Development of Photovoltaic Solar Cell Technology

Transcription

1 Development of Photovoltaic Solar Cell Technology Junhao Chu Shanghai Institute of Technical Physics, Chinese Academy of Sciences Abstract The status and development of photovoltaic (PV) techniques such as crystalline silicon (c-si) PV solar cells, thin film solar cells (TFSC), and newconcept PV solar cells are briefly introduced in this paper. The future of PV technologies is also discussed. Introduction The greenhouse effect, which is induced by the burning of fossil fuels, will cause global warming, which in turn will be a major and remarkable factor that affects the environment. Photovoltaic (PV) solar cells convert sunlight into electrical energy, making it a clean energy. Rapid progress on photovoltaic solar cells has been made recent years. Though the PV applications have faced some problems due to economic or political factors, we believe that the future of the PV industry and PV applications will be still bright. Photovoltaic solar cells are not only a new active innovational research area, but also a high potential and real economic opportunity [1]. Two problems we hope to solve in the near future are as follows: first, the application of PVs should be encouraged; secondly, PV techniques should be improved. Though China is the largest producer of commercial PV technology, it uses very little of what is produced; its 2008 share of the global market was only 0.17%. Module production in China 2010 was 10 GW, however the new installed capacity in China is 520 MW, 3% in the word (16.6 GW). The accumulated capacity is 893MW, 2.2% in the word (40GW) by In 2011, the module production in China is 23 GW, the new installed capacity in China is 2.7 GW, 9.4% in the word (28.5 GW). The accumulated capacity is 3.6GW, 5.3% in the word (68GW) by It is clear that while the amount produced in China is large, the amount applied within China is small. Enlarging the application scale is an urgent task in China. In terms of technology, there has been considerable and rapid progress in PV production over the past several years, and PV science and technology remains a very active area of research and innovation. The solar PV technology research in two ways: to improve the conversion efficiency regardless of the costs, and then, later, try to reduce costs; to search for more low-cost photovoltaic conversion materials and cell constructions, and then, through research, gradually increase the conversion rate. There are presently three different types of PV solar cell technologies which are improving their respective technique levels. Crystalline Silicon (c-si) PV Solar Cells Crystialline silicon (c-si) solar cells were also called Chinese cells due to large scale Chinese production of these cells by Chinese companies such as Trina, Suntech, Yingli etc. Continued research and innovation 10 AAPPS BULLETIN

2 will be essential to the development of more affordable materials. Widespread use of PV technology is limited by the high cost of silicon. Most PVs currently being manufactured and used worldwide are made of solargrade silicon (Si). It is true that although the cost of solar-grade silicon has dropped (from $ /kg in December 2007 to $30/kg in December 2011), silicon is still the most costly component of PV production, accounting for nearly 60% of the cost of producing a single solar cell module. For poly-silicon materials, in order to reduce the cost of silicon materials in addition to the modified Siemens method, physical methods to purify silicon materials from metallurgical grade to the level of solar cell grade have recently been developed. Currently, c-si technology dominates the marketplace, with about 87% of all solar array installations worldwide comprised of c-si cells. The practical limit for c-si cell efficiency is around 26%. However, to get close to this level of efficiency there is sufficient space for improvement. Most importantly, while c-si PV technology is the most efficient of all PV technologies, with an average of 16-20% efficiency in converting absorbed sunlight into electricity, the current cost of c-si technology is still too expensive for widespread use. It needs to use physical methods to purify silicon materials and continue to reduce the thickness of the silicon wafer that makes up such a large part of each first-generation c-si PV cell. In addition, one may do two general things to reduce costs: (1) incorporate light trapping in the cell structure and (2) improve the surface passivation and the bulk passivation. There are several potential methods to improve c-si cell performance. Of course, many other concepts are already incorporated in the production of cells: reducing surface reflection, proper formation of emitter, formation of emitter and base contact regions, low serious resistance from metal, metal contact and grid itself, and the lowest front metal shading. While decreasing the wafer thickness makes good economic sense, it creates a new set of challenges. Specifically, it increases the chance that not all photons will be absorbed, with some being reflected back into the atmosphere, thereby decreasing the cell s solar conversion efficiency. There are two ways that cells can be modified in an effort to avoid this problem: (1) light trapping and (2) improved surface passivation. Light trapping increases the amount of absorbed light that gets trapped, for example, by creating an inverted pyramid silicon surface or adding mirror-like metal layers to the back of the device. Recently black Si techniques have also been investigated. Passivation renders the surface of the semiconductor chemically and electrically passive, thereby reducing its chemical reactivity and improving the probability that the charge carriers (i.e., the electrons and holes) will be transported into the external circuit. There are a few examples of how these techniques have been used to make c-si cells more efficient. For example, by adding passivating layers of SiO 2 to both the front and rear surfaces, Fraunhofer ISE developed a 21.6% efficient LFC-PERC (laser-fired contact - passivated emitter, rear cell) solar cell. SunPower and Sanyo have both used passivation techniques to reduce the cost and improve the efficiency of their respective proprietary c-si PV technologies. SunPower has reported 21% efficiency, and Sanyo has reported 22.3% efficiency. There is some recent research on other materials besides SiO 2 that could be used for passivation. For example, researchers at RWTH have been studying the use of silicon nitride (SiNx) as a passivating material; scientists at Applied Materials, Inc., have been investigating passivation using a combination of SiO 2 and SiNx; and researchers from Eindhoven University of Technology, Netherlands, the Institute for Solar Energy Research at Leibniz University, Hannover, Germany, and the Tokyo Institute of Technology, Japan, have been studying the use of aluminum oxide (Al 2O 3) films as a passivating material. To date, the world record using such passivation materials with c-si cells is Sanyo s reported 22.3% efficiency. Recently many institutions and companies are focusing on Silicon HIT structures, which have a high efficiency hetero-junction with intrinsic thin-layer (HIT) solar cells. The efficiency can reach 22.8% [2] or 23.7% [3]. Thin Film Solar Cells (TFSC) Thin-film PV (TFPV) technologies are used in second-generation solar cells, which include amorphous silicon (a-si)-based TFSCs as well as TFSCs comprised of other chemical compounds, such as polycrystalline copper indium (di)selenide (CIS), copper indium gallium (di)selenide (CIGS), cadmium telluride (CdTe), or epitaxial1 layers of indium gallium phosphide (InGaP) and gallium arsenide (GaAs). Most TFPV cells have relatively low efficiencies compared to c-si cells, with typical efficiencies in the 12-17% efficiency range. Most TFPV production modules are about 6-14% efficient. However, there is wide variation in TFPV efficiency, which depends on, among other things, whether a cell is multi-junction or not. Multi-junction PV cells contain multiple layers of different semiconductor materials, with each type of material absorbing a different wavelength of light. Increasing the range of light wavelengths that can be absorbed increases the amount of solar energy that can August 2012 Vol. 22 No. 4 11

3 be converted into electrical energy. Some multiple epitaxial layer cells (e.g., triple junction cells) have efficiencies greater than 40%. Some examples of new layers of materials being explored for their use in multijunction TFPV devices by scientists at the National Renewable Energy Laboratory, USA, are galliumindium-nitride-arsenide (GaInNAs) and boron-galliumindium-arsenide-phosphide (BGaInAs). In addition to their lower efficiency (compared to silicon wafer PV cells), stability poses another key challenge for TFPV. Many thin-film solar cells are inherently unstable because their materials degenerate over time when exposed to light. Lack of stability can be materials-related (e.g., while CIGS is relatively stable, a-si is susceptible to light-induced degradation), adhesion-related (i.e., peeling can occur between layers or from the substrate), or capsulation-related (i.e, moisture can penetrate the encapsulated module through laminated edges). Despite their efficiency and stability drawbacks, thinfilm solar cells require fewer raw materials and are less expensive to build than c-si cells. They also have a shorter energy payback period. That is, all solar cell modules require energy to produce, but TFPV modules pay back that energy more quickly in terms of generated electricity compared to silicon wafer technology. It takes about three years for a typical wafer silicon cell to pay back all the energy that was required to make the cell in the first place, compared to only about one year for a thin-film silicon-based solar cell. Added to their lower cost and shorter energy payback period, their flexibility and light weight make them suitable for applications not possible with the more conventional silicon wafer technology (e.g., for space, military, building-integrated installations). While TFPVs are not as popular as c-si PVs, a growing percentage of PV installations are thin-film based. For example, a 1.3 MW power station has been built in Germany using CdTe-based solar cells (the Dimbach Solar Park), and an amorphous silicon-based thin-film system is on the roof of the Beijing New Capital Museum, Beijing, China, which provides 300kW of power using flexible panels developed by United Solar Ovonic, a subsidiary of Energy Conversion Devices, Rochester Hills, MN. By 2012, it is predicted that 75% of all PV installations will be c-si-based (down from the current figure of about 87%), and the remaining 25% will be thin film-based. TFPV will become a more popular choice in the future. In fact, in the United States, most silicon-based PV production is with thin-film, not wafer, technology (i.e., by Uni-Solar). There are three inorganic thin-film PV technologies that have reached large-scale manufacturing in the past seven years: First Solar s cadmium telluride (CdTe) TFPV cells, Uni-Solar s amorphous-silicon TFPV cells, and Global Solar s copper indium gallium diselenide (CIGS) TFPV cells. At least one of these companies (First Solar) has experienced major cost breakthroughs as well. The efficiency of CdTe cells module reach 17%, and it is the cheapest one of all solar cell products. Years ago, some TFPVs demonstrated relatively high efficiencies in research lab settings; however, their actual efficiencies following large-scale manufacturing are much lower. CIGS, for example, is a difficult material to manufacture and so, while CIGS TFPVs may be 20.3% efficient in the lab, actual modules manufactured for sale were less than 10% efficient in However the situation changes rapidly. In 2011 the efficiency reached 17.1% by Nanosolar, 17.4% by Solibro, 17.3 by Miasole, 14.8% by Stion, 15.8% by Avancis. In 2012 Solar Fronties announced 17.8% efficiency for 30cm square cells, and an efficiency of 16.3% for modules. Rapid progress has occurred in recent years. The new installation of CIGS cells was 1.2 GW in 2011, and it will be even 2.3 GW more by Many companies have ambitions plans. Flexible CdTe and CIGS TF solar cells are developed such as in G24 Innovations Konarka New flexible solar modules are integrated, rather than installed into existing or new buildings. Roll-to-roll production of the first PV modules results in a flexible lightweight PV material that suits different applications. The efficiency of flexible CdTe solar cells is 12.4% from ETH Zurich. Monolithically interconnected flexible solar modules with an efficiency of 8.0% (31.9 cm2) were developed by interconnecting 11 solar cells in series [4]. However no industrial production has yet occurred. Further work should be done in the areas of large area up-scaling, efficiency improvement, module development, performance stability, etc. Using cheaper materials Zn and Sn as substitutes for In and Ga for CIGS to make a Cu 2ZnSn(Se,S) 4 solar cell is a meaningful project. The best results can be found in reference [5]. New-Concept PV Solar Cells There are three types of new-concept solar cells including wide spectra concentrated PV (CPV), dyesensitized cells (DSC) and organic cells. They can also be referred to as third-generation solar cells: 12 AAPPS BULLETIN

4 CPV is made by using semiconductor multi-junction solar cells combined with a concentrated light system;different wavelength sunlight can be absorbed by cells, therefore, to increase their efficiency. The word record for GaInP/GaAs/GaInNAs ternary junction cells is 43.5% (concentrated by 418 times sunlight) from Solar Junction UA (Prog. Photovolt: Res. Appl. 2012, 20,12?20). III-V compounds and II-VI compounds are good materials for CPV preparation. There are lots of theoretical and experimental works in the world. Some new tests by a c-si junction combined with II-VI CdTe junction shows a strong possibility to improve performance (Sivananth laboratory US). Dye-sensitized solar cells (DSSC). First engineered by Michael Grätzel in 1991, DSSC (also known as Grätzel cells) use dye molecules to absorb incoming light; the photon-excited dye generates an electric current in a separate, non-silicon-based semiconductor material (e.g., titanium dioxide, TiO 2). It should be pointed out that DSSC are less expensive than c-si cells, but they also have a lower efficiency than both c- Si and TFPV cells (with a maximum reported efficiency of just over 11%) and suffer from some of the same stability problems that other TFPV cells have. It is also noted that not only are DSSCs less expensive than c-si cells, they are also easier to produce. In China, researchers are investigating whether lithium iodide (LiI), the traditional electrolyte material used in DSSCs, can be replaced by the less expensive aluminum iodide (AlI 3) to bring their cost down even further. University of Bath, UK, Massachusetts-based Konarka Technologies UK-based G24 Innovations (G24i), Fraunhofer Institute for Solar Energy Systems ISE, Germany, and elsewhere are continuing to investigate new ways to increase the efficiency of DSSC technology. The greatest reported solar energy conversion efficiency of a DSSC is 11.4%, and that record was for a single cell in a research laboratory setting. Efficiencies in practical settings are lower, on the order of 8-9%. Conducting polymer solar cells, like DSSC, are potentially cheaper to manufacture than other types of PV technologies because they use semiconducting polymer materials instead of silicon, but they are not as efficient. They are also not very stable, with typical longevity for conducting polymer cells being only 3 to 5 years. Konarka recently reported 6.4% efficiency, which is the highest performance by 2009 for a conducting polymer solar cell in a research laboratory setting. However, the company s reported rooftop efficiency is only 3.3%. There also are some of the research approaches being taken in an effort to increase the efficiency of polymer solar cells. These include using different types of polymers and nanostructured materials. New results is 8.9% from Chemistry institute, CAS (Adv. Mater. 2012, 24, ) and 9.1% from Polyera UA ( ). Molecular (or small molecule ) organic cells, as with the other new-concept technologies, molecular organic cells are potentially cheaper than silicon wafer and thinfilm technologies, both because of the materials used and because of their manufacturing processes. However, their solar energy conversion efficiency is only about 4% to 6% (Heliatek GmbH, Dresden, Germany), which is the record for small molecule organic solar cells at the moment. Their limited efficiency is due in part to the fact that most currently available molecular semiconductor materials absorb only a narrow wavelength of light and therefore are not capable of harvesting much light. The Future of PV Technologies The five next steps for advancing solar cell technologies are important: (1) Improve existing processing technologies for firstgeneration c-si solar cells. Even if the cost of silicon were to significantly drop, it would affect only 60% of the cost of a c-si solar cell module. The processing and manufacturing of the materials and cells still need to be improved as well in order to bring the remaining 40% of the cost of c-si technology down. (2) Develop new materials for both thin-film and newconcept PV technologies, not just in order to bring the costs down but also for performance reasons. While searching for Earth-abundant materials, we need to find a way to replace some of the more expensive, less abundant and potentially environmentally harmful materials that are currently being used with less expensive, more abundant, non-toxic materials. (3) Continue to develop new structures, processes and concepts for both thin-film and new-concept PV technologies in order to improve the performance of these potentially less expensive technologies. (4) Combine physicist efforts with those of scientists in other disciplines (e.g., physics, materials science and engineering). (5) For all the solar cells the basic scientific problems should be focused as follows: inner electric field construction; effective carriers transport mechanism; carriers mobility, lifetime, diffusion length; August 2012 Vol. 22 No. 4 13

5 processing, materials characters, device structures and the controlling of the opto-electric transition; The relationship between the crystal microstructure and materials performance. electron excitation, transportation and recombination and optoelectronic excitation dynamics. It should be mentioned that silicon is still the best material in terms of efficiency and that, within the next ten years or so, Si-based cells will become more affordable because they will be built with less Si than they are today, and will be with more high efficiency. However, some of the other thin-film materials, like CiGS, CdTe, CZTS and thin-film multi-junction cells by III-V, II-VI and silicon thin film, will become more popular over time. The molecular organic solar cells in particular will become very important in the future because of their low cost, although their efficiency and particularly their stability will need to be improved. The fact that the efficiency and stability of molecular organic solar cells will need to be improved before they can become a viable option for large-scale use led into further discussion about the many basic scientific challenges that still remain. References [1] Junhao Chu, Progress of photovoltaic cells for solar power, 1 st Annual Chemical Sciences and Society Symposium (CS3), Kloster Seeon, Germany July 23-25, [2] Takahiro Mishima, MikioTaguchi, HitoshiSakata, EijiMaruyama, Development status of high-efficiency HIT solar cells, Solar Energy Materials & Solar Cells 95 (2011) [3] Kinoshita T, Fujishima D, Yano A, Ogane A, Tohoda S, Matsuyama K, Nakamura Y, Tokuoka N, Kanno H, Sakata H, Taguchi M, Maruyama E. The approaches for high efficiency HIT solar cell with very thin (<100mm) silicon wafer over 23%. 26th EUPVSC Proceedings 2011; [4] J. Perrenoud, et al., Solar Energy Materials & Solar Cells 95 (2011) S8-S12. [5] D. A. R. Barkhouse, O. Gunawan, T. Gokmen, T. K. Todorov, D. B. Mitzi, Device characteristics of a 10.1% hydrazine-processed Cu 2ZnSn(Se,S) 4 solar cell, Prog. Photovolt: Res. Appl. 20 (2012) AAPPS BULLETIN