ar
X
iv
:p
hy
sic
s/0
51
10
65
v2
[
ph
ys
ics
.m
ed
-p
h]
2
3 N
ov
20
05
Experimental study of a liquid Xenon PET
prototype module
M.-L. Gallin-Martel 1, P. Martin, F. Mayet, J. Ballon,
G. Barbier, C. Barnoux, J. Berger, D. Bondoux, O. Bourrion,
J. Collot, D. Dzahini, R. Foglio, L. Gallin-Martel, A. Garrigue,
S. Jan 2, P. Petit, P. Stassi, F. Vezzu, E. Tournefier 3
Laboratoire de Physique Subatomique et de Cosmologie, CNRS/IN2P3 et
Universite´ Joseph Fourier, 53, avenue des Martyrs, F-38026 Grenoble cedex,
France
submitted to Proc. of the 7th International Workshops on Radiation Imaging
Detectors, 4-7 july 2005, Grenoble, France
Abstract
A detector using liquid Xenon (LXe) in the scintillation mode is studied for Positron
Emission Tomography (PET). The specific design aims at taking full advantage of
the liquid Xenon properties. It does feature a promising solution insensitive to any
parallax effect. This work reports on the spatial resolution capabilities of the first
LXe prototype module, equipped with a Position Sensitive Photo-Multiplier Tube
(PSPMT) operating in the VUV range (178 nm).
Key words: Positron emission tomography (PET), Medical imaging equipment
PACS : 87.58.Fg ; 87.62.+n
1 Corresponding author : Tel.: +33-476-284-128; fax: +33-476-284-004; e-mail: ml-
gallin@lpsc.in2p3.fr.)
2 Present address : Service Hospitalier Fre´de´ric Joliot (SHFJ), CEA, F-91401 Orsay,
France
3 Present address : Laboratoire d’Annecy-le-Vieux de Physique des Particules,
CNRS/IN2P3, BP 110, F-74941 Annecy-le-Vieux cedex, France
Preprint submitted to Nucl. Instrum. Meth. A 2 February 2008

1 Introduction
Positron Emission Tomography (PET) is one of the leading techniques of nu-
clear medicine to access to metabolic and functional information. PET is used
for various medical and biological applications, such as oncology, cardiology as
well as pharmacology. Experimental efforts on a host of techniques have been
made in the field of PET imaging, in particular towards the development of
new generation high resolution PET cameras dedicated to small animal imag-
ing [1,2,3]. A couple of years ago, we proposed to use liquid Xenon in an axial
geometry for a scintillation based PET [4,5,6]. LXe shows rather attractive
features when compared to commonly used crystals : the density 3 gcm−3 is
not so small for a liquid (close to the value of NaI), the decay time is short 3 -
30 ns, the range being due to the various scintillation modes of the Xe atom,
the LXe light yield is very high (60 103 UV/MeV in average [7,8]).
This paper reports on the performances of the first LXe prototype module, in
terms of spatial resolution capabilities.
2 The geometry of the Liquid Xenon PET camera
The active part of this project of LXe camera is a ring featuring an internal
diameter of 8 cm (see Fig. 1). It is filled with liquid Xenon and placed in a
cryostat composed of thin aluminum walls. Sixteen identical modules of the
type shown on Fig. 2, are immersed in this ring. Each module presents a
2× 2 cm2 cross-section in the transaxial plane of the camera. The axial field
of view is 5 cm. A module is optically subdivided by one hundred 2× 2 mm2
MgF2-coated aluminum UV light guides. The UV light is collected on both
sides of a module by two Position Sensitive Photo-Multiplier Tubes (PSPMT).
The (x,y) positions measured by the photo-tubes determine which light guides
have been fired. For each module, the axial coordinate (z) is provided by the
following ratio of the PSPMTi (i=1,2) dynode signals :
Z = (PSPMT1 − PSPMT2)/(PSPMT1 + PSPMT2)
allowing to measure the three coordinates without any parallax error.
3 Experimental set-up
Following the layout displayed on Fig. 3, the Xenon is liquefied in the com-
pressor, then transferred to a container inside the cryostat. The temperature
2

inside the cryostat is kept around 165 K via a liquid nitrogen heat exchanger.
The temperature is constant to better than a few tenths of a degree. The
Xenon container is a stainless steel cylinder 50 mm long and 40 mm in diam-
eter, closed at each end with a suprasil 3 mm thick window. A 22Na source
is mounted on a small carriage moving along the z direction. A LYSO crystal
coupled to a photomultiplier tube completes the experimental set-up to make
the coincidence signal. The VUV photons are then collected with one PSPMT
at each end. PSPMT with the required specifications, i.e high Quantum Effi-
ciency (QE) at 178 nm (QE = 20 %) and still working at low temperatures
(165 K), are not commercially available yet. Hamamatsu provided us with two
prototype tubes, from the R8520-06-C12 series [9], having five aluminum strips
deposited on their window to improve the resistivity of the photocathode at
165 K.
The read out electronics operates at room temperature and is composed of
standard NIM and CAMAC modules. The data acquisition (DAQ in Fig. 3)
software performs barycentre online calculation.
4 Experimental results
4.1 PSMPT spatial resolution and x and y localization
To evaluate the PSPMT spatial resolution, a deuterium light source is used to
produce a constant number of photons in a wide wavelength range. To level
down the number of photons emitted, dedicated light attenuators are placed
in front of the light source. Then to focus the light pulse on a specific area
defined on the PSPMT surface, the light emitted by the source is collected
via an optical fibre going through an opaque plastic disk placed in front of
the PSPMT window. This disk exhibits a matrix of holes equally spaced. The
number of photoelectrons is derived from the dynode signal distribution. The
x and y barycentre distributions are computed on-line by the DAQ and derived
of the anode signals. The PSPMT resolution (FWHM) in x and y as a function
of the number of photoelectrons (Npe) is illustrated by Fig. 4. The resolution
is increasing with Npe, ranging from 0.32 mm down to 0.18 mm for Npe ≥
300. This experimental study concludes that the PSPMT spatial resolution
is fine for our application. The resolution is at the level of 1 mm in the x
and y directions which is in very good agreement with the simulation [10].
The second step of this analysis was to study the light guide separation in the
transaxial plane. It has been evaluated using an 241Am α source located at one
end of the Xenon container. The advantage of using an a source in the liquid is
to give intense point-like sources of photons, at a well defined distance of the
3

PSPMT. Three different matrix of light guides were used for the tests, with
cross sections of 2 × 2, 5 × 5 and 2 × 5 mm2, within an overall cross section
of 20× 20 mm2. In the 2 x 5 configuration, the module had therefore only 40
cells of 48 mm in length. The walls of these cells are made of a double 35 µm
thick aluminum foil, double because specular reflection has been guarantied
on one side only : the basic material is a 35 µm aluminum foil, with a thin
polyethylene film glued on one side to reduce the crookedness of its surface,
followed with evaporation of aluminum again and MgF2 to make the actual
reflecting surface. The Fig. 5 obtained with a guide matrix of 2×5 mm2 shows
a very satisfying light guides separation in the (x,y) transverse plane (see Fig.
5).
4.2 Localization and resolution along the z axis
The specific design of the liquid Xenon PET prototype employs the Depth Of
Interaction (DOI) approach to solve the problem of parallax errors. It permits
a continuous measurement of the z coordinate. An experimental test bench
(see section 3 and Fig. 3) has been built to measure the module prototype
resolution in z. The z coordinate is deduced from the amplitude of the dynode
signals measured on the right and left PSPMT located at each side of the
module, as defined in section 2. The resolution at a z position will be given
by the FWHM of the obtained distributions. The resolution as a function of
the source localization is illustrated by Fig. 6. The resolution is better at the
module extremities rather than at the central position where it is about 10
mm. This result is not as good as expected but a refined simulation of the
light collection in optical guides has been done and three configuration for
the light collection units (present setup with the PSPMT, PSPMT immersed
in the LXe, windowless Avalanche Photo-Diode APD immersed in the LXe)
have been compared [10]. The objective is to increase the amount of collected
light. This analysis by simulation concludes that a higher resolution can be
achieved by :
• increasing the reflectivity of the light guide using other manufacturing pro-
cesses (the current value for the reflectivity is 0.78),
• immersing the PSPMT in the liquid (not recommended by Hamamatsu),
• using high quantum efficiency Avalanche Photodiodes (to be investigated
soon).
4

5 Conclusion
First test of a liquid Xenon TEP prototype module were carried out. The pre-
liminary results of the experimental study allow us to determine the intrinsic
performance of this camera. The localization in the transaxial plane is very
satisfying : the resolution of the PSPMT is better than 0.3 mm in the x and
y directions. Efforts are to be made for the localization in the axial direction
since the resolution in z is not only positon dependent but exhibits a poor av-
erage value of about 8 mm. Simulation of the light collection in optical guides
and in various photodetector devices (PSPMT at the LXe temperature but
not immersed in the liquid, PSPMT immersed in the LXe, windowless APD
immersed in the LXe) concludes that a higher resolution can be achieved [10].
Two next steps are foreseen. At first, the light guide reflectivity (currently
0.78) is to be improved, a new process is under study. Then the VUV light
collection on each module end would be better by using high quantum effi-
ciency windowless APD. We aim to investigate now these two possibilities.
Acknowledgments :
This work has been made possible thanks to the financial grants allocated
by the Rhoˆne-Alpes region through its “Emergence” science program, and
by the CNRS/INSERM via its IPA program dedicated to the small animal
imaging. We are also indebted to Jean-Franc¸ois Le Bas and Daniel Fagret of
the medical department of the Joseph Fourier University of Grenoble for the
support and motivation they brought to this project. We also wish to thank
the technical staff of LPSC and in particular : Y. Carcagno, P. Cavalli, E.
Lagorio, G. Mondin, A. Patti and E. Perbet.
References
[1] S. Weber and A. Bauer, European Journal of Nuclear Medicine and Molecular
Imaging 2004, in press
[2] Y. Yang et al., Phys. Med. Biol. 49 (2004) 2527
[3] K. Wienhard et al., IEEE Trans. Nucl. Sci. 49 (2001) 104
[4] J. Collot et al., Proc. of the IXth Intern. Conference on Calorimetry in High
Energy Physics (CALOR 2000), Oct. 2000, Annecy (France), Eds. B. Aubert
et al. (Frascati Physics Series Vol 21), pp. 305
[5] S. Jan et al., to appear in Proc. of International Conference Imaging
Technologies in Biomedical Sciences (ITBS 2001), May 2001, Milos Island
(Greece)
[6] S. Jan, PhD Thesis, Universite´ Joseph Fourier (Grenoble, France), Sept. 2002
5

[7] T.Doke et al., Nucl. Instr. and Meth. A 291 (1990) 617
[8] J. Seguinot et al., Nucl. Instr. and Meth. A 354 (1995) 280
[9] Hamamatsu Photonics, 8 Rue du Saule Trapu, Parc du Moulin de Massy, 91300
Massy, France
[10] M-L Gallin-Martel et al., A liquid Xenon PET : prototype cell results, to be
sumitted to Nucl. Instr. and Meth.
6

Fig. 1. Transaxial view of the LXe µPET
Fig. 2. Schematic of an elementary module of the LXe camera : the module dimen-
sion are 2 × 2 × 5 cm3.
7

Na22 source
liquid N2


























cell
DAQ
Cryostat
165 K
Xe LYSO
PM2 PM1
cryogenic system
(compressor)
liquid N2
Fig. 3. The experimental set-up.
8

Fig. 5. X and Y localization using the 2 × 5 mm2 light guides configuration.
130 140 150 160 170
re
so
lu
tio
n
(m
m)
0
2
4
6
8
10
12
re
so
lu
tio
n
(m
m)
Source position (mm)
re
so
lu
tio
n
(m
m)
Fig. 6. Axial resolution as a function of the source localization
10