- Email: [email protected]
Development and characterization of a 3D GaAs X-ray detector for medical imaging
Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research A journal homepage: www.elsevier.com/locate/nima
Development and characterization of a 3D GaAs X-ray detector for medical imaging Eric Gros d’Aillon a,n, Marie-Laure Avenel a, Daniel Farcage b, Loïck Verger a a b
CEA, LETI, MINATEC-Campus, 17 rue des Martyrs, F-38054 GRENOBLE, France CEA, DEN-DPC, F-91191 GIF SUR YVETTE, France
art ic l e i nf o
a b s t r a c t
Article history: Received 9 January 2013 Received in revised form 19 March 2013 Accepted 14 June 2013 Available online 22 June 2013
Conventional semiconductor X-ray detectors for medical imaging have either a planar or a pixelated structure. The options available for detection materials are limited by the natural trade-off between the absorption of incident photons and the collection of free charge carriers with these two structures. This trade-off can be avoided by using az 3D structure, in which electrodes are drilled into the detector's volume. This article describes a prototype 3D semiconductor detector, using semi-insulating GaAs. A laser drilling technique was used to create electrodes in the volume of the material. The holes created were characterized by scanning electron microscopy. Electrode contacts were created using electroless Au deposition. The manufacturing process and the first gamma counting results obtained with 241Am and 57 Co sources are presented. The system is capable of individual photon-counting without energy discrimination but requires further development to improve efficiency. & 2013 Elsevier B.V. All rights reserved.
Keywords: 3D Semiconductor X-ray detectors GaAs Laser drilling Medical imaging
1. Introduction Medical imaging systems, such as room temperature X-ray detectors in the 20–100 keV range, extensively use wide band gap semiconductor compounds [1]. Using a planar geometry, the photo-generated carriers must pass through the entire thickness of the detector in order to be collected by the electrodes. Due to the charge carrier transport properties of these materials, the allowable thickness for the sensor layer is limited, thus decreasing potential sensor sensitivity. This limitation could be overcome by using 3D geometry [2,3], where electrodes are drilled into the sensor's thickness. Charge carriers are transported perpendicular to the sensor thickness, as shown in Fig. 1. Most work relating to 3D detectors has been done on silicon sensors [4]. However, silicon is not ideal for X-ray detection in the 20–100 keV range as it has quite low absorption above 25 keV. SiC and GaAs 3D detectors have also been studied by Pellegrini et al. for high energy physic [5] and medical imaging [6] applications. The goal of this study was to develop a prototype 3D-GaAs detector working in counting mode which could be used for digital radiography. For this specific application, typical X-ray tube voltage is 70 kV and the mean X-rays energy is 50 keV. We previously studied the geometry required for this type of detector by numerical simulation [7]. Main conclusion of this study was that, as using planar geometry, the charge carrier (electrons or
n
Corresponding author. Tel.: +33 4 38 78 20 87. E-mail address: [email protected] (E. Gros d’Aillon).
0168-9002/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nima.2013.06.052
holes) with the poorer transport properties limits the device performances, because both carrier types must reach the respective collecting electrodes. The optimal pitch between anodes and cathodes is thus set by the minimum mean drift distance of each charge carrier. This simulation was validated using a 3D semiconductor detector prototype using CdTe:Cl [8], which also served as a proof of concept for developing the 3D-structure. With the same geometry as in [8] with a 1 mm thick GaAs sensor could lead to 80% absorption efficiency at 50 keV, including 5% fill factor loss due to electrodes. The present article describes a 3D X-ray detector composed of semi-insulating (SI) GaAs. Furthermore, we aim to develop a sensor which could be an industrial product, that is why we have chosen to use readily available commercial SI-GaAs and conventional. The semi-insulating GaAs samples are described in the first section; the hole drilling method using laser machining and its characterization are developed in the second section; and the final section details the characteristics of the device. 2. Manufacturing a 3D GaAs detector 2.1. Reasons for choosing GaAs To select the most appropriate material for X-ray photon detection, we focused on three parameters: atomic number, resistivity and charge carrier mobility. Gallium arsenide, with its medium atomic number (31 for Ga and 33 for As), high resistivity (107 Ω cm for SI-GaAs) and high charge carrier mobility appeared
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
127
Table 1 Characteristics of the laser used to drill holes.
Fig. 1. Diagram of a 3D detector.
to be a good candidate. A 1.3 mm thickness of GaAs is necessary to stop 90% of X-rays at 50 keV. Higher Z semi-conductor detectors (such as CdTe, HgI2, PbI2 or TlBr) were also candidates, but GaAs has several advantages over these despite its lower absorption efficiency. These advantages include its wide use in optoelectronic and high-frequency components and its availability as wide surface sections (up to 8″ wafers), which reduces costs. The processing technologies for GaAs have also been extensively developed. However, the lifetime of both electrons and holes in GaAs is very low, probably due to EL2 defects [9], which limit the allowable sensor thickness for an acceptable charge collection efficiency. Several solutions have been proposed to solve this, including chrome doping [10] or the use of thick epitaxial layers [11,12]. For this study, we wished to use readily available materials, we therefore chose SI-GaAs samples from Freiberger (Germany) grown by the Vertical Gradient Freeze (VGF) method. Good spectrometric performances have been obtained using SI-GaAs in previous studies [13], and a 200 μm thick GaAs sensor bonded to a Medipix ASIC [15] has been used for X-ray computing tomography in counting mode [14].
2.2. Hole drilling In our prototype detector, it was necessary to drill high aspect ratio holes though the whole thickness of the sample. Several hole drilling techniques were available to us. Deep Reactive Ion Etching (DRIE) is commonly used to drill semiconductors in the microelectronic and optoelectronic industries. DRIE is highly material dependent, as the chemical species chosen for etching must be selective. Its main advantage is that it allows a large number of holes to be made in a single etch. To create deep holes with straight edges, etching and passivation phases could be alternated (Bosch process). This has previously been done to create 300-μm holes with a 20:1 aspect ratio in a silicon support [16]. However, inductively coupled plasma etching (ICP) [17,18] may be a better option for creating deep holes in GaAs, although this technique has not yet proven to be suitable for creating holes deeper than 200 μm with a high aspect ratio. Other possible techniques include laser drilling, which is material-independent, but is a serial process. Laser-machining in combination with a plasma etch for damage removal has proved to produce an array of holes with an aspect ratio of 100:1 in a 1 mm thick silicon substrate [19]. Previously, we successfully used laser drilling in our laboratory to create a large number of holes with high (50:1) aspect ratio (diameter 20 μm, thickness 1 mm) in cadmium–telluride [8]. For this reason, we chose to use laser drilling to create holes in GaAs in this study. Unfortunately, alternating laser drilling and etch steps was not possible using our set-up. A pulsed laser was used in percussion mode, as detailed in Table 1. The drilling rate was assessed by measuring the hole depth with respect to the number of laser pulses, and was found to be 2.7 μm/shot in the conditions studied.
Type
Pulsed Nd-YLF
Mode Wavelength Pulse energy Rate Pulse width Time to drill a single hole
Percussion in air 263 nm 600 mJ 3 kHz 40 ns 500 ms
Table 2 Characteristics of the two sample batches.
Objective SI-GaAs Thickness Pattern Holes diameter Holes pitch
Batch 1
Batch 2
Drilling characterization VGF from Freiberger 400 mm 3 3 holes matrix Entrance 80 mm, exit 20 mm 350 mm
Device characterization VGF from Freiberger 600 mm 3 3 holes matrix Entrance 120 mm, exit 50 mm 150 mm
2.3. Sample description This study was focused to a single detection cell, with 3 3 holes, although our ultimate goal is to produce a large-area sensor. Two batches of detection cells were drilled, the characteristics of which are shown in Table 2. The first batch was used to tune the laser drilling technique and to check for laser-induced defects. The second batch had wider holes, to simplify bonding, and reduced pitch between holes, to reduce the mean drift distance for charge carriers. The charge carrier mean free path at operation voltage in this material is unknown because the charge carrier life-time has not been measured. This batch was used to manufacture a device. The thickness of the used supports enables an effective absorption of 40 keV photons, but is too thin to provide good quantum efficiency at higher energies. This must be considered as a first step before the thicker samples can be tested. A sample from batch 1 was polished in the direction of slice for observation using a scanning electron microscope (SEM) (Hitachi 4100) (Fig. 2a). The holes showed good reproducibly but had a tapered shape, especially over the first 200 μm, with an entry diameter of 80 μm and an exit diameter of 20 μm. This tapering could be due to the shape of the laser beam or to absorption of the laser by ablated matter. Changing the laser power or the number of shots does not correct the tapering effect. The large entrance hole diameter limits the minimum pitch and reduces the efficiency of the device, at least up to the depth where the holes become smaller. For thicker GaAs devices, (up to 1.3 mm as mentioned in Section 2.1), the process described in [8] could be used. For this study and as a first step, we have chosen to drill larger holes in order to limit the bonding difficulties. Optionally, considering that the tapering is mainly located on the entrance face, a thick sacrificial layer could be used on the entrance surface, and removed after the drilling process. The surface of the holes also displayed micro-cracks (Fig. 2b and c). We attempted to remove these cracks on several GaAs samples by chemical etching with a sulphuric acid–phosphoric acid mix or citric acids. In all cases the cracks were exacerbated by further chemical etching. Laser drilling with nanoseconds shots is a thermal process, therefore giving rise to a thermally and possibly chemically affected area around holes. To analyse such degradations, SEM images with topographical contrast (Secondary Electron mode) and chemical contrast (Back Scattered Electron mode) were
128
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
Fig. 2. Scanning Electron Microscope images of (a) the side of a 3D GaAs sample, (b) the top of holes, and (c) the centre of holes.
Fig. 3. Scanning Electron Microscope view of the side of a 3D GaAs sample. Left panel: topographic contrast (Secondary Electron mode), Right panel: Chemical contrast (Back Scattered Electron mode).
compared (SEM: Hitachi 5500) (Fig. 3). These images revealed no gallium or arsenic deficits as a result of laser machining. To investigate crystallographic degradations, a top view of another sample was observed by Electron Back Scattered Diffraction (EBSD, SEM Oxford/HKL FEG LEO1530) (Fig. 4). This revealed no deviation from the 〈1 0 0〉 direction, indicating that the GaAs single crystal is free from crystalline disorientation after laser machining. Thus, none of our analyses revealed any heat-affected zone on the sample. 3. Device manufacturing and characterization 3.1. Electrode deposition and connection to a read-out electronics Metallic electrodes can be deposited on GaAs by evaporation or cathodic pulverisation of multilayers, such as Ti/Pt/Au, or AuGeNi. For 3D radiation detectors, electrodes must be deposited on the sides of holes throughout their length. To do this, we first tried pulverisation. With this technique, the deposit hardly extends beyond the first 100 μm due to the high aspect ratio of holes. We therefore adopted the same solution as in [8], where the sample was first etched with bromine in methanol before performing electroless deposition of gold contacts from an AuCl3 solution. This contact is not ideally suited for GaAs, but Fig. 5 nevertheless shows that it results in a continuous deposit throughout holes. The top and bottom surfaces of the sample were mechanically polished to remove any deposited Au. After polishing, 50-μm diameter gold wires were manually inserted into holes (Fig. 5, right panel). 3.2. Current–voltage characteristic The current–voltage characteristics of our detector were measured using a Keithley 6517 A electrometer, setting the 8 surrounding
electrodes at a given voltage and measuring the leakage current for the central electrode (Fig. 6, right panel). The fabricated Schottky diode with the used undoped slightly n-type GaAs and the detector structure asymmetry shows a rectifier ability with a reverse branch for positive bias on the surrounding electrodes and a forward branch for negative bias on the surrounding electrodes. Above +20 V the diode seems to breakdown. Below −60 V, the dark current exhibited temporal instabilities, which may be due to the micro-cracks around the top of the holes described above. According to the current–voltage characteristic shown Fig. 6, in order to lower the shot noise, a +20 V bias should have been applied on the surrounding electrodes. However, no gamma photon was counted with these conditions, probably because of the poor hole lifetime in this material. That is why a −60 V bias was applied on the surrounding electrodes for the gamma-ray irradiation detailed below.
3.3. Gamma-ray irradiation A −60 V volt bias was applied to the cathodes, and signal was measured at the anode. The detector was irradiated with two non-collimated gamma sources, 241Am (518 MBq, two main lines: 13.9 keV, 42%, 59.5 keV, 36%) and 57Co (5 MBq, three main lines: 14 keV, 8%, 122 keV, 81%, 136 keV, 10%). The read-out electronics consist of a charge amplifier and a pulse shaper (shaping time: 1 μs), a gain of 2 mV/fC was applied. Data were read from a high pass numeric filter (Krohn-Hite 3945, Butterworth, cutting frequency 100 kHz) and a digital oscilloscope (LeCroy LT374M). Signal amplitude was recorded and binned to produce a histogram, thus producing the detection spectrum
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
129
Fig. 4. Left panel: EBSD of a sample from above. Right panel: zoom on a single hole. The pink colour represents the 〈1 0 0〉 direction. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 5. Left panel: side view of a 3D sample after electrode deposition. Right panel: top view of a device after manual insertion of gold wires.
Fig. 6. Left panel: Current–voltage characteristic of a GaAs 3D detector between −60 V and +20 V. Current is in logartihmic scale.
The sensor is clearly capable of counting photons, but is unable to discriminate between energy sources. The spectra shapes cannot be explained by the shot noise due to leakage current, but could be caused by inefficient charge collection due to poor charge carrier life time or low depletion length near contacts as if only a part of the charge cloud was collected for each photon. Indeed, the gold electroless contact is not really suitable for use with GaAs but was used here because it coated holes more extensively than the other method tested.
4. Conclusion Fig. 7. 241Am and at −60 V.
57
Co spectra measured with the 3D GaAs detector biased
(Fig. 7). Considering the 50 nA dark current, and the 1 μs shaping time, the shot noise due to leakage current is estimated about a 560 electrons rms.
This article describes a proof of concept for 3D detectors made from semi-insulating GaAs, including their development and characterization. Laser machining was used to drill holes with a high aspect ratio throughout the sample's thickness, down to 600 μm. Micro-cracks were apparent near hole sides, and may be responsible for the high dark current measured with the
130
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
prototype. The system was found to have some photon-counting capacity, but was unable to discriminate between energy sources. Production of large-area 3D GaAs X-ray detectors for medical applications will require several stumbling blocks to be overcome: the conical shape of holes must be corrected to reduce the distance between electrodes; a contact with good electric properties and compatible with 3D geometry, must be found. Improvements to hole connection through metal filling of holes and bump-bonding are being considered. References [1] A. Owens, A. Peacock, Compound semiconductor radiation detectors, Nuclear Instruments and Methods in Physics Research Section A 531 (2004) 18–37. [2] S.I. Parker, C.J. Kenney, J. Segal, 3D – A proposed new architecture for solidstate radiation detectors, Nuclear Instruments and Methods in Physics Research Section A 395 (1997) 328. [3] C. Kenney, S. Parker, J. Segal, C. Storment, Silicon detectors with 3-D electrode arrays: fabrication and initial test results, IEEE Transactions on Nuclear Science NS46 (1999) 1224–1236. [4] C. Da Via, M. Boscardin, G.F. Dalla Betta, G. Darbo, C. Fleta, C. Gemme, et al., 3D silicon sensors: design, large area production and quality assurance for the ATLAS IBL pixel detector upgrade, Nuclear Instruments and Methods in Physics Research Section A 694 (2012) 321–330. [5] G. Pellegrini, P. Roy, R. Bates, D. Jones, K. Mathieson, J. Melone, V. O’Shea, K. M. Smith, I. Thayne, P. Thornton, J. Linnros, W. Rodden, M. Rahman, Technology development of 3D detectors for high-energy physics and imaging, Nuclear Instruments and Methods in Physics Research Section A 487 (2002) 19–26. [6] G. Pellegrini, P. Roy, A. Al-Ajili, R. Bates, L. Haddad, M. Horn, K. Mathieson, J. Melone, V. O’Shea, K.M. Smith, I. Thayne, M. Rahman, G. Pellegrini, et al., Technology development of 3D detectors for medical imaging, Nuclear Instruments and Methods in Physics Research Section A 504 (2003) 149–153. [7] M. Ruat, E. Gros d’Aillon, L.Verger, 3D semiconductor radiation detectors for medical imaging: simulation and design, in: Proceeding of the Nuclear Science Symposium Conference Record IEEE, 2008, p. 434. [8] M.L. Avenel, D. Farcage, M. Ruat, L. Verger, E. Gros D’Aillon, Development and characterization of a 3D CdTe:Cl semiconductor detector for medical imaging, Nuclear Instruments and Methods in Physics Research Section A 671 (2012) 144–149.
[9] M. Rogalla, R. Geppert, R. Göppert, M. Hornung, J. Ludwig, Th. Schmid, R. Irsigler, K. Runge, A. Söldner-Rembold, Carrier lifetime under low and high electric field conditions in semi-insulating GaAs, Nuclear Instruments and Methods in Physics Research Section A 410 (1998) p74–p78. [10] G.I. Ayzenshtat, D.L. Budnitskii, D.Yu. Mokeev, V.A. Novikov, E.M. Syresin, O. P. Tolbanov, A.V. Tyazhev, G.A. Shelkov, Gamma-ray detectors based on GaAso Cr4for nanostructural investigations, Russian Physics Journal 51 (10) (2008) 1037–1052. [11] J.C. Bourgoin, A new GaAs material for X-ray imaging, Nuclear Instruments and Methods in Physics Research Section A 460 (2001) 159–164. [12] G.C. Sun, N. Mañez, M. Zazoui, A. Al-Ajili, D.W. Davidson, V. O’Shea, F. Quarati, K.M. Smith, D. Chambellan, O. Gal, Ph. Pillot, M. Lenoir, J.P. Montagne, A. Bchetnia, J.C. Bourgoin, State of the art on epitaxial GaAs detectors, Nuclear Instruments and Methods in Physics Research Section A 546 (2005) 140–147. [13] F. Dubecký, C. Ferrari, D. Korytár, E. Gombia, V. Nečas, Performance of semiinsulating GaAs-based radiation detectors: role of key physical parameters of base materials, Nuclear Instruments and Methods in Physics Research Section A 576 (2007) 27–31. [14] J. Pribil, B. Zat’ko, I. Frollo, F. Dubecký, P. Šcepko, J. Mudron, Quantum Imaging X-ray CT Systems Based on GaAs Radiation Detectors Using Perspective Imaging Reconstruction Techniques, Measurement Science Review 9 (2009) 27–32. [15] S.R. Amendolia, M.G. Bisogni, P. Delogu, M.E. Fantacci, G. Paternoster, V. Rosso and A.Stefanini, Characterization of a mammographic system based on single photon counting pixel arrays coupled to GaAs X-ray detectors, Medical Physics 36 (2009) p1330–p1339. [16] A. Kok, T.E. Hansen, T. Hansen, G.U. Jensen, N. Lietaer, M. Mielnik, P. Storås, High aspect ratio deep RIE for novel 3D radiation sensors in high energy physics applications, in: Proceedings of the Nuclear Science Symposium Conference Record IEEE, 2009, pp. 1623–1627. [17] S. Combrié, S. Bansropun, M. Lecomte, O. Parillaud, S. Cassette, H. Benisty, J. Nagle, Optimization of an inductively coupled plasma etching process of GaInP/GaAs based material for photonic band gap applications, Journal of Vacuum Science and Technology B 23 (2005) 1521–1526. [18] D.S. Rawal, V.R. Agarwal, H.S. Sharma, B.K. Sehgal, R. Muralidharan, H.K. Malik, Study of inductively coupled Cl2/BCl3 plasma process for high etch rate selective etching of via-holes in GaAs, Vacuum 85 (2010) 452–457. [19] M. Chistophersen, B.F. Phlips, Laser-micromachining for 3D silicon detectors in: Proceedings of the Nuclear Science Symposium Conference Record IEEE, 2010, pp.378–381.
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research A journal homepage: www.elsevier.com/locate/nima
Development and characterization of a 3D GaAs X-ray detector for medical imaging Eric Gros d’Aillon a,n, Marie-Laure Avenel a, Daniel Farcage b, Loïck Verger a a b
CEA, LETI, MINATEC-Campus, 17 rue des Martyrs, F-38054 GRENOBLE, France CEA, DEN-DPC, F-91191 GIF SUR YVETTE, France
art ic l e i nf o
a b s t r a c t
Article history: Received 9 January 2013 Received in revised form 19 March 2013 Accepted 14 June 2013 Available online 22 June 2013
Conventional semiconductor X-ray detectors for medical imaging have either a planar or a pixelated structure. The options available for detection materials are limited by the natural trade-off between the absorption of incident photons and the collection of free charge carriers with these two structures. This trade-off can be avoided by using az 3D structure, in which electrodes are drilled into the detector's volume. This article describes a prototype 3D semiconductor detector, using semi-insulating GaAs. A laser drilling technique was used to create electrodes in the volume of the material. The holes created were characterized by scanning electron microscopy. Electrode contacts were created using electroless Au deposition. The manufacturing process and the first gamma counting results obtained with 241Am and 57 Co sources are presented. The system is capable of individual photon-counting without energy discrimination but requires further development to improve efficiency. & 2013 Elsevier B.V. All rights reserved.
Keywords: 3D Semiconductor X-ray detectors GaAs Laser drilling Medical imaging
1. Introduction Medical imaging systems, such as room temperature X-ray detectors in the 20–100 keV range, extensively use wide band gap semiconductor compounds [1]. Using a planar geometry, the photo-generated carriers must pass through the entire thickness of the detector in order to be collected by the electrodes. Due to the charge carrier transport properties of these materials, the allowable thickness for the sensor layer is limited, thus decreasing potential sensor sensitivity. This limitation could be overcome by using 3D geometry [2,3], where electrodes are drilled into the sensor's thickness. Charge carriers are transported perpendicular to the sensor thickness, as shown in Fig. 1. Most work relating to 3D detectors has been done on silicon sensors [4]. However, silicon is not ideal for X-ray detection in the 20–100 keV range as it has quite low absorption above 25 keV. SiC and GaAs 3D detectors have also been studied by Pellegrini et al. for high energy physic [5] and medical imaging [6] applications. The goal of this study was to develop a prototype 3D-GaAs detector working in counting mode which could be used for digital radiography. For this specific application, typical X-ray tube voltage is 70 kV and the mean X-rays energy is 50 keV. We previously studied the geometry required for this type of detector by numerical simulation [7]. Main conclusion of this study was that, as using planar geometry, the charge carrier (electrons or
n
Corresponding author. Tel.: +33 4 38 78 20 87. E-mail address: [email protected] (E. Gros d’Aillon).
0168-9002/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nima.2013.06.052
holes) with the poorer transport properties limits the device performances, because both carrier types must reach the respective collecting electrodes. The optimal pitch between anodes and cathodes is thus set by the minimum mean drift distance of each charge carrier. This simulation was validated using a 3D semiconductor detector prototype using CdTe:Cl [8], which also served as a proof of concept for developing the 3D-structure. With the same geometry as in [8] with a 1 mm thick GaAs sensor could lead to 80% absorption efficiency at 50 keV, including 5% fill factor loss due to electrodes. The present article describes a 3D X-ray detector composed of semi-insulating (SI) GaAs. Furthermore, we aim to develop a sensor which could be an industrial product, that is why we have chosen to use readily available commercial SI-GaAs and conventional. The semi-insulating GaAs samples are described in the first section; the hole drilling method using laser machining and its characterization are developed in the second section; and the final section details the characteristics of the device. 2. Manufacturing a 3D GaAs detector 2.1. Reasons for choosing GaAs To select the most appropriate material for X-ray photon detection, we focused on three parameters: atomic number, resistivity and charge carrier mobility. Gallium arsenide, with its medium atomic number (31 for Ga and 33 for As), high resistivity (107 Ω cm for SI-GaAs) and high charge carrier mobility appeared
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
127
Table 1 Characteristics of the laser used to drill holes.
Fig. 1. Diagram of a 3D detector.
to be a good candidate. A 1.3 mm thickness of GaAs is necessary to stop 90% of X-rays at 50 keV. Higher Z semi-conductor detectors (such as CdTe, HgI2, PbI2 or TlBr) were also candidates, but GaAs has several advantages over these despite its lower absorption efficiency. These advantages include its wide use in optoelectronic and high-frequency components and its availability as wide surface sections (up to 8″ wafers), which reduces costs. The processing technologies for GaAs have also been extensively developed. However, the lifetime of both electrons and holes in GaAs is very low, probably due to EL2 defects [9], which limit the allowable sensor thickness for an acceptable charge collection efficiency. Several solutions have been proposed to solve this, including chrome doping [10] or the use of thick epitaxial layers [11,12]. For this study, we wished to use readily available materials, we therefore chose SI-GaAs samples from Freiberger (Germany) grown by the Vertical Gradient Freeze (VGF) method. Good spectrometric performances have been obtained using SI-GaAs in previous studies [13], and a 200 μm thick GaAs sensor bonded to a Medipix ASIC [15] has been used for X-ray computing tomography in counting mode [14].
2.2. Hole drilling In our prototype detector, it was necessary to drill high aspect ratio holes though the whole thickness of the sample. Several hole drilling techniques were available to us. Deep Reactive Ion Etching (DRIE) is commonly used to drill semiconductors in the microelectronic and optoelectronic industries. DRIE is highly material dependent, as the chemical species chosen for etching must be selective. Its main advantage is that it allows a large number of holes to be made in a single etch. To create deep holes with straight edges, etching and passivation phases could be alternated (Bosch process). This has previously been done to create 300-μm holes with a 20:1 aspect ratio in a silicon support [16]. However, inductively coupled plasma etching (ICP) [17,18] may be a better option for creating deep holes in GaAs, although this technique has not yet proven to be suitable for creating holes deeper than 200 μm with a high aspect ratio. Other possible techniques include laser drilling, which is material-independent, but is a serial process. Laser-machining in combination with a plasma etch for damage removal has proved to produce an array of holes with an aspect ratio of 100:1 in a 1 mm thick silicon substrate [19]. Previously, we successfully used laser drilling in our laboratory to create a large number of holes with high (50:1) aspect ratio (diameter 20 μm, thickness 1 mm) in cadmium–telluride [8]. For this reason, we chose to use laser drilling to create holes in GaAs in this study. Unfortunately, alternating laser drilling and etch steps was not possible using our set-up. A pulsed laser was used in percussion mode, as detailed in Table 1. The drilling rate was assessed by measuring the hole depth with respect to the number of laser pulses, and was found to be 2.7 μm/shot in the conditions studied.
Type
Pulsed Nd-YLF
Mode Wavelength Pulse energy Rate Pulse width Time to drill a single hole
Percussion in air 263 nm 600 mJ 3 kHz 40 ns 500 ms
Table 2 Characteristics of the two sample batches.
Objective SI-GaAs Thickness Pattern Holes diameter Holes pitch
Batch 1
Batch 2
Drilling characterization VGF from Freiberger 400 mm 3 3 holes matrix Entrance 80 mm, exit 20 mm 350 mm
Device characterization VGF from Freiberger 600 mm 3 3 holes matrix Entrance 120 mm, exit 50 mm 150 mm
2.3. Sample description This study was focused to a single detection cell, with 3 3 holes, although our ultimate goal is to produce a large-area sensor. Two batches of detection cells were drilled, the characteristics of which are shown in Table 2. The first batch was used to tune the laser drilling technique and to check for laser-induced defects. The second batch had wider holes, to simplify bonding, and reduced pitch between holes, to reduce the mean drift distance for charge carriers. The charge carrier mean free path at operation voltage in this material is unknown because the charge carrier life-time has not been measured. This batch was used to manufacture a device. The thickness of the used supports enables an effective absorption of 40 keV photons, but is too thin to provide good quantum efficiency at higher energies. This must be considered as a first step before the thicker samples can be tested. A sample from batch 1 was polished in the direction of slice for observation using a scanning electron microscope (SEM) (Hitachi 4100) (Fig. 2a). The holes showed good reproducibly but had a tapered shape, especially over the first 200 μm, with an entry diameter of 80 μm and an exit diameter of 20 μm. This tapering could be due to the shape of the laser beam or to absorption of the laser by ablated matter. Changing the laser power or the number of shots does not correct the tapering effect. The large entrance hole diameter limits the minimum pitch and reduces the efficiency of the device, at least up to the depth where the holes become smaller. For thicker GaAs devices, (up to 1.3 mm as mentioned in Section 2.1), the process described in [8] could be used. For this study and as a first step, we have chosen to drill larger holes in order to limit the bonding difficulties. Optionally, considering that the tapering is mainly located on the entrance face, a thick sacrificial layer could be used on the entrance surface, and removed after the drilling process. The surface of the holes also displayed micro-cracks (Fig. 2b and c). We attempted to remove these cracks on several GaAs samples by chemical etching with a sulphuric acid–phosphoric acid mix or citric acids. In all cases the cracks were exacerbated by further chemical etching. Laser drilling with nanoseconds shots is a thermal process, therefore giving rise to a thermally and possibly chemically affected area around holes. To analyse such degradations, SEM images with topographical contrast (Secondary Electron mode) and chemical contrast (Back Scattered Electron mode) were
128
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
Fig. 2. Scanning Electron Microscope images of (a) the side of a 3D GaAs sample, (b) the top of holes, and (c) the centre of holes.
Fig. 3. Scanning Electron Microscope view of the side of a 3D GaAs sample. Left panel: topographic contrast (Secondary Electron mode), Right panel: Chemical contrast (Back Scattered Electron mode).
compared (SEM: Hitachi 5500) (Fig. 3). These images revealed no gallium or arsenic deficits as a result of laser machining. To investigate crystallographic degradations, a top view of another sample was observed by Electron Back Scattered Diffraction (EBSD, SEM Oxford/HKL FEG LEO1530) (Fig. 4). This revealed no deviation from the 〈1 0 0〉 direction, indicating that the GaAs single crystal is free from crystalline disorientation after laser machining. Thus, none of our analyses revealed any heat-affected zone on the sample. 3. Device manufacturing and characterization 3.1. Electrode deposition and connection to a read-out electronics Metallic electrodes can be deposited on GaAs by evaporation or cathodic pulverisation of multilayers, such as Ti/Pt/Au, or AuGeNi. For 3D radiation detectors, electrodes must be deposited on the sides of holes throughout their length. To do this, we first tried pulverisation. With this technique, the deposit hardly extends beyond the first 100 μm due to the high aspect ratio of holes. We therefore adopted the same solution as in [8], where the sample was first etched with bromine in methanol before performing electroless deposition of gold contacts from an AuCl3 solution. This contact is not ideally suited for GaAs, but Fig. 5 nevertheless shows that it results in a continuous deposit throughout holes. The top and bottom surfaces of the sample were mechanically polished to remove any deposited Au. After polishing, 50-μm diameter gold wires were manually inserted into holes (Fig. 5, right panel). 3.2. Current–voltage characteristic The current–voltage characteristics of our detector were measured using a Keithley 6517 A electrometer, setting the 8 surrounding
electrodes at a given voltage and measuring the leakage current for the central electrode (Fig. 6, right panel). The fabricated Schottky diode with the used undoped slightly n-type GaAs and the detector structure asymmetry shows a rectifier ability with a reverse branch for positive bias on the surrounding electrodes and a forward branch for negative bias on the surrounding electrodes. Above +20 V the diode seems to breakdown. Below −60 V, the dark current exhibited temporal instabilities, which may be due to the micro-cracks around the top of the holes described above. According to the current–voltage characteristic shown Fig. 6, in order to lower the shot noise, a +20 V bias should have been applied on the surrounding electrodes. However, no gamma photon was counted with these conditions, probably because of the poor hole lifetime in this material. That is why a −60 V bias was applied on the surrounding electrodes for the gamma-ray irradiation detailed below.
3.3. Gamma-ray irradiation A −60 V volt bias was applied to the cathodes, and signal was measured at the anode. The detector was irradiated with two non-collimated gamma sources, 241Am (518 MBq, two main lines: 13.9 keV, 42%, 59.5 keV, 36%) and 57Co (5 MBq, three main lines: 14 keV, 8%, 122 keV, 81%, 136 keV, 10%). The read-out electronics consist of a charge amplifier and a pulse shaper (shaping time: 1 μs), a gain of 2 mV/fC was applied. Data were read from a high pass numeric filter (Krohn-Hite 3945, Butterworth, cutting frequency 100 kHz) and a digital oscilloscope (LeCroy LT374M). Signal amplitude was recorded and binned to produce a histogram, thus producing the detection spectrum
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
129
Fig. 4. Left panel: EBSD of a sample from above. Right panel: zoom on a single hole. The pink colour represents the 〈1 0 0〉 direction. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 5. Left panel: side view of a 3D sample after electrode deposition. Right panel: top view of a device after manual insertion of gold wires.
Fig. 6. Left panel: Current–voltage characteristic of a GaAs 3D detector between −60 V and +20 V. Current is in logartihmic scale.
The sensor is clearly capable of counting photons, but is unable to discriminate between energy sources. The spectra shapes cannot be explained by the shot noise due to leakage current, but could be caused by inefficient charge collection due to poor charge carrier life time or low depletion length near contacts as if only a part of the charge cloud was collected for each photon. Indeed, the gold electroless contact is not really suitable for use with GaAs but was used here because it coated holes more extensively than the other method tested.
4. Conclusion Fig. 7. 241Am and at −60 V.
57
Co spectra measured with the 3D GaAs detector biased
(Fig. 7). Considering the 50 nA dark current, and the 1 μs shaping time, the shot noise due to leakage current is estimated about a 560 electrons rms.
This article describes a proof of concept for 3D detectors made from semi-insulating GaAs, including their development and characterization. Laser machining was used to drill holes with a high aspect ratio throughout the sample's thickness, down to 600 μm. Micro-cracks were apparent near hole sides, and may be responsible for the high dark current measured with the
130
E. Gros d’Aillon et al. / Nuclear Instruments and Methods in Physics Research A 727 (2013) 126–130
prototype. The system was found to have some photon-counting capacity, but was unable to discriminate between energy sources. Production of large-area 3D GaAs X-ray detectors for medical applications will require several stumbling blocks to be overcome: the conical shape of holes must be corrected to reduce the distance between electrodes; a contact with good electric properties and compatible with 3D geometry, must be found. Improvements to hole connection through metal filling of holes and bump-bonding are being considered. References [1] A. Owens, A. Peacock, Compound semiconductor radiation detectors, Nuclear Instruments and Methods in Physics Research Section A 531 (2004) 18–37. [2] S.I. Parker, C.J. Kenney, J. Segal, 3D – A proposed new architecture for solidstate radiation detectors, Nuclear Instruments and Methods in Physics Research Section A 395 (1997) 328. [3] C. Kenney, S. Parker, J. Segal, C. Storment, Silicon detectors with 3-D electrode arrays: fabrication and initial test results, IEEE Transactions on Nuclear Science NS46 (1999) 1224–1236. [4] C. Da Via, M. Boscardin, G.F. Dalla Betta, G. Darbo, C. Fleta, C. Gemme, et al., 3D silicon sensors: design, large area production and quality assurance for the ATLAS IBL pixel detector upgrade, Nuclear Instruments and Methods in Physics Research Section A 694 (2012) 321–330. [5] G. Pellegrini, P. Roy, R. Bates, D. Jones, K. Mathieson, J. Melone, V. O’Shea, K. M. Smith, I. Thayne, P. Thornton, J. Linnros, W. Rodden, M. Rahman, Technology development of 3D detectors for high-energy physics and imaging, Nuclear Instruments and Methods in Physics Research Section A 487 (2002) 19–26. [6] G. Pellegrini, P. Roy, A. Al-Ajili, R. Bates, L. Haddad, M. Horn, K. Mathieson, J. Melone, V. O’Shea, K.M. Smith, I. Thayne, M. Rahman, G. Pellegrini, et al., Technology development of 3D detectors for medical imaging, Nuclear Instruments and Methods in Physics Research Section A 504 (2003) 149–153. [7] M. Ruat, E. Gros d’Aillon, L.Verger, 3D semiconductor radiation detectors for medical imaging: simulation and design, in: Proceeding of the Nuclear Science Symposium Conference Record IEEE, 2008, p. 434. [8] M.L. Avenel, D. Farcage, M. Ruat, L. Verger, E. Gros D’Aillon, Development and characterization of a 3D CdTe:Cl semiconductor detector for medical imaging, Nuclear Instruments and Methods in Physics Research Section A 671 (2012) 144–149.
[9] M. Rogalla, R. Geppert, R. Göppert, M. Hornung, J. Ludwig, Th. Schmid, R. Irsigler, K. Runge, A. Söldner-Rembold, Carrier lifetime under low and high electric field conditions in semi-insulating GaAs, Nuclear Instruments and Methods in Physics Research Section A 410 (1998) p74–p78. [10] G.I. Ayzenshtat, D.L. Budnitskii, D.Yu. Mokeev, V.A. Novikov, E.M. Syresin, O. P. Tolbanov, A.V. Tyazhev, G.A. Shelkov, Gamma-ray detectors based on GaAso Cr4for nanostructural investigations, Russian Physics Journal 51 (10) (2008) 1037–1052. [11] J.C. Bourgoin, A new GaAs material for X-ray imaging, Nuclear Instruments and Methods in Physics Research Section A 460 (2001) 159–164. [12] G.C. Sun, N. Mañez, M. Zazoui, A. Al-Ajili, D.W. Davidson, V. O’Shea, F. Quarati, K.M. Smith, D. Chambellan, O. Gal, Ph. Pillot, M. Lenoir, J.P. Montagne, A. Bchetnia, J.C. Bourgoin, State of the art on epitaxial GaAs detectors, Nuclear Instruments and Methods in Physics Research Section A 546 (2005) 140–147. [13] F. Dubecký, C. Ferrari, D. Korytár, E. Gombia, V. Nečas, Performance of semiinsulating GaAs-based radiation detectors: role of key physical parameters of base materials, Nuclear Instruments and Methods in Physics Research Section A 576 (2007) 27–31. [14] J. Pribil, B. Zat’ko, I. Frollo, F. Dubecký, P. Šcepko, J. Mudron, Quantum Imaging X-ray CT Systems Based on GaAs Radiation Detectors Using Perspective Imaging Reconstruction Techniques, Measurement Science Review 9 (2009) 27–32. [15] S.R. Amendolia, M.G. Bisogni, P. Delogu, M.E. Fantacci, G. Paternoster, V. Rosso and A.Stefanini, Characterization of a mammographic system based on single photon counting pixel arrays coupled to GaAs X-ray detectors, Medical Physics 36 (2009) p1330–p1339. [16] A. Kok, T.E. Hansen, T. Hansen, G.U. Jensen, N. Lietaer, M. Mielnik, P. Storås, High aspect ratio deep RIE for novel 3D radiation sensors in high energy physics applications, in: Proceedings of the Nuclear Science Symposium Conference Record IEEE, 2009, pp. 1623–1627. [17] S. Combrié, S. Bansropun, M. Lecomte, O. Parillaud, S. Cassette, H. Benisty, J. Nagle, Optimization of an inductively coupled plasma etching process of GaInP/GaAs based material for photonic band gap applications, Journal of Vacuum Science and Technology B 23 (2005) 1521–1526. [18] D.S. Rawal, V.R. Agarwal, H.S. Sharma, B.K. Sehgal, R. Muralidharan, H.K. Malik, Study of inductively coupled Cl2/BCl3 plasma process for high etch rate selective etching of via-holes in GaAs, Vacuum 85 (2010) 452–457. [19] M. Chistophersen, B.F. Phlips, Laser-micromachining for 3D silicon detectors in: Proceedings of the Nuclear Science Symposium Conference Record IEEE, 2010, pp.378–381.