Amorphous Silicon Oxinitride in Silicon Thin-film Solar Cells

Available online at www.sciencedirect.com ScienceDirect Energy Procedia 44 (2014) 203 – 208 E-MRS Spring Meeting 2013 Symposium D - Advanced Inorganic Materials and Structures for Photovoltaics, 27-31 May 2013, Strasbourg, France Amorphous silicon oxinitride in silicon thin-film solar cells Cordula Walder*, Alex Neumüller, Oleg Sergeev, Martin Kellermann, Karsten von Maydell, Carsten Agert NEXT ENERGY · EWE Research Centre for Energy Technology at the University of Oldenburg, Carl-von-Ossietzky-Str. 15, 26129 Oldenburg, Germany Abstract Silicon oxide is a promising material for silicon thin-film solar cells. As a doped layer it shows low parasitic absorption while as an intrinsic layer it can be used as a high band gap absorber in multijunction solar cells. Whereas doped silicon oxide layers prepared with CO2 are widely investigated, intrinsic layers and the use of N2O as a source gas for oxygen incorporation in silicon thin-film solar cells have not received much attention yet. Therefore we present the optical and electrical properties of doped and intrinsic silicon oxinitride layers at varying N2O flow as well as their performance in single cells. We found that p-doped amorphous silicon oxinitride layers enhance the short circuit current and the open circuit voltage of amorphous silicon single cells while the fill factor is reduced due to higher series resistance. When N2O is added to the intrinsic layer of amorphous silicon single cells the cell performance deteriorates drastically which is attributed to increased defect density. ©2013 2013The TheAuthors. Authors. Published by Elsevier © Published by Elsevier Ltd.Ltd. Selection under responsibility of The Materials Research SocietySociety (E-MRS) Selectionand andpeer-review peer-review under responsibility of European The European Materials Research (E-MRS). Keywords: solar cell, high bandgap, oxide, nitride, p-doped, intrinsic 1. Introduction Much effort has been made in the last years do increase the efficiency of silicon thin-film solar cells. A very critical loss mechanism of the solar cell device is parasitic absorption in the doped layers due to the high defect * Corresponding author. Tel.: +49 441 99906-219. E-mail address: [email protected] 1876-6102 © 2013 The Authors. Published by Elsevier Ltd. Selection and peer-review under responsibility of The European Materials Research Society (E-MRS) doi:10.1016/j.egypro.2013.12.028 204 Cordula Walder et al. / Energy Procedia 44 (2014) 203 – 208 density. Typically, in production the bandgap is widened by carbon to overcome this problem. In the last years pSiOx:H layers, which show improved performance in solar cells, were also investigated intensively [1-3]. The application of silicon oxide to absorber layers has been considered as well. Simulations show that in triple junction solar cells the band gap of the top cell has to be increased to about 2eV to optimize the efficiency [4]. First experiments show promising results [5-7]. We report on our development of intrinsic and doped hydrogenated amorphous silicon oxinitride (a-SiOxNy:H) for the use in silicon based thin-film solar cells. Doped a-SiOxNy:H reveals low parasitic absorption losses and has the potential to increase the open circuit voltage. Intrinsic a-SiOxNy:H is considered as high bandgap topmost absorber in multijunction solar cells. While silicon oxide layers stemming from the use of CO2 as process gas have been extensively studied as transparent contacts [1-3, 8-11] little attention has been paid to using N2O for the same purpose and for intrinsic layers [12-13]. That is why we present the optical and electrical properties of doped and intrinsic a-SiOxNy:H layers at varying N2O flow as well as their performance in single cells. 2. Experimental methods Amorphous silicon oxinitride layers were deposited by Plasma Enhanced Chemical Vapour Deposition (PECVD) using the Cluster Tool from Von Ardenne with separate chambers for the doped and intrinsic layers. The deposition frequency was 13.56 MHz. The remaining deposition parameters are presented in Table 1. Single layers were deposited on Schott Eco glass and p-i-n single cells on commercial Asahi VU transparent conductive oxide (TCO). Silver back contacts were evaporated onto the layers by electron beam. Table 1. Deposition parameters. p-a-SiOxNy:H i-a-Si:H i-a-SiOxNy:H n-a-Si:H (SiH4) [sccm] 40 40 40 40 (H2) [sccm] 80 200 600 200 (N2O) [sccm] 0-50 (B2H6) [sccm] 32 0-15 (PH3) [sccm] 25 P [W] 10 10 20 15 p [mbar] 0.3 1 1 1 del [mm] 25 15 20 20 Tsub [°C] 220 220 220 220 The optical Tauc bandgap of the single layers was obtained by transmission and reflection measurements with the UV-VIS-NIR spectrometer Cary 5000 from Varian. Conductivity was calculated from current voltage measurements in vacuum with the electrometer 6517B from Keithley. The light source for photoconductivity was a blue LED. The sample geometry for conductivity measurements is defined by parallel in plane silver electrodes with 1 mm gap width in between. Illuminated current voltage characteristics (IV) were measured with the WACOM dual lamp solar simulator according to standard test conditions (AM1.5G spectrum, 1000W/m 2, 25°C). We determined external quantum efficiencies (EQE) from differential spectral response measurements. 3. Results and discussion 3.1. p-doped amorphous silicon oxinitride Fig. 1a shows the optical Tauc bandgap ETauc for p-doped amorphous silicon oxinitride layers as a function of the N2O flow (N2O). ETauc increases linearly from about 1.75 eV to about 2.18 eV when (N2O) is raised from 0 Cordula Walder et al. / Energy Procedia 44 (2014) 203 – 208 sccm to 50 sccm. This is accompanied by a decrease in dark conductivity by several orders of magnitude as can be seen in Fig. 1b. This is in accordance with the findings of Yoon et al. [12]. Fig. 1. (a) Tauc bandgap of p-a-SiOxNy.H single layers versus N2O flow; (b) Dark conductivity of p-a-SiOxNy.H single layers versus bandgap. Clearly a tradeoff between bandgap widening and dark conductivity of the p-doped layer has to be found for good cell performance. We chose a p-doped layer with 30 sccm N2O flow which shows a high Tauc bandgap of about 2 eV and still reasonable dark conductivity in the range of 10 -7 S/cm. With this p-layer we produced a single cell featuring the following layers: Asahi VU/p-a-SiOxNy:H/i-a-Si:H/n-a-Si:H/Ag. The total thickness of the p-i-n stack is about 400 nm. In Fig. 2a IV measurements of this cell are depicted together with results from a reference cell with no N2O added to the p-layer. Apparently the addition of N2O to the p-layer leads to less parasitic absorption losses so that the short circuit current is raised by about 1 mA/cm2. This gain in current comes from the blue spectral part where a significant amount of light is lost due to absorption in the p-doped layer with no bandgap widening as demonstrated by the external quantum efficiencies in Fig. 2b. Apart from the gain in short circuit current the open circuit voltage is also slightly increased by widening the p-layer bandgap with N2O. This phenomenon can be explained by the fact that widening the p-layer bandgap introduces a barrier for the holes which therefore remain longer in the intrinsic layer and increase the charge carrier density as well as the splitting of the quasi Fermi levels before they are extracted. Unfortunately the series resistance of the cell with p-aSiOxNy:H is higher than of the reference cell which is indicated by the slope of the IV curve near the open circuit voltage (Fig. 2a). Consequently the fill factor is affected accordingly. Yet as indicated by Yoon et al. [12] this problem may be overcome by increased boron doping of the p-a-SiOxNy:H-layer. The efficiency of the solar cell can be increased further by an optimized back reflector instead of a pure silver back contact. The results show that N2O is an alternative process gas for widening the bandgap of the p-doped layer in p-i-n single cells. Fig. 2. (a) IV curves and (b) EQE of a-Si:H single cells with ( (N2O) = 30 sccm) and without ( (N2O)=0 sccm) bandgap widening of p-layer. 205 206 Cordula Walder et al. / Energy Procedia 44 (2014) 203 – 208 3.2. Intrinsic amorphous silicon oxinitrtide The incorporation of oxygen in the absorber layer is a more challenging task than in the doped layers due to the high impact of defects on the device performance. Taking this into account the N 2O flow is limited to a maximum of 15 sccm. As shown in Fig. 3a the optical Tauc bandgap of intrinsic amorphous silicon oxinitride layers rises almost linearly with the N2O flow. The gas flow ratios cannot be compared directly to those of the doped layers because different H2 gas flows were used. At a gas flow ratio of (N2O)/ (SiH4) ≈ 0.4 ( (N2O)=15sccm) a Tauc bandgap of about 2.1 eV is reached. In contrast, Das et al. reported a 10 times higher gas flow ratio to achieve the same bandgap with CO2 as a source gas instead of N2O [14]. This might be caused by easier dissociation of N2O than CO2 in the plasma or by additional nitrogen incorporation in the network. Fig. 3b shows photo and dark conductivity of i-a-SiOxNy:H-layers. Initially photo and dark conductivity rise with the addition of N2O then decrease by several orders of magnitude. The difference between photo and dark conductivity does not change continuously for different Tauc bandgaps. Das et al. observe a similar decrease of photo and dark conductivity with bandgap widening by CO2 addition but the difference between photo and dark conductivity starts to decrease as well at a bandgap of approximately 1.9 eV [14]. Fig. 3. (a) Tauc bandgap of i-a-SiOxNy.H single layers versus N2O flow; (b) Dark conductivity of i-a-SiOxNy.H single layers versus bandgap. We applied the intrinsic amorphous silicon oxinitride layers described above in p-i-n single cells with the following structure: Asahi VU/p-a-SiOxNy:H/i-a-SiOxNy:H /n-a-Si:H/Ag. Fig. 4 shows the IV parameters of these cells. The points correspond to the mean values and the error bars to the standard deviation of several equal cells on one 10cm x 10cm substrate. As can be seen the cell performance deteriorates drastically once N 2O is added to the intrinsic layer. The short circuit current is expected to decrease with N2O flow since the layer absorption is reduced with increasing bandgap. But the open circuit voltage is meant to increase with N2O flow because the rise in bandgap should allow for a bigger splitting of the quasi Fermi levels. Since the reduction in open circuit voltage is accompanied by a reduction in fill factor, mainly due to reduced parallel resistance, deep defects seem to prohibit good cell performance when N2O is added to the process gases. These defects are probably introduced by nitrogen into the atomic network. Fourier transformed infrared spectra show overlapping SiO and SiN absorption resonances. Additionally, increased absorption at 2100cm-1 is observed showing the poor quality of the material for absorber layers. On the other hand band discontinuities at the interface between the high bandgap p-doped and intrinsic layers also might result in a lower fill factor and open circuit voltage. Authors using CO2 instead of N2O (Inthisang et al. [5] and Sritharathikhun et al. [7]) achieve better results in single solar cells with VOC exceeding 1V. In contrast to the p-doped layer the increased defect density in the intrinsic layer leads to inferior cell performance. Cordula Walder et al. / Energy Procedia 44 (2014) 203 – 208 Fig. 4. (a) Open circuit voltage, (b) fill factor, (c) short circuit current and (d) efficiency of a-SiOxNy:H single cells versus N2O flow in i-layer. 4. Conclusion and Outlook We found that widening the bandgap of the p-layer with N2O increases the short circuit current and the open circuit voltage of a-Si:H single cells while the fill factor is reduced due to larger series resistance. However this drawback can be overcome by increased boron doping of the p-layer. These results show that N2O is an alternative process gas to the commonly used CH4 or CO2 for widening the band gap of the p-a-Si:H layer. Electrical properties of intrinsic a-SiOxNy:H single layers for the use as high band gap absorbers seem to improve slightly with the initial addition of N2O. Yet single cell properties deteriorate drastically once N2O is added to the i-layer. Apparently N2O introduces too many defects which could be related to nitrogen incorporation since better results have been shown using CO2 as a source gas instead. Consequently, it will be interesting to investigate the influence of other bandgap widening process gases such as CO 2 and CH4 in comparison to N2O. Acknowledgements We would like to thank our colleagues from NEXT ENERGY for fruitful discussions and for their help with layer depositions. References [1] Biron R, Pahud C, Haug FJ, Escarré J, Söderström K, Ballif C. Window layer with p doped silicon oxide for high Voc thin-film silicon-i-p solar cells. J Appl Phys 2011; 110:124511-124517. [2] Cuony P, Alexander DTL, Perez-Wurfl I, Despeisse M, Bugnon G, Boccard M, Söderström T, Hessler-Wyser A, Hébert C, Ballif C. Silicon filaments in silicon oxide for next-generation photovoltaics. Advanced Materials 2012; 24:1182-1186. [3] Söderström K, Bugnon G, Biron R, Pahud C, Meillaud F. Thin-film silicon triple-junction solar cell with 12.5% stable efficiency on innovative flat light-scattering substrate. J Appl Phys 2012; 112:114503-114506. 207 208 Cordula Walder et al. / Energy Procedia 44 (2014) 203 – 208 [4] Yunaz I, Yamada A, Konagai M. Theoretical analysis of amorphous silicon alloy based triple junction solar cells. Jpn J Appl Phys 2007; 46:L1152. [5] Inthisang S, Sriprapha K, Miyajima S, Yamada A, Konagai M. Hydrogenated amorphous silicon oxide solar cells fabricated near the phase transition between amorphous and microcrystalline structures. Jpn J Appl Phys 2009; 48:122402. [6] Sriprapha K, Piromjit C, Limmanee A, Sritharathikhun J. Development of thin film amorphous silicon oxide/microcrystalline silicon double-junction solar cells and their temperature dependence. Solar Energy Materials and Solar Cells 2011; 95:115-118. [7] Sritharathikhun J, Moollakorn A, Kittisontirak S, Limmanee A, Sriprapha K. High quality hydrogenated amorphous silicon oxide film and its application in thin film silicon solar cells. Current Applied Physics 2011; 11:S17-S20. [8] Sichanugrist P, Sasaki T, Asano A, Ichikawa Y, Sakai H. Amorphous silicon oxide and its application to metal/n-i-p/ITO type a-Si solar cells. Solar Energy Materials and Solar Cells 1994; 34:415-422. [9] Janssen R, Janotta A, Dimova-Malinovska D, Stutzmann M. Optical and electrical properties of doped amorphous silicon suboxides. Phys Rev B 1999; 60:13561-13572. [10] Janotta A, Janssen R, Schmidt M, Graf T, Stutzmann M, Görgens L, Bergmaier A, Dollinger G, Hammerl C, Schreiber S, Stritzker B. Doping and its efficiency in a-SiOx:H. Phys Rev B 2004; 69:115206. [11] Xiao L, Astakhov O, Carius R, Lambertz A, Grundler T, Finger F. Defects and structure of μc-SiOx:H deposited by PECVD. Phys Status Solidi C 2010; 7:941-944. [12] Yoon K, Kim Y, Park JJ, Shin CH, Baek S, Jang J, Iftiquar SM, Yi J. Preparation and characterization of p-type hydrogenated amorphous silicon oxide film and its application to solar cell. Journal of Non-Crystalline Solids 2011; 357:2826-2832. [13] Brinkmann N, Sommer D, Micard G, Hahn G, Terheiden B. Electrical, optical and structural investigation of plasma-enhanced chemicalvapor-deposited amorphous silicon oxynitride films for solar cell applications. Solar Energy Materials and Solar Cells 2013; 108:180-188. [14] Das D, Iftiquar SM, Barua AK. Wide optical-gap a-SiO:H films prepared by rf glow discharge. Journal of Non-Crystalline Solids 1997; 210:148-154. View publication stats