The aim of the thin film silicon solar cell industry for production capacities of several Gigawatt in recent years is strongly related to the impressive development of microcrystalline silicon (µc-Si:H). The application of µc-Si:H as absorber layer in stacked solar cell devices has given new hope for highly stabilized efficiencies of thin film silicon solar modules. The quality and process technology of µc-Si:H have made considerable progress. As the µc-Si:H gets more advanced, new and old interest is in high efficiency & stable amorphous silicon (a-Si:H) material for top cells. Also any means which allow reducing the thickness of the top cell while still delivering sufficient current to match with the bottom cell are intensely investigated. Apparently all kinds of optical manipulations are of interest here unless one wants to reduce the total cell thickness and sacrifice some maximum efficiency against stability and production cost. We will present our latest developments in thin film silicon solar cells with new types of window layers, intermediate reflectors, anti-reflective coatings, all in connection with our high quality a-Si:H and µc-Si:H absorber layer materials.

The conversion efficiency of thin-film silicon (TF Si) solar cells needs to be raised up to achieve high level of competitiveness in the PV market. Challenging goals of the production expansion and the targeting high conversion efficiencies of TF Si solar cells up to 20 % by 2025 have been released [1]. These goals require intensive R&D activities and breakthroughs on the material-, interface-, and the complete solar-cell device level.. Novel absorber layers, transparent conductive oxides (TCOs) with advanced nano-textures, oxide based doped layers, dielectric back reflectors and other innovative solutions are under investigation by several groups to improve the conversion efficiencies and decrease the production costs of TF Si solar cells and PV modules. However, clear directions and requirements for achieving high efficiency goals still need to be identified.In this contribution we investigate the requirements and define the directions for optical design of thin-film silicon solar cells to achieve high stabilized efficiencies above 20 %. Challenges concerning layers, interfaces and device structures by using ultra-thin absorber layers are identified. The investigation is focused on optical improvements, for the electrical parameters it is assumed that the values of currently achieved record solar cells can be preserved (or improved) in optically optimized device concepts. Detailed optical analysis and optimization of single-, double- and triple-junction devices are carried out by means of 1-D semi-coherent optical simulator SunShine [2] and other 2-D and 3-D simulation tools, which were well calibrated and verified on the existing state-of-the-art TF Si solar cells. Starting from the existing state-of-the art devices, considering realistic optical properties of layers and interfaces, we show step by step what is required to achieve the stabilized efficiencies of above 15 % for single-junction a-Si:H cell (with absorber thickness of only 100 nm) and above 20 % for tandem micromorph cell (a-Si:H of 100 nm and uc-Si:H of 800 nm). We demonstrate that the following optical improvements are required:- introduction of anti-reflecting coatings or nano-structures at front interfaces (air/glass, glass/TCO, TCO/p),- all front interfaces can be perfectly flat in our device design,- introduction of a special light scatterer at the back side (crucial requirement),- significantly reduced optical losses in the supporting layers (crucial requirement),- wavelength-selective intermediate reflector in multi-junction device.Possible directions towards realizations of the requirements are indicated and discussed.[1] K. Kurokawa et al., “Accelerated and Extended Japanese PV Technology Roadmap “PV2030 +” released by NEDO in 2009.[2] J. Krc et al., Prog. Photovot. Res. Appl. 11 (2003) 15.

Hydrogenated nanocrystalline silicon (nc Si:H) has become a promising candidate to replace hydrogenated amorphous silicon-germanium alloy (a-SiGe:H) in multijunction thin film silicon solar cells. In view of its indirect bandgap, the nc-Si:H layer must be much thicker than its amorphous counterparts to effectively absorb the incident radiation. Typical thickness for a nc Si:H based multijunction cell is 2-5 µm, compared with <0.5 µm for a corresponding amorphous silicon (a-Si:H) and a-SiGe:H based device. For commercial viability, the nc-Si:H layer must be deposited at a high rate. In order to explore the limit of nc-Si:H technology, it is important to fabricate solar cells without constraint of manufacturing. In parallel, one must also determine the corresponding limit that incorporates manufacturing constraint of deposition time. Thus, one must investigate the highest efficiency attainable with and without manufacturing constraint of deposition time. In this paper, we report on this two-pronged strategy of fabricating large area, high efficiency a-Si:H/nc-Si:H/nc-Si:H solar cells at low and high rates.Triple-junction solar cells were fabricated in a large area batch reactor on Ag/ZnO back reflector coated stainless steel substrate, using a Modified Very High Frequency (MVHF) excitation process. For both low- and high-rate cases, we optimized the deposition parameters, such as pressure, gas flow, dilution, and power. The deposition rate for the nc-Si:H layers for the high-rate case was ~1.0-1.5 nm/s. We found that the deposition parameters were more relaxed for fabricating large-area solar cells for the low-rate case. We did SIMS analysis on the optimized films, and found the impurity concentrations were one order of magnitude lower than normally observed. In particular, the oxygen concentration is reduced to ~1018 cm-3. This is among the lowest oxygen concentration reported in literature. The low impurity content is attributed to superior cathode hardware and the optimized deposition recipe. We fabricated large area (aperture area 400-464 cm2) cells, and encapsulated the cells using standard flexible encapsulants. We have light soaked the high-rate cells, and are currently conducting light soak tests on the low-rate samples. The highest stable efficiency attained for the high rate cells is 10.6%, as confirmed by NREL. The initial efficiency of the low-rate samples is 11.8-12.2% as measured under our Spire solar simulator. We will present the details of the research done to develop the high- and low-rate devices.

Recently, in the quest for higher efficiencies for thin film solar cells, much emphasis is placed on light trapping or absorption enhancement techniques, such as the use of plasmonic or diffractive back contacts [1] and luminescent concentrators [2], to concentrate the incident light in cells with a thinner absorber layer or with a smaller area. Even in less advanced schemes, merely optical concentration of light can yield higher efficiencies. There are only very few reports on the effects of concentration in thin film silicon-based solar cells. Due to the presence of midgap states, a fast decline in fill factor was observed in earlier work. However, with the advent of more stable and lower defect density protocrystalline silicon materials as well as high quality micro-/nanocrystalline silicon materials, as well as the increasing concentration ratios obtained by novel light management techniques, it is worth revisiting the performance of cells with these absorber layers under moderately concentrated sunlight. We determined the behavior of the external J-V parameters of pre-stabilized substrate-type (n-i-p) amorphous and microcrystalline solar cells under moderate concentrations, between 1 sun and 20 suns, while maintaining the cell temperature at 25°C. It should be noted that these cells already comprise some sort of light concentration due to the use of textured surfaces. It was found that the cell efficiency of both the amorphous and the microcrystalline cells increased with increasing concentration, showing an optimum at approximately 5 suns. Furthermore, the enhancement in efficiency for the microcrystalline cells was larger than for the amorphous cells. From this we conclude that the carrier transport mechanism in microcrystalline cells is less of drift type than that in amorphous cells. We show that the Voc’s up to 0.63 V can be reached in microcrystalline cells while FF’s only decrease by 10%. The effects have also been computed using the device simulator ASA, showing qualitative agreement. We conclude that it is meaningful to design an optical concentration ratio of 5 to 10 suns for thin film silicon solar cells[1] V.E. Ferry, M.A. Verschuuren, H.B.T. Li, E. Verhagen, R.J. Walters, R.E.I. Schropp, H.A. Atwater, A. Polman, “Light Trapping in Ultrathin Plasmonic Solar Cells”, Optics Express 18 102 (2010) A237.[2] W.G.J.H.M. van Sark et al., “Luminescent solar concentrators – A review of recent results”, Optics Express16 (2008) 21773.

Multi-junction solar cell is an effective approach toward high efficiency and improved stability in thin film solar cell applications. Hydrogenated amorphous silicon germanium (a-SiGe:H) has received much attention due to its high absorption coefficient and adjustable bandgap. The property of a-SiGe:H thin films has been characterized and optimized in our previous work [1]. However, one of the major challenge is the bandgap discontinuity between the a-SiGe:H i-layer and the a-Si:H doped layer. Such discontinuity at both the p/i the i/n interfaces can be alleviated by graded interface [2]. In this work, we applied the concept of bandgap profiling to our single-junction a-SiGe:H cells. Together with other cell optimization, the cell efficiency of 8.59% has been attained.The a-SiGe:H solar cells were deposited by a 27.12 MHz radio-frequency plasma-enhanced chemical vapor deposition (PECVD) system. The germane concentration and the hydrogen dilution ratio were varied during the deposition of the a-SiGe:H thin films. The bandgap was calculated from Tauc’s plot by analyzing the transmittance spectra measured in UV/VIS spectroscopy. The conductivity of the thin films and characteristics of the solar cells were measured by AM1.5G illuminated I-V measurement system.Our results show that the graded bandgap at both the p/i and i/n interfaces enhanced the cell efficiency as compared to cells with constant bandgap of 1.55 eV. According to the results, the increase of the efficiency was due to the enhanced Voc and FF, which were arising from a better short-wavelength absorption and carrier transport, respectively. To assess the effect of i/n grading, the width of the i/n region was systematically varied from 0 to 20nm while maintain the same absorber thickness of 200nm. As the thickness of i/n grading increases, the cell efficiency increases from 7.8% to 8.3% despite a slight reduction of Jsc. The graded i/n region might facilitate the transportation of deep holes; nevertheless, further increase in i/n grading deteriorates the cell efficiency, which may due to the suppression of long-wavelength absorption. Finally, the optimal thickness of p-layer, n-layer and grading structure were integrated for fabricating the solar cells. From our results, the a-SiGe:H cell efficiency of 8.59% was achieved with Voc = 748 mV, Jsc = 16.31 mA/cm2, FF = 70.38 %.This work was sponsored by the Center for Green Energy Technology at the National Chiao Tung University and the National Science and Technology Program for Energy.[1]C.M. Wang, Y.T. Huang, K.H. Yen, H.J. Hsu, H.W. Zan and C.C. Tsai, Mat. Res. Soc. Symp. Proc., Spring meeting, San Francisco (2010)[2]S. Guha, J. Yang, A. Pawlikiewicz, T. Glatfelter, R. Ross, and S. Ovshinsky, Appl. Phys. Lett. 54, 2330-2332 (1989)

The conversion efficiency of thin film silicon solar cells has been improved by employing the narrow-gap hydrogenated microcrystalline silicon (μc-Si:H) in combination with the wide-gap hydrogenated amorphous silicon (a-Si:H) based on a multijunction concept [1,2]. Since the μc-Si:H is an indirect band gap material, relatively thick absorber layer (>2 μm) is necessary for efficient infrared light absorption, which in turn requires the high-rate deposition technique for industrial production. To achieve high deposition rate and high efficiency, we have developed a novel deposition process using high-pressure depletion regime in plasma-enhanced chemical vapor deposition [3,4]. This technique allows growing μc-Si:H at high rates (> 2 nm/s) while preserving excellent film qualities in terms of denser microstructure and less post-oxidation behavior. As a result, efficiencies of 8-9% have been demonstrated for the μc-Si:H single junction solar cells at deposition rates between 2 and 3 nm/s.Despite the successful material combination of a-Si:H and μc-Si:H, the stabilized efficiency of the a-Si:H/μc-Si:H tandem solar cells is still limited as low as ~12%. The one of the major limitations of efficiency is the weak infrared absorption in μc-Si:H bottom cell. To extend the spectral sensitivities of solar cells into longer wavelengths, we proposed the application of hydrogenated microcrystalline silicon-germanium alloys (μc-Si_1-xGe_x:H) as a narrower-gap bottom-cell absorber in multijunction structures such as a-Si:H/μc-Si_1-xGe_x:H and a-Si:H/μc-Si:H/μc-Si_1-xGe_x:H. In the previous work [5], we have demonstrated efficient (~7-8%) μc-Si_1-xGe_x:H (x~0.1-0.17) single junction p-i-n solar cells with markedly higher short-circuit current densities than for μc-Si:H (x=0) solar cells due to enhanced infrared absorption. Nevertheless, the photocarrier collection degrades severely when increasing either Ge content (x>0.2) or cell thickness (t_i >1 μm). We attributed the inferior performance of such solar cells to the creation of the Ge-related native defect acceptors that strongly distort the built-in electric field in the p-i-n solar cells. Recently, we have developed a counter doping technique that compensates the Ge-related acceptor states for further improvement of μc-Si_1-xGe_x:H solar cells.In this contribution, we review our research and progresses in μc-Si:H and μc-Si_1-xGe_x:H thin film materials, deposition process and solar cell devices. Apart from the optimization of these materials, the improvements of light trapping and TCO layer, which are also crucial in boosting the photocurrent of the thin film solar cells, will be addressed.[1] J. Meier et al., Solar Energy Material and Solar Cells, 66, 73 (2001).[2] K. Yamamoto et al., Solar Energy, 77, 939 (2004).[3] M. Kondo et al. J. Non-Cryst. Solids, 266-269, 84 (2000).[4] T. Matsui et al. Jpn. J. Appl. Phys. Part 2, 42, L901 (2003).[5] T. Matsui et al., Prog. Photovolt: Res. Appl. 18, 48 (2010).

Spatial distribution of dangling-bond defects in high-rate-growth microcrystalline silicon (µc-Si:H) thin films by plasma-enhanced chemical-vapor deposition (PECVD) has been investigated using high precision electron-spin resonance (ESR) system with specially designed TM-mode cavity where planar sample is measured without removing substrate. We have found that dangling-bond defect is distributed uniformly in the bulk region independent of crystallite size and high density dangling-bond is located at the surface region (~ 10-12 cm-2) in µc-Si:H films. Both the bulk-defect density and surface-areal-defect density are increased when increasing the growth rate (up to 6.7 nm/sec) through an elevation of electron temperature in the plasma during film growth. Presence of large number of surface defects is one of the crucial causes for deteriorating n-i-p type solar-cell performance through the photocarrier recombination as well as current leakage at the p/i interface. To overcome this issue in µc-Si:H based solar cells grown at high rate, we have attempted to use a novel interface-treatment method in fabrication process of n-i-p type solar cells, e. g., thin silicon layer with low defect density (compress layer) is deposited on the surface of µc-Si:H grown at high rate followed by thermal annealing. The compress layer deposition with post thermal annealing is a key to reduce surface-dangling-bond density, in which the areal dangling-bond density located at the surface is decreased down to less than 30% comparing with as-deposited state. Importance of starting procedure in µc-Si:H growth at high rate has also been indicated, since n/i-interface properties are determined at this moment in n-i-p type solar cell. We have found that SiH4-gas introduction-time constant into H2 plasma plays an important role in controlling the structural and optoelectronic properties of the n/i interface.Consequently, a high conversion efficiency of 9.27% has been demonstrated in µc-Si:H based n-i-p solar cells whose intrinsic layer is grown at high rate of 2.3 nm/sec thanks to the presence of 50 nm-thick compress layer with post thermal annealing together with the control of SiH4-gas-introduction scheme during the initial growth stage of intrinsic layer.

Real time spectroscopic ellipsometry (SE) has been developed to monitor cassette roll-to-roll deposition of thin film hydrogenated silicon (Si:H) n-i-p solar cells and their backreflector (BR) layers on flexible polymer substrates. Monitoring is performed at a single spot at the center of the substrate width using an SE range of 0.75 - 5 eV. The methodology is first demonstrated in growth studies from nucleation to final thickness for magnetron sputtered ZnO films on top of opaque Ag in the BR structure. The methodology is then extended to plasma-enhanced chemical vapor deposition (PECVD) of the i and p-layers in succession on the BR/n-layer stack. Roll-to-roll substrate motion is initiated first, followed by real time SE data collection; finally, the plasma is ignited so that film nucleation can be observed. The film thickness is observed to increase with time until a steady state is reached, after which the bulk layer thickness at the monitoring point is constant with time. This occurs when the elapsed deposition time equals the time required for the substrate to travel from the trailing edge of the deposition zone to the monitoring point. Although a constant substrate speed is selected such that the final film thickness is achieved in the time required for the substrate to move through the entire deposition zone, this speed does not permit study of film growth that occurs after the substrate passes the monitoring point. To address this deficiency, the substrate speed is reduced only over an initial length of the roll such that the final film thickness of interest is reached at the monitoring point. In this way, real time SE can be used to analyze the entire layer on an initial length of the roll before its full length is coated. Furthermore, the use of a moving substrate in real time SE enables new capabilities in process analysis. The thickness evolution of ZnO during sputtering shows reasonable agreement with a simulation assuming that the deposition flux varies in accordance with a simple inverse square of the target-substrate distance as the substrate moves through the deposition zone. The thickness evolution of the i- and p-layers during PECVD can be simulated by assuming constant deposition rate throughout the deposition zone. Fitting of the thickness evolution in sputtering can be further improved by introducing gas phase scattering of deposition species, whereas fitting in PECVD can be improved by assuming a non-zero decay length of reactive species beyond the cathode width. After the various multilayer fabrication steps, SE has also been applied for large area mapping of the coated polymer so as to determine the thickness uniformity of the layers across the width of the substrate. The overall goal of this work is to apply optical probes in order to develop optimum deposition procedures for thin film Si:H solar cell structures and to evaluate their uniformity in roll-to-roll multi-chamber deposition.

The crystalline volume fraction of the intrinsic hydrogenated microcrystalline (µc-Si:H) absorber layer of a thin-film silicon solar cell is a crucial material parameter which strongly influences the performance of solar cells. To get a better understanding and control of the deposition process of µc-Si:H the implementation of in-situ diagnostic tools is of great importance. Raman spectroscopy is one of the most frequently used techniques used to obtain a measure of the crystalline volume fraction. However in a capacitively coupled plasma-enhanced chemical vapor deposition (PECVD) process the implementation of Raman spectroscopy is difficult to accomplish, particularly in large area systems. Due to the need of low angle optical access to the growing film it is necessary to pierce the electrode surface.We present a novel showerhead-electrode design that enables in-situ Raman measurements during PECVD deposition. In this paper the optical feed through was shielded electrically to guarantee the homogeneous deposition of a µc-Si:H thin film. We show that with this electrode it is possible to deposit homogenous intrinsic absorber layers and to measure the Raman crystallinity of the film in situ. To suppress the influence of the plasma emission on the recorded spectra a lock-in technique was used. With this setup, the signal-to-noise ratio is increased drastically.A high laser power density is favorable to get sufficiently large Raman signals but local heating also has to be avoided. The local temperature was simulated as a function of pulse intensity and length and the size of the laser spot. The optimum conditions enabling useful Raman signals at a minimum heating were applied experimentally to verify the reduced heating of the growing film. Using this setup it was possible to measure the evolution of crystallinity of the film during the growth with a time resolution of less than 20 seconds which corresponds to an additional film thickness of 5 nm up to 20 nm during each measurement depending on the studied deposition rate. By analyzing the ratio of the Stokes- and anti-Stokes-scattering intensity the thermodynamic temperature of the growing film was determined. An increase of film temperature due to plasma heating was observed for the deposition of µc-Si:H with an excitation frequency of 13.56 MHz and a growth rate of about 2.5 Å/s.

We have grown device-quality epitaxial silicon thin films at high growth rate (GR) up to 700 nm/min, using hot-wire chemical vapor deposition (HWCVD) from silane. These growth rates exceed those used for industrial amorphous and nanocrystalline thin-film Si solar cells by a factor of about 20. Such layers have potential as the absorber layer in film crystal silicon solar cells deposited on seed layers on inexpensive substrates such as display glass. To obtain high GR, we explored a parameter space suggested by our model of the GR dependence on the hot-wire to substrate distance (d), system pressure (p), and silane flow [1]. In-situ spectroscopic ellipsometry (SE) was used to determine the growth rate and epitaxial quality in real time. The GR depends linearly on 1/d over a range of low silane pressures and sample-filament distances; however, deviations at high p and small d indicate that changes in the gas-phase and filament surface chemistries affect the production of growth radicals. In this regime, good quality epitaxy is possible as long as the substrate temperature is above about 620°C. For all deposition parameters we explored, the bulk epitaxy quality depends primarily on the initial moments of growth. Therefore, for our best test devices, we deposited about 50 nm of high quality epitaxial layer at a GR ~ 150 nm/min, before adjusting to the high GR conditions. To demonstrate layer quality in a high GR epitaxial solar cell, we deposited a 2.3 µm epitaxial silicon layer at 700 nm/min on a RCA-cleaned, (100) oriented, n+ (Si:As) silicon wafer. A simple mesa structure (wafer/epi Si/I a-Si/p+ a-Si:H/ITO) was processed into a solar cell by wet and dry etching techniques. The finished device had an open-circuit voltage of 0.424 V before any hydrogenation treatment. We will discuss further device improvements and strategies for increasing the GR beyond 1 micron/min.This work was supported by the U.S. DOE Solar Energy Technology Program under Contract No. DE-AC36-08GO28308.1. I. T. Martin, C. W. Teplin, J. R. Doyle, H. M. Branz, and P. Stradins. J. Appl. Phys. 107, 054906 (2010).

Plasma-enhanced chemical vapour deposition (PECVD) of p-i-n type hydrogenated amorphous silicon (a-Si:H) based solar cells in a single plasma reactor offers advantages of low cost compared to multi-chamber processes that use separate reactors to deposit the p-, i- and n-layers, respectively. It is desirable that the deposition o