1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Development and characterization of a 3D
CdTe:Cl semiconductor detector for medical
imaging
Marie-Laure Avenela, Daniel Farcageb, Marie Ruata, Loïck Vergera and Eric Gros d’Aillona
a
CEA, LETI, MINATEC-Campus, 17 rue des Martyrs, F-38054 GRENOBLE.
CEA, DEN-DPC, F-91191 GIF SUR YVETTE
Corresponding author: Eric Gros d’Aillon – Present address: CEA, LETI, MINATEC-Campus, 17 rue des
Martyrs, F-38054 GRENOBLE, FRANCE
Phone: +33 (0)4-38-78-20-87
e-mail: eric.grosdaillon@cea.fr
b
Abstract – Conventional semiconductor radiation detectors for medical imaging use either a planar
structure or a pixelated structure. These structures exhibit a natural trade-off between the absorption of
incident photons and the collection of free charge carriers, resulting in a limited choice of detection
materials. Such a trade-off can be avoided by using a 3D structure in which electrodes are drilled into the
detection volume. A prototype 3D semiconductor detector has been developed, using CdTe:Cl. A laser
drilling technique was used to create electrodes in the volume of the material. The electrodes were
contacted using electroless Au deposition. The manufacturing process and the first spectrometric results
obtained with 241Am and 57Co irradiation are presented below. Synchrotron X-ray irradiation was also
performed at the European Synchrotron Radiation Facility (ESRF) at an incident energy of 60 keV. An
individual photon-counting ability was exhibited. These results will be used as a proof of concept for
investigating 3D detectors in the medical-imaging energy range.
Keywords – 3D semiconductor Detectors, Medical Imaging, CdTe, Laser Drilling
1.
INTRODUCTION
3D detection geometry was first introduced by Parker et al. in 1997 [1]. Whereas in conventional planar or
pixelated detection structures electrodes are deposited onto the surface of the detector, 3D structures are made of
a three-dimensional array of cylindrical electrodes placed in the detection volume and generally extend through the
whole thickness of the detector. These electrodes are alternately biased as cathodes and anodes, as shown in Fig.
1.
Figure 1: Diagram of a 3D detector
In the 3D geometry, photons are usually absorbed along the thickness of the detector, whereas charge collection
occurs in the perpendicular direction, in a plane parallel to the surface of the detector.
Doubled-sided 3D silicon sensors [2] [3] have been widely studied for high-energy particle detection. The silicon
is processed using Reactive Ion Etching (RIE) [4] and laser drilling [5] of electrodes. 3D silicon structures have
also been studied for medical imaging applications in the 10-60 keV energy range [6]. The silicon detection
layers are usually a few hundred microns thick. However, thicker detection layers are needed to achieve good
absorption above 60 keV. Silicon is thus not suitable for the medical imaging applications targeted here,
involving an energy level of up to 140 keV. Therefore, an alternative material, such as semi-insulating GaAs,
was considered. However, CdTe:Cl was chosen as a proof of concept for 3D-structure development and as a
reference point in terms of achievable performances in comparison to existing CdTe planar/pixelated structures.
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
The detector manufacturing process is described below, from the laser drilling of holes into the CdTe:Cl to the
electroless deposition of Au contacts along the surface of the holes, read-out circuit connection and
characterization under 241Am, 57Co gamma-ray and synchrotron X-ray irradiation.
2.
CdTe:Cl 3D DETECTOR MANUFACTURING
The CdTe:Cl was provided by Acrorad. The following transport properties were considered: µelectrons = 1000
cm²/V/s, µholes = 50 cm²/V/s, τelectrons= τ holes = 1 µs [7], that allow for both electrons and holes collection.
Two samples were made.
Sample 1 aimed at checking the laser drilling technique. It consisted of a 60 000-hole matrix drilled into a 1 mmthick CdTe detector. A hole diameter of 20 µm at a pitch of 100 µm was achieved, i.e., a 50:1 aspect ratio for the
holes.
Sample 2 reduced the constraints on the hole diameter and pitch and was our prototype for radiation detection
characterization. Several arrays of 3x3 holes were drilled in a 1.6 mm thick-CdTe:Cl detector with a hole
diameter of 100 µm at a pitch of 350 µm (chosen after simulation [8]).
2.1. LASER-MACHINING
Semiconductor etching techniques, like dry etching, chemical etching or RIE, are mainly developed for silicon
[9]. In this case, holes were laser drilled. Laser-machining is material independent but is a serial process.
The laser drilling was performed at the CEA-Saclay facility. An Nd-YLF pulsed laser was used, exhibiting the
following parameters:
- Wavelength = 263 nm, enabling a small spot size and small hole diameter
- Average power = 1.95 W
- Rate = 3 kHz
- Pulse width = 41 ns
The holes, through the whole material thickness, were made by laser beam percussion. During this process, the
laser drilling creates a plasma which contains material particles; these particles can be deposited onto the
material surface, near the drilled holes. One hole is drilled in 500 ms.
Sample 1 was observed using an optical microscope and Scanning Electron Microscope (SEM) imaging after Au
contacts had been deposited onto the surface of the holes. The sample was cleaved so that the shape of the holes
could be studied. Images of Sample 1 are presented in Fig. 2.
Figure 2: Sample 1 SEM and optical micrographs of (a) the top surface and (b, c) the thickness of the
detector
Fig.2-a shows a detailed view of the top surface of the holes of the sample 1 obtained using an SEM. The
sample was polished after Au deposition and exhibits no extra fragment material on the surface. Fig.2-b shows
an optical micrograph of the thickness of the detector with contacted holes well-defined over the whole thickness
of the detector. Mechanical stress lines caused by the cleavage of the sample can be observed in between the
holes. The general shape of the holes is slightly conical, a feature arising from the laser drilling technique. It can
also be observed that the Au was properly deposited over the whole surface and the length of the holes. Then,
Sample 1 was polished along the thickness of the detector and characterized by Scanning Electron Microscopy
(Fig.2-c). We can see slightly tapered holes with a hole entrance diameter bigger than the hole exit diameter. A
high repeatability of the laser drilling process can also be inferred.
Sample 2 was also observed under optical microscopy. The holes were of an enhanced conical shape (not shown)
with an entrance diameter of approximately 100 µm and an exit diameter ranging from 35 to 45 µm. The holes
were more tapered due to the larger entrance diameter chosen. The chosen hole diameter and pitch for Sample 2
facilitated connection to the read-out electronics, enabling this sample to be used for initial characterization
under irradiation.
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
2.2. ELECTRODE DEPOSITION AND CONNECTION TO READ-OUT ELECTRONICS
Prior to Au deposition, a chemical etch with Br in methanol was performed to remove the material fragments
created during the laser drilling process. Au contacts were then deposited on the surface of the holes by
electroless deposition from a solution of AuCl3. Finally, the top and bottom surfaces of the sample were polished
mechanically to remove the Au deposited on these surfaces.
Sample 2 was connected to the read-out electronics using gold wires with diameters of 25 µm, inserted manually
into the holes. A suitable electrical contact with acceptable resistivity was expected.
The polarity applied to the wires can be chosen to create different “pixel” arrangements. Fig. 3 shows the two
biasing configurations used in this work. From now on, these will be referred to as “diamond-shaped biasing”
and “square-shaped biasing”. The diamond-shaped biasing arises from connecting the electrodes alternately as
cathodes and anodes. In the square-shaped biasing configuration, all the first neighbors of an anode are cathodes.
The “pixel” surface is thus doubled.
Figure 3: Schematic representation of the detector biasing configurations
2.3. CURRENT-VOLTAGE CHARACTERISTICS
Current – Voltage characteristics were obtained for the square-shaped biased detector, which was tested under
synchrotron X-ray irradiation. The measured values are shown in Fig. 4 (absolute values and logarithmic scale).
Figure 4: Current-Voltage characteristics of the square-shaped biased CdTe:Cl 3D detector
A linear regime is observed in the 0 to -4 V range (inset figure). A numerical resolution method is required to
extract the resistivity of a 3D detection structure. The simulation of the detector in this bias configuration was
performed with COMSOL [11]. The simulated resistivity of 3D intrinsic CdTe:Cl was 6.109 Ω.cm, in agreement
with the measured value [7].
A slope break can be observed in the -4 to -8 V range.
At bias values over -8 V, the leakage current values increased drastically, resulting in values of up to 300 nA at
-20V, the chosen nominal bias. Indeed, the charge carrier transport properties in CdTe require a high bias
voltage.
Such high values of leakage current might arise from crystalline defects created at the surface of the hole during
the laser-drilling process. Damaged areas may be removed with an additional dry or chemical etch. With such
high values of leakage current, good spectrometric detection is unlikely. A noise of 1400 electrons RMS (i.e. a
FWHM of 25%) was expected at 60 keV. Conventional planar or pixelated CdTe detectors are known for their
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
low leakage current. Damage caused by 3D micro-machining is likely to degrade the natural spectroscopic
abilities of CdTe:Cl.
3.
241
Am AND 57Co GAMMA RAY IRRADIATION
3.1. MEASUREMENT SET-UP
The detector was connected to a previously developed spectrometric measurement bench [10] and irradiated with
both 241Am and 57Co non-collimated sources.
A negative bias of 20V was applied to the cathodes, while the signal was collected at the anode. The read-out
electronics consisted of a charge amplifier and a pulse shaper (shaping time = 1µs) with a gain of 12 mV/fC and
a discriminator threshold of 10 keV. The signal amplitude was recorded and binned into a histogram, thus
building the detection spectrum. The spectra obtained are presented in the next section.
3.2. RESULTS
Fig. 5 shows the measured spectra obtained under 241Am irradiation (60 keV incident photon energy and 3.7
MBq activity) and 57Co irradiation (122 keV main incident photons energy and 0.2 MBq activity) with 3D
CdTe:Cl, biased in both diamond-shaped and square-shaped biasing configurations. The bias voltage was -20 V.
Figure 5: 241Am and 57Co spectra measured in both square-shaped and diamond-shaped biasing
configurations for 3D CdTe:Cl detector
A photon-counting ability was revealed in both biasing configurations. The energy discrimination between the
60 keV and the 122 keV peaks was good. For the diamond-shaped detector, the FWHM of 29% (=17 keV) at 60
keV and 35% (=43 keV) at 122 keV were extracted from the spectra. For the square-shaped detector, the FWHM
of 30% (=18 keV) at 60 keV and 37% (=45 keV) at 122 keV were extracted from the spectra. The very large
electronic noise explains the poor resolution. Spectrum tails at 60 keV and 122 keV were also observed. These
tails might have been caused by inhomogeneous regions of charge collection, fluorescence in the CdTe or
photoelectron generation in the electrodes. The difference in the absorption depth in the material between 60 and
122 keV photons might explain the difference in the energy resolution. Given the CdTe absorption cross sections
[12], the 60 keV photons probe mainly the first few hundred microns of the detector whereas the 122 keV
photons probe the whole thickness. This might explain the higher noise in the 122keV spectra
The applied bias was then swept between -5 V and -50 V. For each spectrum, the peak channel and total counts
were extracted. The values obtained are shown in Fig. 6-a and 6-b respectively, as a function of the applied bias
(absolute values). The detector was irradiated with a 241Am source. Both the diamond-shaped and square-shaped
biasing configurations were compared.
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
Figure 6: Peak channel (a) and total counts (b) as a function of applied bias under 241Am irradiation
Fig.6-a shows a stabilization of the peak channel in both bias configurations above -10 V, which indicates full
charge collection. As shown in Fig. 6-a, the square-shaped bias configuration is slightly more efficient than the
diamond-shaped bias configuration. However, the electric field is expected to be more favorable in the diamondshaped configuration. The geometry effect is not explained yet.
Fig.6-b shows a stabilization of the counts number for both bias configurations in the -20 to -50 V range. At -20
V, the ratio of total counts in the square-shaped bias configuration to the total counts in the diamond-shaped bias
configuration is 1.7. This seems to be consistent with the square-shaped “pixel” surface being twice the size of
the diamond-shaped “pixel” surface. But these results are not fully explained. Similar results have been obtained
under 57Co irradiation (not shown).
Photon counting and 241Am / 57Co peak energy discrimination were demonstrated using a non-optimized design
(electrode diameter and pitch have to be optimized). At an equivalent applied bias, the square-shaped bias
configuration was more efficient due to its doubled surface. However, the signal amplitude was similar in both
bias configurations. Low electric field and diffusion-dominated regions were expected in the square-shaped
configurations, resulting in enhanced trapping and reduced charge collection efficiency.
4.
SYNCHROTRON X-RAY IRRADIATION
This experiment aimed to map the detection response of 3D CdTe:Cl detectors in a square-shaped bias
configuration. Highly–focused, bright synchrotron light was the most appropriate for this experiment. The beam
can be collimated down to a few hundred square microns, while maintaining a high photon flux.
The results obtained were compared to simulated efficiency maps [8].
4.1. MEASUREMENT SET-UP
The experiments were performed on the ID6 beamline at the European Synchrotron Radiation Facility (ESRF)
[13].
The beam energy was adjusted to 60 keV (radiography energy). A theoretical absorption of 90% of 60 keV
photons in the first 600 µm of 1.6 mm CdTe:Cl detector is calculated thanks to the absorption cross sections
given by the NIST database [12]. The beam was collimated down to 20x20 µm² with a tungsten parallel double
shutter. The expected input rate was 105 photons/s (not measured onto the detector surface). Two motorized
linear stages, assembled perpendicularly, were used to move the detector during the scan. The read-out
electronics were identical to those used in the previous experiments. The measurement set-up was calibrated by
injecting 2.2 fC into the readout circuit through a 1-pF capacitor with an Agilent 33250 arbitrary waveform
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
generator, what corresponds to the energy deposit by a 60 keV photon. The 3D CdTe:Cl detector was biased at 20V in a square-shaped biasing configuration.
4.2. EXPERIMENTAL RESULTS AND VALIDATION OF THE SIMULATION MODEL
A 32*32-position top-surface scan of the detector at a pitch of 25 µm was performed. At each measurement
point, the spectrum was collected for 10 s. The total count map, shown in Fig. 7, was calculated as the spectrum
integral above 30 keV. Under 30 keV, the counted photons are due to the fluorescence in CdTe.
Figure 7: Total count map of 60 keV X-ray irradiated 3D CdTe:Cl detector
The 3*3 electrode array is clearly visible. The measured pitch between electrodes on the scan map is 350 µm,
corresponding to the actual pitch. A noisy area can be observed between 600 and 800 µm on the vertical scale
and all along the horizontal scale, corresponding to photonic noise induced by the beam line monochromator
instability during this part of the scan. The photonic noise reduced the total number of counts in the spectrum,
while the full energy peak position remained unchanged. Away from this region, incident photon absorption
efficiency is homogeneous over the whole surface of the detector.
For each measured spectrum, the 60 keV peak channel was extracted and was normalized thanks to the
measurement set-up calibration. It corresponds to the experimental Charge Induction Efficiency (CIE). Fig. 8
plots the normalized peak channel for each scan position. The colormap scale ranges from 0 to 1.
Figure 8: Normalized peak channel map. Colormap scale ranging from 0 to 1
The normalized peak channel is greater than 0.8 over the scan surface, and the detector response is homogenous.
The same results are shown in Fig. 9-a (with a colormap scale shrunk to the 0.8 to 1 range) and compared to a
simulated CIE map (Fig. 9-b). 2D simulations of the square-shaped bias configuration detector were obtained
using the finite element method with COMSOL [11]. The CIE was calculated as the ratio of the induced charge
Q to the total deposited charge Q0, in one computation at every point of the simulated surface, using the adjoint
method [8]. A direct comparison was made between the experimental and simulated results, resulting in a firstorder assumption that the normalized peak channel accounts for the experimental CIE. The simulated map
resolution was 1 µm, whereas the experimental resolution was 25 µm. Considering the above-described
approximations, a high level of consistency was observed between the simulated and experimental efficiency
maps.
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
Figure 9: (a) Experimental normalised peak channel map; (b) Simulated CIE map and contour lines for
applied bias. Colormap scale in the 0.8 to 1 range in both figures
Inhomogeneous efficiency regions can be seen on the experimental scan, arising from the chemical and physical
aggression of the material during the hole drilling and connection process. These regions might also stem from a
misalignment of the detector’s beam axis and absorption axis, resulting in a blurred detector efficiency map. The
four-leaf-clover shape around the central anode, which can be observed in both the measured and the simulated
CIE maps, corresponds to the electric field distribution in the square-shaped biased 3D structure. The highest
CIE values can be found around the cathodes in both the simulated and the experimental maps. When photons
impinge close to the cathodes, only electrons contribute to the signal, whereas both electrons and holes
contribute to the signal for photons impinging close to the anode. This feature arises from the fact that the
transport properties of the electrons are better than those of the holes in the CdTe:Cl.
These results enabled the validation of the chosen simulation method. They also showed that detection was
homogenous across the volume of the 3D detector.
5.
CONCLUSION
A first proof of concept for 3D detectors made of a material other than silicon is presented in this paper. 3D
CdTe:Cl single cell prototype detectors were developed and characterized. The technical feasibility of laser
drilling and contacting holes in a 1 to 1.6 mm-thick CdTe:Cl detector was confirmed. High leakage current
values, leading to measurement noise, were recorded. However, photon-counting abilities were exhibited, along
with a good separation of 60 keV and 122 keV photoelectric peaks. The FWHM of 17-18 keV at 60 keV and 4345 keV at 122 keV were measured. Two bias configurations were tested, resulting in square-shaped and
diamond-shaped “pixel” structures. The square-shaped bias configuration resulted in increased counts at an
equivalent bias voltage, due to the double detection surface. However, the measured signal amplitude was not
significantly different between the two bias configurations. Synchrotron X-ray irradiation of the prototype was
also conducted. Homogeneous absorption and charge collection efficiency were revealed, which was consistent
with FEM simulation.
Further researches will involve the realization of a real detector for medical imaging. A parallel process to drill
the holes, like a dry etch or the use of several laser beams, and a better holes connection with metal filling of
holes and bump-bonding are being considered. Semi-insulating GaAs (bulk material) is also considered as an
alternative to CdTe. The transport properties of GaAs do not enable it to be used in conventional planar or
pixelated structures. 3D GaAs detectors could be considered for medical imaging applications.
ACKNOWLEDGEMENTS
We would like to thank Cyril Ponchut and the ID6 line for their help throughout our ESRF experiments.
REFERENCES
[1]
[2]
[3]
[4]
S.I. Parker, C.J. Kenney, J. Segal, “3D - A proposed new architecture for solid-state radiation detectors”, Nucl. Instr.and Meth. Phys.
Res. A, 395 (1997) 328
G. Pellegrini, M. Lozano, M. Ullán, R. Bates, C. Fleta, D. Pennicard, “First double-sided 3-D detectors fabricated at CNM-IMB”, Nucl.
Instr.and Meth. Phys. Res. A, 592 (2008) 38
A. Zoboli, M. Boscardin, L. Bosisio, G.-F.D. Betta, C. Piemonte, S. Ronchin, N. Zorzi, “Double-Sided, Double-Type-Column 3-D
Detectors: Design, Fabrication, and Technology Evaluation”, IEEE Trans. Nucl. Sci. 55(5) (2008) 2775
C. Kenney, S Parker, J. Segal, C. Storment, “Silicon detectors with 3-D electrode arrays: fabrication and initial test results”, IEEE
Trans. Nucl. Sci. 46(4) (1999) 1224
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
M.Chistophersen, B.F.Phlips, “Laser-micromachining for 3D silicon detectors” in: Nuclear Science Symposium Conference Record
IEEE (2010) 378
G. Pellegrini, P.Roy, A.El-Ajili, R.Bates, L.Haddad, M.Horn, K.Mathieson, J.Melone, V.O’Shea, K.M.Smith, I.Thayne, M.Rahman,
“Technology development for high-energy physics and imaging”, Nucl. Instr.and Meth. Phys. Res. A, 487 (2002) 19
E. Gros d’Aillon, M.C. Gentet, G.Montemont, J.Rustique, L.Verger, “Simulation and experimental results on monolithic CdZnTe
gamma-ray detectors”, IEEE Trans. Nucl. Sci. 52 (2005) 3096
M. Ruat, E.Gros d’Aillon, L.Verger, “3D semiconductor radiation detectors for medical imaging: Simulation and design”, in: Nuclear
Science Symposium Conference Record IEEE (2008) 434
E. Kukharenka, M.M.Farooqui, L.Grigore, “Electroplating moulds using dry film thick negative photoresist”, J. Micromech. Microeng.
13 (2003) 67
L.Verger , M.C.Gentet, L.Gerfault, R.Guillemaud, C.Mestais, O.Monnet, G.Petroz, J.P.Rostaing, J.Rustique, “Performance and
perspectives of a CdZnTe-based gamma camera for medical imaging”, IEEE Trans. Nucl. Sci. 51 (2004) 3111
COMSOL commercial simulation tool v.3.5 http://www.comsol.com
http://www.nist.gov/pml/data/xcom/index.cfm
http://www.esrf.eu/Accelerators/Groups/InsertionDevices/Radiations/ID6movies