nirs.go.jp; taiga@nirs.go.jp

M. Nitta

Chiba University, 1-33 Yayoicho, Inage-ku,

Chiba 263-8522, Japan

Radiol Phys Technol (2014) 7:57–66

DOI 10.1007/s12194-013-0231-4

positron emitters is very low compared with that in com-

mon PET scans using radioisotope tracers. Because of the

low radioactivity, natural radioactivity originating from176Lu (natural abundance: 2.59 %) [30–32] is dominant

among true coincidences in small OpenPET scanning.176Lu emits gamma rays with three energies, 307, 202, and

88 keV, followed by b- disintegration with a maximum

energy of 597 keV. The b- and one of the gamma rays

causes background coincidences. To avoid the background

coincidences, we intend to replace the LGSO scintillators

with GSO scintillators for a better OpenPET because the

latter has much less radioactive contamination. However,

using GSO scintillators is expected to the lower energy

resolution and spatial resolution due to the smaller quantity

of scintillation photons of GSO. Therefore, we decided to

employ the PS-PMT with a super-bialkali photocathode,

which has higher quantum efficiency than does a conven-

tional bialkali PS-PMT to compensate for the reduced

number of scintillation photons. Performance comparisons

between the super-bialkali type and the regular bialkali

type PMT were done for a DOI based on phoswich detector

[33] and several kinds of scintillators [34], where

improvements were observed.

Our purpose in this work was to evaluate the perfor-

mance of the DOI detectors composed of a 4-layer GSO

scintillator array with attached super bialkali (SBA) and

bialkali (BA) type PS-PMTs. The performance items

selected were the ability to identify the scintillating crys-

tals, the energy resolution, and the timing resolution for a

combination of GSO and SBA and a combination of GSO

and BA.

2 Materials and methods

2.1 The four-layer DOI detector

The four-layer DOI detector was composed of an array of

GSO scintillators (Hitachi Chemical Co., Tokyo, Japan) and

a PS-PMT (Fig. 1a). Each layer was a 16 9 16 array of

2.9 9 2.9 9 7.5 mm3 crystals. The DOI identification was

performed by a reflector control method. Specular reflectors

(Sumitomo 3 M, Ltd., Tokyo, Japan) were inserted at a

border of the scintillators, with different patterns for each

layer to project 3-dimensionally (3D) distributed interacting

points onto a 2D flood histogram. At the borders without the

reflector, crystals were glued in place by optical adhesive

(KE420, Shin-Etsu Chemical Co. Ltd., Tokyo, Japan). The

reflector patterns and the expected 2D position histograms

are depicted in Fig. 1b–j. The entire 4-layer scintillator array

was also covered with the specular reflector. The PS-PMT

was a multi-anode type (8 9 8 channels) H8500 PMT

(Hamamatsu Photonics K.K., Hamamatsu, Japan), and the

effective area was 49 9 49 mm2. The photocathodes were

super bialkali and bialkali types, with respective quantum

efficiencies at the GSO emission peak of about 30 and 20 %.

Quantum efficiencies and GSO emission spectra are shown

in Fig. 2, as obtained from the manufacturer’s specification

sheet and [35]. The geometric configurations for these two

types of PS-PMTs were equivalent except for the numbers of

dynodes: 8 dynodes for the super-bialkali type and 12

dynodes for the bialkali type. An interacting point was cal-

culated by a conventional method, an Anger method by the

use of a resistor chain, which provided signals at four cor-

ners, A, B, C, and D. Thus, interacting points in (X,Y) coor-

dinates were derived from the following formulas:

X ¼ �ðAþ CÞ þ ðBþ DÞðAþ Bþ C þ DÞ ;

Y ¼ �ðAþ BÞ þ ðC þ DÞðAþ Bþ C þ DÞ :

We evaluated the energy resolution, peak-to-valley

ratio, distance-to-width ratio, and timing resolution for

DOI detectors composed of GSO with the super-bialkali

type PS-PMT (SBA) and the bialkali type PS-PMT (BA).

2.2 Measurement of the detector performance

2.2.1 Experimental setups

We measured the energy spectra of each crystal, i.e., 1024

(4 layers 9 16 9 16) spectra, and we calculated the energy

resolutions for energies of 511, 662, 835, and 1275 keV for

both SBA and BA. The DOI detector was irradiated by

point sources of 22Na (511 and 1275 keV), 137Cs

(662 keV), and 54Mn (835 keV) placed at 10 cm from the

surface of the crystal array. The data for 22Na were used for

evaluation of the crystal separation ability, that is, the

peak-to-valley ratio and the distance-to-width ratio.

For timing resolution measurement, we employed a

BaF2 scintillator with a fast rise time (about 10 ns) as a

reference. The 22Na point source was placed between the

DOI and the BaF2 scintillator at 10 cm from the crystal

array surface and 3 cm from the BaF2 surface.

2.2.2 Acquisition systems

The data acquisition system consisted mainly of the nuclear

instrument modules (NIM) and the computer-automated

measurement and control system (Fig. 3). A dynode signal,

inverted and amplified linearly, entered the constant-fraction

discriminator (CFD), which provided a trigger signal for a

gate generator. A charge-sensitive analog-to-digital con-

vertor (ADC) was used for four anode signals, i.e., A, B, C,

and D in Sect. 2.1, and their incoming timings were adjusted

58 Y. Hirano et al.

by cable delays to be coincident with an incoming gate sig-

nal. During the conversion and ADC readout with a network

crate controller CC/NET (TOYO Corporation, Tokyo,

Japan), the trigger logic pulses were vetoed by the busy

signal coming from the crate controller. The anode signals

were amplified linearly to compensate for the low gain of the

SBA due to its small number of dynodes. For the BA, because

saturation of the ADCs was observed, the four corner signals

were attenuated before reaching the amplifier. Because

fluctuation of amplifier gain influences energy resolution, we

still used the amplifier even with the BA to get a fair com-

parison of the energy resolution. The voltage supplied to the

PS-PMTs was -1000 V.

As to the timing resolution measurement, we added a set

of NIM modules for the BaF2 scintillator with an attached

PMT (H6610, Hamamatsu Photonics K.K., Hamamatsu,

Japan). The anode output was divided into two signals: one

entered the CFD, and the other entered the ADC. A coin-

cidence between the BaF2 scintillator and the DOI detector

provided a trigger. Also, the trigger logic pulse was the

common start signal of a time-to-digital converter (TDC)

with 25 ps resolution time (C009, HOSHIN, Kawasaki,

Japan). The TDC stop signals were delayed CFD outputs of

the BaF2 scintillator and the DOI detector. A histogram of

subtraction of the two TDC outputs represented the timing

spectrum.

(c) 1st layer

(g) (h) (i) (j)

(d) 2nd layer (e) 3rd layer (f) 4th layer

Air

Optical adhesionReflector

PS-PMT (H8500)8 x 8 anode array

Expected 2D flood histogram

1st

2nd

3rd

4th

(b)(a)

Reflector

GSO:2.9 x 2.9 x 5.0mm3

16x16 array x 4 layers

Fig. 1 Conceptual scheme of the DOI detector (a). Expected pattern diagram of the DOI detector (b) and expected pattern diagrams of each

layer (c–f). Reflector arrangements of each layer (g–j)

0

5

10

15

20

25

30

35

40

200 300 400 500 600 700 800

Qua

ntum

Eff

icie

ncy

[%]

Wavelength [nm]

1.0

0.8

0.6

0.4

0.2

0.0

intensity [a.u]

QE (bialkali)QE (super bialkali)

emission spectrum (GSO)

Fig. 2 Quantum efficiencies (QEs) of super bialkali and standard

bialkali photocathodes (as specified by the Electron Tube Division,

Hamamatsu Photonics K.K.) and the emission spectrum of the GSO

scintillator [35]

A DOI detector with super bialkali photocathode 59

2.2.3 Analysis

Procedures for extracting the energy spectra of each crystal

were as follows: we gated an energy spectrum of all

acquired events for 511 keV, and we made a 2D flood

histogram with 512 9 512 matrix size. Then, we defined

the circular regions of interest (ROIs) within a 10-pixel

radius on each hot spot. We subsequently constructed

energy spectra inside these ROIs. Finally, the 511-keV

photopeaks were fitted by use of a Gaussian function and

an exponential function assumed as the background, and

the full width at half-maximum (FWHM) was calculated

for individual crystals. These ROIs were also used for

energies of 662, 835, and 1275 keV. The mean and stan-

dard deviation (SD) of each layer were evaluated; peak-to-

valley ratios were defined as

peak-to-valley ratio ¼ peak height

valley height;

and were calculated by the following process: we defined

polygon of ROIs including a row of 1st and 2nd (3rd and

4th) layer crystals. In this detector, crystal responses

originating from two different layers were aligned in a row.

From the profiles of 16 rows, the peak-to-valley ratio, a

ratio between the peak value and the minimum value

between neighboring peaks, was calculated. In addition, to

evaluate the crystal separation ability, we employed

another metric described in [36]: the distance-to-width

ratio, defined by the following formula:

distance -to -width ratio =average of distance between peaks

average of width ðFWHM):

A higher value of the distance-to-width ratio means

better separation ability. This metric includes two

parameters, distance and width. The distance and width

were derived from fitting of multi-Gaussian functions to a

profile. Distance-to-width and peak-to-valley ratios of SBA

and BA were compared for each row. Also, the total mean

and SD were evaluated.

Timing spectra for individual crystals were constructed

by use of events inside the circular ROI and the 3-sigma

energy window for 511 keV, determined by fits in the

energy spectra analysis. We calculated the timing resolu-

tion in FWHM (ns) by fitting with a Gaussian function to

the timing spectrum, and the mean and SD for each layer

were evaluated.

3 Results

3.1 2D flood histograms

The 2D flood histograms of the SBA and BA are shown in

Fig. 4. Clear separation of hot spots corresponding to

crystals was obtained for both types. Each spot for the SBA

was slightly sharper than that for the BA, especially at the

edges. The highest intensity was obtained for the first layer

responses, and for layers farther from the incident surface,

the intensities gradually became lower.

3.2 Energy spectrum and resolution

Representative 22Na energy spectra of each layer together

with fit results are shown in Fig. 5. The energy resolutions

in FWHM (%) are also described on the photopeaks.

Energy resolutions (mean ± SD) at 511 keV of the 1st,

2nd, 3rd and 4th layers for the SBA were 12.1 ± 0.4,

12.7 ± 0.4, 15.5 ± 1.8, and 11.5 ± 1.1, respectively. For

AmplifierAttenuator(for BA)

Attenuator(for BA)

ADC

GateGenerator

InvertedAmplifier

DOIdetector

CC/NETCFD

veto

dynode

anode delay

gate

Amplifier ADC

Coincidence

InvertedAmplifier

DOIdetector

CC/NET

TDCCFD

CFDDividerBaF2

veto

dynode

anode delay

gate

startstopstop

GateGenerator

Na-22

Na-22

(a) Measurment of energy resolution, peak-to-valley ratios, and distance-to-width ratios

(b) Measurment of timing resolution

10cm

10cm

3cm

Fig. 3 Electrical setups for

measurements of energy

resolution and crystal

identification performance

(a) and timing resolution (b).

The difference between the

measurement systems for the

SBA and for the BA was the

installation of an attenuator for

measurements of the latter

60 Y. Hirano et al.

the BA, these were 12.8 ± 0.5, 13.5 ± 0.6, 16.7 ± 2.5,

and 12.9 ± 1.5 %. The SDs were derived from fluctuation

of each crystal energy resolution. Figure 6 shows energy

resolutions of the first layer for the energies of 511, 662,

835, and 1275 keV. The energy resolutions for the SBA

were 12.0, 10.6, 9.45, and 7.68 %, respectively. For the

BA, these were, respectively, 12.7, 11.2, 10.1, and 8.41 %.

Improvement in energy resolution of several percent was

obtained with the SBA.

3.3 Peak-to-valley and distance-to-width ratios

Profiles of rows at the center and edge are shown in Figs. 7

and 8. A row included responses from two layers (1st and

2nd or 3rd and 4th). The peak-to-valley and distance-to-

width ratios of each row are shown in Fig 9. Both ratios

were improved.

3.4 Timing resolution

The average timing resolutions of the 1st, 2nd, 3rd, and 4th

layers for the SBA were 2.7 ± 0.1, 2.7 ± 0.1, 2.6 ± 0.1,

and 2.4 ± 0.1 ns, respectively. Those for the BA were

2.7 ± 0.2, 2.8 ± 0.1, 2.6 ± 0.1, and 2.5 ± 0.2 ns. No

significant improvement was achieved.

3.5 Summary of performance results

The detector performance comparisons are summarized in

Table 1. The SBA had better values for all performance

metrics. The values in Table 1 are averages derived from

all crystals. However, when we had less than 200 entries

for the energy and timing spectra, we did not use those data

for calculation of the averages.

4 Discussion

We measured the performance of the DOI detectors with

the super-bialkali type photocathode PS-PMT (SBA) and

the standard bialkali type photocathode PS-PMT (BA).

Better values were obtained for the SBA for all perfor-

mance items.

Considering the available quantum efficiencies based on

the manufacturer’s specification sheet and the emission

spectra of GSO [35, 37], we expected that about 1.27 times

more scintillation photons could be obtained by use of the

BSA. Energy resolution is proportional to 1=ffiffiffiffi

Np

, where N

is the number of observed scintillation photons. Thus, the

energy resolution of SBA is expected to be BA=ffiffiffiffiffiffiffiffiffi

1:27p

.

From the energy resolutions obtained at several energies

(Fig. 6), about 1.15 times more scintillation photons were

available, which was less than the expected value. How-

ever, for this detector, the energy resolution depends not

only on the number of observed photons, but also on the

anode uniformity and/or electrical noise. The coefficients

of variance of 64 anode gains were 24 and 11 % for the

SBA and BA, respectively. The difference between the

expected value and the experimental value is possibly

explained by the large deviation of the anode gains of the

SBA. Also, the energy resolution depends on non-statistical

factors such as the variation in the scintillation light col-

lection, the scintillation light attenuation in the scintillator.

These are discussed in [34], where super-bialkali and bi-

alkali type photocathodes were compared with use of

several kinds of scintillators; the expected energy resolu-

tions were not observed, just as happened in our present

work. Regarding energy resolution among layers, the

energy resolution of the 4th layer was the best because the

scintillation photons propagated to the PMT with less

(a) SBA (b) BA

Fig. 4 2D flood histograms for the SBA (a) and BA (b). The highest intensity was obtained for the 1st-layer responses, and for layers farther

from the incident surface, the intensities gradually became lower

A DOI detector with super bialkali photocathode 61

reflection and absorption than for the other layers. The

energy resolution of the 3th layer, however, was the worst

due to spillover from the 2nd layer. Two photopeaks

originating from different layers were overlapped, resulting

in poor energy resolution. For the 4th layer, although

spillover was also observed, the two photopeaks were

calculated separately.

Regarding the crystal identification ability, the SBA had

a better performance, up to as much as 5 %, for peak-to-

valley and distance-to-width ratios. Some hot spots, for

example, those of row 1, were improved visually (Fig. 4),

which reflected the increased ratios (Fig. 9). Deviations of

the distance-to-width ratios were smaller than those of the

peak-to-valley ratios. In our detector, peaks originated

from two layers aligned in a row, which caused a large

deviation because of the difference in peak heights. Thus,

the distance-to-width ratios were suitable as a crystal

identification metric for this detector because two param-

eters, i.e., width and distance, were considered in the

metric. As a result, we observed a smaller deviation than

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0

200

400

600

800

1000

1200

1400

1600

1800

SBA layer 1 center SBA layer 2 center

SBA layer 3 center SBA layer 4 center

BA layer 1 center BA layer 2 center

BA layer 3 center BA layer 4 center

SBA layer 1 edge SBA layer 2 edge

SBA layer 3 edge SBA layer 4 edge

0

200

400

600

800

1000

1200

0

100

200

300

400

500

600

700

800

0

100

200

300

400

500

0 1000 2000 3000 4000 5000 6000 7000 8000 0 1000 2000 3000 4000 5000 6000 7000 8000

0 1000 2000 3000 4000 5000 6000 7000 8000 0 1000 2000 3000 4000 5000 6000 7000 8000

0

200

400

600

800

1000

0

100

200

300

400

500

600

700

0

50

100

150

200

250

300

350

400

450

0

50

100

150

200

250

300

(a) GSO+SBA (b) GSO+BA

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

0

200

400

600

800

1000

1200

1400

0

100

200

300

400

500

600

700

800

0

50

100

150

200

250

300

350

0

200

400

600

800

1000

1200

1400

0

100

200

300

400

500

600

700

800

900

0

100

200

300

400

500

0

20

40

60

80

100

120

140

160

180

200

220

12.3% 13.1%

13.4% 15.0%10.2% 15.8%

12.6% 13.9%

12.5%13.3%

17.6% 13.9%

12.4% 12.5%

14.6% 11.2%

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 1000 2000 3000 4000 5000 6000 7000 8000

0 1000 2000 3000 4000 5000 6000 7000 8000

0 1000 2000 3000 4000 5000 6000 7000 8000

0 1000 2000 3000 4000 5000 6000 7000 8000

BA layer 1 edge BA layer 2 edge

BA layer 3 edge BA layer 4 edge

Pulse Height (A.U) Pulse Height (A.U)

Pulse Height (A.U) Pulse Height (A.U) Pulse Height (A.U) Pulse Height (A.U)

Pulse Height (A.U) Pulse Height (A.U)

Pulse Height (A.U) Pulse Height (A.U)

Pulse Height (A.U) Pulse Height (A.U)

Pulse Height (A.U) Pulse Height (A.U)

Pulse Height (A.U) Pulse Height (A.U)

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

center

edge

Fig. 5 Representative energy spectra for SBA (a) and BA (b) at the center and edge, together with fitting function and energy resolution (in

FWHM) of photopeaks

62 Y. Hirano et al.

for the peak-to-valley ratio. However, the distance-to-

width ratio depends on the accuracy of the multi-Gaussian

fits. In some peaks, the fitting curves were not consistent

with the experimental histograms.

Because of the smaller number of dynodes for the SBA

than for the BA, we expected that the SBA would have

better timing resolution. However, almost the same reso-

lution was observed. There might not have been a suffi-

ciently fast rise time for the GSO scintillator to allow an

effect from the difference in the number of dynodes to be

seen. Also, the smaller gain in PMT possibly caused fluc-

tuation of the timing. Similar to the energy resolution, the

timing resolution of the 4th layer was the best. Because the

4th-layer scintillators were closest to the photocathode,

fluctuation of the time for the scintillation photons to reach

the photocathode could be suppressed.

We confirmed advantages for the SBA in the DOI

detector. However, the improvements were minor. One of

the reasons is the larger deviation of anode gains of SBA

than that of BA. An anode gain correction for PS-PMTs

will be needed to enhance the advantage of SBA. In

addition, the emission peak of the GSO scintillators did not

match the 400-nm wavelength, which was the maximum

quantum efficiency (35 %) of the SBA. The wavelength

also gives a large difference between the SBA and BA. A

scintillator with a short emission wavelength would be

expected to give a better performance. For example,

LaBr3(Ce), YAP, and BaF2 have suitable emission spectra.

These maximum emission wavelengths are 380 [38], 370

[39], and 310 nm [40], respectively. Considering their

6

7

8

9

10

11

12

13

14

300 500 700 900 1100 1300

Energy [keV]

Ene

rgy

Res

olut

ion

[%]

835

662

511

1275

GSO+SBA

GSO+BA

Fig. 6 Energy resolutions of the 1st layer for energies of 511 keV

(22Na), 662 keV (137Cs), 835 keV (54Mn) and 1275 keV (22Na); their

respective energy resolutions were 12.0, 10.6, 9.45, and 7.68 % for

the SBA and 12.7, 11.2, 10.1, and 8.41 % for the BA. Energy

resolutions were better with the SBA because more scintillation

photons were observed

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.80

500

1000

1500

2000

2500

3000

3500

4000

0

500

1000

1500

2000

2500

3000

3500

0

500

1000

1500

2000

2500

3000

3500

4000

0

500

1000

1500

2000

2500

3000

3500

4000

4500

X-coordinate(A.U)-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

X-coordinate(A.U)

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

X-coordinate(A.U)

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

X-coordinate(A.U)

Cou

nts

Cou

nts

Cou

nts

Cou

nts

GSO+SBA layers 1-2 row 1 GSO+BA layers 1-2 row 1

GSO+SBA layers 1-2 row 8 GSO+BA layers 1-2 row 8

row 8

row 1

layers 1-2

(a)

(c)

(b)

(d)

Fig. 7 Representative profiles of the layers 1–2. Row 1 of SBA (a), row 1 of BA (b), row 8 of SBA (c), and row 8 of BA (d) are shown together

with fitting lines of multi-Gaussian functions

A DOI detector with super bialkali photocathode 63

emission spectra, about 1.4, 1.4, and 1.5 times more scin-

tillation photons will be available. In addition, these scin-

tillators have fast decay constants of less than 30 ns (YAP:

30 n, BaF2: 0.6–0.7 ns, LaBr3(Cs): 16 n). For these scin-

tillators, a better performance would be obtained by use of

SBA instead of BA.

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0

200

400

600

800

1000

1200

1400

1600

1800

2000p5

0

200

400

600

800

1000

1200

1400

1600

1800

0

200

400

600

800

1000

1200

1400

1600

1800

p5

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

X-coordinate(A.U)

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

X-coordinate(A.U)

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

X-coordinate(A.U)

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

X-coordinate(A.U)C

ount

sC

ount

s

Cou

nts

Cou

nts

GSO+SBA layers 3-4 row 1 GSO+BA layers 3-4 row 1

GSO+SBA layers 3-4 row 8 GSO+BA layers 3-4 row 8row 8

row 1

layers 3-4

(a)

(c)

(b)

(d)

Fig. 8 Representative profiles of layers 3–4. Row 1 of SBA (a), row 1 of BA (b), row 8 of SBA (c), and row 8 of BA (d) are shown together

with fitting lines of multi-Gaussian functions

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 160.0

0.5

1.0

1.5

3.0

2.5

3.0

3.5

4.0

0.0

0.5

1.0

1.5

3.0

2.5

3.0

3.5

4.0

peak

-to-

valle

y ra

tiope

ak-t

o-va

lley

ratio

row

row

layers 1-2 layers 1-2

layers 3-4layers 3-4

(b)(a)

(d)(c)

0.0

0.5

1.0

1.5

3.0

2.5

3.0

3.5

4.0

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

row

0.0

0.5

1.0

1.5

3.0

2.5

3.0

3.5

4.0

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

row

Dis

tanc

e-to

-wid

th r

atio

Dis

tanc

e-to

-wid

th r

atio

row12

layers 1-2

3579

111315

468

10121416

row12

layers 3-4

3579

111315

468

10121416

GSO+SBAGSO+BA

GSO+SBAGSO+BA

GSO+SBAGSO+BA

GSO+SBAGSO+BA

Fig. 9 Metrics of crystal identification ability; distance-to-width ratios for the layers 1–2 (a), layers 3–4 (c), and peak-to-valley ratios for layers

1–2 (b) and layers 3–4 (d)

64 Y. Hirano et al.

5 Conclusions

The performance of SBA and BA by use of the 4-layer DOI

detector with a GSO scintillator array was compared.

Although improvements in the performance with SBA

instead of BA were observed for all items, energy resolu-

tion, timing resolution, and crystal identification abilities,

these improvements were minor. Combination of scintil-

lators with short emission wavelength and SBA would give

more improvements.

Acknowledgments The authors thank Dr. H. Ishibashi of Hitachi

Chemical and Mr. M. Nakamura of the Electron Tube Division,

Hamamatsu Photonics K.K., for providing data on emission spectra

and quantum efficiencies, respectively.

Conflict of interest The authors declare that they have no conflict

of interest.

References

1. Yamashita T, Watanabe M, Shimizu K, Uchida H. High resolu-

tion block detectors for PET. IEEE Trans Nucl Sci. 1990;37:

589–93.

2. Liu H, Omura T, Watanabe M, Yamashita T. Development of a

depth of interaction detector for c-rays. Nucl Instrum Methods A.

2001;459:182–90.

3. Fremout AAR, Chen R, Bruyndonckx P, Tavernier SPK. Spatial

resolution and depth-of-interaction studies with a pet detector

module composed of LSO and an APD array. IEEE Trans Nucl

Sci. 2002;49:31–8.

4. Carrier C, Martel C, Schmitt D, Lecomte R. Design of a high

resolution positron emission tomograph using solid state scintil-

lation detectors. IEEE Trans Nucl Sci. 1988;35:685–90.

5. Schmand M, Eriksson L, Casey ME, Andreaco MS, Melcher C,

Wienhard K, Flugge G, Nutt R. Performance results of a new

DOI detector block for a high resolution PET—LSO research

tomograph HRRT. IEEE Trans Nucl Sci. 1998;45:3000–6.

6. Yamamoto S, Ishibashi H. A GSO depth of interaction detector

for PET. IEEE Trans Nucl Sci. 1998;45:1078–82.

7. Saoudi A, Pepin CM, Dion F, Bentourkia M, Lecomte R, Andreaco

M, Casey M, Nutt R, Dautet H. Investigation of depth-of-interac-

tion by pulse shape discrimination in multicrystal detectors read out

by avalanche photodiodes. IEEE Trans Nucl Sci. 1999;46:462–7.

8. Seidel J, Vaquero JJ, Siegel S, Gandler WR, Green MV. Depth

identification accuracy of a three layer phoswich PET detector

module. IEEE Trans Nucl Sci. 1999;46:485–90.

9. Ohi J, Tonami H. Investigation of a whole-body DOI-PET sys-

tem. Nucl Instrum Methods A. 2007;571:223–6.

10. Murayama H, Ishibashi H, Uchida H, Omura T, Yamashita T.

Depth encoding multicrystal detectors for PET. IEEE Trans Nucl

Sci. 1998;45:1152–7.

11. Tsuda T, Murayama H, Kitamura K, Yamaya T, Yoshida E,

Omura T, Kawai H, Inadama N, Orita N. A four-layer depth of

interaction detector block for small animal PET. IEEE Trans Nucl

Sci. 2004;51:2537–42.

12. Ito M, Lee JS, Park M-J, Sim K-W, Hong SJ. Design and sim-

ulation of a novel method for determining depth-of-interaction in

a PET scintillation crystal array using a single-ended readout by a

multi-anode PMT. Phys Med Biol. 2010;55:3827–41.

13. Nishikido F, Inadama N, Oda I, Shibuya K, Yoshida E, Yamaya T,

Kitamura K, Murayama H. Four-layer depth-of-interaction PET

detector for high resolution PET using a multi-pixel S8550 ava-

lanche photodiode. Nucl Instrum Methods A. 2010;621:570–5.

14. Moses WW, Derenzos SE, Melchert CL, Manentet RA. A room

temperature LSO/PIN photodiode PET detector module that mea-

sures depth of interaction. IEEE Trans Nucl Sci. 1995;42:1085–9.

15. Miyaoka RS, Lewellen TK, Yu H, McDaniel DL. Design of a

depth of interaction (DOI) PET detector module. IEEE Trans

Nucl Sci. 1998;45:1069–73.

16. Shao Y, Silverman RW, Farrell R, Cirignano L, Grazioso R, Shah

KS, Vissel G, Clajus M, Tumer TO, Cherry SR. Design studies of

a high resolution PET detector using APD arrays. IEEE Trans

Nucl Sci. 2000;47:1051–7.

17. Dokhale PA, Silverman RW, Shah KS, Grazioso R, Farrell R,

Glodo J, McClish MA, Entine G, Tran V-H, Cherry SR. Perfor-

mance measurements of a depth-encoding PET detector module

based on position-sensitive avalanche photodiode read-out. Phys

Med Biol. 2004;49:4293–304.

18. Salvador S, Huss D, Brasse D. Design of a high performances

small animal PET system with axial oriented crystals and DOI

capability. IEEE Trans Nucl Sci. 2009;56:17–23.

19. Delfino EP, Majewski S, Raylman RR, Stolin A. Towards 1 mm

PET resolution using DOI modules based on dual-sided SiPM

readout. IEEE Nucl Sci Symp Conf Rec. 2010:3442–49.http://

ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5874446&url=

http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%

3Farnumber%3D5874446.

20. Levin CS. Design of a high-resolution and high-sensitivity scin-

tillation crystal array for PET with nearly complete light collec-

tion. IEEE Trans Nucl Sci. 2002;49:2236–43.

21. Clement D, Frei R, Loude J-F, Morel C. Development of a 3D

position sensitive scintillation detector using neural networks.

IEEE Nucl Sci Symp Conf Rec. 1998;3:1448–52.

22. Bruyndonckx P, L0eonard S, Tavernier S, Lemaıtre C, Devroede

O, Wu Y, Krieguer M. Neural network-based position estimators

for PET detectors using monolithic LSO blocks. IEEE Trans Nucl

Sci. 2004;51:2520–5.

23. Lerche WC. Depth of interaction detection for c-ray imaging.

Nucl Instrum Methods A. 2004;600:624–34.

24. Maas MC, Schaart DR, van der Laan DJ, Bruyndonckx P,

Lemaıtre C, Beekman FJ, van Eijk CWE. Monolithic scintillator

Table 1 Comparison of SBA and BA performance values

1st layer 2nd layer 3rd layer 4th layer

Energy resolution (%) (FWHM)

SBA 12.1 ± 0.4 12.7 ± 0.4 15.5 ± 1.8 11.5 ± 1.1

BA 12.8 ± 0.5 13.5 ± 0.6 16.7 ± 2.5 12.9 ± 1.5

Timing resolution (ns) (FWHM)

SBA 2.7 ± 0.1 2.7 ± 0.1 2.6 ± 0.1 2.4 ± 0.1

BA 2.7 ± 0.2 2.8 ± 0.1 2.6 ± 0.1 2.5 ± 0.2

1st and 2nd layers 3rd and 4th layers

Peak-to-valley ratio

SBA 1.86 ± 1.13 1.54 ± 0.85

BA 1.78 ± 1.05 1.52 ± 0.82

Distance-to-width ratio

SBA 1.57 ± 0.49 1.41 ± 0.51

BA 1.47 ± 0.59 1.33 ± 0.51

A DOI detector with super bialkali photocathode 65

PET detectors with intrinsic depth-of-interaction correction. Phys

Med Biol. 2009;54:1893–908.

25. Schaart DR, van Dam HT, Seifert S, Vinke R, Dendooven P,

Lohner H, Beekman FJ. A novel, SiPM-array-based, monolithic

scintillator detector for PET. Phys Med Biol. 2009;54:3501–12.

26. van Dam SS, Vinke R, Dendooven P, Lohner H, Beekman FJ,

Schaart DR. A practical method for depth of interaction deter-

mination in monolithic scintillator PET detectors. Phys Med Biol.

2011;56:4135–45.

27. Yamaya T, Inaniwa T, Minohara S, Yoshida E, Inadama N,

Nishikido F, Shibuya K, Lam C-F, Murayama H. A proposal of

an open PET geometry. Phys Med Biol. 2011;53:757–73.

28. Yamaya T, et al. Development of a small prototype for a proof-

of-concept of OpenPET imaging. Phys Med Biol. 2011;56:

1123–37.

29. Yoshida E, Kinouchi S, Tashima H, Nishikido F, Inadama N,

Murayama H, Yamaya T. System design of a small OpenPET

prototype with 4-layer DOI detectors. Radiol Phys Technol.

2012;29:92–7.

30. Bircher C, Shao Y. Use of internal scintillator radioactivity to

calibrate DOI function of a PET detector with a dual-ended-

scintillator readout. Med Phys. 2012;39:777–87.

31. Yamamoto S, Horii H, Hurutani M, Matsumoto K, Senda M.

Investigation of single, random, and true counts from natural

radioactivity in LSO-based clinical PET. Ann Nucl Med.

2005;12:109–14.

32. Watson CC, Casey ME, Eriksson L, Mulnix T, Adams D,

Bendriem B. NEMA NU 2 performance tests for scanners with

intrinsic radioactivity. J Nucl Med. 2004;45:822–6.

33. Vaquero JJ, Udlas JM, Seidel J, Espana S, Desco M. Effects of

the super bialkali photocathode on the performance

characteristics of a position-sensitive depth-of-interaction PET

detector module. IEEE Trans Nucl Sci. 2010;57:2437–41.

34. Yamamoto S, Watabe H, Kato K, Hatazawa J. Performance

comparison of high quantum efficiency and normal quantum

efficiency photomultiplier tubes and position sensitive photo-

multiplier tubes for high resolution PET and SPECT detectors.

Med Phys. 2012;39:6900–7.

35. Shimizu S, Sumiya K, Ishibashi H, Senguttvan N, Redkin BS,

Ishii M, Kobayashi M, Susa K, Murayama H. Effect of Mg-, Zr-,

and Ta-doping on scintillation properties of Gd2SiO5:Ce crystal.

IEEE Trans Nucl Sci. 2003;50:7149–74.

36. Lau FW, Vandenbroucke A, Reynolds PD, Olcott PD, Horowitz

MA, Levin CS. Analog signal multiplexing for PSAPD-based

PET detectors: simulation and experimental validation. Phys Med

Biol. 2010;55:7149–74.

37. Valais I, Michail C, David S, Nomicos CD, Panayiotakis GS,

Kandarakis I. A comparative study of the luminescence proper-

ties of LYSO:Ce, LSO:Ce, GSO:Ce and BGO single crystal

scintillators for use in medical X-ray imaging. Physica Med.

2008;24:122–5.

38. Saint-Goban BrilanCeTM data sheet. http://www.detectors.saint-

gobain.com/uploadedFiles/SGdetectors/Documents/Product_

Data_Sheets/BrilLanCe380-data-sheet.pdf. Accessed 14 Aug

2013.

39. Moszynski M, Kapusta M, Wolski D, Klamra W, Cederwall B.

Properties of the YAP:Ce scintillator. Nucl Instrum Methods A.

1998;404:157–65.

40. Saint-Goban BaF2 data sheet http://www.detectors.saint-gobain.

com/uploadedFiles/SGdetectors/Documents/Product_Data_Sheets/

BaF2-Data-Sheet.pdf. Accessed 1 Aug 2013.

66 Y. Hirano et al.