Solar

Textured ZnO enables high-efficiency silicon solar modules

Issue 4 and Volume 2.

One of the most critical processes for high-efficiency amorphous-microcrystalline thin film silicon solar cells involves the front-side transparent conductive oxide (TCO). This very thin layer scatters the incoming light which increases the fraction of light absorbed in the layer. To date, most production a-Si thin film solar cell designs, both single- and tandem-junction, have relied on commercially available tin oxide (SnO) deposited on glass. We describe a study that shows how improving light-trapping properties with ZnO films enables higher efficiency in large-size thin film solar modules. The experiment used magnetron-sputtered and wet chemically etched zinc oxide (ZnO) films for the front contact. The films were deposited using an Applied ATON physical vapor deposition (PVD) type tool that can be integrated into Applied Materials’ a-Si/µc-Si SunFab module production line.

One of the most critical processes for high-efficiency amorphous-microcrystalline thin film silicon solar cells involves the front-side transparent conductive oxide (TCO). This very thin layer scatters the incoming light which increases the fraction of light absorbed in the layer. To date, most production a-Si thin film solar cell designs, both single- and tandem-junction, have relied on commercially available tin oxide (SnO) deposited on glass. We describe a study that shows how improving light-trapping properties with ZnO films enables higher efficiency in large-size thin film solar modules. The experiment used magnetron-sputtered and wet chemically etched zinc oxide (ZnO) films for the front contact. The films were deposited using an Applied ATON physical vapor deposition (PVD) type tool that can be integrated into Applied Materials’ a-Si/µc-Si SunFab module production line.

 

ZnO:Al TCO film preparation

1×1.3 m2 (Gen5) glass substrates with front contact coatings deposited in a horizontal in-line sputtering system were used for the study. The aluminum-doped zinc oxide (ZnO:Al) layer was magnetron sputtered using rotatable ceramic ZnO:Al2O3 targets in an Ar atmosphere. Substrate temperatures above 250°C were used for the deposition process. Post-deposition, the ZnO film was etched in diluted hydrochloric acid. A key etch process parameter was the exposure time, which correlates with the removed film thickness, as well as increased scattered light and sheet resistance. The film sheet resistance was measured by using a 4-point probe. Surface texture was characterized using scanning electron microscopy (SEM) and atomic force microscopy (AFM).

Both a-Si single-junction and a-Si/μc-Si tandem-junction devices were fabricated on these etched ZnO:Al substrates, using an AKT 15K PECVD system. Details on the CVD process are described elsewhere [3]. Back contacts were sputtered in an in-line Applied ATON 1600 coater using ZnO/Al and ZnO/Ag back contacts for single and tandem-junction devices, respectively. Efficiencies were measured using a class-A large-area solar simulator. In order to determine the gain in cell performance caused by the light-trapping effect, quantum efficiency (QE) was used as the measurement.

Film uniformity. Good deposition uniformity is critical to larger-area module performance. The sheet resistance had an average value of 4.4 Ω/? with ±3.1% uniformity. The transmission (average between 400 and 1100nm) of this sheet had an average value of 83.7% with ±0.2 % uniformity.

Characterization of textured surface. Because the etch process step is a critical step, an etching series on 100x100mm small area was performed, in which the etch time was varied from 5−150s. Figure 1 shows representative SEM graphs of this etch series. Increasing etch time clearly resulted in larger crater size, with the smaller craters showing relatively less light-scattering ability. We determined that the surface morphology at 80s exposure time was similar to structures that are known to yield excellent light trapping [2].

Figure 1. SEM pictures of etch series.

Measured by AFM, the root mean square roughness (RMS) increases almost linearly with the etch exposure time, until it suddenly drops to zero (Fig. 2) indicating that the ZnO film is completely removed. However, RMS and haze both increase until ~120s, indicating that the light scattering ability increases from almost 0% haze value to somewhere above 80% before decreasing.

Figure 2. Dependence of resistance, roughness and haze on etch exposure time.

Because the front contact etched film needs to conduct the generated charge carriers, its sheet resistance is part of the series resistance of the modules, and therefore it should be below 20 Ωsq. The sheet resistance of this series, as shown in Fig. 2, rises very slowly during the first 80s then increases drastically. This is probably due to the very thin interconnections in between the remaining film structures when over etched. The trade-off between electrical and optical properties, i.e., between resistivity and absorption, defines the etch process window with best values at ~80s for this set of parameters.

The haze value shown in Fig. 3 is an “integrated” white haze, measured with the so-called Pauschmeter. Figure 4 shows the ‘spectral’ haze from multiple etch exposures; this increases with crater-size and RMS. At 550nm wavelength the spectral haze was equivalent to the “integrated” haze.

Figure 3. Spectral haze curves of etch series.
Figure 4. QE curve of TJ cells with different ZnO surface roughness (etch series).

Up to 80s etch time, the typical trend of textured surfaces was observed. However, the spectral behavior changed for longer exposure times, with a maximum around the 500nm wavelength. This means that short wavelengths were less scattered by surfaces etched at long exposures. This also could be explained by the formation of holes with a diameter greater than the wavelength. Very long exposure times in the etch process offered an additional positive effect in that the transparency became even better.

Figure 5. Characteristic IV curve of Gen5 module.

The differential spectral response (DSR) measurement allows a more precise analysis of top and bottom cell current; the resulting quantum efficiency (QE) curves (Fig. 4) indicate the wavelength range, and thereby the location in the tandem cell layer stack where light absorption and charge carrier generation was improved. The improvement of the textured surface was most pronounced in the long wavelength range (above 700nm). At the shortest etch exposure time (15s) there were still many interferences, which were evidence of a high fraction of directly transmitted, not scattered, light at the ZnO-Si interface. The interferences disappeared with longer exposure time, indicating that the light was trapped in the silicon layer stack and was reflected multiple times.

In the short wavelength range, increasing exposure time improved the fraction of photons gathered by the top cell. The rough surface improved light coupling because the average refractive index at the ZnO-Si interface is graded, reducing reflection. The reduced absorption losses in the ZnO was due to the thinner film improving the QE curve for wavelengths below 400nm.

Solar module performance. Following the process development on a small cell area, Gen 5 module optimization was performed on the ZnO TCO. The I-V curve of a Gen 5 module with ZnO TCO is shown in Fig. 5. The maximum initial power measured at standard test condition was 156W, which corresponds to an aperture efficiency of 11.5%.

 

References

1. A. Löffl, S. Wieder, B. Rech, O. Kluth, C. Beneking, and H. Wagner, Proc. 14th European Photovoltaic Solar Enery Conference, Vol. II, Barcelona, Spain (1997), p. 2089

2. O. Kluth, B. Rech, L. Houben, S. Wieder, G. Schöpe, C. Beneking, H. Wagner, A. Löffl and H. W. Schock, Thin Solid Films 351, 247(1999).

3. S. Klein, M. Rohde, T. Stolley, K. Schwanitz, S. Buschbaum, Proc. 23rd European Photovoltaic Solar Energy Conference, 1-5 September 2008, Valencia, Spain

Philipp Obermeyer received his degree in physics from the RWTH Aachen University, Germany and is R&D Project Manager in the Display and SunFab Solar Group (DSS) at Applied Materials GmbH & Co. KG, Siemensstraße 100, 63755 Alzenau, Germany; email [email protected].

Daniel Severin received his Ph.D. from the University of Duisburg-Essen, Germany in physics and is a solar R&D project manager at Applied Materials GmbH & Co. KG.

Markus Kress received his Ph.D. in physics from the University of Frankfurt Main, Germany and is a project manager for solar R&D at Applied Materials GmbH & Co. KG.

Stefan Klein received his PhD in physics from the University of Stuttgart and the University of Munich, Germany and is a manager of solar R&D at Applied Materials GmbH & Co. KG.

Ursula Schmidt received her Ph.D. in physics from the University of Kaiserslautern, Germany and is a solar R&D project manager at Applied Materials GmbH & Co. KG.

Stephen Wieder received his Ph.D. in physics from the University of Aachen, Germany and is director of solar R&D at Applied Materials GmbH & Co. KG.