- Email: [email protected]
Light-induced Recovery of Effective Carrier Lifetime in Boron-doped Czochralski Silicon at Room Temperature
Available online at www.sciencedirect.com
ScienceDirect Energy Procedia 92 (2016) 801 – 807
6th International Conference on Silicon Photovoltaics, SiliconPV 2016
Light-induced recovery of effective carrier lifetime in boron-doped Czochralski silicon at room temperature Hiroaki Ichikawaa*, Isao Takahashia, Noritaka Usamia, Katsuhiko Shirasawab, Hidetaka Takatob a
b
Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, JAPAN Fukushima Renewable Energy Institute, AIST (FREA), Machiike-dai, Koriyama, Fukushima 963-0215, JAPAN
Abstract We report on the enhancements of effective carrier lifetime by light-induced “recovery” instead of “degradation” in oxygen-rich boron-doped Czochralski-grown silicon (Cz-Si) wafers at room temperature. The highest measured lifetime of ߬=550 ߤ ݏwas obtained under light illumination at room temperature. We found that the increasing rate of the lifetimes depends on the firing condition and the illumination intensity. These dependencies are very similar to those of permanent recovery caused by hydrogenation of B-O defects or so called regeneration induced by illumination at higher temperatures. Therefore, it can be presumed that the significant increase of minority carrier lifetime is caused by hydrogenation of recombination active defects. However, the defect-inactive state is not stable and the lifetime was found to decrease slowly after the illuminated recovery step. The thermal activation energy of this defect reactivation reaction was obtained to be 0.36 eV from the Arrhenius plot, which is much smaller than that of the thermal dissociation of the hydrogenated B-O complex of 1.25 eV. This result indicates that the measured defect transitions are markedly different from those of B-O related defects. Possible mechanisms to control defect deactivation are discussed based on a given fractional concentration of charged hydrogen. By combining this deactivation process and a regeneration process, a very high lifetime value could be achieved in Cz-Si for solar cells. © by by Elsevier Ltd.Ltd. This is an open access article under the CC BY-NC-ND license © 2016 2016The TheAuthors. Authors.Published Published Elsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer review by the scientific conference committee of SiliconPV 2016 under responsibility of PSE AG. Peer review by the scientific conference committee of SiliconPV 2016 under responsibility of PSE AG. Keywords: degradation, regeneration, passivation, hydrogen, silicon solar cell
* Corresponding author. Tel.: +81-52-789-3243; fax: +81-52-789-4516. E-mail address: [email protected]
1876-6102 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer review by the scientific conference committee of SiliconPV 2016 under responsibility of PSE AG. doi:10.1016/j.egypro.2016.07.071
802
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
1. Introduction It is well known that the condition of carrier injection and temperature gives great impact on defect transitions in the silicon bulk. Therefore, light-induced effect has been intensively investigated, especially degradation caused by the boron-oxygen (B-O) related defect in the Cz-Si material [1-3]. Light-induced degradation (LID) occurs due to defect transitions of recombination inactive B-O complexes to electrically active defects under carrier injection as well as illumination. The presence of active defects leads to a decrease of the minority carrier lifetime and the efficiency for use of silicon solar cells. In 2006, Herguth et al. [4] discovered a permanent recovery of the minority carrier lifetime via hydrogenation of B-O complexes under illumination at elevated temperature (regeneration). In 2013, Hallam et al. [5] assumed that it is very critical for passivation of B-O complexes to control the charge state of monatomic hydrogen and the Fermi-level position in the band gap with varying carrier injection level and temperature. Based on these studies, it is assumed that there are three parameters that play a key role in lifetime degradation and recovery in silicon wafer: (i) temperature (ii) carrier injection and (iii) charge state of atoms in the silicon bulk In this work, we introduce the effective carrier lifetime to show light-induced recovery instead of degradation in boron-doped Cz-Si wafers at room temperature. The increasing rate of the lifetime was found to depend on the firing condition and the illumination intensity. These dependencies suggest that these defect transitions are caused by carrier injection with controlling the charge states of atoms and defects. The mechanism and the possibility of this light-induced defect transitions are discussed focusing on the carrier injection and the charge state of atoms and defects, especially monatomic hydrogen in the silicon bulk. 2. Experiment 1-2 ȍ cm boron doped Cz-Si wafers (156 mm×156 mm×180 ȝm) were processed as follows. First, saw damages on the surface were removed by KOH etching (50 % KOH) at 90 oC for 10 min. After standard RCA cleaning process, gettering step with phosphorus diffusion was done at 850 oC to getter metallic impurities such as Fe and Cu which could limit the lifetime. Thereafter, the phosphorus silicate glass (PSG) as well as the n+-region on the surface was removed using HF and KOH, followed by additional RCA cleaning step. Then the surface was passivated by depositing AlOx layers on both sides of the samples via thermal atomic layer deposition. The thickness of AlOx layers were varied in the range of 5-20 nm in order to change the atomic hydrogen concentration inside the silicon bulk. After that, 80 nm thick silicon nitride layer was deposited on top of the AlOx layer via plasma-enhanced chemical vapor deposition (PE-CVD). An additional annealing step at 450 oC was performed for 30 min to activate the passivating properties of the AlOx/SiNx:H stacks. Finally, some samples were fired at high temperature around 550 oC in a furnace. During these thermal steps, hydrogen diffuses from SiNx:H layer to the silicon bulk. The resulting samples were once pre-annealed in dark condition at 220 oC for 15 min to deactivate B-O defects. The samples were then exposed to 10 mW/cm2 illumination for 12 h at room temperature to activate B-O defects and kept in darkness in order to undergo complete electron detrapping and achieve the equilibrium charge state. For illuminated recovery, an illumination treatment of 100 mW/cm2 90 min at room temperature was performed followed by keeping the samples in dark condition for 90 min again. In order to evaluate the defect concentration inside the wafer, the effective carrier lifetime was measured after each step at a fixed injection level of 1.0×1015 cm-3 using Sinton instruments WCT120 lifetime tester [6]. The measured lifetime values were converted into normalized defect concentrationܰሺݐሻ ൌ ͳΤ߬ሺݐሻ െ ͳȀ߬௧௩ , with ߬ሺሻ and ߬௧௩ are the lifetime at time t and the lifetime in the light-induced defect inactive state, respectively.
803
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
3. Results 3.1. Defect deactivation under illumination and reactivation in darkness We found that defect deactivation occurred only for samples which were kept under illumination at room temperature. The changes in minority carrier lifetimes are shown in Fig. 1. During defect deactivation process, the samples with firing showed a huge increase of minority carrier lifetime under illumination. However, when the samples were removed from illumination, the lifetime decreased slowly with exponential trends. It is important to note that both defect related reactions have strong dependence on the light illumination and the firing condition. In this respect, they seem to be similar to hydrogenation of B-O defect as well as regeneration.
Effective Lifetime (μs)
600
Under Illumination
In Darkness
w/ firing w/o firing
500 400 300 200 100
0
50 100 150 Treatment Time (min)
Fig. 1. Increase of minority carrier lifetime under illumination of 100 mW/cm2 and decrease in darkness. The lines are single exponential fits to the data.
3.2. Role of hydrogen in defect deactivation
Effective Lifetime (μs)
In order to confirm the role of atomic hydrogen in defect deactivation, we prepared samples with different thickness of the AlOx layer (5 nm and 20 nm). According to Dekkers et al. [7], hydrogen transmissibility of the AlOx layer changes by more than one order of magnitude between 5 and 10 nm layer thickness. This means that AlOx layer behaves as a barrier layer for atomic hydrogen diffusion from SiNx:H layer to the silicon bulk during firing step. Fig. 2 shows the change in measured effective lifetimes over the process sequence. The minority carrier lifetime for all fired samples is increased by illuminated recovery step. This result indicates that the presence of atomic hydrogen in the silicon bulk is critical for improvement of carrier lifetime during the illuminated recovery step. However, the improvement of the lifetime is also observed for the sample with 5 nm AlOx layer without firing. These results seem to indicate that atomic hydrogen is diffused during an additional annealing step at 450 oC after deposition of SiNx:H layer. Furthermore, 5 nm AlOx layer allows adequate atomic hydrogen to diffuse into the silicon bulk without firing. 140 120
600
w/ firing w/o firing
500
100
400
80
300
60
200
40
100
20 0
w/ firing w/o firing
0 Pre-anneal 1 Degradation2 Illum.Recovery3
In Darkness 4
0
0 Pre-anneal 1 Degradation2 Illum.Recovery3
In Darkness 4
Fig. 2. Effective carrier lifetimes of the wafers with different thickness of AlOx layer of (Left) 5 nm and (Right) 20 nm.
804
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
3.3. Estimation of the thermal activation energy of the defect reactivation reaction In order to evaluate the activation energy of the defect reactivation, immediately after illumination recovery, the defect reactivation reaction in the dark was then monitored at several temperature (20 to 80 oC). The evolution of normalized defect concentration with reactivation time is given in Fig. 3. The corresponding Arrhenius plot to estimate the thermal activation energy of the reactivation reaction is shown in Fig. 4. The thermal activation energy of the defect reactivation reaction is obtained to be 0.36 eV. This value is much smaller than the thermal activation energy of the dissociation reaction of the hydrogenated B-O defect complex of 1.25 eV [8]. By taking this result into account, it seems natural to conclude that defect deactivation and reactivation observed in this work are not related to hydrogenation of B-O defects but other kinds of defect transitions occurred in the silicon bulk, the interface or on the surface of the wafer. -3
1
20 ºC 40 ºC 60 ºC 80 ºC
4.0
Reactivation Rate (1/μs)
Norm. Defect Concentration (1/μs)
5.0x10
3.0 2.0 1.0 0.0 0
5
10
15
20
25
30
6 4 2
0.1 4 2
0.01 2.7 2.8
2.9
3.0
3.1
3.2
3.3
3.4 3.5
Inverse Reactivation Temperature (1000/K)
Reactivation Time (min) Fig. 3. Evolution of normalized defect concentration after illuminated recovery at different temperatures. The lines are single exponential fits to the data.
6
Fig. 4. Arrhenius correlation between temperature and defect reactivation rate gives thermal activation energy of 0.36 eV.
3.4. The charge state of hydrogen varied by the carrier injection The charge state of hydrogen in the silicon is controlled via Fermi-level shifting in the band gap. According to the model of Herring et al. and Hallam et al. [9,10], the fractional concentration of charged hydrogen species in the p-type silicon with varying illumination intensity are shown below [(1)-(3)], where fs is the fractional concentration correspond to charged hydrogen of the charge states s, vs, and Zs are the number of possible sites per unit cell, and the effective partition function, respectively. Ed and Ea are the donor level and the acceptor level introduced by monatomic hydrogen which are temperature-independent and EFn is the quasiFermi-level of electrons.
f0
§ v Z §¨¨ Ed E Fn ·¸¸ v Z §¨¨ E Fn Ea ·¸¸ · ¨1 e© kT ¹ e© kT ¹ ¸ ¨ v0 Z 0 ¸ v0 Z 0 © ¹
f
f0
§ E d E Fn kT
v Z ¨¨© e v0 Z 0
· ¸¸ ¹
§ E Fn E a ·
1
(1)
(2)
v Z ¨¨© kT ¸¸¹ (3) e v0 Z 0 The values of constants used in the equations (1)-(3) are shown in Table 1, where m0 is the electron mass (9.11 × 10í31 kg), and mde and mdh are the density of states effective mass for electrons and holes. f
f0
805
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807 Table 1. Values of constants used in (1)-(3) [10,11] Constant
Value
mde
0.33 m0
mdh
0.56 m0
Ed
Ec - 0.16 eV
Ea
(Ev+Ec)/2 - 0.07eV
v+
8
v0
4
v-
2
Z+
1
Z0
1
Z-
1
-1
10
0
H
-3
10
-
H
+
H
-5
-7
10
-9
1.0 Suns
10
0.1 Suns
Fractional Concentration
The simulated distribution of the fractional concentration of the charged hydrogen as a function of the carrier injection level is shown in Fig. 5. The fractional concentration varies by many orders of magnitude with varying the illumination intensity. Especially, at the injection level between 1014~1015 cm-3 correspond to the illumination intensity of 0.1~1.0 Suns (dashed pink lines in Fig. 5), minority carrier injection decreases the fractions of H0 and H+ by one and two orders of magnitude. Meanwhile, almost all monatomic hydrogen is charged negatively. The fractions of hydrogen species corresponding to each illumination intensity and carrier injection level are listed in Table 2. Carrier injection level is determined by the result of QSS-PC measurements.
10
10
8
10
10
12
10
10
14
10
16
18
10
-3
10
20
Carrier Injection Level (cm ) Fig. 5. Calculated fractional concentration of three charge states of monatomic hydrogen in the silicon bulk at 300 K. Table 2. The fractions of hydrogen species corresponding to each illumination intensity. Illumination Intensity (Suns) 1.0 0.5 0.3 0.1
Carrier Injection Level (cm-3) 1.4×1015 5.1×1014 3.2×1014 8.3×1013
H0 Fraction 4.04×10-6 1.13×10-5 1.80×10-5 6.76×10-5
H- Fraction 9.99×10-1 9.99×10-1 9.99×10-1 9.99×10-1
H+ Fraction 4.04×10-5 1.11×10-4 2.79×10-4 3.90×10-3
806
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
3.5. Dependencies of defect deactivation on the illumination intensity Minority carrier injection has a strong impact on charge states and reactivity of monatomic hydrogen and defects [11]. Fig. 6 shows the experimental data of normalized defect concentration under different illumination intensities at room temperature. With increasing intensity the defect deactivation process is accelerated. Based on the relation between carrier injection level and hydrogen fraction listed in Table 2, if H0 or H+ contribute to defect deactivation, the defect deactivation rate must be decreased with increasing illumination intensity. From this result, we assume three possible mechanisms and species which control and limit the light-induced recovery process: A) Charge state of active defects also changes by carrier injection, and it controls the defect deactivation rate. H0 and/or H+ have a negative effect on defect deactivation.
C)
Defect deactivation is occurred only by carrier injection, and electron emission rate of the deactivated defects is very low. Norm. Defect Concentration (1/μs)
B)
-3
5x10
1.0 Suns 0.5 Suns 0.3 Suns 0.1 Suns
4 3 2 1 0
0
5 10 15 20 Deactivation Time (s)
25
Fig. 6. Evolution of normalized defect concentration with different illumination intensities. The lines are single exponential fits to the data.
4. Summary Light-induced recovery of effective carrier lifetime at room temperature was observed. The samples showed a huge increase of minority carrier lifetime as high as ߬=550 ߤ ݏunder illumination and a slow exponential decrease in darkness. Different deactivation behaviors were observed in the samples with 5 nm and 20 nm thickness of AlOx layer, indicating a key role of hydrogen in deactivation. The difference in thermal activation energy of the defect reactivation reaction implies that these defect transitional phenomena are caused by hydrogenation of electricalactive defects located in the silicon bulk, the interface or on the surface of the wafer and not related to B-O defects. The deactivation rate is accelerated with increasing illumination intensity. Therefore, the deactivation rate mainly depends on the charge states of interstitial atoms and bonds in the silicon bulk such as Si, B, O, H, and so on. For a given fractional concentration of charged hydrogen, three possible mechanisms of defect deactivation that limit the process in light-induced recovery are to be considered: charge state of the defects, fractional concentrations of H0 or H+ and the electron emission rate of defects. The optimization of defect deactivation could be achieved with improved control of charge states, which appears to be of critical importance in taking both advantages of lightinduced recovery and regeneration. References [1] J. Schmidt, K Bothe, R Hezel. Structure and transformation of the metastable centre in CZ-silicon solar cells, 3rd World Conference on Photovoltaic Energy Conversion, Osaka, Japan, 2003; 2887-2892.
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807 [2] K. Bothe and J. Schmidt. Electronically activated boron-oxygen-related recombination centers in crystalline silicon, In: J. Appl. Phys. 2006; 99 (1), 013701. [3] J. Adey, R. Jones, D. Palmer, P. Briddon, S. Öberg. Degradation of Boron-Doped Czochralski-Grown Silicon Solar Cells, In: Phys. Rev. Lett. 2004; 93 (5), 055504. [4] A. Herguth. Avoiding boron-oxygen related degradation in highly boron doped CZ silicon, Proc. 21st EUPVSEC, Dresden, Germany, (Munich: WIP 2006), p.530-537. [5] B. J. Hallam, S. R. Wenham, P. G. Hamer, M. D. Abbott, A. Sugianto, C. E. Chan, A. M. Wenham, M. G. Eadie, G. Xu. Hydrogen Passivation of B-O Defects in Czochralski Silicon, In: Energy Procedia 2013; 38, 561–570. [6]R. A. Sinton, A. Cuevas. Contactless determination of current-voltage characteristics and minority-carrier lifetimes in semiconductors from quasi-steady-state photoconductance data, Appl. Phys. Lett. 1996; 69, 2150-2512. [7] HFW. Dekkers, G. Beaucarne, M. Hiller, H. Charifi, A. Slaoui. Molecular hydrogen formation in hydrogenated silicon nitride, Applied Physics Letters 2006; 89 (21), 1914. [8] S. Wilking, C. Beckh, S. Ebert, A. Herguth, G. Hahn. Influence of bound hydrogen states on BO- regeneration kinetics and consequences for high-speed regeneration process, Solar Energy Materials & Solar Cells 2014; 131, 2-8. [9] C. Herring, N. Johnson, C. Van de Walle. Energy levels of isolated interstitial hydrogen in silicon, Phys.Rev.B 2001; 64 (12), 125209-1– 125209-27. [10] Brett J. Hallam, Phill G. Hamer, Stuart R. Wenham, Malcolm D. Abbott, Adeline Sugianto, Alison M. Wenham, Catherine E. Chan, GuangQi Xu, Jed Kraiem, Julien Degoulange, and Roland Einhaus. Advanced Bulk Defect Passivation for Silicon Solar Cells, IEEE Journal of Photovoltaics 2014; 4 (1), 88-95. [11] Chang Sun, Fiacre E. Rougieux, Daniel Macdonald. A unified approach to modelling the charge state of monatomic hydrogen and other defects in crystalline silicon, In: J. Appl. Phys. 2015; 99 (1), 013701.
807
ScienceDirect Energy Procedia 92 (2016) 801 – 807
6th International Conference on Silicon Photovoltaics, SiliconPV 2016
Light-induced recovery of effective carrier lifetime in boron-doped Czochralski silicon at room temperature Hiroaki Ichikawaa*, Isao Takahashia, Noritaka Usamia, Katsuhiko Shirasawab, Hidetaka Takatob a
b
Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, JAPAN Fukushima Renewable Energy Institute, AIST (FREA), Machiike-dai, Koriyama, Fukushima 963-0215, JAPAN
Abstract We report on the enhancements of effective carrier lifetime by light-induced “recovery” instead of “degradation” in oxygen-rich boron-doped Czochralski-grown silicon (Cz-Si) wafers at room temperature. The highest measured lifetime of ߬=550 ߤ ݏwas obtained under light illumination at room temperature. We found that the increasing rate of the lifetimes depends on the firing condition and the illumination intensity. These dependencies are very similar to those of permanent recovery caused by hydrogenation of B-O defects or so called regeneration induced by illumination at higher temperatures. Therefore, it can be presumed that the significant increase of minority carrier lifetime is caused by hydrogenation of recombination active defects. However, the defect-inactive state is not stable and the lifetime was found to decrease slowly after the illuminated recovery step. The thermal activation energy of this defect reactivation reaction was obtained to be 0.36 eV from the Arrhenius plot, which is much smaller than that of the thermal dissociation of the hydrogenated B-O complex of 1.25 eV. This result indicates that the measured defect transitions are markedly different from those of B-O related defects. Possible mechanisms to control defect deactivation are discussed based on a given fractional concentration of charged hydrogen. By combining this deactivation process and a regeneration process, a very high lifetime value could be achieved in Cz-Si for solar cells. © by by Elsevier Ltd.Ltd. This is an open access article under the CC BY-NC-ND license © 2016 2016The TheAuthors. Authors.Published Published Elsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer review by the scientific conference committee of SiliconPV 2016 under responsibility of PSE AG. Peer review by the scientific conference committee of SiliconPV 2016 under responsibility of PSE AG. Keywords: degradation, regeneration, passivation, hydrogen, silicon solar cell
* Corresponding author. Tel.: +81-52-789-3243; fax: +81-52-789-4516. E-mail address: [email protected]
1876-6102 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer review by the scientific conference committee of SiliconPV 2016 under responsibility of PSE AG. doi:10.1016/j.egypro.2016.07.071
802
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
1. Introduction It is well known that the condition of carrier injection and temperature gives great impact on defect transitions in the silicon bulk. Therefore, light-induced effect has been intensively investigated, especially degradation caused by the boron-oxygen (B-O) related defect in the Cz-Si material [1-3]. Light-induced degradation (LID) occurs due to defect transitions of recombination inactive B-O complexes to electrically active defects under carrier injection as well as illumination. The presence of active defects leads to a decrease of the minority carrier lifetime and the efficiency for use of silicon solar cells. In 2006, Herguth et al. [4] discovered a permanent recovery of the minority carrier lifetime via hydrogenation of B-O complexes under illumination at elevated temperature (regeneration). In 2013, Hallam et al. [5] assumed that it is very critical for passivation of B-O complexes to control the charge state of monatomic hydrogen and the Fermi-level position in the band gap with varying carrier injection level and temperature. Based on these studies, it is assumed that there are three parameters that play a key role in lifetime degradation and recovery in silicon wafer: (i) temperature (ii) carrier injection and (iii) charge state of atoms in the silicon bulk In this work, we introduce the effective carrier lifetime to show light-induced recovery instead of degradation in boron-doped Cz-Si wafers at room temperature. The increasing rate of the lifetime was found to depend on the firing condition and the illumination intensity. These dependencies suggest that these defect transitions are caused by carrier injection with controlling the charge states of atoms and defects. The mechanism and the possibility of this light-induced defect transitions are discussed focusing on the carrier injection and the charge state of atoms and defects, especially monatomic hydrogen in the silicon bulk. 2. Experiment 1-2 ȍ cm boron doped Cz-Si wafers (156 mm×156 mm×180 ȝm) were processed as follows. First, saw damages on the surface were removed by KOH etching (50 % KOH) at 90 oC for 10 min. After standard RCA cleaning process, gettering step with phosphorus diffusion was done at 850 oC to getter metallic impurities such as Fe and Cu which could limit the lifetime. Thereafter, the phosphorus silicate glass (PSG) as well as the n+-region on the surface was removed using HF and KOH, followed by additional RCA cleaning step. Then the surface was passivated by depositing AlOx layers on both sides of the samples via thermal atomic layer deposition. The thickness of AlOx layers were varied in the range of 5-20 nm in order to change the atomic hydrogen concentration inside the silicon bulk. After that, 80 nm thick silicon nitride layer was deposited on top of the AlOx layer via plasma-enhanced chemical vapor deposition (PE-CVD). An additional annealing step at 450 oC was performed for 30 min to activate the passivating properties of the AlOx/SiNx:H stacks. Finally, some samples were fired at high temperature around 550 oC in a furnace. During these thermal steps, hydrogen diffuses from SiNx:H layer to the silicon bulk. The resulting samples were once pre-annealed in dark condition at 220 oC for 15 min to deactivate B-O defects. The samples were then exposed to 10 mW/cm2 illumination for 12 h at room temperature to activate B-O defects and kept in darkness in order to undergo complete electron detrapping and achieve the equilibrium charge state. For illuminated recovery, an illumination treatment of 100 mW/cm2 90 min at room temperature was performed followed by keeping the samples in dark condition for 90 min again. In order to evaluate the defect concentration inside the wafer, the effective carrier lifetime was measured after each step at a fixed injection level of 1.0×1015 cm-3 using Sinton instruments WCT120 lifetime tester [6]. The measured lifetime values were converted into normalized defect concentrationܰሺݐሻ ൌ ͳΤ߬ሺݐሻ െ ͳȀ߬௧௩ , with ߬ሺሻ and ߬௧௩ are the lifetime at time t and the lifetime in the light-induced defect inactive state, respectively.
803
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
3. Results 3.1. Defect deactivation under illumination and reactivation in darkness We found that defect deactivation occurred only for samples which were kept under illumination at room temperature. The changes in minority carrier lifetimes are shown in Fig. 1. During defect deactivation process, the samples with firing showed a huge increase of minority carrier lifetime under illumination. However, when the samples were removed from illumination, the lifetime decreased slowly with exponential trends. It is important to note that both defect related reactions have strong dependence on the light illumination and the firing condition. In this respect, they seem to be similar to hydrogenation of B-O defect as well as regeneration.
Effective Lifetime (μs)
600
Under Illumination
In Darkness
w/ firing w/o firing
500 400 300 200 100
0
50 100 150 Treatment Time (min)
Fig. 1. Increase of minority carrier lifetime under illumination of 100 mW/cm2 and decrease in darkness. The lines are single exponential fits to the data.
3.2. Role of hydrogen in defect deactivation
Effective Lifetime (μs)
In order to confirm the role of atomic hydrogen in defect deactivation, we prepared samples with different thickness of the AlOx layer (5 nm and 20 nm). According to Dekkers et al. [7], hydrogen transmissibility of the AlOx layer changes by more than one order of magnitude between 5 and 10 nm layer thickness. This means that AlOx layer behaves as a barrier layer for atomic hydrogen diffusion from SiNx:H layer to the silicon bulk during firing step. Fig. 2 shows the change in measured effective lifetimes over the process sequence. The minority carrier lifetime for all fired samples is increased by illuminated recovery step. This result indicates that the presence of atomic hydrogen in the silicon bulk is critical for improvement of carrier lifetime during the illuminated recovery step. However, the improvement of the lifetime is also observed for the sample with 5 nm AlOx layer without firing. These results seem to indicate that atomic hydrogen is diffused during an additional annealing step at 450 oC after deposition of SiNx:H layer. Furthermore, 5 nm AlOx layer allows adequate atomic hydrogen to diffuse into the silicon bulk without firing. 140 120
600
w/ firing w/o firing
500
100
400
80
300
60
200
40
100
20 0
w/ firing w/o firing
0 Pre-anneal 1 Degradation2 Illum.Recovery3
In Darkness 4
0
0 Pre-anneal 1 Degradation2 Illum.Recovery3
In Darkness 4
Fig. 2. Effective carrier lifetimes of the wafers with different thickness of AlOx layer of (Left) 5 nm and (Right) 20 nm.
804
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
3.3. Estimation of the thermal activation energy of the defect reactivation reaction In order to evaluate the activation energy of the defect reactivation, immediately after illumination recovery, the defect reactivation reaction in the dark was then monitored at several temperature (20 to 80 oC). The evolution of normalized defect concentration with reactivation time is given in Fig. 3. The corresponding Arrhenius plot to estimate the thermal activation energy of the reactivation reaction is shown in Fig. 4. The thermal activation energy of the defect reactivation reaction is obtained to be 0.36 eV. This value is much smaller than the thermal activation energy of the dissociation reaction of the hydrogenated B-O defect complex of 1.25 eV [8]. By taking this result into account, it seems natural to conclude that defect deactivation and reactivation observed in this work are not related to hydrogenation of B-O defects but other kinds of defect transitions occurred in the silicon bulk, the interface or on the surface of the wafer. -3
1
20 ºC 40 ºC 60 ºC 80 ºC
4.0
Reactivation Rate (1/μs)
Norm. Defect Concentration (1/μs)
5.0x10
3.0 2.0 1.0 0.0 0
5
10
15
20
25
30
6 4 2
0.1 4 2
0.01 2.7 2.8
2.9
3.0
3.1
3.2
3.3
3.4 3.5
Inverse Reactivation Temperature (1000/K)
Reactivation Time (min) Fig. 3. Evolution of normalized defect concentration after illuminated recovery at different temperatures. The lines are single exponential fits to the data.
6
Fig. 4. Arrhenius correlation between temperature and defect reactivation rate gives thermal activation energy of 0.36 eV.
3.4. The charge state of hydrogen varied by the carrier injection The charge state of hydrogen in the silicon is controlled via Fermi-level shifting in the band gap. According to the model of Herring et al. and Hallam et al. [9,10], the fractional concentration of charged hydrogen species in the p-type silicon with varying illumination intensity are shown below [(1)-(3)], where fs is the fractional concentration correspond to charged hydrogen of the charge states s, vs, and Zs are the number of possible sites per unit cell, and the effective partition function, respectively. Ed and Ea are the donor level and the acceptor level introduced by monatomic hydrogen which are temperature-independent and EFn is the quasiFermi-level of electrons.
f0
§ v Z §¨¨ Ed E Fn ·¸¸ v Z §¨¨ E Fn Ea ·¸¸ · ¨1 e© kT ¹ e© kT ¹ ¸ ¨ v0 Z 0 ¸ v0 Z 0 © ¹
f
f0
§ E d E Fn kT
v Z ¨¨© e v0 Z 0
· ¸¸ ¹
§ E Fn E a ·
1
(1)
(2)
v Z ¨¨© kT ¸¸¹ (3) e v0 Z 0 The values of constants used in the equations (1)-(3) are shown in Table 1, where m0 is the electron mass (9.11 × 10í31 kg), and mde and mdh are the density of states effective mass for electrons and holes. f
f0
805
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807 Table 1. Values of constants used in (1)-(3) [10,11] Constant
Value
mde
0.33 m0
mdh
0.56 m0
Ed
Ec - 0.16 eV
Ea
(Ev+Ec)/2 - 0.07eV
v+
8
v0
4
v-
2
Z+
1
Z0
1
Z-
1
-1
10
0
H
-3
10
-
H
+
H
-5
-7
10
-9
1.0 Suns
10
0.1 Suns
Fractional Concentration
The simulated distribution of the fractional concentration of the charged hydrogen as a function of the carrier injection level is shown in Fig. 5. The fractional concentration varies by many orders of magnitude with varying the illumination intensity. Especially, at the injection level between 1014~1015 cm-3 correspond to the illumination intensity of 0.1~1.0 Suns (dashed pink lines in Fig. 5), minority carrier injection decreases the fractions of H0 and H+ by one and two orders of magnitude. Meanwhile, almost all monatomic hydrogen is charged negatively. The fractions of hydrogen species corresponding to each illumination intensity and carrier injection level are listed in Table 2. Carrier injection level is determined by the result of QSS-PC measurements.
10
10
8
10
10
12
10
10
14
10
16
18
10
-3
10
20
Carrier Injection Level (cm ) Fig. 5. Calculated fractional concentration of three charge states of monatomic hydrogen in the silicon bulk at 300 K. Table 2. The fractions of hydrogen species corresponding to each illumination intensity. Illumination Intensity (Suns) 1.0 0.5 0.3 0.1
Carrier Injection Level (cm-3) 1.4×1015 5.1×1014 3.2×1014 8.3×1013
H0 Fraction 4.04×10-6 1.13×10-5 1.80×10-5 6.76×10-5
H- Fraction 9.99×10-1 9.99×10-1 9.99×10-1 9.99×10-1
H+ Fraction 4.04×10-5 1.11×10-4 2.79×10-4 3.90×10-3
806
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807
3.5. Dependencies of defect deactivation on the illumination intensity Minority carrier injection has a strong impact on charge states and reactivity of monatomic hydrogen and defects [11]. Fig. 6 shows the experimental data of normalized defect concentration under different illumination intensities at room temperature. With increasing intensity the defect deactivation process is accelerated. Based on the relation between carrier injection level and hydrogen fraction listed in Table 2, if H0 or H+ contribute to defect deactivation, the defect deactivation rate must be decreased with increasing illumination intensity. From this result, we assume three possible mechanisms and species which control and limit the light-induced recovery process: A) Charge state of active defects also changes by carrier injection, and it controls the defect deactivation rate. H0 and/or H+ have a negative effect on defect deactivation.
C)
Defect deactivation is occurred only by carrier injection, and electron emission rate of the deactivated defects is very low. Norm. Defect Concentration (1/μs)
B)
-3
5x10
1.0 Suns 0.5 Suns 0.3 Suns 0.1 Suns
4 3 2 1 0
0
5 10 15 20 Deactivation Time (s)
25
Fig. 6. Evolution of normalized defect concentration with different illumination intensities. The lines are single exponential fits to the data.
4. Summary Light-induced recovery of effective carrier lifetime at room temperature was observed. The samples showed a huge increase of minority carrier lifetime as high as ߬=550 ߤ ݏunder illumination and a slow exponential decrease in darkness. Different deactivation behaviors were observed in the samples with 5 nm and 20 nm thickness of AlOx layer, indicating a key role of hydrogen in deactivation. The difference in thermal activation energy of the defect reactivation reaction implies that these defect transitional phenomena are caused by hydrogenation of electricalactive defects located in the silicon bulk, the interface or on the surface of the wafer and not related to B-O defects. The deactivation rate is accelerated with increasing illumination intensity. Therefore, the deactivation rate mainly depends on the charge states of interstitial atoms and bonds in the silicon bulk such as Si, B, O, H, and so on. For a given fractional concentration of charged hydrogen, three possible mechanisms of defect deactivation that limit the process in light-induced recovery are to be considered: charge state of the defects, fractional concentrations of H0 or H+ and the electron emission rate of defects. The optimization of defect deactivation could be achieved with improved control of charge states, which appears to be of critical importance in taking both advantages of lightinduced recovery and regeneration. References [1] J. Schmidt, K Bothe, R Hezel. Structure and transformation of the metastable centre in CZ-silicon solar cells, 3rd World Conference on Photovoltaic Energy Conversion, Osaka, Japan, 2003; 2887-2892.
Hiroaki Ichikawa et al. / Energy Procedia 92 (2016) 801 – 807 [2] K. Bothe and J. Schmidt. Electronically activated boron-oxygen-related recombination centers in crystalline silicon, In: J. Appl. Phys. 2006; 99 (1), 013701. [3] J. Adey, R. Jones, D. Palmer, P. Briddon, S. Öberg. Degradation of Boron-Doped Czochralski-Grown Silicon Solar Cells, In: Phys. Rev. Lett. 2004; 93 (5), 055504. [4] A. Herguth. Avoiding boron-oxygen related degradation in highly boron doped CZ silicon, Proc. 21st EUPVSEC, Dresden, Germany, (Munich: WIP 2006), p.530-537. [5] B. J. Hallam, S. R. Wenham, P. G. Hamer, M. D. Abbott, A. Sugianto, C. E. Chan, A. M. Wenham, M. G. Eadie, G. Xu. Hydrogen Passivation of B-O Defects in Czochralski Silicon, In: Energy Procedia 2013; 38, 561–570. [6]R. A. Sinton, A. Cuevas. Contactless determination of current-voltage characteristics and minority-carrier lifetimes in semiconductors from quasi-steady-state photoconductance data, Appl. Phys. Lett. 1996; 69, 2150-2512. [7] HFW. Dekkers, G. Beaucarne, M. Hiller, H. Charifi, A. Slaoui. Molecular hydrogen formation in hydrogenated silicon nitride, Applied Physics Letters 2006; 89 (21), 1914. [8] S. Wilking, C. Beckh, S. Ebert, A. Herguth, G. Hahn. Influence of bound hydrogen states on BO- regeneration kinetics and consequences for high-speed regeneration process, Solar Energy Materials & Solar Cells 2014; 131, 2-8. [9] C. Herring, N. Johnson, C. Van de Walle. Energy levels of isolated interstitial hydrogen in silicon, Phys.Rev.B 2001; 64 (12), 125209-1– 125209-27. [10] Brett J. Hallam, Phill G. Hamer, Stuart R. Wenham, Malcolm D. Abbott, Adeline Sugianto, Alison M. Wenham, Catherine E. Chan, GuangQi Xu, Jed Kraiem, Julien Degoulange, and Roland Einhaus. Advanced Bulk Defect Passivation for Silicon Solar Cells, IEEE Journal of Photovoltaics 2014; 4 (1), 88-95. [11] Chang Sun, Fiacre E. Rougieux, Daniel Macdonald. A unified approach to modelling the charge state of monatomic hydrogen and other defects in crystalline silicon, In: J. Appl. Phys. 2015; 99 (1), 013701.
807