Light-induced Degradation of Silicon Solar Cells with ... - Core

Available online at www.sciencedirect.com

ScienceDirect Energy Procedia 77 (2015) 599 – 606

5th International Conference on Silicon Photovoltaics, SiliconPV 2015

Light-induced degradation of silicon solar cells with aluminium oxide passivated rear side Karin Krauss, Fabian Fertig, Dorothee Menzel, Stefan Rein FraunhoferInstitute for Solar Energy Systems ISE, Heidenhofstraße 2, 79110 Freiburg, Germany

Abstract Light-induced degradation (LID) has been shown to significantly affect the performance of multicrystalline silicon (mc-Si) solar cells with aluminium oxide (AlOx) passivated rear side. Within this work, the impact of LID on the conversion efficiency of different silicon solar cell architectures with and without AlOx passivation is investigated. Under conditions representing realistic module operation, significant light-induced degradation of up to 'K = -2.9 %rel in conversion efficiency has been observed for multicrystalline silicon (mc-Si) solar cells with AlOx passivation. This degradation has been found to be higher than the degradation of, both, a mc-Si aluminium back surface field (Al-BSF) solar cell and, remarkably, a Czochralski-grown silicon (Cz-Si) solar cell with AlOx passivation. For a more detailed investigation of the interaction of mc-Si and AlOx passivation during degradation, a photoluminescence-based “effective defect” imaging has been performed on AlOx-passivated mc-Si lifetime samples. The local effective defect lifetime related to recombination due to LID-induced defects is found to vary strongly in the range of Weff,defect = 5 to 75 μs and, furthermore, areas with low effective lifetime could be identified as areas with relatively high dislocation density. © Published by Elsevier Ltd. This © 2015 2015The TheAuthors. Authors. Published by Elsevier Ltd. 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 2015 under responsibility of PSE AG. Peer review by the scientific conference committee of SiliconPV 2015 under responsibility of PSE AG Keywords: Light-induced degradation; multicrystalline silicon; aluminium oxide; passivated emitter and rear cell; solar cells

1. Introduction In the past, the majority of studies on light-induced degradation (LID) within boron-doped p-type silicon have been focused on the formation of a boron-oxygen (BO) complex [1–5] and/or the dissociation of iron-boron (FeB) pairs [6–10]. Czochralski-grown silicon (Cz-Si) has been studied more extensively than multicrystalline silicon (mcSi) since it generally features relatively high oxygen concentrations in the range of [Oi] = 5×1017 to 1018 cm-3 leading to significant LID due to BO defect formation [11,12]. The concentration of interstitial iron which is generally higher in mc-Si due to in-diffusion from the crucible and its coating [13,14] can be reduced efficiently by gettering during phosphorus diffusion [15]. Concerning the BO complex, the process of regeneration is well-known [16].

1876-6102 © 2015 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 2015 under responsibility of PSE AG doi:10.1016/j.egypro.2015.07.086

600

Karin Krauss et al. / Energy Procedia 77 (2015) 599 – 606

However, Ramspeck et al. have reported LID on cells with a plasma-enhanced chemical vapour deposited (PECVD) aluminium oxide (AlOx) passivated rear side that can neither be explained by BO defect formation nor FeB pair dissociation and can even be higher for mc-Si compared to Cz-Si [17]. Furthermore, the authors have reported LID on mc-Si aluminium back-surface field (Al-BSF) solar cells being significantly higher at elevated temperatures [18], which could also not be solely attributed to BO defect formation and/or FeB pair dissociation. In a recent study investigating Al-BSF cells and cells with PECVD AlOx passivated rear sides, the authors found the defect causing this significantly higher LID is unlikely to be a sole bulk defect and may also be related to changes in the passivation quality of the rear surface [19]. The passivation quality of atomic-layer-deposited (ALD) AlOx [20,21] on p-type Si wafers has been reported to improve under light soaking at temperatures T 50°C indicating an increase of the fixed-charge density Qf [22,23]. However, the authors are not aware of light-soaking studies on PECVD AlOx passivation on p-type silicon [24], which is applied in Ref. [19] and this work. This work investigates LID under conditions practically relevant for outdoor module operation. Within this work the key findings reported in Ref. [19], where AlOx passivated mc-Si cells have been shown to be strongly affected under such conditions, are shown and discussed in more detail. Furthermore, the degradation of bulk lifetime in mcSi wafers and the stability of the rear-side passivation quality are investigated in order to determine the origin of the observed LID effect. 2. LID on cell level 2.1. Experimental Setup front-side contact

BOSCO

ARC

mc

BOSCO

PERC

Cz

Al-BSF

PERC

anneal 1 (dark, 200°C, 20min) storage 1 (dark, 25°C, 120h)

LID 1 (0.2 suns, < 40°C, 480h) storage 2 (dark, 25°C, 72h) flash 300 times anneal 2 (dark, 200°C, 20min)

LID 2 (0.15 suns, 70°C, 520h)

´classical´ defect states BO

base

FeB

ann.

Fei

ann.

FeB

deg.

Fei

deg.

FeB

deg.

Fei

emitter

via rear-side contact

passivation

front-side contact

PERC

ARC

base

rear-side contact

emitter

passivation

front-side contact

ann.

Fei

reg.

Fei

Al-BSF

ARC

a

BSF

base

emitter

rear-side contact

b Fig. 1. (a) Experimental setup for systematical exclusion of BO defect formation and FeB pair dissociation and subsequent degradation under conditions emulating critical module operation in the field. (b) Sketch of the three different cell concepts applied for the LID study on cell level. Graphs are taken from Ref. [19].

In order to quantify the impact of the classical BO formation and FeB dissociation, the experimental setup shown in Fig. 1a is applied. For a detailed description of the individual steps up to the second annealing step and their effect on the defect states of BO and FeB, please refer to Ref. [19]. After a second annealing step, the BO complex should be inactive (annealed state) and FeB pairs should be dissociated. The cells are then submitted to a second light soaking step (“LID 2”) at elevated temperatures, the temperature and carrier injection conditions chosen in a way to emulate conditions which are realistical to occur during module operation. The temperature is elevated to


601

T | 70°C representing a more critical but still realistic temperature for cells integrated in modules under field operation. Furthermore, the irradiation level is reduced to E | 0.15 suns, being lower than the irradiation levels chosen in Refs. [17,18], in order to achieve an injection level in the cells which corresponds more closely to the injection level under operation at the maximum power point (MPP) rather than under open-circuit conditions. The cells are characterized by means of current-voltage (I-V) and suns-Voc [25] measurements in-between the steps as well as time-dependently during both light soaking treatments. Fig. 1b shows the investigated cell concepts. All cells feature a homogeneous front side emitter and screen-printed metallisation. Both, the PERC (Passivated emitter and rear cell) [26] and the BOSCO (BOth Sides COllecting and COntacted) [27] cells’ rear side is passivated with a PECVD-AlOx. Furthermore the BOSCO cell features a rear-side emitter interconnected to the front side by diffused vias and disconnected from the rear base contact. All mc-Si cells are processed on neighboring wafers from standard (i.e., not high-performance [28]) industrial mc-Si material from a centre block with a base resistivity of Ub § 1.1 :cm with the processes outlined in Ref. [29]. For the Cz-Si cell, an industrial wafer with Ub § 2.5 :cm has been used. For more details regarding the cell processes applied please refer to Refs. [19,29]. The choice of cell concepts included in the study is motivated by the following considerations: i) Sensitivity towards degradation of bulk minority carrier diffusion length Lbulk: The BOSCO cell’s interconnected double-sided emitter allows for double-sided carrier collection. It has been shown that this leads to a relatively low sensitivity of the cell performance towards a decrease of Lbulk compared to PERC and Al-BSF cells. For detailed explanation and an in-depth comparison of the cell concepts please refer to Ref. [29]. Therefore, the BOSCO cell should be the least and the PERC cell the most sensitive to a LID causing a degradation of Lbulk. ii) Sensitivity towards changes in rear-side passivation quality: Both, the BOSCO and the PERC cells’ rear side feature a PECVD AlOx [24]. Therefore, changes in the rear side passivation quality should affect all cells except for the Al-BSF cell. iii) Impact of mc-Si: In order to differentiate between defects affecting both mc-Si and Cz-Si and a specific mcSi related defect causing LID, the Cz-Si PERC cell is included as a reference. 2.2. Results Upon the classical preparation of the different defect states of LID related to FeB/Fei and BO, all cells behave typically. While FeB pair dissociation only leads to minor changes in cell performance, BO formation leads to an efficiency degradation of 'K = -4.5 %rel for the Cz-Si cell and an efficiency degradation of less than 'K < |-1 %rel| for all mc-Si cells. The experimental results of “LID 1” are shown in Ref. [19] together with a detailed discussion of the impact of “LID 1” on cell performance. In summary, up to this point in the experiment, the conclusion could be drawn that mc-Si cells are affected less by LID than Cz-Si cells due to lower oxygen concentration. However, the behaviour significantly changes with modification of temperature and illumination level. Fig. 2 shows the relative degradation of the I-V parameters during “LID 2” at T = 70°C and E = 0.15 suns. The Cz-Si cell first degrades with a maximum LID of 'K = -10.6 %rel after t = 70 h resulting in a lower absolute efficiency than the mc-Si PERC cell at this point, before it recovers almost completely. Both, the mc-Si PERC and the BOSCO cells degrade significantly more than the Al-BSF cell during “LID 2”. After t = 390 h of illumination, an efficiency decrease of 'K = -2.8 / -2.9 %rel is observed for the BOSCO / PERC and of 'K = -1.3 %rel for the Al-BSF cell, respectively,. Furthermore, it can be observed that for all cells the losses in efficiency originate to a high degree from a degradation of the fill factor which at t = 390 h amounts to 'FF = -1.8 %rel and -1 %rel for the PERC/BOSCO and the Al-BSF mc-Si cell, respectively. The Cz-Si cell’s fill factor degrades up to 'FF = -4.3 %rel after t = 70 h and then recovers. The LID being pronounced most strongly in the fill factor is a result of the injection dependence of the Shockly-Read-Hall recombination induced by the LID, which can be observed by suns-Voc measurements as discussed in Ref. [19]. As a further result of the injection-dependent recombination induced by “LID 2”, the opencircuit voltage is affected least resulting in maximum degradation of 'Voc | -1 %rel or less. As the BOSCO and PERC cells, i.e., the cell structures with the lowest and the highest sensitivity to bulk diffusion length, respectively, degrade to a similar degree, the degradation phenomenon observed upon “LID 2” seems to be unlikely caused by a sole bulk defect. Possible explanations for a higher defect concentration observed in the cell structures with

602


passivated rear side compared to the Al-BSF cell may be potential aluminium gettering during Al-BSF formation, which could decrease a defect concentration in the Al-BSF cell resulting in lower LID. Furthermore, BOSCO and PERC cells feature an AlOx passivated rear side. Therefore, a degradation of surface passivation quality could be another possible explanation. To date, only an increase in passivation quality of non-PECVD deposited AlOx samples upon light soaking has been reported [30,22]. A decrease in passivation quality could explain the distinct behaviour of both, the PERC and the BOSCO cell. With respect to the AlOx-passivated Cz-Si cell, the recovery in cell performance can at least partly attributed to a permanent regeneration of the BO defect [16]. If the passivation is partly responsible for the Cz-Si cell’s first strong performance degradation, it seems to recover, too.

Cz - PERC mc - PERC mc - BOSCO mc - Al-BSF

Fig. 2. Relative degradation of efficiency, short-circuit current density, open-circuit voltage and fill factor during “LID 2” (with reference to values after anneal 1, compare Fig. 1a) for all cells. Graphs taken from Ref. [19] and accompanying “Online Supporting Information”.

3. Effective defect imaging on lifetime samples 3.1. Preparation of lifetime samples and experimental setup In order to investigate the impact of the observed LID on the effective minority charge carrier lifetime of mc-Si passivated with PECVD AlOx, test samples have been prepared. To quantify the effects, similar mc-Si material from an edge block with a base resistivity of Ub § 1.7 :cm has been used to process symmetrical lifetime samples with planarised surfaces passivated with PECVD AlOx. The processing sequence was designed in a way to ensure maximum analogy to the cell process. That is, the samples have been textured and phosphorous diffused. After removal of the phosphorus silicate glass, the samples have been treated with potassium hydroxide to etch off the phosphorus emitter and planarise the texture. Subsequently, the wafers have been passivated with PECVD AlOx on

603


both sides, stacked with PECVD silicon oxide (SiO) and silicon nitride (SiN) capping layers and submitted to fast firing in an industrial belt furnace to activate the passivation. For defect manipulation, the same treatments have been applied as on cell level, see section 2 and sketch in Fig. 1a. Lifetime has been measured spatially resolved by means of lifetime-calibrated photoluminescence (PL) measurements [31,32] in-between the different steps and time-dependently during the “LID 1” and “LID 2” treatment. Since a distinct injection-dependent recombination has been observed in the cell experiment, PL images have been recorded at an illumination level of GPL = 1 sun representing an injection level corresponding to opencircuit and GPL = 0.3 suns representing a lower injection levels close to the MPP. The PL signal has been calibrated to effective carrier lifetime via quasi-steady-state photo conductance (QSSPC) measurements [33] according to Ref. [31]. The measured effective carrier lifetime Weff is related to the lifetime in the mc-Si bulk Wbulk and the lifetime given by surface recombination Wsurf via 1

W eff

§ 1 1 · ¸¸ . ¨¨ © W bulk W surf ¹

(1)

To allow imaging of an “effective defect” which is activated upon the degradation, an effective defect lifetime

Weff,defect is defined, which is derived from the initial lifetime Weff,initial after the first annealing step and the degraded lifetime Weff,degraded after the different defect manipulation treatments via 1

W eff, defect

§ 1 1 ·¸ ¨ . ¨W ¸ W eff, degraded eff, initial © ¹

(2)

Note, that Weff,defect only reflects recombination due to instable defects which are activated upon the performed treatment but is completely free of any recombination via stable defects in the bulk or at the surface as these are subtracted out. Moreover, it is emphasized that the instable defects leading to Weff,defect may be located either in the

Effective defect lifetime Weff,defect (Ps)

Effective defect lifetime map

(c) Effective defect lifetime Weff,defect (Ps)

(b) Effective ifetime Weff,degraded (s)

Effective lifetime Weff,initial (s)

GPL = 0.3 suns

(a)

Effective ifetime Weff,degraded (s)

Lifetime map after t = 560h LID2

Effective lifetime Weff,initial (s)

GPL = 1 sun

Lifetime map initial state

Fig. 3. Effective lifetime map for PL images taken at (top) GPL = 1sun and (bottom) GPL = 0.3 suns (a) after anneal 1 (compare Fig.1a) and (b) after LID2 (illumination for t = 560 h at 70°C). (c) Effective defect lifetime after t = 560 h LID2 calculated from (a) and (b) according to Eq. (2).

604


bulk or at the surface or is a combined effect, as Weff,defect reflects an effective lifetime according to Eq. 1. Imaging of the local effective defect lifetime can be done if spatially-resolved lifetime measurements are performed via lifetime-calibrated PL in the initial and the degraded state according to Eq. 2. 3.2. Results

A B

B

A

B

Wdefect (Ps)

B

A

PL Signal a.u.

A

Defect lifetime

Fig. 3a shows the lifetime-calibrated PL images of a test sample in the initial state after the first anneal. For a PL intensity of GPL = 1 sun, the square-root harmonic mean of the local lifetime is Weff,mean = 46 s with maximum local lifetimes up to Weff | 200 s. At the end of LID1 (t = 530 h, lifetime image not depicted), the square root harmonic harmonic mean is reduced to Weff,mean = 27 s, which is even further reduced to Weff,mean = 13 s after LID2 (t = 560 h). At the end of LID 2 the maximum local lifetime is reduced to values Weff < 100 s as shown in Fig. 3b. The effective defect lifetime of all defects activated upon LID 2 is determined according to Eq. 2 from the PL images in the initial and the degraded state. As can be seen in Fig. 3c, areas with the lowest LID-induced recombination show a local defect lifetime Weff,defect > 75 s, while in areas with high LID-induced recombination the local defect lifetime reaches values well below Weff,defect < 10 s. For areas with low effective defect lifetime, no significant injection dependence of Weff,defect can be observed between GPL = 1 sun and GPL = 0.3 suns. However, for areas with Weff,defect > 50 s at GPL = 1 sun, the defect lifetime decreases when lowering the PL intensity to GPL = 0.3 sun. That is, the grains with high defect lifetime show a pronounced injection-dependence in the LID-induced recombination, which is in accordance with the observed behaviour on cell level. If only one dominating defect was activated by LID, the strong lateral variation observed in the effective defect lifetime would be related to a strong lateral variation in the defect concentration. However, a superposition of more than one defect in the bulk and/or at the surface with differing individual defect lifetimes could be the reason for the resulting lumped effective defect lifetime Weff,defect with strong lateral variations. In the top row of Fig. 4, two areas with neighbouring grains of high and low effective defect lifetime are depicted enlarged. In comparison with the PL image of the as-cut wafer shown in the bottom row of Fig. 4, it can be seen that grains with high effective defect lifetime are free from visible dislocations whereas the grains with low effective

Fig. 4. (Top) Effective defect map after t = 560 h of LID 2 and enlarged areas A and B, depicting grains with relatively high and low effective defect lifetimes. (Bottom) As-cut PL images of the wafer with corresponding outtakes of areas A and B.


605

defect lifetime show a high density of dislocations (area B) or are cornered by dislocation clusters (area A). The correlation between areas of low defect lifetime and high dislocation density may be due to the following reasons: i) Dislocation clusters are generally decorated by impurities that can form highly recombination-active bulk defects during the performed light-soaking that affect the bulk lifetime within the dislocation clusters and in neighbouring grains. In this case, low defect lifetimes would be due to bulk defects. ii) In areas of high dislocation density surface passivation quality may be instable. For example, changes in micro-roughness compared with grains with lower dislocation densities could affect the interface between mc-Si and the PECVD-AlOx. In this case, low defect lifetimes would be due to surface passivation effects on the mc-Si surface. However, the dislocations in mc-Si seem to have a strong impact on the effective lifetime degradation of AlOxpassivated mc-Si wafers and, therefore, play a significant role in the LID of mc-Si solar cells with AlOx passivation. 4. Conclusion Within this work, light-induced degradation (LID) under conditions commonly applied for the activation of the classical boron-oxygen defect, and, conditions relevant under module operation for mc-Si cells have been investigated. It has been shown that, under conditions practically relevant for module operation, mc-Si cells with PECVD aluminium oxide (AlOx) passivated rear sides degrade significantly more than both, a Cz-Si cell with PECVD AlOx passivation and a mc-Si Al-BSF cell without AlOx passivation. Furthermore, the same degree of conversion efficiency degradation of 'K = -2.8 and -2.9 %rel has been observed for a BOSCO and a PERC cell, respectively, despite their significantly different sensitivity to bulk diffusion length Lbulk. That is, a sole bulk defect seems to be unlikely to be responsible for the observed degradation and the rear side passivation seems to play a role as well. In order to further investigate the interaction of the rear-side passivation and the mc-Si bulk, an effective defect imaging including bulk and surface recombination has been performed on lifetime samples. The derived effective defect maps shows a strong variation in the local effective defect lifetime in the range of Weff,defect = 5 to 75 s. Areas with low effective lifetime and, thus, high concentration of LID-induced defects could be identified as areas with relatively high dislocation density. Acknowledgements The authors acknowledge U. Jäger for allocation of the investigated Cz cell, P. Saint-Cast, Hannes Höffler and Jonas Haunschild for fruitful discussions and K. Ortstein and N. Hamamera for support with the experiments. K. Krauss gratefully acknowledges the scholarship of the German Federal Environmental Foundation (“Deutsche Bundesstiftung Umwelt”). References [1] [2] [3] [4] [5] [6] [7]

Rein S, Rehrl T, Warta W, Glunz SW, Willeke G. Electrical and thermal properties of the metastable defect in boron-doped Czochralski silicon (Cz-Si). In: Proceedings of the 17th European Photovoltaic Solar Energy Conference; 2001, p. 1555–1560. Rein S, Diez S, Falster R, Glunz SW. Quantitative correlation of the metastable defect in Cz-silicon with different impurities. In: Proceedings of the 3rd World Conference on Photovoltaic Energy Conversion; 2003, p. 1057–1060. Schmidt J, Bothe K. Structure and transformation of the metastable boron- and oxygen-related defect center in crystalline silicon. Phys. Rev. B 2004; 69:24107, doi:10.1103/PhysRevB.69.024107. Voronkov VV, Falster R. Latent complexes of interstitial boron and oxygen dimers as a reason for degradation of silicon-based solar cells. J. Appl. Phys. 2010; 107:53509, doi:10.1063/1.3309869. Bothe K, Sinton R, Schmidt J. Fundamental boron-oxygen-related carrier lifetime limit in mono- and multicrystalline silicon. Prog. Photovolt: Res. Appl. 2005; 13:287–96, doi:10.1002/pip.586. Graff K. Metal impurities in silicon-device fabrication. 2nd ed. Berlin, New York: Springer; 2000. Istratov AA, Hieslmair H, Weber ER. Iron and its complexes in silicon. Applied Physics A: Materials Science & Processing 1999; 69:13– 44, doi:10.1007/s003390050968.

606 [8] [9] [10] [11] [12] [13]

[14]

[15] [16] [17] [18]

[19] [20]

[21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33]

Karin Krauss et al. / Energy Procedia 77 (2015) 599 – 606 Istratov AA, Hieslmair H, Weber ER. Iron contamination in silicon technology. Applied Physics A: Materials Science & Processing 2000; 70:489–534, doi:10.1007/s003390051074. Macdonald DH, Geerligs LJ, Azzizi A. Iron detection in crystalline silicon by carrier lifetime measurements for arbitrary injection and doping. J. Appl. Phys. 2004; 95:1021, doi:10.1063/1.1637136‫ޤ‬. Rein S, Glunz SW. Electronic properties of interstitial iron and iron-boron pairs determined by means of advanced lifetime spectroscopy. J. Appl. Phys. 2005; 98:113711, doi:10.1063/1.2106017. Rein S, Warta W, Glunz SW. Investigation of carrier lifetime in p-type Cz-silicon: specific limitations and realistic prediction of cell performance. In: Proceedings of the 28th IEEE Photovoltaic Specialists Conference; 2000, p. 57–60. Schmidt J, Lim B, Walter D, Bothe K, Gatz S, Dullweber T, Altermatt PP. Impurity-Related Limitations of Next-Generation Industrial Silicon Solar Cells. IEEE J. Photovoltaics 2013; 3:114–8, doi:10.1109/JPHOTOV.2012.2210030. Schubert MC, Schon J, Schindler F, Kwapil W, Abdollahinia A, Michl B, Riepe S, Schmid C, Schumann M, Meyer S, Warta W. Impact of Impurities From Crucible and Coating on mc-Silicon Quality—the Example of Iron and Cobalt. IEEE J. Photovoltaics 2013; 3:1250–8, doi:10.1109/JPHOTOV.2013.2279116. Kwapil W, Zuschlag A, Reis I, Schwirtlich I, Meyer S, Zierer R, Krain R, Kießling FM, Schumann M, Schmid C, Riepe S. Influence of Crucible and Coating on the Contamination of Directionally Solidified Silicon: First Results of the German Research Network "SolarWinS". In: Proceedings of the 27th European Photovoltaic Solar Conference and Exhibition; 2012, p. 627–635. Bentzen A, Holt A, Kopecek R, Stokkan G, Christensen JS, Svensson BG. Gettering of transition metal impurities during phosphorus emitter diffusion in multicrystalline silicon solar cell processing. J. Appl. Phys. 2006; 99:93509, doi:10.1063/1.2194387. Herguth A, Schubert G, Kaes M, Hahn G. A New Approach to Prevent the Negative Impact of the Metastable Defect in Boron Doped CZ Silicon Solar Cells. In: Proceedings of the 4th World Conference on Photovoltaic Energy Conversion; 2006, p. 940–943. Ramspeck K, Zimmermann S, Nagel H, Metz A, Gassenbauer Y, Birkmann B, Seidl A. Light Induced Degradation of Rear Passivated mcSi Solar Cells. In: Proceedings of the 27th European Photovoltaic Solar Conference and Exhibition; 2012, p. 861–865. Fertig F, Krauß K, Geisemeyer I, Broisch J, Höffler H, Odden JO, Soiland A, Rein S. Fully Solderable Large-Area Screen-Printed Al-BSF p-Type Mc-Si Solar Cells from 100 % Solar Grade Feedstock Yielding h > 17 %: Challenges and Potential on Cell and Module Level. In: Proceedings of the 27th European Photovoltaic Solar Conference and Exhibition; 2012, p. 1031–1038. Fertig F, Krauß K, Rein S. Light-induced degradation of PECVD aluminium oxide passivated silicon solar cells. Phys. Status Solidi RRL 2015; 9:41–6, doi:10.1002/pssr.201409424. Agostinelli G, Delabie A, Vitanov P, Alexieva Z, Dekkers H, Wolf S de, Beaucarne G. Very low surface recombination velocities on ptype silicon wafers passivated with a dielectric with fixed negative charge. Solar Energy Materials and Solar Cells 2006; 90:3438–43, doi:10.1016/j.solmat.2006.04.014. Hoex B, Heil, S. B. S., Langereis E, van de Sanden, M. C. M., Kessels, W. M. M. Ultralow surface recombination of c-Si substrates passivated by plasma-assisted atomic layer deposited Al[sub 2]O[sub 3]. Appl. Phys. Lett. 2006; 89:42112, doi:10.1063/1.2240736. Liao B, Stangl R, Mueller T, Lin F, Bhatia CS, Hoex B. The effect of light soaking on crystalline silicon surface passivation by atomic layer deposited Al2O3. J. Appl. Phys. 2013; 113:24509, doi:10.1063/1.4775595. Hezel R, Jaeger K. Low-Temperature Surface Passivation of Silicon for Solar Cells. J. Electrochem. Soc. 1989; 136:518–23, doi:10.1149/1.2096673. Saint-Cast P, Kania D, Hofmann M, Benick J, Rentsch J, Preu R. Very low surface recombination velocity on p-type c-Si by high-rate plasma-deposited aluminum oxide. Appl. Phys. Lett. 2009; 95:151502, doi:10.1063/1.3250157. Sinton RA, Cuevas A. A quasi-steady-state open-circuit voltage method for solar cell characterization. In: Proceedings of the 16th European Photovoltaic Solar Energy Conference; 2000, p. 1152–1155. Blakers AW, Wang A, Milne AM, Zhao J, Green MA. 22.8% efficient silicon solar cell. Appl. Phys. Lett. 1989; 55:1363, doi:10.1063/1.101596. Fertig F, Krauß K, Greulich J, Clement F, Biro D, Preu R, Rein S. The BOSCO solar cell - a both sides collecting and contacted structure. Phys. Status Solidi RRL 2014; 8:381–4, doi:10.1002/pssr.201409143. Yang YM, Yu A, Hsu B, Hsu WC, Yang A, Lan CW. Development of high-performance multicrystalline silicon for photovoltaic industry. Prog. Photovolt: Res. Appl. 2013, doi:10.1002/pip.2437. Fertig F, Greulich J, Krauß K, Clement F, Biro D, Preu R, Rein S. The BOSCO Solar Cell: Simulation and Experiment. IEEE J. Photovoltaics 2014; 4:1243–51, doi:10.1109/JPHOTOV.2014.2333875. Wolf S de, Demaurex B, Descoeudres A, Ballif C. Very fast light-induced degradation of a-Si:H/c-Si(100) interfaces. Phys. Rev. B 2011; 83:233301, doi:10.1103/PhysRevB.83.233301. Herlufsen S, Schmidt J, Hinken D, Bothe K, Brendel R. Photoconductance-calibrated photoluminescence lifetime imaging of crystalline silicon. phys. stat. sol. (RRL) 2008; 2:245–7, doi:10.1002/pssr.200802192. Giesecke JA, Schubert MC, Michl B, Schindler F, Warta W. Minority carrier lifetime imaging of silicon wafers calibrated by quasi-steadystate photoluminescence. Solar Energy Materials and Solar Cells 2011; 95:1011–8, doi:10.1016/j.solmat.2010.12.016. Sinton RA, Cuevas A, Stuckings M. Quasi-Steady-State Photoconductance, A New Method for Solar Cell Material and Device Characterization. In: Conference Record of the Twenty Fifth IEEE Photovoltaic Specialists Conference - 1996; 1996, p. 457–460.