- Email: [email protected]
Process monitoring of texture-etched high-rate ZnO:Al front contacts for silicon thin-film solar cells
Thin Solid Films 532 (2013) 66–72
Contents lists available at SciVerse ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Process monitoring of texture-etched high-rate ZnO:Al front contacts for silicon thin-film solar cells Gabrielle Jost ⁎, Tsvetelina Merdzhanova 1, Thomas Zimmermann 2, Jürgen Hüpkes 3 Forschungszentrum Jülich GmbH, IEK-5 Photovoltaik, 52425 Jülich, Germany
a r t i c l e
i n f o
Available online 22 December 2012 Keywords: ZnO:Al Process monitoring Sputter deposition Thin-film Texture Solar cell Optimization Light trapping
a b s t r a c t In this study angular resolved scattering measurements are used to monitor the surface texture properties of sputter-etched ZnO:Al and to identify the influences of process parameter changes on the surface texture. Variations in pressure and temperature are possible drifts in the industrial fabrication that can influence the ZnO:Al properties. Therefore – with respect to industrial relevance – this study focuses on the influence of these parameters. Deposition parameter dependent trends in the angular resolved scattering are studied. Correlations between changes in the results of the angular resolved scattering measurements and the performance of a-Si:H/μc-Si:H tandem solar cells are shown. Additionally, it is demonstrated that the measurements can be used to optimize the surface texture of ZnO: Al for single-junction μc-Si:H solar cells by adjusting the deposition conditions according to the previously identified trends. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Sputtered and texture-etched ZnO:Al is a promising front contact in silicon thin-film solar cells [1–3]. For thin-film devices a well-defined surface texture determines the performance significantly [4,5]. The consistency of the texture is one of the predominant drawbacks in industrial production of sputter-etched ZnO:Al front contacts because of its strong dependence on the deposition conditions. Hence, even small changes or process instabilities can lead to changes in the surface texture [6]. Since changes in process temperature and pressure do not always occur deliberately, but are due to inhomogeneity especially in industrial large area fabrication, means of process control need to be established to facilitate a stable front contact quality. A common process control approach is the optical haze measurement which has been found to be insufficient to describe texture changes adequately with respect to the solar cell performance [7]. Atomic force microscope imaging allows a more detailed analysis of the surface texture [8] but this method does not fulfill the requirements for fast and cost effective process control. Angular resolved scattering (ARS) is an inexpensive, fast optical measurement that gives detailed information regarding the surface texture. Using two optical models the ARS results – called angular intensity distribution (AID) – can be interpreted as an image of inclination angles or
⁎ Corresponding author. Tel.: +49 2461 61 6310; fax: +49 2461 61 3735. E-mail addresses: [email protected] (G. Jost), [email protected] (T. Merdzhanova), [email protected] (T. Zimmermann), [email protected] (J. Hüpkes). 1 Tel.: +49 2461 61 3177; fax: +49 2461 61 3735. 2 Tel.: +49 2461 61 3242; fax: +49 2461 61 3735. 3 Tel.: +49 2461 61 2594; fax: +49 2461 61 3735. 0040-6090/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tsf.2012.11.147
feature sizes on the surface [9,10]. Besides the ability to describe the textured surface the AID has also proven to be a predictive tool for solar cell current. It has been shown already that the AID data can be used to correlate the large angle scattering intensity with the short-circuit current densities of thin-film silicon solar cells [7,11,12]. In the presented study ARS measurements are applied as a monitoring tool for high-rate sputtered, texture-etched ZnO:Al. Process parameter dependent trends in the AID are identified. a-Si:H/μc-Si:H tandem cell parameters are presented to confirm that the detectable differences in the ARS monitoring are significantly influencing the solar cell performance. Furthermore an example is presented showing that the known correlation between AID and single-junction μc-Si:H cell performance and the identified trends can be combined to optimize the surface texture of ZnO:Al for single-junction μc-Si:H solar cells. 2. Experimental details All samples were deposited on commercially available float glass ipawhite (from Interpane Glas Industrie AG). In a first production step all float glass substrates were coated with an approximately 150 nm thick SiOxNy-interlayer. This interlayer has three functions: first of all, it is a barrier-layer to prevent Na+-ions from diffusing out of the glass substrate. Secondly, the interlayer works as an anti-reflective coating between the glass substrate and the subsequently deposited transparent conductive oxide (TCO) layer. Finally, the interlayer also serves as a seed-layer for the growth of the TCO layer. The SiOxNy-film is sputter-deposited in a reactive process using two planar Sispa® targets (by Heraeus Materials Technology GmbH & Co. KG). Oxygen and nitrogen were supplied to the chamber during the sputtering process. The excitation mode was magnetron-assisted mid-frequency at
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
40 kHz with a power output of 4 kW. To simulate the use of multiple cathodes in an industrial in line system the samples were deposited in the so called dynamic mode with a moving carrier passing the cathodes multiple times. In a second production step ZnO:Al was sputtered onto the interlayer using two 0.5 wt.% aluminum-doped ceramic tube targets. These rotatable tube targets were operated in magnetron configuration at midfrequency mode with 40 kHz. The power output was kept constant at 10 kW to facilitate high-deposition rates of up to 62.6 nm∗m/min. Similar to the SiOxNy-coating the ZnO:Al deposition was executed in dynamic mode with the substrate carrier passing the targets multiple times during the deposition. While output power and gas flow stayed constant during the deposition processes, the temperature and pressure values were varied from one deposition to the next to simulate drifts and inhomogeneity in an industrial large area fabrication. These parameter changes were repeated for two series of different thickness of which the first, thicker series achieved film-thicknesses around 1280 nm (10 passes), whereas the second, thinner series yielded thicknesses around 900 nm (7 passes). All ZnO:Al layers and interlayers were deposited in a vertical inline sputter deposition system VISS 300 by von Ardenne Anlagentechnik Dresden GmbH [6]. For further deposition parameter details see Table 1. After deposition the ZnO:Al films were texture-etched in 0.5% HCl to roughen the surface. The etch step was conducted at room temperature for 120 s for all thick and 40 s for all thin ZnO:Al films. The textured ZnO:Al substrates were investigated using a scanning electron microscope (SEM) to gain insight into the surface features. Furthermore, the samples were inspected optically using an angular resolved light scattering set-up. The ARS set-up uses a green laser with a wavelength of 550 nm. The laser beam is chopped at a frequency of 230 Hz and then split into a reference beam and a measurement beam. The reference beam is led directly onto a silicon photodiode whereas the measurement beam passes through an aperture, striking the smooth glass side of the sample perpendicularly. A planar silicon photodiode is used to detect the transmitted, scattered light on the other side of the sample. The detector is free to move on a circular path at a fix distance around the sample. The step size of the moving detector can be set to 1° leading to an equivalent angular resolution of the measured signal. The measured photodiode current was used to calculate the normalized angular intensity distribution (AID) taking the solid angle into account. The investigated ZnO:Al films were applied as front contact layer in a-Si:H/μc-Si:H tandem solar cells to study their influence on the solar cell performance. The silicon absorber layers were deposited in a single-chamber process. Details regarding the silicon deposition system and process are described elsewhere [13,14]. Room-temperature sputtered ZnO:Al and silver were employed as back contact and the area of the tandem solar cell was defined by laser scribing. The IV-characteristic of the solar cells were measured in a class A sun simulator using an AM 1.5 spectrum. Additionally, external quantum efficiency (EQE) measurements were conducted to identify the current distribution in the tandem device.
Table 1 Process parameters or parameter intervals of SiOxNy interlayer deposition and ZnO:Al deposition. Process parameter
Unit
SiOxNy
ZnO:Al
Heater temperature Pressure Power ΦAr ΦO2 ΦN2 Vsubstrate carrier Passes
[°C] [Pa] [kW] [sccm] [sccm] [sccm] [mm/s]
25 0.3 4 200 18 100 3 2
430–500 0.3–1 10 200 – – 8 7 or 10
67
3. Results and discussion 3.1. Temperature series Fig. 1 shows the SEM images of the deposited ZnO:Al surfaces after the texture-etch. The ZnO:Al samples in the left column show a heater temperature series from 430 °C to 500 °C with all deposited layers having a thickness around 1280 nm-thick series. The right column of SEM-images shows the same temperature series from 430 °C–500 °C for thin deposited films with a thickness of around 900 nm. The thick layer series shows a development in surface topography leading from rather smooth surfaces with small crater-like features at 430 °C to a very rough surface with large craters at 450 °C. Raising the process temperature even further the features seem to flatten again, leading to an almost smooth surface at 500 °C. For the thin layer series the behavior is similar. Small craters at 430 °C are increasing in diameter while the process temperature is increased leading to very distinct craters at 450 °C. Further increase in process temperature facilitates even larger craters, however, this is accompanied by a general flattening of the surface features. Fig. 2 shows the AID results from 0° to 90° scattering angle of the two temperature series. Fig. 2A) displays the thick series whereas Fig. 2B) shows the thin series. The gray arrows within the graphs indicate the order of the measured samples from low temperature to high deposition temperature. The thick samples in Fig. 2A) have a decreasing ARS signal intensity at large angles (≥35°) with increasing deposition temperature. At the same time the scattered light intensity at angles below 30° increases and the maximum or peak value of scattered light intensity shifts to smaller angles with increasing temperature. The thin samples in Fig. 2B) show a similar behavior. With increasing deposition temperature the scattered light intensities at angles below 30° increases and the peak value shift towards smaller angles. The scattered intensity at large angles, however, shows a different trend. Beginning with a low temperature of 430 °C the scattered light intensity at large angles at first increases with temperature, reaching a maximum for the sample deposited at 450 °C. Increasing the temperature further leads to a decrease in scattered light intensity at large angles alike the thick series. The SEM finding may be explained using the modified Thornton Model by Kluth et al. [15] or the further advanced study by Berginski et al. [16]. Both studies state that the deposition temperature has a significant impact on the growth of sputter deposited ZnO:Al. With increasing temperature the films become denser. The difference in density becomes visible after etching when the surfaces reveal different textures. These textures are grouped into three types: type A or I (low temperature, low density) reveals very small features and reveals almost no roughness after etching, type B or II (medium temperature, medium density) shows a homogeneous distribution of large feature on the surface, type C or III (high temperature, high density) shows even larger features than B but with heterogeneous distribution. The SEM images in Fig. 1 show for both temperature series – thin and thick – the evolution from type A surface at 430 °C to a type B surface at 450 °C and further to a type C surface at 500 °C. The angular resolved scattering results can be interpreted using several optical models [9,10,17]. This paper will focus on the following two models for interpretation. The first model is based on the raytracing theory [9]. Hence, every single ray that strikes the textured ZnO:Al surface will undergo scalar scattering in accordance with Snell's law. The measured angular resolved scattering intensities are therefore the superposition of all single scattering events in the illuminated surface area. Large or “flat” surface angles will be visible in the small angle range of the scattered intensity whereas small or “steep” angles appear in the large angle intensities. The second model treats the ZnO: Al surface as a superposition of diffraction gratings with different lattice constants [10]. In this model large grating constants or “large surface features” will appear in the small angle range of scattered intensities
68
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
Fig. 1. Scanning electron microscope images of ZnO:Al surfaces after the texture-etch in 0.5% HCl. The left column of images shows the thick layers depending on the deposition temperature, the right column the thin layers depending on the deposition temperature.
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
69
measurements reveal detailed surface information regarding surface feature size and inclination angles upon interpretation. Furthermore, not only differences are detectable but also temperature correlated trends in the AID are identifiable. 3.2. Pressure series
Fig. 2. Angular intensity distribution of light scattered at texture-etched ZnO:Al surfaces. A) thick ZnO:Al layer deposited at different temperatures; B) thin ZnO:Al layer deposited at different temperatures.
whereas small lattice constants or “small surface features” appear in the large angle range of scattered intensities. Applying these two models to the measured AID, the peak shift to smaller angles can be interpreted as the transition to larger features with increasing deposition temperature for both series. At the same time the intensity loss at large angles with increasing temperature can be interpreted as the disappearance of small surface features or vice versa “steep” surface angles. The drop in large angle scattered intensity between 450 °C and 430 °C in the thin temperature series cannot be attributed to an increase in feature size as the crater diameter obviously decreases. The presence of flatter surface angles seems to be a more reasonable explanation. This interpretation is in agreement with the model of Kluth et al. since they predict that a reduction in temperature will lead to a type A structure at some point. This type A structure characteristically exhibits laterally small features and additionally small vertical dimensions. The missing vertical dimensions accompany flatter surface angles leading to a decrease in the scattered intensity at large angles. The fact that the thick series does not show the same behavior can be attributed to the fact that the lowest probed temperature of 430 °C is not low enough for this configuration to create a type A surface texture. Concluding it can be determined that even very small differences in deposition temperature of only 10 °C, which can easily occur in large area fabrication, can already lead to significant changes in the surface topography after etching. These differences in topography are only partially accessible via SEM imaging. ZnO:Al samples that appear very similar in the SEM images e.g. the thin 430 °C and 440 °C surface show differences in the ARS measurements. The more cost effective ARS
Fig. 3 shows the SEM images of differently deposited ZnO:Al surfaces after the texture-etch. The samples in the left column show a pressure series from 0.3 Pa to 1 Pa with all deposited layers having a thickness around 1280 nm-thick series. The right column of SEM-images belongs to the same pressure series from 0.3 Pa to 1 Pa for thin deposited samples with a thickness of around 900 nm. The thick layer series shows no obvious differences between the samples at different deposition pressure. Apparently, all three samples reveal a surface covered with similar large crater structures. For the thin layer series the behavior is different. Starting at a low pressure of 0.3 Pa with a homogeneously crater covered surface, the crater size and distribution appears consistent when shifting the pressure to 0.5 Pa but the craters seem to flatten. Increasing the pressure even further to 1 Pa then leads to steeper features again but with decreased lateral size. Fig. 4 presents the angular resolved scattering results of the two pressure series. Fig. 4A) displays the thick series whereas Fig. 4B) shows the thin series. Similar to the SEM results the thick series shows no significant differences in the AID either. All three samples show a strong peak at small angles (around 8°) and about one order of magnitude in intensity lower scattering at large angles (≥35°). The thin series in Fig. 4B) on the other hand shows different angular intensity distribution for each of the pressure samples. The 0.3 Pa sample shows a strong peak at small angles (around 15°) and about a factor of six lower scattering intensities at large angles. Changing the pressure from 0.3 Pa to 0.5 Pa the large angle scattering decreases while the small angle scattering stays constant. Raising the pressure even further from 0.5 Pa up to 1 Pa the small angle peak diminishes significantly and the large angle scattering increases again. The SEM findings in the pressure series can partially be explained by the Modified Thornton Model [15]. According to Kluth et al. an increase in process pressure with otherwise constant deposition conditions will lead to a material shift from a type C (large crater with inhomogeneous distribution) to type B (large craters with homogeneous distribution) and finally to a type A (small features with homogeneous distribution). For the thick sample series in Fig. 3), all samples belong to type B structure. Apparently, the change in process pressure is not significant enough to cause differences in the etching behavior of the samples. Within the investigated boundaries of 0.3 Pa to 1 Pa the thick sample series is unaffected by pressure changes. For the thin sample series there are clearly visible differences in the SEM images. However, the behavior does not follow the theory of the Modified Thornton Model. With increasing the pressure from 0.3 Pa to 0.5 Pa, type B at 0.3 Pa does not change into a type C structure but a flattened type B structure. Increasing the pressure further to 1 Pa then leads to the transition to a type C structure as expected according to the Modified Thornton Model. The aforementioned two optical models are applied again to interpret the measured AID. For the thick sample series the peak at the rather small angle of about 8° indicates the presence of large features and or “flat” angles on the surface. All three pressure samples show the same angular intensity distribution indicating that they are similar. The differences in the thin sample series on the other hand indicate no significant change in the feature size when shifting from 0.3 Pa to 0.5 Pa, as the peak position stays almost constant at 12°. However, the decrease in large angle scattering intensity suggests the presence of less steep angles on the surface or the flattening of the features. Increasing the pressure further then leads to a diminishing of the small angle peak, indicating that the surface features reduce in size. At the same time the large angle scattering rises again, suggesting that there are steeper angles on the surface.
70
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
Fig. 3. SEM images of ZnO:Al surfaces after the texture-etch in 0.5% HCl. The left column of images shows the thick layers depending on the deposition pressure, the right column the thin layers depending on the deposition pressure.
Concluding it can be stated that the pressure changes seem to affect the deposited series to a different degree. The thick sample series undergoes pressure changes during deposition within our investigated constraints without any changes to the surface topography after etching. Meanwhile applying the same pressure changes, there are dramatic changes detectable for the thin sample series. Hence, the thick sample series proves to be more stable than the thin sample series in the pressure regime. 3.3. Solar cell parameters The ZnO:Al films used in the tandem solar cells as a front TCO contacts were the samples of the thin layer pressure series which were etched for different etch times than the reference samples in 3.2. The etch duration was adapted to 42 s, 46 s and 29 s for the ZnO:Al films deposited at 0.3 Pa, 0.5 Pa and 1 Pa respectively to ensure the same thickness after etching rather than the same etch time. The angular intensity distribution as seen in Fig. 4B was not significantly altered. The solar cell parameters of the a-Si:H/μc-Si:H tandem cells with a total silicon absorber thickness of 1.96 μm that were deposited on the thin pressure series samples are summarized in Table 2. The solar cell parameters show only small differences in the solar cell performance of the 0.3 Pa and 0.5 Pa samples whereas the 1 Pa sample deviates strongly in all characteristics. The current distribution
shows almost the same top-cell current density for all three solar cell types, whereas the bottom cell current density deviates by 0.7 mA/cm 2 between the two lower pressure samples and the 1 Pa sample. Additionally, the open circuit voltage drops by more than 30 mV when shifting to a high pressure of 1 Pa. This leads to an about 1% abs. lower solar cell efficiency of the 1 Pa sample in comparison with the two low pressure samples. Fig. 5 shows the external quantum efficiency spectra of the tandem solar cells deposited on top of the thin ZnO:Al pressure series. The EQE of the top cell shows almost no change for the three different ZnO:Al TCOs. The same applies for the bottom cell current up to a wavelength of 700 nm. Beyond this wavelength the EQE of the ZnO:Al types differs significantly. The low pressure (0.3 Pa) and medium pressure (0.5 Pa) samples show only very slight deviations whereas the high pressure sample (1 Pa) shows a strongly reduced EQE for wavelengths above 700 nm. The solar cell results show a strong correlation with the AID of the texture-etched ZnO:Al front contacts. The ability to trap light within the solar cell is one of the most important qualities of a textured front contact. Especially for wavelength above 700 nm a pathlength enhancement through light trapping is essential to allow sufficient absorption. The fact, that in the case of the tandem devices shown the bottom cell EQE is only affected for wavelength above 700 nm, indicates that the differences in structure that were detected as differences in the AID are affecting the light trapping behavior of the
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
71
A
Fig. 5. External quantum efficiency spectra of a-Si:H/μc-Si:H solar cells with pressure series of thin deposited ZnO:Al films applied as front contact.
B A
B
Fig. 4. Angular intensity distribution of light scattered at texture-etched ZnO:Al surfaces. A) thick ZnO:Al layer deposited at various process pressures; B) thin ZnO:Al layer deposited at various process pressures.
solar cell. Therefore angular resolved scattering measurements are an effective method to detect differences in the front contact surface texture that may lead to differences in the device performance. In the following section it will be demonstrated, how this measurement can not only be used to monitor the quality of a front contact but how to use it to optimize the front contact surface texture. 3.4. Texture optimization Fig. 6 illustrates the value of the angular intensity distribution at 60° scattering angle of the thick ZnO:Al temperature series a) and the thin ZnO:Al temperature series b) of Section 3.1 as a function of the heater temperature TH during deposition. Fig. 6A) shows the Table 2 Solar cell parameters of a-Si:H/μc-Si:H tandem cells with thin deposited ZnO:Al pressure series applied as front contact. Solar cell parameter
Eta FF Voc Isc Isc,top Isc,bottom
Unit
[%] [%] [V] [mA/cm2] [mA/cm2] [mA/cm2]
Process pressure in ZnO:Al deposition 3 μbar
5 μbar
10 μbar
11.9 72.2 1.401 11.7 11.5 12.1
11.6 72.4 1.397 11.5 11.4 12.0
10.8 70.6 1.365 11.2 11.4 11.3
Fig. 6. Angular intensity distribution of thick (A) and thin (B) ZnO:Al temperature series at 60° scattering angle plotted as a function of the heater temperature TH during deposition.
highest AID at 60° for the samples deposited at the lowest deposition temperature probed in this study — at 430 °C. With further increased deposition temperature the AID at 60° decreases. In contrast Fig. 6B) shows a maximum of scattered intensity at 60° for the thin sample deposited at 450 °C. Samples deposited at higher or lower temperature reveal lower scattering intensities at this angle. It is known from previous studies that the integrated scattered intensity at angles between 45°–90° – often simply referred to as large angle scattering – correlates with the short-circuit current density of single-junction μc-Si:H solar cells [7]. It was also shown that rather than an integration of all large angles, a single sampling point in terms of a discrete large angle can represent the same correlation between scattering intensity and short-circuit current density. In [12] the discrete scattering intensity of 60° was correlated with the short-circuit current density, with higher AID (60°) values leading to higher short-circuit current densities. Taking Fig. 6 into account there is a distinct maximum of AID (60°) over the deposition temperature. Hence, to improve the surface texture of the sputtered ZnO:Al for high currents in a μc-Si:H solar cell by varying the deposition temperature, the angular resolved scattering measurement can be used to find optimum by searching the maximum of AID (60°). 4. Conclusion In this paper angular resolved scattering measurements as process monitoring tool for texture-etched ZnO:Al in silicon thin-film solar cells were investigated. It was shown that ARS measurements can be used to
72
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
monitor the surface texture changes that occur due to process parameter changes in temperature and pressure. The detected differences in angular resolved scattering measurements are found to correlate with the changes in short-circuit current densities. Finally, process parameter dependent trends were identified in the angular intensity distribution that can be used to modify or optimize the surface texture in order to increase the short-circuit current density in single-junction μc-Si:H solar cells.
[5] [6] [7] [8] [9] [10] [11]
Acknowledgments The authors would like to thank H. Siekmann, J. Worbs, D. Weigand, A. Bauer and G. Schöppe for their technical assistance; A. Doumit and H.P. Bochem for SEM imaging and C. Zahren for device characterization. Financial support by the German Federal Ministry (BMU, grants 0327625, 0325299A and 0325356B) is gratefully acknowledged. References [1] J. Nomoto, T. Hirano, T. Miyata, T. Minami, Thin Solid Films 520 (2011) 1400. [2] J.K. Rath, Y. Liu, M.M. de Jong, J. de Wild, J.A. Schuttauf, M. Brinza, R.E.I. Schropp, Thin Solid Films 518 (2010) e129. [3] J.-S. Cho, Y.-J. Jeong, S.-H. Park, K.H. Yoon, Sol. Energy Mater. Sol. Cells 95 (2011) 190. [4] J. Bailat, D. Dominé, R. Schlüchter, J. Steinhauser, S. Fay, F. Freitas, C. Bücher, L. Feitknecht, X. Niquille, T. Tscharner, A. Shah, C. Ballif, in: Conference Record of
[12]
[13] [14]
[15] [16] [17]
the 2006 IEEE 4th World Conference on Photovoltaic Energy Conversion, 2006, p. 1533. J. Müller, B. Rech, J. Springer, M. Vanecek, Solar Energy 77 (2004) 917. J. Hüpkes, H. Zhu, J.I. Owen, G. Jost, E. Bunte, Thin Solid Films (2012), http://dx.doi.org/10.1016/j.tsf.2011.09.031. H. Schade, P. Lechner, R. Geyer, H. Stiebig, B. Rech, O. Kluth, in: Conference Record of 31st IEEE — Photovoltaic Specialist Conference, 2005, p. 1436. A. Krasnov, Sol. Energy Mater. Sol. Cells 94 (2010) 1648. M. Schulte, K. Bittkau, B.E. Pieters, S. Jorke, H. Stiebig, J. Hüpkes, U. Rau, Proc. SPIE 7717 (2010) 77170N. K. Bittkau, M. Schulte, T. Beckers, R. Carius, Proc. SPIE 7725 (2010) 77250N-3. W. Dewald, V. Sittinger, B. Szyszka, D. Wippler, P. Obermeyer, F. Hamelmann, H. Stiebig, F. Säuberlich, D. Severin, S. Klein, M. Rohde, U. Schmidt, in: Proceedings of European Photovoltaic Solar Energy Conference and Exhibition, WIP Renewable Energies, Hamburg, 2011, p. 2704. G. Jost, E. Bunte, J. Worbs, H. Siekmann, J. Hüpkes, in: Proceedings of European Photovoltaic Solar Energy Conference and Exhibition, WIP Renewable Energies, Valencia, 2010, p. 3078. T. Merdzhanova, J. Woerdenweber, T. Zimmermann, U. Zastrow, A.J. Flikweert, H. Stiebig, W. Beyer, A. Gordijn, Sol. Energy Mater. Sol. Cells 98 (2012) 146. J. Woerdenweber, T. Merdzhanova, T. Zimmermann, A.J. Flikweert, H. Stiebig, W. Beyer, A. Gordijn, J. Non-Cryst. Solids (2012), http://dx.doi.org/10.1016/j.jnoncrysol.2011. 12.041. O. Kluth, G. Schöpe, J. Hüpkes, C. Agashe, J. Müller, B. Rech, Thin Solid Films 442 (2003) 80. M. Berginski, J. Hüpkes, M. Schulte, G. Schöpe, H. Stiebig, B. Rech, M. Wuttig, J. Appl. Phys. 101 (2007) 074903. D. Dominé, F.-J. Haug, C. Battaglia, C. Ballif, J. Appl. Phys. 107 (2010) 044504.
Contents lists available at SciVerse ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Process monitoring of texture-etched high-rate ZnO:Al front contacts for silicon thin-film solar cells Gabrielle Jost ⁎, Tsvetelina Merdzhanova 1, Thomas Zimmermann 2, Jürgen Hüpkes 3 Forschungszentrum Jülich GmbH, IEK-5 Photovoltaik, 52425 Jülich, Germany
a r t i c l e
i n f o
Available online 22 December 2012 Keywords: ZnO:Al Process monitoring Sputter deposition Thin-film Texture Solar cell Optimization Light trapping
a b s t r a c t In this study angular resolved scattering measurements are used to monitor the surface texture properties of sputter-etched ZnO:Al and to identify the influences of process parameter changes on the surface texture. Variations in pressure and temperature are possible drifts in the industrial fabrication that can influence the ZnO:Al properties. Therefore – with respect to industrial relevance – this study focuses on the influence of these parameters. Deposition parameter dependent trends in the angular resolved scattering are studied. Correlations between changes in the results of the angular resolved scattering measurements and the performance of a-Si:H/μc-Si:H tandem solar cells are shown. Additionally, it is demonstrated that the measurements can be used to optimize the surface texture of ZnO: Al for single-junction μc-Si:H solar cells by adjusting the deposition conditions according to the previously identified trends. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Sputtered and texture-etched ZnO:Al is a promising front contact in silicon thin-film solar cells [1–3]. For thin-film devices a well-defined surface texture determines the performance significantly [4,5]. The consistency of the texture is one of the predominant drawbacks in industrial production of sputter-etched ZnO:Al front contacts because of its strong dependence on the deposition conditions. Hence, even small changes or process instabilities can lead to changes in the surface texture [6]. Since changes in process temperature and pressure do not always occur deliberately, but are due to inhomogeneity especially in industrial large area fabrication, means of process control need to be established to facilitate a stable front contact quality. A common process control approach is the optical haze measurement which has been found to be insufficient to describe texture changes adequately with respect to the solar cell performance [7]. Atomic force microscope imaging allows a more detailed analysis of the surface texture [8] but this method does not fulfill the requirements for fast and cost effective process control. Angular resolved scattering (ARS) is an inexpensive, fast optical measurement that gives detailed information regarding the surface texture. Using two optical models the ARS results – called angular intensity distribution (AID) – can be interpreted as an image of inclination angles or
⁎ Corresponding author. Tel.: +49 2461 61 6310; fax: +49 2461 61 3735. E-mail addresses: [email protected] (G. Jost), [email protected] (T. Merdzhanova), [email protected] (T. Zimmermann), [email protected] (J. Hüpkes). 1 Tel.: +49 2461 61 3177; fax: +49 2461 61 3735. 2 Tel.: +49 2461 61 3242; fax: +49 2461 61 3735. 3 Tel.: +49 2461 61 2594; fax: +49 2461 61 3735. 0040-6090/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tsf.2012.11.147
feature sizes on the surface [9,10]. Besides the ability to describe the textured surface the AID has also proven to be a predictive tool for solar cell current. It has been shown already that the AID data can be used to correlate the large angle scattering intensity with the short-circuit current densities of thin-film silicon solar cells [7,11,12]. In the presented study ARS measurements are applied as a monitoring tool for high-rate sputtered, texture-etched ZnO:Al. Process parameter dependent trends in the AID are identified. a-Si:H/μc-Si:H tandem cell parameters are presented to confirm that the detectable differences in the ARS monitoring are significantly influencing the solar cell performance. Furthermore an example is presented showing that the known correlation between AID and single-junction μc-Si:H cell performance and the identified trends can be combined to optimize the surface texture of ZnO:Al for single-junction μc-Si:H solar cells. 2. Experimental details All samples were deposited on commercially available float glass ipawhite (from Interpane Glas Industrie AG). In a first production step all float glass substrates were coated with an approximately 150 nm thick SiOxNy-interlayer. This interlayer has three functions: first of all, it is a barrier-layer to prevent Na+-ions from diffusing out of the glass substrate. Secondly, the interlayer works as an anti-reflective coating between the glass substrate and the subsequently deposited transparent conductive oxide (TCO) layer. Finally, the interlayer also serves as a seed-layer for the growth of the TCO layer. The SiOxNy-film is sputter-deposited in a reactive process using two planar Sispa® targets (by Heraeus Materials Technology GmbH & Co. KG). Oxygen and nitrogen were supplied to the chamber during the sputtering process. The excitation mode was magnetron-assisted mid-frequency at
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
40 kHz with a power output of 4 kW. To simulate the use of multiple cathodes in an industrial in line system the samples were deposited in the so called dynamic mode with a moving carrier passing the cathodes multiple times. In a second production step ZnO:Al was sputtered onto the interlayer using two 0.5 wt.% aluminum-doped ceramic tube targets. These rotatable tube targets were operated in magnetron configuration at midfrequency mode with 40 kHz. The power output was kept constant at 10 kW to facilitate high-deposition rates of up to 62.6 nm∗m/min. Similar to the SiOxNy-coating the ZnO:Al deposition was executed in dynamic mode with the substrate carrier passing the targets multiple times during the deposition. While output power and gas flow stayed constant during the deposition processes, the temperature and pressure values were varied from one deposition to the next to simulate drifts and inhomogeneity in an industrial large area fabrication. These parameter changes were repeated for two series of different thickness of which the first, thicker series achieved film-thicknesses around 1280 nm (10 passes), whereas the second, thinner series yielded thicknesses around 900 nm (7 passes). All ZnO:Al layers and interlayers were deposited in a vertical inline sputter deposition system VISS 300 by von Ardenne Anlagentechnik Dresden GmbH [6]. For further deposition parameter details see Table 1. After deposition the ZnO:Al films were texture-etched in 0.5% HCl to roughen the surface. The etch step was conducted at room temperature for 120 s for all thick and 40 s for all thin ZnO:Al films. The textured ZnO:Al substrates were investigated using a scanning electron microscope (SEM) to gain insight into the surface features. Furthermore, the samples were inspected optically using an angular resolved light scattering set-up. The ARS set-up uses a green laser with a wavelength of 550 nm. The laser beam is chopped at a frequency of 230 Hz and then split into a reference beam and a measurement beam. The reference beam is led directly onto a silicon photodiode whereas the measurement beam passes through an aperture, striking the smooth glass side of the sample perpendicularly. A planar silicon photodiode is used to detect the transmitted, scattered light on the other side of the sample. The detector is free to move on a circular path at a fix distance around the sample. The step size of the moving detector can be set to 1° leading to an equivalent angular resolution of the measured signal. The measured photodiode current was used to calculate the normalized angular intensity distribution (AID) taking the solid angle into account. The investigated ZnO:Al films were applied as front contact layer in a-Si:H/μc-Si:H tandem solar cells to study their influence on the solar cell performance. The silicon absorber layers were deposited in a single-chamber process. Details regarding the silicon deposition system and process are described elsewhere [13,14]. Room-temperature sputtered ZnO:Al and silver were employed as back contact and the area of the tandem solar cell was defined by laser scribing. The IV-characteristic of the solar cells were measured in a class A sun simulator using an AM 1.5 spectrum. Additionally, external quantum efficiency (EQE) measurements were conducted to identify the current distribution in the tandem device.
Table 1 Process parameters or parameter intervals of SiOxNy interlayer deposition and ZnO:Al deposition. Process parameter
Unit
SiOxNy
ZnO:Al
Heater temperature Pressure Power ΦAr ΦO2 ΦN2 Vsubstrate carrier Passes
[°C] [Pa] [kW] [sccm] [sccm] [sccm] [mm/s]
25 0.3 4 200 18 100 3 2
430–500 0.3–1 10 200 – – 8 7 or 10
67
3. Results and discussion 3.1. Temperature series Fig. 1 shows the SEM images of the deposited ZnO:Al surfaces after the texture-etch. The ZnO:Al samples in the left column show a heater temperature series from 430 °C to 500 °C with all deposited layers having a thickness around 1280 nm-thick series. The right column of SEM-images shows the same temperature series from 430 °C–500 °C for thin deposited films with a thickness of around 900 nm. The thick layer series shows a development in surface topography leading from rather smooth surfaces with small crater-like features at 430 °C to a very rough surface with large craters at 450 °C. Raising the process temperature even further the features seem to flatten again, leading to an almost smooth surface at 500 °C. For the thin layer series the behavior is similar. Small craters at 430 °C are increasing in diameter while the process temperature is increased leading to very distinct craters at 450 °C. Further increase in process temperature facilitates even larger craters, however, this is accompanied by a general flattening of the surface features. Fig. 2 shows the AID results from 0° to 90° scattering angle of the two temperature series. Fig. 2A) displays the thick series whereas Fig. 2B) shows the thin series. The gray arrows within the graphs indicate the order of the measured samples from low temperature to high deposition temperature. The thick samples in Fig. 2A) have a decreasing ARS signal intensity at large angles (≥35°) with increasing deposition temperature. At the same time the scattered light intensity at angles below 30° increases and the maximum or peak value of scattered light intensity shifts to smaller angles with increasing temperature. The thin samples in Fig. 2B) show a similar behavior. With increasing deposition temperature the scattered light intensities at angles below 30° increases and the peak value shift towards smaller angles. The scattered intensity at large angles, however, shows a different trend. Beginning with a low temperature of 430 °C the scattered light intensity at large angles at first increases with temperature, reaching a maximum for the sample deposited at 450 °C. Increasing the temperature further leads to a decrease in scattered light intensity at large angles alike the thick series. The SEM finding may be explained using the modified Thornton Model by Kluth et al. [15] or the further advanced study by Berginski et al. [16]. Both studies state that the deposition temperature has a significant impact on the growth of sputter deposited ZnO:Al. With increasing temperature the films become denser. The difference in density becomes visible after etching when the surfaces reveal different textures. These textures are grouped into three types: type A or I (low temperature, low density) reveals very small features and reveals almost no roughness after etching, type B or II (medium temperature, medium density) shows a homogeneous distribution of large feature on the surface, type C or III (high temperature, high density) shows even larger features than B but with heterogeneous distribution. The SEM images in Fig. 1 show for both temperature series – thin and thick – the evolution from type A surface at 430 °C to a type B surface at 450 °C and further to a type C surface at 500 °C. The angular resolved scattering results can be interpreted using several optical models [9,10,17]. This paper will focus on the following two models for interpretation. The first model is based on the raytracing theory [9]. Hence, every single ray that strikes the textured ZnO:Al surface will undergo scalar scattering in accordance with Snell's law. The measured angular resolved scattering intensities are therefore the superposition of all single scattering events in the illuminated surface area. Large or “flat” surface angles will be visible in the small angle range of the scattered intensity whereas small or “steep” angles appear in the large angle intensities. The second model treats the ZnO: Al surface as a superposition of diffraction gratings with different lattice constants [10]. In this model large grating constants or “large surface features” will appear in the small angle range of scattered intensities
68
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
Fig. 1. Scanning electron microscope images of ZnO:Al surfaces after the texture-etch in 0.5% HCl. The left column of images shows the thick layers depending on the deposition temperature, the right column the thin layers depending on the deposition temperature.
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
69
measurements reveal detailed surface information regarding surface feature size and inclination angles upon interpretation. Furthermore, not only differences are detectable but also temperature correlated trends in the AID are identifiable. 3.2. Pressure series
Fig. 2. Angular intensity distribution of light scattered at texture-etched ZnO:Al surfaces. A) thick ZnO:Al layer deposited at different temperatures; B) thin ZnO:Al layer deposited at different temperatures.
whereas small lattice constants or “small surface features” appear in the large angle range of scattered intensities. Applying these two models to the measured AID, the peak shift to smaller angles can be interpreted as the transition to larger features with increasing deposition temperature for both series. At the same time the intensity loss at large angles with increasing temperature can be interpreted as the disappearance of small surface features or vice versa “steep” surface angles. The drop in large angle scattered intensity between 450 °C and 430 °C in the thin temperature series cannot be attributed to an increase in feature size as the crater diameter obviously decreases. The presence of flatter surface angles seems to be a more reasonable explanation. This interpretation is in agreement with the model of Kluth et al. since they predict that a reduction in temperature will lead to a type A structure at some point. This type A structure characteristically exhibits laterally small features and additionally small vertical dimensions. The missing vertical dimensions accompany flatter surface angles leading to a decrease in the scattered intensity at large angles. The fact that the thick series does not show the same behavior can be attributed to the fact that the lowest probed temperature of 430 °C is not low enough for this configuration to create a type A surface texture. Concluding it can be determined that even very small differences in deposition temperature of only 10 °C, which can easily occur in large area fabrication, can already lead to significant changes in the surface topography after etching. These differences in topography are only partially accessible via SEM imaging. ZnO:Al samples that appear very similar in the SEM images e.g. the thin 430 °C and 440 °C surface show differences in the ARS measurements. The more cost effective ARS
Fig. 3 shows the SEM images of differently deposited ZnO:Al surfaces after the texture-etch. The samples in the left column show a pressure series from 0.3 Pa to 1 Pa with all deposited layers having a thickness around 1280 nm-thick series. The right column of SEM-images belongs to the same pressure series from 0.3 Pa to 1 Pa for thin deposited samples with a thickness of around 900 nm. The thick layer series shows no obvious differences between the samples at different deposition pressure. Apparently, all three samples reveal a surface covered with similar large crater structures. For the thin layer series the behavior is different. Starting at a low pressure of 0.3 Pa with a homogeneously crater covered surface, the crater size and distribution appears consistent when shifting the pressure to 0.5 Pa but the craters seem to flatten. Increasing the pressure even further to 1 Pa then leads to steeper features again but with decreased lateral size. Fig. 4 presents the angular resolved scattering results of the two pressure series. Fig. 4A) displays the thick series whereas Fig. 4B) shows the thin series. Similar to the SEM results the thick series shows no significant differences in the AID either. All three samples show a strong peak at small angles (around 8°) and about one order of magnitude in intensity lower scattering at large angles (≥35°). The thin series in Fig. 4B) on the other hand shows different angular intensity distribution for each of the pressure samples. The 0.3 Pa sample shows a strong peak at small angles (around 15°) and about a factor of six lower scattering intensities at large angles. Changing the pressure from 0.3 Pa to 0.5 Pa the large angle scattering decreases while the small angle scattering stays constant. Raising the pressure even further from 0.5 Pa up to 1 Pa the small angle peak diminishes significantly and the large angle scattering increases again. The SEM findings in the pressure series can partially be explained by the Modified Thornton Model [15]. According to Kluth et al. an increase in process pressure with otherwise constant deposition conditions will lead to a material shift from a type C (large crater with inhomogeneous distribution) to type B (large craters with homogeneous distribution) and finally to a type A (small features with homogeneous distribution). For the thick sample series in Fig. 3), all samples belong to type B structure. Apparently, the change in process pressure is not significant enough to cause differences in the etching behavior of the samples. Within the investigated boundaries of 0.3 Pa to 1 Pa the thick sample series is unaffected by pressure changes. For the thin sample series there are clearly visible differences in the SEM images. However, the behavior does not follow the theory of the Modified Thornton Model. With increasing the pressure from 0.3 Pa to 0.5 Pa, type B at 0.3 Pa does not change into a type C structure but a flattened type B structure. Increasing the pressure further to 1 Pa then leads to the transition to a type C structure as expected according to the Modified Thornton Model. The aforementioned two optical models are applied again to interpret the measured AID. For the thick sample series the peak at the rather small angle of about 8° indicates the presence of large features and or “flat” angles on the surface. All three pressure samples show the same angular intensity distribution indicating that they are similar. The differences in the thin sample series on the other hand indicate no significant change in the feature size when shifting from 0.3 Pa to 0.5 Pa, as the peak position stays almost constant at 12°. However, the decrease in large angle scattering intensity suggests the presence of less steep angles on the surface or the flattening of the features. Increasing the pressure further then leads to a diminishing of the small angle peak, indicating that the surface features reduce in size. At the same time the large angle scattering rises again, suggesting that there are steeper angles on the surface.
70
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
Fig. 3. SEM images of ZnO:Al surfaces after the texture-etch in 0.5% HCl. The left column of images shows the thick layers depending on the deposition pressure, the right column the thin layers depending on the deposition pressure.
Concluding it can be stated that the pressure changes seem to affect the deposited series to a different degree. The thick sample series undergoes pressure changes during deposition within our investigated constraints without any changes to the surface topography after etching. Meanwhile applying the same pressure changes, there are dramatic changes detectable for the thin sample series. Hence, the thick sample series proves to be more stable than the thin sample series in the pressure regime. 3.3. Solar cell parameters The ZnO:Al films used in the tandem solar cells as a front TCO contacts were the samples of the thin layer pressure series which were etched for different etch times than the reference samples in 3.2. The etch duration was adapted to 42 s, 46 s and 29 s for the ZnO:Al films deposited at 0.3 Pa, 0.5 Pa and 1 Pa respectively to ensure the same thickness after etching rather than the same etch time. The angular intensity distribution as seen in Fig. 4B was not significantly altered. The solar cell parameters of the a-Si:H/μc-Si:H tandem cells with a total silicon absorber thickness of 1.96 μm that were deposited on the thin pressure series samples are summarized in Table 2. The solar cell parameters show only small differences in the solar cell performance of the 0.3 Pa and 0.5 Pa samples whereas the 1 Pa sample deviates strongly in all characteristics. The current distribution
shows almost the same top-cell current density for all three solar cell types, whereas the bottom cell current density deviates by 0.7 mA/cm 2 between the two lower pressure samples and the 1 Pa sample. Additionally, the open circuit voltage drops by more than 30 mV when shifting to a high pressure of 1 Pa. This leads to an about 1% abs. lower solar cell efficiency of the 1 Pa sample in comparison with the two low pressure samples. Fig. 5 shows the external quantum efficiency spectra of the tandem solar cells deposited on top of the thin ZnO:Al pressure series. The EQE of the top cell shows almost no change for the three different ZnO:Al TCOs. The same applies for the bottom cell current up to a wavelength of 700 nm. Beyond this wavelength the EQE of the ZnO:Al types differs significantly. The low pressure (0.3 Pa) and medium pressure (0.5 Pa) samples show only very slight deviations whereas the high pressure sample (1 Pa) shows a strongly reduced EQE for wavelengths above 700 nm. The solar cell results show a strong correlation with the AID of the texture-etched ZnO:Al front contacts. The ability to trap light within the solar cell is one of the most important qualities of a textured front contact. Especially for wavelength above 700 nm a pathlength enhancement through light trapping is essential to allow sufficient absorption. The fact, that in the case of the tandem devices shown the bottom cell EQE is only affected for wavelength above 700 nm, indicates that the differences in structure that were detected as differences in the AID are affecting the light trapping behavior of the
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
71
A
Fig. 5. External quantum efficiency spectra of a-Si:H/μc-Si:H solar cells with pressure series of thin deposited ZnO:Al films applied as front contact.
B A
B
Fig. 4. Angular intensity distribution of light scattered at texture-etched ZnO:Al surfaces. A) thick ZnO:Al layer deposited at various process pressures; B) thin ZnO:Al layer deposited at various process pressures.
solar cell. Therefore angular resolved scattering measurements are an effective method to detect differences in the front contact surface texture that may lead to differences in the device performance. In the following section it will be demonstrated, how this measurement can not only be used to monitor the quality of a front contact but how to use it to optimize the front contact surface texture. 3.4. Texture optimization Fig. 6 illustrates the value of the angular intensity distribution at 60° scattering angle of the thick ZnO:Al temperature series a) and the thin ZnO:Al temperature series b) of Section 3.1 as a function of the heater temperature TH during deposition. Fig. 6A) shows the Table 2 Solar cell parameters of a-Si:H/μc-Si:H tandem cells with thin deposited ZnO:Al pressure series applied as front contact. Solar cell parameter
Eta FF Voc Isc Isc,top Isc,bottom
Unit
[%] [%] [V] [mA/cm2] [mA/cm2] [mA/cm2]
Process pressure in ZnO:Al deposition 3 μbar
5 μbar
10 μbar
11.9 72.2 1.401 11.7 11.5 12.1
11.6 72.4 1.397 11.5 11.4 12.0
10.8 70.6 1.365 11.2 11.4 11.3
Fig. 6. Angular intensity distribution of thick (A) and thin (B) ZnO:Al temperature series at 60° scattering angle plotted as a function of the heater temperature TH during deposition.
highest AID at 60° for the samples deposited at the lowest deposition temperature probed in this study — at 430 °C. With further increased deposition temperature the AID at 60° decreases. In contrast Fig. 6B) shows a maximum of scattered intensity at 60° for the thin sample deposited at 450 °C. Samples deposited at higher or lower temperature reveal lower scattering intensities at this angle. It is known from previous studies that the integrated scattered intensity at angles between 45°–90° – often simply referred to as large angle scattering – correlates with the short-circuit current density of single-junction μc-Si:H solar cells [7]. It was also shown that rather than an integration of all large angles, a single sampling point in terms of a discrete large angle can represent the same correlation between scattering intensity and short-circuit current density. In [12] the discrete scattering intensity of 60° was correlated with the short-circuit current density, with higher AID (60°) values leading to higher short-circuit current densities. Taking Fig. 6 into account there is a distinct maximum of AID (60°) over the deposition temperature. Hence, to improve the surface texture of the sputtered ZnO:Al for high currents in a μc-Si:H solar cell by varying the deposition temperature, the angular resolved scattering measurement can be used to find optimum by searching the maximum of AID (60°). 4. Conclusion In this paper angular resolved scattering measurements as process monitoring tool for texture-etched ZnO:Al in silicon thin-film solar cells were investigated. It was shown that ARS measurements can be used to
72
G. Jost et al. / Thin Solid Films 532 (2013) 66–72
monitor the surface texture changes that occur due to process parameter changes in temperature and pressure. The detected differences in angular resolved scattering measurements are found to correlate with the changes in short-circuit current densities. Finally, process parameter dependent trends were identified in the angular intensity distribution that can be used to modify or optimize the surface texture in order to increase the short-circuit current density in single-junction μc-Si:H solar cells.
[5] [6] [7] [8] [9] [10] [11]
Acknowledgments The authors would like to thank H. Siekmann, J. Worbs, D. Weigand, A. Bauer and G. Schöppe for their technical assistance; A. Doumit and H.P. Bochem for SEM imaging and C. Zahren for device characterization. Financial support by the German Federal Ministry (BMU, grants 0327625, 0325299A and 0325356B) is gratefully acknowledged. References [1] J. Nomoto, T. Hirano, T. Miyata, T. Minami, Thin Solid Films 520 (2011) 1400. [2] J.K. Rath, Y. Liu, M.M. de Jong, J. de Wild, J.A. Schuttauf, M. Brinza, R.E.I. Schropp, Thin Solid Films 518 (2010) e129. [3] J.-S. Cho, Y.-J. Jeong, S.-H. Park, K.H. Yoon, Sol. Energy Mater. Sol. Cells 95 (2011) 190. [4] J. Bailat, D. Dominé, R. Schlüchter, J. Steinhauser, S. Fay, F. Freitas, C. Bücher, L. Feitknecht, X. Niquille, T. Tscharner, A. Shah, C. Ballif, in: Conference Record of
[12]
[13] [14]
[15] [16] [17]
the 2006 IEEE 4th World Conference on Photovoltaic Energy Conversion, 2006, p. 1533. J. Müller, B. Rech, J. Springer, M. Vanecek, Solar Energy 77 (2004) 917. J. Hüpkes, H. Zhu, J.I. Owen, G. Jost, E. Bunte, Thin Solid Films (2012), http://dx.doi.org/10.1016/j.tsf.2011.09.031. H. Schade, P. Lechner, R. Geyer, H. Stiebig, B. Rech, O. Kluth, in: Conference Record of 31st IEEE — Photovoltaic Specialist Conference, 2005, p. 1436. A. Krasnov, Sol. Energy Mater. Sol. Cells 94 (2010) 1648. M. Schulte, K. Bittkau, B.E. Pieters, S. Jorke, H. Stiebig, J. Hüpkes, U. Rau, Proc. SPIE 7717 (2010) 77170N. K. Bittkau, M. Schulte, T. Beckers, R. Carius, Proc. SPIE 7725 (2010) 77250N-3. W. Dewald, V. Sittinger, B. Szyszka, D. Wippler, P. Obermeyer, F. Hamelmann, H. Stiebig, F. Säuberlich, D. Severin, S. Klein, M. Rohde, U. Schmidt, in: Proceedings of European Photovoltaic Solar Energy Conference and Exhibition, WIP Renewable Energies, Hamburg, 2011, p. 2704. G. Jost, E. Bunte, J. Worbs, H. Siekmann, J. Hüpkes, in: Proceedings of European Photovoltaic Solar Energy Conference and Exhibition, WIP Renewable Energies, Valencia, 2010, p. 3078. T. Merdzhanova, J. Woerdenweber, T. Zimmermann, U. Zastrow, A.J. Flikweert, H. Stiebig, W. Beyer, A. Gordijn, Sol. Energy Mater. Sol. Cells 98 (2012) 146. J. Woerdenweber, T. Merdzhanova, T. Zimmermann, A.J. Flikweert, H. Stiebig, W. Beyer, A. Gordijn, J. Non-Cryst. Solids (2012), http://dx.doi.org/10.1016/j.jnoncrysol.2011. 12.041. O. Kluth, G. Schöpe, J. Hüpkes, C. Agashe, J. Müller, B. Rech, Thin Solid Films 442 (2003) 80. M. Berginski, J. Hüpkes, M. Schulte, G. Schöpe, H. Stiebig, B. Rech, M. Wuttig, J. Appl. Phys. 101 (2007) 074903. D. Dominé, F.-J. Haug, C. Battaglia, C. Ballif, J. Appl. Phys. 107 (2010) 044504.