Study of Radiation-Induced Defects in p-Type Si1−xGex Diodes before and after Annealing

In this work, electrically active defects of pristine and 5.5 MeV electron irradiated p-type silicon–germanium (Si1−xGex)-based diodes were examined by combining regular capacitance deep-level transient spectroscopy (C-DLTS) and Laplace DLTS (L-DLTS) techniques. The p-type SiGe alloys with slightly different Ge contents were examined. It was deduced from C-DLTS and L-DLTS spectra that the carbon/oxygen-associated complexes prevailed in the pristine Si0.949Ge0.051 alloys. Irradiation with 5.5 MeV electrons led to a considerable change in the DLT spectrum containing up to seven spectral peaks due to the introduction of radiation defects. These defects were identified using activation energy values reported in the literature. The double interstitial and oxygen complexes and the vacancy, di-vacancy and tri-vacancy ascribed traps were revealed in the irradiated samples. The interstitial carbon and the metastable as well as stable forms of carbon–oxygen (CiOi* and CiOi) complexes were also identified for the electron-irradiated SiGe alloys. It was found that the unstable form of the carbon–oxygen complex became a stable complex in the irradiated and the subsequently annealed (at 125 °C) SiGe samples. The activation energy shifts in the radiation-induced deep traps to lower values were defined when increasing Ge content in the SiGe alloy.


Introduction
Silicon-germanium alloys are promising materials for the fabrication of photocells and powering space applications [1]. This alloy is also employed in the production of high-frequency heterojunction bipolar transistors for operation in the near THz range [2]. Silicon-germanium provides a novel approach to the formation of high-conversion efficiency and highly scalable thermoelectric materials. Silicon-germanium alloys have recently been reported [3] to function well as lithium-ion battery anodes. This alloy is also prospective for the fabrication of microelectronic and optoelectronic devices such as high-speed temperature sensors, Hall effect transducers and γ-ray detectors [4,5]. Therefore, the spectrum of carrier traps is a desirable characteristic for material quality evaluation.
Silicon-germanium material-based devices are capable of operating in harsh radiation environments [6,7]. Silicon-germanium-based pixel detectors with enhanced radiation tolerance are promising for applications in the future High-Luminosity Large Hadron Collider [8]. However, there are difficulties in growing bulk SiGe single-crystals due to the differences in the physical properties of silicon and germanium such as density and melting temperature. For example, single crystals In this work, the pristine and electron-irradiated Si 1−x Ge x diodes with an n + p structure were examined. The diodes were fabricated using SiGe substrates grown using the Czochralski technique. The diode basis was formed from the p-type material (doped with boron), containing either 1%, 1.4% or 5.1% of Ge. For comparison, the diodes made of pure Si used the same (as the SiGe alloys) boron-doping parameters. Irradiation with 5.5 MeV electrons was performed at room temperature using a linear accelerator with electron fluxes of 2 × 10 12 cm −2 s −1 . The Si as well as the Si 0.99 Ge 0.01 alloy diodes were irradiated with fluence of 2 × 10 15 cm −2 . The Si 0.986 Ge 0.014 and Si 0.949 Ge 0.051 alloy-based diodes were irradiated with fluences of 5 × 10 13 cm −2 and 2 × 10 14 cm −2 , respectively. The irradiated samples were consequently annealed at 125 • C for 15 min to investigate the changes of DLT spectra under heat treatment.
The DLT spectra were recorded using a commercial HERA-DLTS 1030 instrument (PhysTech GmbH, Moosburg an der Isar, Upper Bavaria, Germany). The DLTS measurements were performed using a routine C-DLTS regime. These DLT spectra were examined in the temperature range of 15-280 K. The majority carrier trap spectra were recorded at reverse bias voltage (U R ) of 3 V and injection pulses (t p ) of 10 ms duration. Each spectrum was analysed by combining correlation functions and the Laplace method.

Recorded DLT Spectra and Extracted Trap Parameters
Up to seven spectral peaks (assigned to the E 1 -E 7 traps, as illustrated in Figure 1b) were observed within C-DLT spectra recorded on the 5.5 MeV electron irradiated Si diode when using a fluence of Φ = 2 × 10 14 e/cm 2 . Figure 1a shows the barrier capacitance changes with temperature (C b -T) in the 5.5 MeV electron-irradiated and subsequently annealed Si samples. It can be noticed in Figure 1a that an onset within the C b -T curves was obtained for the as-irradiated and subsequently annealed Si samples. The shift of the onset may have appeared due to the irradiation-and annealing-induced transformations and density variations of carrier trap species, which caused freezing of carriers within the temperature range under consideration [19,21]. A few peaks (for instance, E 6 ) changed their position under annealing relatively to an abscise scale implying the intricate transformation of traps assigned to this spectral peak. Such a spectral range was carefully examined (Figure 1d) using routine and Laplace transform DLTS (L-DLTS). It was clarified that the E 6 peak can be composed of two peaks E 6-1 and E 6-2 , just after irradiation. These peaks were ascribed to traps with slightly different activation energies, and their values can be evaluated using Arrhenius plots (as shown within inset (i) for Figure 1d). The concentration and activation energy of traps attributed to the E 6-2 peak increased after annealing as can be deduced from Figure 1b,d. It is worth mentioning that only the E 6-2 peak remained after annealing (instead of the E 6-1 and E 6-2 as well as E 5 peaks), and its amplitude was close to the sum of the E 6-1 and E 6-2 as well as E 5 peaks before annealing. This implies that the changes in the E 5 and E 6-1 as well as E 6-2 spectral peaks actually represented transformations of the defects due to the annealing.
within the temperature range under consideration [19,21]. A few peaks (for instance, E6) changed their position under annealing relatively to an abscise scale implying the intricate transformation of traps assigned to this spectral peak. Such a spectral range was carefully examined (Figure 1d) using routine and Laplace transform DLTS (L-DLTS). It was clarified that the E6 peak can be composed of two peaks E6-1 and E6-2, just after irradiation. These peaks were ascribed to traps with slightly different activation energies, and their values can be evaluated using Arrhenius plots (as shown within inset (i) for Figure 1d). The concentration and activation energy of traps attributed to the E6-2 peak increased after annealing as can be deduced from Figure 1b,d. It is worth mentioning that only the E6-2 peak remained after annealing (instead of the E6-1 and E6-2 as well as E5 peaks), and its amplitude was close to the sum of the E6-1 and E6-2 as well as E5 peaks before annealing. This implies that the changes in the E5 and E6-1 as well as E6-2 spectral peaks actually represented transformations of the defects due to the annealing.
The evolution of radiation defects in p-type Si introduced by electron beam is rather well understood [18,22]. Thereby, identification of the most resolved traps in Si can be reliably implemented based on activation energy values reported in the literature. Parameters for all the identified Si traps are presented in Table 1. The trap with the activation energy of 0.080 eV (E1 in Table 1) is attributed to the double interstitial and oxygen (I2O) complex [18]. The 0.100 eV (E2) level can be assigned to a triple vacancy (V3) [18]. The origin of the E3 trap is not clear, however, it might be related to vacancy (V) [23]. The The highlighted spectral range inherent for the E 5 , E 6-1 and E 6-2 trap appearances. Inset (i) Arrhenius plots for traps E 6-1 and E 6-2 . Here, τ denotes the carrier lifetime relative to emission; υ th is the carrier thermal velocity, and N V stands for the effective density of hole states in the valence band.
The evolution of radiation defects in p-type Si introduced by electron beam is rather well understood [18,22]. Thereby, identification of the most resolved traps in Si can be reliably implemented based on activation energy values reported in the literature. Parameters for all the identified Si traps are presented in Table 1.
The trap with the activation energy of 0.080 eV (E 1 in Table 1) is attributed to the double interstitial and oxygen (I 2 O) complex [18]. The 0.100 eV (E 2 ) level can be assigned to a triple vacancy (V 3 ) [18]. The origin of the E 3 trap is not clear, however, it might be related to vacancy (V) [23]. The trap with Materials 2020, 13, 5684 4 of 10 activation energy of 0.190 eV (E 4 ) is associated with V 2 + V 3 complex [18,22]. The trap with activation energy of 0.285 eV (E 5 ) is associated with a carbon interstitial (C i ) [18], while the close energy duplet of 0.360 (E 6-1 ) and 0.371 (E 6-2 ) are attributed to the metastable and stable forms of the carbon-oxygen (C i O i * and C i O i ) complexes [22,24], respectively. After 15 min annealing at 125 • C, the unstable form of the carbon-oxygen complex seems to become the stable complex according to reactions [22]: The density of traps ascribed to the stable form of the carbon-oxygen (E 6-2 ) complex should hold the density of constituents, represented by the sum of E 5 , E 6-1 and E 6-2 DLTS peaks before annealing. The alternative sequence of reactions in formation of the stable E 6-2 complex would be as follows [22]: The annealing out of E 1 and E 5 traps together with an increase in the amplitude of the E 6-2 peak (Figure 1b) supports the predicted sequences of the reactions denoted in Equations (1) and (2).
The slight addition of Ge to get the Si 0.99 Ge 0.01 alloy should not drastically modify the spectrum of Si radiation defects introduced using the same irradiation conditions (Φ = 2 × 10 15 cm −2 ). Indeed, the structure of the DLT spectrum (illustrated in Figure 2a) and its changes after annealing of the electron-irradiated Si 0.99 Ge 0.01 diode resembled that obtained in Figure 1. Other parameters for all the revealed traps in the Si 0.99 Ge 0.01 alloy are presented in Table 2. These results indicate that values of the activation energy, ascribed to the E 1 -E 7 traps in Figure 1, are shifted to the low-energy range relative to those obtained for the Si of the same type and doping level.
Another slight addition of Ge to obtain the Si 0.986 Ge 0.014 alloy should further modify the spectrum of Si radiation defects. Figure 2b shows the barrier capacitance changes with temperature (C b -T) in the 5.5 MeV electron-irradiated and subsequently annealed Si 0.986 Ge 0.014 samples. However, together with the onset within the C b -T curves, observed in Figure 1a, variations of the slope of the C b -T curves can be noticed for the pristine, as-irradiated and subsequently annealed Si 0.986 Ge 0.014 samples. These changes in the C b -T curve onsets and slopes can be ascribed to irradiation and annealing-induced transformations and various trap species density variations that cause freezing of carriers within the temperature range under consideration [19,21]. Indeed, the structure of the DLT spectrum ( Figure 2c) and its changes after annealing of the electron irradiated Si 0.986 Ge 0.014 diode resembles that obtained in Figure 1b. Moreover, it was observed that the E 3 peak was composed of two peaks E 3-1 and E 3-2 , as obtained after irradiation of the Si 0.986 Ge 0.014 alloy diode. The latter (E 3-1 and E 3-2 ) peaks were ascribed to traps with slightly different activation energies, which had values that were determined by using Arrhenius plots (Figure 2d). The E 3-1 and E 3-2 peaks disappeared after annealing, thereby indicating the transformation of vacancy related defects.   The L-DLTS technique was additionally employed to separate the E 6-1 and E 6-2 traps more precisely. The trap activation energy values were evaluated using Arrhenius plots (Figure 2d). Parameters for all the unveiled traps are presented in Table 2. However, values of the activation energy ascribed to the E 1 -E 7 traps, illustrated in Figure 2c and listed in Table 2, were close to those extracted from the Si 0.99 Ge 0.01 spectra (Figure 2a). However, these activation energy values were slightly different from those obtained for the Si diodes. The activation energy values extracted for Si 0.986 Ge 0.014 diodes were again shifted to the low-energy range relative to those obtained for Si of the same conductivity type.
The rather different DLTS characteristics (relative to those of pure Si as well as of 1% and 1.4% Ge-containing SiGe alloy) were obtained for the 5.5 MeV electron-irradiated and subsequently annealed Si 0.949 Ge 0.051 material diodes. Figure 3a illustrates the barrier capacitance changes with temperature (C b -T) in the pristine, in the 5.5 MeV electron-irradiated and the subsequently annealed Si 0.949 Ge 0.051 samples. The change in the slope of the C b -T curves was inherent for all the pristine, the as-irradiated and the subsequently annealed samples. Again, the onsets within the C b -T curves and slightly different slopes seem to appear due to the irradiation-and annealing-induced transformations and various traps species density variations [19,21]. The L-DLTS technique was additionally employed to separate the E6-1 and E6-2 traps more precisely. The trap activation energy values were evaluated using Arrhenius plots (Figure 2d). Parameters for all the unveiled traps are presented in Table 2. However, values of the activation energy ascribed to the E1-E7 traps, illustrated in Figure 2c and listed in Table 2, were close to those extracted from the Si0.99Ge0.01 spectra (Figure 2a). However, these activation energy values were slightly different from those obtained for the Si diodes. The activation energy values extracted for Si0.986Ge0.014 diodes were again shifted to the low-energy range relative to those obtained for Si of the same conductivity type.
The rather different DLTS characteristics (relative to those of pure Si as well as of 1% and 1.4% Ge-containing SiGe alloy) were obtained for the 5.5 MeV electron-irradiated and subsequently annealed Si0.949Ge0.051 material diodes. Figure 3a illustrates the barrier capacitance changes with temperature (Cb-T) in the pristine, in the 5.5 MeV electron-irradiated and the subsequently annealed Si0.949Ge0.051 samples. The change in the slope of the Cb-T curves was inherent for all the pristine, the as-irradiated and the subsequently annealed samples. Again, the onsets within the Cb-T curves and slightly different slopes seem to appear due to the irradiation-and annealing-induced transformations and various traps species density variations [19,21].   The L-DLTS method was chosen to separate the traps inherent for low-temperature range (40-80 K) owing to the L-DLTS elevated resolution (up to 2 MeV). The DLT spectra recorded on the as-irradiated and annealed Si 0.949 Ge 0.051 samples are illustrated in Figure 3b. However, here, the DLT spectra covered the temperature range >50 K. The barrier capacitance of the Si 0.949 Ge 0.051 diodes vanished in the low temperature range (Figure 3a), and the application of the capacitance DLTS was then impossible [19]. The activation energies of the traps were evaluated using Arrhenius plots (Figure 3c). The spectra with two prevailing peaks, namely, E 4 and E 6-2 , were again recorded after annealing at 125 • C for 15 min, similar to the regularity observed for the Si 0.986 Ge 0.014 sample. The DLTS signatures for all the traps observed in Si 0.949 Ge 0.051 samples are listed in Table 2. It was obtained that the activation energy values in the Si 0.949 Ge 0.051 alloy were shifted even more to the low-energy range relative to those obtained for the Si, Si 0.99 Ge 0.01 and Si 0.986 Ge 0.014 samples.
The comparison of the DLT spectra obtained in the as-irradiated (a) and annealed (b) Si and SiGe alloy samples is generalized in Figure 4 to clarify the activation energy shifts. The L-DLTS method was chosen to separate the traps inherent for low-temperature range (40-80 K) owing to the L-DLTS elevated resolution (up to 2 MeV). The DLT spectra recorded on the asirradiated and annealed Si0.949Ge0.051 samples are illustrated in Figure 3b. However, here, the DLT spectra covered the temperature range >50 K. The barrier capacitance of the Si0.949Ge0.051 diodes vanished in the low temperature range (Figure 3a), and the application of the capacitance DLTS was then impossible [19]. The activation energies of the traps were evaluated using Arrhenius plots (Figure 3c). The spectra with two prevailing peaks, namely, E4 and E6-2, were again recorded after annealing at 125 °C for 15 min, similar to the regularity observed for the Si0.986Ge0.014 sample. The DLTS signatures for all the traps observed in Si0.949Ge0.051 samples are listed in Table 2. It was obtained that the activation energy values in the Si0.949Ge0.051 alloy were shifted even more to the low-energy range relative to those obtained for the Si, Si0.99Ge0.01 and Si0.986Ge0.014 samples.
The comparison of the DLT spectra obtained in the as-irradiated (a) and annealed (b) Si and SiGe alloy samples is generalized in Figure 4 to clarify the activation energy shifts. The tendency of changes in the activation energy values of traps in Si1−xGex alloy as a function of Ge content is sketched in Figure 4c, based on DLT spectroscopy data obtained in this work. Additionally, the Laplace DLT spectra were re-plotted according to the re-calculation procedure described in Reference [26]. These L-DLT spectra are illustrated in Figure 4d to highlight the shifts in energy levels ascribed to V2 and CiOi defects in the as-irradiated diodes. It can be inferred from Figure  4a,b, that the activation energy values decreased significantly with enhancement of the Ge content in the SiGe alloy, irrespective of the irradiation and annealing procedures. This activation energy The tendency of changes in the activation energy values of traps in Si 1−x Ge x alloy as a function of Ge content is sketched in Figure 4c, based on DLT spectroscopy data obtained in this work. Additionally, the Laplace DLT spectra were re-plotted according to the re-calculation procedure described in Reference [26]. These L-DLT spectra are illustrated in Figure 4d to highlight the shifts in energy levels ascribed to V 2 and C i O i defects in the as-irradiated diodes. It can be inferred from Materials 2020, 13, 5684 8 of 10 Figure 4a,b, that the activation energy values decreased significantly with enhancement of the Ge content in the SiGe alloy, irrespective of the irradiation and annealing procedures. This activation energy variation was also independent of the DLTS peak amplitude changes. These results can be understood as an indication that the levels moved closer to the valance band [12,27]. The peak shifts to the higher energy range with an increase in the Ge content was obtained for n-type SiGe alloys [28][29][30] conversely to those investigated in this work-p-type SiGe alloys. This effect can be explained either through an occupation of the radiation defect core by Ge atoms in the SiGe alloy due to the lattice parameter change or via lattice bond length variations which affect the conduction and valence band parameters of the SiGe alloy [12]. The opposite tendency within the observed DLTS peak shifts for the p-type and n-type SiGe alloys might be alternatively explained through the shifts of the Fermi level and the consequent filling of the radiation defect states.

Summary
The deep trap spectra in the pristine, 5.5 MeV electron irradiated and the 125 • C annealed p-type SiGe alloys with slightly different Ge content were examined. It was deduced from C-DLTS and L-DLTS spectra that the carbon/oxygen associated complexes prevailed in the pristine Si 0.949 Ge 0.051 alloys. Irradiation with 5.5 MeV electrons led to considerable change in the DLT spectrum containing up to seven spectral peaks due to the introduction of the radiation defects. These defects have been identified using activation energy values reported in the literature. The double interstitial and oxygen (I 2 O) complexes and the vacancy, di-vacancy and tri-vacancy ascribed traps were revealed in the irradiated samples. The interstitial carbon and the metastable as well as stable forms of the carbon-oxygen (C i O i * and C i O i ) complexes were also identified for the irradiated SiGe alloys. It was found that the carbon-oxygen metastable complexes (C i O i * ) were transformed into stable-state complexes (C i O i ) under 125 • C annealing for 15 min of the irradiated samples. It was determined that the activation energy shifts of radiation-induced deep traps to low values were defined by an increase in the Ge content of the SiGe alloy.