Comparative study by IBIC of Si and SiC diodes irradiated with high energy protons

Accepted Manuscript Comparative study by IBIC of Si and SiC diodes irradiated with high energy protons J. Garcia Lopez, M.C. Jimenez-Ramos, M. Rodriguez-Ramos, J. Ceballos, F. Linez, J. Raisanen PII: S0168-583X(15)01275-6 DOI: http://dx.doi.org/10.1016/j.nimb.2015.12.029 Reference: NIMB 61617 To appear in: Nucl. Instr. and Meth. in Phys. Res. B Received Date: 29 September 2015 Revised Date: 7 December 2015 Accepted Date: 15 December 2015 Please cite this article as: J. Garcia Lopez, M.C. Jimenez-Ramos, M. Rodriguez-Ramos, J. Ceballos, F. Linez, J. Raisanen, Comparative study by IBIC of Si and SiC diodes irradiated with high energy protons, Nucl. Instr. and Meth. in Phys. Res. B (2015), doi: http://dx.doi.org/10.1016/j.nimb.2015.12.029 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Comparative study by IBIC of Si and SiC diodes irradiated with high energy protons J. Garcia Lopez*1, 2, M. C. Jimenez-Ramos2, M. Rodriguez-Ramos2, J. Ceballos3, F. Linez4, J. Raisanen5 1Dpto. Física Atómica, Molecular y Nuclear, Universidad de Sevilla. 41080 Sevilla, Spain 2CNA ( U. Sevilla, J. Andalucia, CSIC), Av. Thomas A. Edison 7, 41092 Sevilla, Spain 3Institute of Microelectronics of Seville, IMSE-CNM (CSIC/University of Seville), Seville 41092, Spain 4Department of Applied Physics, Aalto University, 02150 Espoo, Finland 5Department of Physics, University of Helsinki, Helsinki 00014, Finland Abstract The transport properties of a series of Si and SiC diodes have been studied using the Ion Beam Induced Charge (IBIC) technique. Structural defects were induced into the samples during the irradiation with 17 MeV protons. The experimental values of the Charge Collection Efficiency (CCE) vs bias voltages have been analyzed using a modified driftdiffusion model, which takes into account the recombination of carriers in the neutral and depletion regions. From these simulations, we have obtained the values of the carrier’s lifetime for pristine and irradiated diodes, which are found to degrade faster in the case of the SiC samples. However, the decrease of the CCE at high bias voltages is more important for the Si detectors, indicative of the lower radiation hardness of this material compared to SiC. The nature of the proton-induced defects on Si wafers has been studied by Positron Annihilation Spectroscopy (PAS) and Doppler Broadening Spectroscopy (DBS). The results suggest that the main defect detected by the positrons in p-type samples is the divacancy while for n-type at least a fraction of the positron annihilate in another defect. The concentration of defects is much lower than the number of vacancies predicted by SRIM. Keywords Si diodes, SiC detector, IBIC, proton irradiation, PAS, Doppler broadening *Corresponding author Tel.: +34 954460553, fax: +34 954460145, email: [email protected] 2 1. Introduction Silicon has a high-quality and robust native oxide that provides a well-controlled interface to the underlying silicon, while protecting the bulk from environmental contamination. This property of silicon is unique among semiconductors and it is a key ingredient to understand why this material is prevailing for the fabrication of electronic devices and detectors. However, in hostile environments, like nuclear reactors, particle accelerators and outer space, radiation damage is the main driver in the search for new materials [1]. Silicon carbide (SiC) is one of the most attractive semiconductor materials for the development of high-temperature, high-power and high-frequency devices [2]. Moreover, SiC has been already used as radiation detector [3] and have the advantage of much higher radiation tolerance compared to silicon [4]. In this paper we present a comparative study by Ion Beam Induced Charge (IBIC) of the radiation hardness of Si and 4H-SiC detectors irradiated with high energy protons. Besides, the nature of the proton-induced defects on Si wafers from which the diodes were fabricated has been studied by Positron Annihilation Spectroscopy (PAS) and Doppler Broadening Spectroscopy (DBS). The present work has been done in the framework of the IAEA Coordinated Research Project (CRP) No: F11016 [5], whose main objective is to obtain a deeper theoretical knowledge and experimental data on defects created in semiconductor and insulator materials by the use of accelerator-based light and heavy ion irradiation. To create the damage we have employed 17 MeV protons, which is the higher proton energy available among the participants of this CRP, with the distinctive feature that these ions traverse completely the samples, without leaving residual hydrogen in the crystals. As the defect profile is homogeneous, during an IBIC experiment all minority and majority carriers will have a certain probability to be trapped across the path to their respective electrodes and the detection properties, like the charge collection efficiency (CCE), can be evaluated in terms of the lifetime of both charge carriers. On the contrary, the irradiations carried out in other laboratories (ANSTO, RBI, JAEA and SNL) along this CRP utilized heavier ions and/or lower energies [6-9], which generate a much shallow and non-constant damage profile. In these conditions, most of the majority carriers generated by IBIC will move across a defectfree region, while the lifetime of minority carriers will depend on the carrier position. Therefore, this will not be an adequate parameter to describe the radiation hardness of the 3 detector. Instead, the capture coefficients αn,p will provide the key parameters for the characterization of the effects of radiation damage in the semiconductors [10]. The methodology and experiments concerning the CCE degradation in Si and SiC diodes after irradiation with MeV ions (others that 17 MeV protons) are described in [11, 12], respectively. 2. Experimental 2.1.Materials The Si samples consist of 300 µm thick n- and p-type Floating Zone Si diodes and wafers with bulk doping concentration of ~1012/cm3 fabricated by the Helsinki Institute of Physics (HIP). The full depletion voltage were ~ 30 and 40 V for the n- and p-type diodes, respectively, and the saturated leakage current was about 10 nA for both types. A more complete description of these samples can be found in [11, 13, 14]. The SiC diodes were provided by the Japanese Atomic Energy Agency (JAEA). Basically they consist on a ∼ 47 µm thick n-type 4H-SiC epilayer grown on 4H-SiC substrate. The Schottky contact, with dimensions 1mm x 1mm x 80 nm, was created by Ni sputtering on top of the SiC epilayer while the ohmic contact was Ni sintered at 950ºC at the substrate back side. The doping concentration of the epilayer, estimated at JAEA from the C-V measurements, is 5x1014 cm-3. The relationship between depletion layer thickness (w) and reverse bias voltage (VB) is given by w (µm) = 1.50013 V0.5, as obtained in [15] for SiC diodes from the same batch. 2.2. Proton irradiation To create structural damage, the samples were irradiated using the external beamline of the 18/9 compact Cyclotron of the National Accelerator Centre (CNA) in Sevilla (Spain). The Si and SiC diodes were placed in air at 15 cm from the 150 µm thick kapton exit window. Although the cyclotron delivers 18 MeV protons, the actual proton energy at the sample’s surface after traversing the kapton foil and the air was 17 MeV, as calculated using the SRIM code [16]. Since the proton beam current cannot be measured directly on 4 the diode, the proton fluence was indirectly determined by knowing the ratio between the amount of beam impinging on a 10 mm diameter carbon collimator placed right after the exit window and connected to a Brookhaven 1000c current integrator, and the amount of beam passing through this collimator. A second collimator made of 2 mm thick Al with 0.2 mm diameter hole was placed in front of the Si diodes in order to limit the irradiated volume, which otherwise would increase the leakage current to unacceptable levels. Due to the wide bandgap of the 4H-SiC material (Eg = 3.23 eV) compared to Si (Eg = 1.17 eV), the increase of the leakage current of the SiC diodes after irradiation is not a big concern and, therefore, their full volume was irradiated by removing the Al collimator. The fluences given to the Si and SiC detectors range between 1012 p/cm2 and 1013 p/cm2. The vacancy profile induced by the 17 MeV proton beam, evaluated by the SRIM code, was practically flat for all the samples, with ~ 200 vacancies/cm/ion in the Si diodes assuming a displacement energy of 21 eV and ~ 150 vacancies/cm/ion in the SiC diodes supposing a displacement energy of 35 eV and 21 eV for Si and C, respectively. For PAS and BD measurements, the n-type and p-type FZ Si wafers used to fabricate the diodes were irradiated in air, at 8 cm from the 125 µm thick Al exit window. In order to guarantee the lateral homogeneity of the beam over the 8x8 mm2 surface of the samples, a 50 µm thick W foil was placed immediately after the exit window. At 4 cm from this foil, a graphite collimator with a 1 cm diameter hole was placed to define the beam size. As the beam slightly diverges from the collimator to the sample position due to the interaction with the air, the actual beam size and homogeneity at the sample’s distance was measured with a radiochromic film. The beam energy calculated by SRIM at the sample surface was 15.7 MeV, which produces practically the same vacancy profile as 17 MeV protons. To determine the proton fluence, the wafers were stuck with carbon tape to an electrically floating graphite plate connected to a current integrator. Irradiation was performed for the fluences 1015, 4.26×1015 and 8.5×1015 p/cm2. All the analysis for diodes and wafers were carried out several days after irradiation, during which the samples were kept at room temperature, so only permanent damage induced by the displacement of atomic nuclei will be considered. 5 2.3.IBIC measurements IBIC analysis of the Si diodes was performed in the microbeam line of the CNA with a 4.07 MeV proton beam. Using a slightly focused 10x10 µm2 beam with a low count rate (<200 Hz), the damaged areas were first localized through the 1x1 mm2 IBIC mappings, as shown in Fig. 1, and then point measurements were performed in the center of the perturbed regions to extract the data. The IBIC signal was recorded as a function of the applied bias voltage using a Canberra 2003BT preamplifier, a Tennelec TC245 amplifier with a shaping time of 2 µs and the OMDAQ ADC/MCA system from Oxford Microbeams Ltd. It is important to emphasize that, despite the use of the 200 µm collimator, the increase of the leakage current Ic after the 17 MeV proton irradiation is not negligible. Then, the actual bias voltage VB applied to the Si diodes can be notably different from the voltage indicated by the front panel digital display of the bias supply (Vread) and is determined by VB = Vread – IcxRL, where RL = 110 MΩ is the detector load resistor. As the range of 4.07 MeV protons in Si is ~150 µm (approximately half of the detector thickness), by setting the reverse bias voltage it was possible to work in conditions in which all the charge carriers are generated within the depletion region (VB≥10 V, see Fig. 1 in [14]) and can be collected by its electric field (drift component). On the other hand, using lower bias values, a fraction of the IBIC signal was produced by the minority carriers diffusing from the electroneutral substrate to the active region of the detector (diffusion component). In the case of the SiC diodes the full volume was irradiated with the high energy protons, thus there was no need to use a microbeam for the IBIC analysis. Instead, alpha spectrometry measurements were carried out using a triple alpha source (244Cm, 241Am, 239Pu with average energy emissions 5.795 MeV, 5.478 MeV and 5.143 MeV, respectively) with activity ~ 1 µCi. The voltage breakdown for these SiC diodes is about -450 V [12], so the maximum bias applied during this work was -400V, corresponding to a depletion layer thickness of 30 µm. The generated charges in the diodes were collected using the same electronic chain as for the Si samples, but the shaping time of the amplifier was set to 0.75 µs to optimize the energy resolution. As pointed out in [12, 15], a characteristic of the signals obtained with 6 the SiC detector is that, even working at optimal conditions, at low bias the peaks widen and it becomes very difficult or impossible to assign them a well-defined energy, as can be observed in Fig. 2, where at -400 V the signal from each individual isotope can be clearly distinguished in contrast to the measurement at -60 V. This effect is more pronounced for irradiated samples and for the alpha particles with a higher initial energy. For that reason, the experimental results shown later in this work corresponds to the alpha peak emitted by the 239Pu isotope. The projected range of 5.14 MeV alpha particles in SiC is about 16 µm (Fig. 3), approximately half of the maximum depletion width of our diodes. For VB ≥ 125 V the depletion zone is ≥ 16 µm, and all the charge carriers are generated in the electrically active region. At lower bias, the IBIC signal will contain also a diffusion component. From this point of view, the IBIC experiments performed on the SiC sensors using 5.14 MeV alpha particles are equivalent to those carried out in the Si detectors using 4.07 MeV protons. 2.4. PAS and DBS measurements The positron lifetime measurements were performed at Aalto University using a conventional lifetime spectrometer [17] with a time resolution of 260 ps. Two identical samples were sandwiched with 10 µCi 22Na positron source wrapped in 1.5 µm Al foil. The maximum positron energy from the source is 541 keV meaning that in Si 94 % of the positrons annihilate in the 300 µm. The measurements were performed at 300 K. More than 2×106 counts for each spectrum was collected. The average positron lifetime is τave=∑i Ii τi with i, the different states in which positron annihilates with the lifetime τi and the intensity Ii. The Doppler broadening measurements were performed using a monoenergetic beam of positrons with E=0.5-25 keV. The highest energy corresponds to a mean probed depth of 3 µm in silicon samples. The Doppler-broadened annihilation line at 511 keV was measured with a high-purity Ge detector with a resolution of 1.2 keV. The momentum distribution was characterized with the usual shape parameters S and W, defined as the fractions of positrons annihilating with electrons in the low-momentum range 0–3.4×10-3 m0c and in the high-momentum range 11–27×10-3 m0c, respectively. To avoid any surface effects the 7 Doppler broadening measurements were performed for positron energies from 20 to 25 keV corresponding to depths of 2-3 µm. 3. Results and discussion The measurement of the CCE as a function of the reverse bias voltage is an accurate method to assess the performance and the radiation tolerance of the detectors [18]. The variation of the CCE vs VB is usually evaluated in the frame of the drift-diffusion model proposed by Breese [19-23]  = ∫w idx dE E0 1 + i E 1dxe dx dE diff pL wx R w − − ∫ (1) where Ei, Rp and dE/dx are the initial energy, the projected range and the ionizing energy loss of the incident ion in the material of the diode, respectively; w is the depletion width of the device, which is assumed in the following to be smaller than the ion penetration range, and Ldiff is the minority-carrier diffusion length, i.e. the average distance that a minority carrier can cover by diffusion before it recombines. Ldiff is related to the carrier lifetime τmin by minmin τ ×= DLDiff , where Dmin is the diffusion coefficient. The first integral (drift term) represents the contribution from the charge carriers generated within the depletion region which are fully collected as drift current. The second integral (diffusion term) represents the contribution from the minority carriers created in the electroneutral substrate that reach the edge of the depletion region as diffusion current. The experimental CCE for pristine samples and after various dose of implantation is shown in Fig. 4 for the p and n-type Si samples and in Fig. 5 for the SiC diodes. The reported CCE value is the ratio of the centroid peak position (for a given VB and irradiated fluence) and the position of the centroid in a pristine (non-irradiated) sample measured at the full saturation voltage for Si and at VB = -400 V for SiC. As observed, the complete charge collection is merely obtained for pristine Si and SiC samples when the depletion region is larger than the projected range of the probing ion (VB ≥ 10 V for Si and VB ≥ 125 V for SiC). For the irradiated diodes, the experimental CCE is always < 100%, even for the highest bias employed. This result is in contradiction with the simple drift-diffusion model 8 described in eq. 1, which anticipates 100% efficiency when w > Rp (i.e. when only the drift term applies) and evidences that carrier recombination takes place also into the drift region due to the radiation induced defects, despite the important electric field present (~105 V/cm for SiC). To evaluate more accurately the CCE data, we have recently developed an upgraded driftdiffusion (D-D) model, which takes into account the possibility of carrier recombination in the depletion and neutral regions of the detector [14]. In this D-D model the drift term becomes:  =∫w idx dE E0 1[      +     ]   (2) where the term into square brackets represents the probability for a charge carrier created at coordinate x to reach its respective electrode without recombination. Here Lmin,maj = (µτ)min,majԐ(x) are the drift lengths for minority (majority) carriers, µmin,maj are the mobility and τmin,maj are the lifetimes. For the calculations, the electric field was assumed to decrease linearly from its maximum value at the surface to zero at the edge of the depletion region or at the back contact, ℰ=#$% &( − , as happens for perfect one-sided abrupt junctions (Si samples) and Schottky diodes (SiC) [24]. The diffusion term is also modified to take into account the probability for recombination of the minority carriers arriving by diffusion at x = w and drifting toward the surface electrode: *+,- =   .    i E 1dye dy dE diff pL wy R w − − ∫ (3) The stopping power dE/dx was obtained from SRIM, while the mobility and diffusion coefficients were considered to be constant for Si and SiC with the accepted values for low doping concentration (µe (Si)= 1350 cm2/Vs; µh (Si)= 450 cm2/Vs; De (Si)= 36 cm2/s; Dh (Si)= 12 cm2/s; µe (SiC)= 900 cm2/Vs; µh (SiC)= 120 cm2/Vs and Dh (SiC)= 3 cm2/s). The open symbols in Fig. 4 and 5 represent the calculated CCE from eq. 2 and 3 using the least square fitting procedure, where the free parameters were the lifetimes for electrons and holes. In SiC the fitting of the CCE data was almost independent of the actual electron 15 [14] J. Garcia Lopez, M.C. Jimenez-Ramos, M. Rodriguez-Ramos, J. Forneris and J. Ceballos. “An upgraded drift-diffusion model for evaluating the carrier lifetimes in radiation-damaged semiconductor detectors”. Nucl. Instr. and Meth. B (2015), http://dx.doi.org/10.1016/j.nimb.2015.09.012 [15] Z. Pastuovic. IAEA Technical report, “IBIC analysis of the CCE degradation measured in partly damaged n-type silicon-carbide Schottky diodes prepared by the JAEA”. December 2013. [16] J.F. Ziegler, J.P. Biersack, M.D. Ziegler, “SRIM – The Stopping and Range of Ions in Matter”, Ion Implantation Press, 2008. http://www.lulu.com/content/1524197. [17] F. Tuomisto and I. Makkonen, Rev. Mod. Phys. 85, 1583 (2013) [18] G. Casse, Nucl. Instr. and Meth. A 612 (2010) 464-469 [19] M. B. H. Breese. J. Appl. Phys, 74, 3789 (1993) [20] M.B.H. Breese, E. Vittone, G. Vizkelethy and P.J. Sellin. Nucl. Instr. and Meth. B 264 (2007) 345-360 [21] M. De Napoli, F. Giacoppo, G. Raciti and E. Rapisarda Nucl. Instr. and Meth. A 608 (2009) 80-85 [22] S. Onoda et al. IEEE Trans. Nucl. Sci. 54, 6, (2007) [23] F. Nava, E. Vittone, P. Vanni, G. Verzellesi, P.G. Fuochi, C. Lanzieri, and M. Glaser. Nucl. Instr. and Meth. A 505 (2003) 645 [24] Semiconductor Detector System. Helmuth Spieler. Oxford University Press. 2011. [25] E. Gaubas,, T. Ceponis, A. Uleckas, J. Vaitkus, J. Raisanen, , Nucl. Instr. and Meth. A 612 (2010) 559–562 [26] F. Nava, G. Wagner, C. Lanzieri, P. Vanni and E. Vittone. Nucl. Instr. and Meth. A 510 (2003) 273-280 [27] M. Huhtinen. Nucl. Instr. and Meth. A 491 (2002) 194-215 [28] M.J. Boschini, P.G. Rancoita and M. Tacconi (2014), SR-NIEL Calculator: Screened Relativistic (SR) Treatment for Calculating the Displacement Damage and Nuclear Stopping Powers for Electrons, Protons, Light- and Heavy- Ions in Materials (version 2.1.1); [Online] available at INFN sez. Milano-Bicocca, Italy [2015, September ]: http://srniel.org/. 16 [29] N. Iwamoto et al. IEEE Trans. Nucl. Sci. 58, 1, (2011) [30] V. Ranki, A. Pelli, K. Saarinen, Phys. Rev. B 69, 115205 (2004) [31] S. Mäkinen, H. Rajainmäki and S. Linderoth, Phys. Rev. B 42, 11566 (1990) [32] M.J. Puska, C. Corbel and R.N. Nieminen, Phys. Rev. B 42, 11566 (1990) [33] M. Hakala, M.J. Puska and R.N. Nieminen, Phys. Rev. B 57, 7621 (1998) [34] S. Mäkinen and M.J. Puska, Phys. Rev. B 40, 12523 (1989) [35] R. Poirier, V. Avalos, S. Dannefaer, F. Schiettekatte and S. Roorda, Nucl. Instr. and Meth. B 206 (2003) 85-89 [36] H. Kauppinen, C. Corbel, J. Nissilä, K. Saarinen, and P. Hautojärvi, Phys. Rev. B 57, 12911 (1998) [37] J. Mäkinen, C. Corbel, P. Hautojärvi, P. Moser, and F. Pierre, Phys. Rev. B 39, 10162 (1989) [38] S. Väyrynen, P. Pusa, P. Sane, P. Tikkanen, J. Räisänen, K. Kuitunen, F. Tuomisto, J. Härkönen, I.Kassamakov, E. Tuominen and E. Tuovinen, Nucl. Instr. and Meth. A572 (2007) 978-984 [39] W. Shockley, W. T. Read, Physical Review 87, (1952), 835-842 6. Acknowledgements This work has been done under the IAEA Coordinated Research Project F11016. We thank all members of this CRP, especially E. Vittone from University of Torino, for useful discussions. The help of Dr. F. Tuomisto, from Aalto University, for the interpretation of the positron results is greatly appreciated. 17 7. Figure captions. Fig. 1. 1x1 mm2 CCE mapping for n-type Si diode around the spot irradiated to 5 x 1012 p/cm2 recorded using a low bias voltage (V=5 volts) to increase the sensitivity. The damaged region is clearly visible near the bright line coming from the guard ring. Fig. 2. Alpha spectra of a triple alpha source recorded with a pristine SiC diode at different bias voltage. Fig. 3. Ionization profile of 5.143 MeV alpha particles on SiC. Fig. 4a. Experimental (dots) and calculated (triangles) CCE values using the modified D-D model for pristine (black) and irradiated (red) p-type diodes. Fig. 4b. Experimental (dots) and calculated (squares) CCE values using the modified D-D model for pristine (black) and irradiated (red, blue) n-type diodes. Fig. 5. Experimental (squares) and calculated (dots) CCE values using the modified D-D model for pristine (black) and irradiated (blue, red) SiC diodes. For calculations we have considered τe = τh. Fig. 6. C-V curves for SiC diodes. Pristine (black line) and after irradiation with 1012 p/cm2 (red line). Figure 7: Average positron lifetime measured for n and p-type Si samples (on the top) as a function of the fluence. At the middle and at the bottom the lifetime components and their respective intensities (solid line component 1 and dashed line component 2) are presented. Figure 8: S vs W plot obtained for the 15.8 MeV-H+ irradiated samples. The S and W values reported are the average values of all S and W parameters measured in the positron energy range 20-25 keV. The dotted and dashed lines represent the slopes between the Si bulk (black diamond) and the monovacancy (V1) and the divacancy (V2), respectively, as calculated by Hakala [33]. 18 19 20 21 22 23 24