Radiation-induced point- and cluster-related defects with strong impact on damage properties of silicon detectors

https://doi.org/10.1016/j.nima.2009.09.065Get rights and content

Abstract

This work focuses on the investigation of radiation-induced defects responsible for the degradation of silicon detector performance. Comparative studies of the defects induced by irradiation with 60Co-γ rays, 6 and 15 MeV electrons, 23 GeV protons and reactor neutrons revealed the existence of point defects and cluster-related centers having a strong impact on damage properties of Si diodes. The detailed relation between the “microscopic” reasons as based on defect analysis and their “macroscopic” consequences for detector performance is presented. In particular, it is shown that the changes in the Si device properties (depletion voltage and leakage current) after exposure to high levels of 60Co-γ doses can be completely understood by the microscopically investigated formation of two point defects, a deep acceptor and a shallow donor, both depending strongly on the oxygen concentration in the silicon bulk. Specific for hadron irradiation are the annealing effects which decrease (increase) the originally observed damage effects as seen by the changes of the depletion voltage and these effects are known as “beneficial” and “reverse” annealing, respectively. A group of three cluster-related defects, revealed as deep hole traps, proved to be responsible specifically for the reverse annealing. Their formation is not affected by the oxygen content or silicon growth procedure suggesting that they are complexes of multi-vacancies located inside extended disordered regions.

Introduction

One of the most challenging applications for silicon detectors arises from their use in the inner tracking region of forthcoming colliding beam experiments as, e.g. Large Hadron Collider (LHC at the European research center CERN) and especially its planned upgrade (S-LHC), the International Linear Collider (ILC), or high-brilliance photon sources like the European XFEL foreseen also for the next decade [1], [2], [3]. Segmented silicon sensors (micro-strip and pixel devices) are at present the most precise electronic tracking detectors in high-energy physics experiments (HEP). This paper addresses the understanding of radiation damage effects resulting from the non-ionizing energy loss (NIEL) leading to displacement damage in the silicon bulk. Surface- and interface-related effects caused by ionization are on the other hand of considerable importance only for applications in environments with high X-ray doses (as e.g. in the European XFEL), a topic not addressed here.

For the S-LHC the detectors closest to the beam have to perform at hadron fluences up to several 1016 cm−2 under complex, long-term operation scenarios [4], [5], [6]. The limitations for their practical application in the hadron colliders are caused by irradiation-induced defects leading to changes in the effective doping concentration (Neff) full depletion voltage (Vdep), the reverse current at the depletion voltage (Idep) and the degradation in the charge collection efficiency (CCE) [5], [6], [7], [8], [9], [10], [11], [12], [13], [14]. These device properties are subject to changes not only during irradiation but also during beam-off periods. Especially the long-term annealing effects in Neff increase the initial depletion voltage and therefore they are of extreme importance envisioning extended operational periods of several years. In order to avoid a non-tolerable increase in the depletion voltage, the detectors have to be cooled not only during operation but also for the beam-off periods throughout the entire lifetime of the experiments. It would be of considerable importance if this cold storage could be avoided. One encouraging result for improving the radiation tolerance was obtained by the CERN-RD48 collaboration by performing an oxygen enrichment of float-zone (FZ) wafers used for detector fabrication. This defect engineering attempt was motivated by the assumption that a large O-concentration would inhibit the formation of the V2O-defect thought to be the main reason for the observed change in the effective doping. A cost effective realization of oxygen enrichment up to several 1017 cm−3 was achieved by an in-diffusion of oxygen from the Si–SiO2 interface (during treatments at high temperatures) after wafer oxidation. This procedure is called the DOFZ process (Diffusion Oxygenated Float Zone) and its main benefit is a considerable reduction in damage effects after gamma and charged hadron irradiation [15], [16], [17], [18]. In contrast to charged hadron and γ-irradiation, neutron damage seems to be less dependent on the O-concentration. Although the obtained increase in the radiation tolerance may be sufficient to meet the requirements for LHC, the DOFZ process cannot be a solution for the much more demanding S-LHC application because of the intolerable increase in the depletion voltage up to more than 1000 V and the decrease in the charge collection efficiency (CCE) below the necessary threshold for the readout electronics. Therefore, further efforts for proper defect (and device) engineering leading to a radiation tolerance above the present level are indispensable.

Any promising attempt for radiation hardening of the material as well as improvements by modifying the detector processing will rely on a thorough knowledge of the generation of electrically active defects, which are responsible for the observed changes in the device properties at their operating temperature. This goal is addressed in our work focusing on a detailed investigation of specific radiation-induced defects (point- and cluster-related) and their direct correlation with device performance parameters. Pion-induced damage, dominating in the innermost layers of the tracking area, will result from both isolated point defects (mainly due to Coulomb interactions) and densely packed displacement regions (clusters) caused by energetic recoils from hadron reactions. A separation of both components can be best undertaken by studying damage caused by γ-irradiation (point defects) compared to neutron-induced damage (clustered displacements), while the mixture of both as envisioned in pion damage can be represented by studying the effects after high energetic proton irradiation. In addition several low-MeV electron irradiations proved to be useful for understanding the bridge between pure point defect-related and cluster-dominated defects.

Section snippets

Experimental details and techniques

Several kinds of n-type silicon material, presently discussed as candidates foreseen for use at S-LHC, have been investigated for this purpose:

Float-zone silicon FZ, produced by Wacker Siltronic [19], orientation 〈1 1 1〉, 300 μm thick, resistivity 3–4  cm, effective doping concentration Nd∼1012 cm−3. The standard processed p+-n silicon diodes (labeled as STFZ) have an oxygen content of [O]<1016 cm−3. O-enriched diodes (labeled as DOFZ) were obtained using an in-diffusion of oxygen from the Si–SiO2

Defect properties and detector performance

Many electrically active defects, induced by irradiation, are detected by DLTS and TSC experiments. Most of them (VO, V2, Ci, CiOi, CiCs, IO2) were already investigated in detail and no correlation with the “macroscopic” behavior of the diodes could be established [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46]. The main characteristics of defects, from the electrical point of view, are the emission rates of carriers in the conduction and valence bands given byen,p(T)=cn,p

Point defects, predominantly after γ-irradiation

Knowing the already huge efforts spent on investigating defects after hadron irradiation, with no real success in finding the defects responsible for the detector deterioration, we have decided to start our investigations with the simplest case, i.e. in situations where only point defects are generated. This was achieved by performing irradiation with 60Co-γ rays.

The diodes used for these investigations were processed on high-resistivity silicon oxygen lean (STFZ) and oxygenated (DOFZ) material

Defect studies after hadron irradiation

Contrary to low energetic particles, primarily electrons as secondary particles after γ-irradiation (photo- and Compton-effect), producing mainly point defects, more energetic particles, especially hadrons (pions, protons, neutrons or heavy ions) and to some extent also larger energy electrons (see Section 6) may lead to a cascade of many successive displacements, resulting in the formation of a large diversity of extended defects as densely packed conglomerates (“clusters”) of vacancies and

Studies after electron irradiation

As discussed in Section 2, irradiations with several MeV electrons should already result in the generation of cluster-related defects as seen after hadron irradiations but depending on the energy the ratio between cluster and point defects will vary. In order to test this assumption we reinvestigated some TSC measurements after 6 and 15 MeV electron irradiation. The spectra as seen directly after irradiation (no annealing performed in this case) are shown in Fig. 14a (6 MeV electron irradiation

Summary

Future colliding beam experiments at, e.g. the S-LHC with hadron fluences up to several 1016 cm−2 in the innermost layers of the tracking area will place an unprecedented challenge for the radiation tolerance of silicon sensors not yet met by present day devices. Improvements have been shown to be possible by certain modifications of the silicon material, process technology and operational conditions. But success oriented endeavors need to be based on a detailed understanding of the damage

Acknowledgements

This work was carried out in the frame of the CERN-RD50 collaboration and funded partly by the CiS Hamburg Project under Contract no. SSD 0517/03/05, the BMBF under Contract no. 05HC6GU1 and by the Romanian Ministry of Education and Research under the Core Program, contract 45N/2009. We gratefully thank E. Nossarzewska at ITME, Warsaw, for her professional work in preparing the EPI-diodes and performing the spreading resistance measurements. Likewise the help of A. Barcz at ITE, Warsaw, was

References (72)

  • G. Grübel

    Comptes Rendus Physique

    (2008)
  • M. Moll

    Nucl. Instr. and Meth. Phys. Res. A

    (2006)
  • G. Kramberger

    Nucl. Instr. and Meth. Phys. Res. A

    (2007)
  • E. Fretwurst

    Nucl. Instr. and Meth. Phys. Res. A

    (2005)
  • E. Fretwurst et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2003)
  • E. Tuovinen et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2006)
  • E. Fretwurst et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2005)
  • J. Härkönen et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2004)
  • G. Kramberger et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2007)
  • G. Kramberger et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2005)
  • V. Khomenkov et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2006)
  • Z. Li et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2001)
  • G. Lindström

    Nucl. Instr. and Meth. Phys. Res. A

    (2001)
  • G. Lindström

    Nucl. Instr. and Meth. Phys. Res. A

    (2001)
  • G. Lindström et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2006)
  • M. Huhtinen

    Nucl. Instr. and Meth. Phys. Res. A

    (2002)
  • I. Pintilie et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2006)
  • M. Moll et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (1997)
  • J. Hermansson et al.

    Phys. B: Condens. Matter

    (2001)
  • I. Pintilie et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2003)
  • I. Pintilie et al.

    Phys. B: Condens. Matter

    (2003)
  • E. Fretwurst et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2007)
  • B.C. MacEvoy

    Nucl. Instr. and Meth. Phys. Res. A

    (1997)
  • M. Bruzzi

    Nucl. Instr. and Meth. Phys. Res. A

    (2005)
  • M. Scaringella et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2007)
  • M. Moll et al.

    Nucl. Instr. and Meth. Phys. Res. B

    (2002)
  • F. Gianotti, M.L. Mangano, T. Virdee, Physics potential and experimental challenges of the LHC luminosity upgrade, CERN...
  • M.A. Vos

    Nucl. Instr. and Meth. Phys. Res. A

    (2008)
  • G. Lindström et al.

    Nucl. Instr. and Meth. Phys. Res. A

    (2006)
  • A. Ruzin et al.

    IEEE Trans. Nucl. Sci.

    (1999)
  • Siltronic AG, 81737 Munich,...
  • Okmetic Oyj, 01510 Vantaa,...
  • Institute of Electronic Materials Technology (ITME), 01919 Warsaw,...
  • CiS Research Institute for Micro Sensors and Photovoltaics GmbH, 99099, Erfurt,...
  • A. Barcz, M. Zielinski, E. Nossarzewska, G. Lindström, Appl. Surface Sci. 203–204 (2003) 396; A. Barcz, Institute of...
  • E. Nossarzewska, Institute of Electronic Materials Technology (see [21]), private...
  • Cited by (91)

    • Investigation of high resistivity p-type FZ silicon diodes after <sup>60</sup>Co γ-irradiation

      2024, Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
    • Effect of ionizing radiation on quasi-floating gate transistors

      2023, AEU - International Journal of Electronics and Communications
    • Evaluation of HPK n<sup>+</sup>-p planar pixel sensors for the CMS Phase-2 upgrade

      2023, Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
    • Defect characterization studies on irradiated boron-doped silicon pad diodes and Low Gain Avalanche Detectors

      2023, Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
    • Microscopic origin of the acceptor removal in neutron-irradiated Si detectors - An atomistic simulation study

      2022, Acta Materialia
      Citation Excerpt :

      Besides, similar values of gc have been reported under neutron irradiation independently of substrate resistivity, in both p- and n-type substrates, and for materials with different O contents [47]. Consequently, the increase of Neff with fluence has been attributed to the formation of radiation-induced electrically active intrinsic defects, with a joint contribution of several donor and deep acceptor levels, the latter ones being predominant [12,13]. This supports the idea that in neutron-irradiated p-type Si Neff is determined by two competing mechanisms: an initial acceptor removal at low and medium fluences caused by B deactivation by Si I's, and the formation of deep acceptors at large fluences associated to the accumulation of intrinsic defects.

    • Atomistic simulations of acceptor removal in p-type Si irradiated with neutrons

      2022, Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms
    View all citing articles on Scopus
    View full text