The photovoltaic (PV) industry was founded on and, if you will, fueled by crystalline-silicon (cSi) solar cells and modules.
by James Gee, Advent Solar, Albuquerque, NM USA
The photovoltaic (PV) industry was founded on and, if you will, fueled by crystalline-silicon (cSi) solar cells and modules. Crystalline-silicon is not expected a priori to be an ideal material for photovoltaic energy conversion. It has an indirect bandgap, which means that it has relatively weak optical absorption so that thick (typically >140µm today) substrates are required to yield respectable efficiencies. The cost of such substrates has helped motivate a great deal of research and development in thin-film semiconductors and other material systems for PV.
However, the head start and strength of cSi as a PV material is due to its extensive technological development from the electronics industry. Crystalline-silicon is, perhaps, the best understood technological material today. The result is that crystal growth techniques are well established and the effects of dopants, impurities, crystal defects, gettering, and passivation are all relatively well understood. Similarly, there is extensive knowledge regarding processing silicon. For example, a wide variety of processing techniques for doping, annealing, chemical and physical vapor deposition, etching, patterning, contacting, etc. are available and well established—including a large vendor base. Crystalline-silicon PV products have also earned an excellent reputation for reliability and durability with an outstanding energy-conversion efficiency.
While the very rapid growth of the PV industry has led to several constraints in the supply chain, the growth of the polysilicon feedstock, in particular, has provided thin-film PV products, particularly CdTe and more recently a-Si and alloys, an opportunity to enter the market. The supply constraints are a temporary issue, though, associated with the rapid growth of the industry, and there are already indications that the supply constraints for polysilicon feedstock will be significantly alleviated during 2009. Crystalline-silicon PV has a record of continuous cost reduction and, with its higher efficiencies, will maintain a strong market position in a large and technologically diversified industry. Crystalline-silicon represented ~89% of the PV products shipped in 2007 .
As will be emphasized in this paper, the large technological base shared with the electronics industry provides a new generation of opportunities for novel solar technology improvements using crystalline-silicon.
Conventional vs. technology architecture-based designs
Most cSi products today use a conventional design that would be familiar to the engineers that first developed cSi PV cells and modules 30 years ago. ?Conventional? in this paper refers to stand-alone, p-type silicon wafers with an n-type diffusion on the front surface, screen-printed Ag grids, and an Al-alloyed back-surface field and contact on the rear surface. The front grid has a well-known performance tradeoff of reducing series resistance versus reducing optical obscuration losses. Collectively, the optical and resistive losses of the front grid are >8%.
Perhaps, it is less known that the presence of contacts on the front and rear surfaces impacts the module technology, compromising both performance and cost . Conventional solar cells are ?strung? in electrical series with soldering and flat Cu ribbon wire with special tools (stringers). The cross section of the wire is, however, limited—thicker wires are too stiff and thin, wide wires would obscure too much light. The net result is that the interconnect resistance losses can account for another 4% reduction in performance. The stringing process itself is non-planar in geometry that is difficult to automate that limits throughput per tool. The process is also difficult to use with thin cells because the resulting series string of cells is fragile.
While conventional cSi PV technology will make further performance and cost improvements, it will take a change in the basic architecture of the cell and module for substantial improvements. Improvements in product generations involving sophisticated technologies are frequently referred to technology architectures whereby components, standards, and interfaces are all coordinated for best system performance .
Back-contact cells provide a new crystalline-silicon photovoltaic architecture that provides advantages at the cell, module, and even the system level. And while the first-generation cSi PV technology took advantage of the semiconductor processing capabilities of the electronics industry, the next-generation cSi technology will take advantage of concepts and processes from the packaging portion of the electronics industry.
Architectural approach to cell/module design
An architectural approach addresses performance, manufacturability, and scalability in the design of the cell and module together. Scalability here refers to the ability to scale the substrate (size or thickness), the module size and features, and the manufacturing footprint (throughput per tool). The net result is that the new architecture should be able to address future market requirements more easily.
Figure 1. Illustration of a monolithic module assembly (MMA).
The electronics packaging industry provides an example of an architecture that scaled more easily for performance and cost. Initially, thin wires were bonded to the semiconductor die. The process was serial and made for inefficient chip designs due to the necessity of routing all external connections to the edge of the semiconductor die. Flip-chip technology distributes the bonding points throughout the semiconductor die. These bonding points are coated with a solder and the die is then aligned over a matching circuit so that the die is electrically attached at all the points simultaneously in a reflow oven. Flip-chip process provides performance, cost, and scaling advantages by changing the fundamental geometry of the semiconductor die to better match the final product (packaged electronic component).
The performance and manufacturing limitations of conventional solar cells arise from the front-grid geometry. Back-contact cells provide a fundamentally new approach for optimizing the cell and module design. The interconnects for back-contact cells are no longer constrained by optical losses. Hence, wide and thin interconnects can be used that minimize both electrical resistance and stiffness (stress). The negative- and positive-polarity contacts are now on the same surface, so the assembly process now has a more planar geometry that can use more rapid pick-and-place style tools.
The planar geometry will also be more compatible with thin cells and allows for closer spacing of the cells inside the module. The coplanar contacts also allow for a completely new assembly process where the solar cell interconnection is made during the lamination step. This process requires a module backsheet with a circuit and an electrical attachment material that bonds during a lamination cycle (Fig. 1). In effect, the stringer process step and tool is now replaced with the roll-to-roll processing used for fabrication of the backsheet with the electrical circuit. The process has been called monolithic module assembly (MMA) due to the integration of the encapsulation and electrical assembly into a single step.
Advent Solar has introduced an integrated back-contact cell and module architecture called Ventura Technology. It combines a back-contact cell technology using emitter-wrap-through (EWT) with a highly automated, planar manufacturing process (MMA). The EWT cell is a back-contact cell that is particularly advantageous with solar-grade Si substrates (see ?Emitter wrap-through back-contact solar cells? on p. 28).
Figure 2. Schematic view of an EWT cell using backside distributed contacts.
The important addition to the EWT cell in the architectural integration with MMA is the distributed contacting points (Fig. 2). These distributed points reduce the resistance in the solar cell and minimize the use of and the stress from the expensive solar cell metallization (Ag). This cell/module design is an example of the type of results from the architectural integration—the rapid transfer of current from the expensive solar metallization to the less expensive metal in the backsheet circuit minimizes cost while maximizing performance at the same time. Early production cells with efficiencies >17% have been obtained with the new geometry. More importantly, the architectural integration of EWT and MMA maintains more of the cell power in the finished module. Encapsulation loss refers to the ratio of the module versus cell power. This loss term is typically ~4?5% for modules using conventional cells. Early production modules using the new technology have achieved less than half this value; the various production and performance advantages are summarized in the table on. p. 33.
Science-based reliability methodology
The MMA process involves a new assembly method with a new electrical contacting material—conductive adhesives. PV modules are expected to achieve very long lifetimes of 25 years. The flexibility of conductive adhesives may be an advantage for long lifetimes, and they have been used in some particularly demanding applications (e.g., automotive electronics). Conventional module technology has undergone a great deal of accelerated environmental stress testing and field testing that has given the community confidence in its reliability.
Industry has developed consensual standards for accelerated testing to help ensure consistent and reliable product performance. While test coupons and full-sized modules with the new MMA process and design have passed all major components (thermal cycling, humidity freeze, and damp heat) of the IEC certification test sequence, a more detailed study was initiated to develop greater confidence in long-term field reliability.
Qualification of a product with new processes and materials can be accelerated with use of a science-based reliability test methodology (see Fig. 3). Traditionally, reliability testing involves large statistical samples for correlation of accelerated testing to field testing. A science-based methodology uses modeling and simulation to help identify performance characteristics, special test structures for isolating various components, identification of failure modes and mechanisms, development of predictive reliability models, and development of new quality assurance tests and procedures .
Work in progress includes development of a detailed finite-element model of the MMA module construction, characterization of the viscoelastic properties of the module materials, development of test coupons for adhesion and resistance of various components in the module stack, development of failure analysis techniques for examining the new product, and development of new quality assessment procedures. The result of such a program is greater confidence in the reliability of the new product and in ability to maintain the process in production.
The evolution of an industry and it products are generally expected to favor more integrated design of components for maximum system cost/performance. A new approach has been described for optimizing back-contact cSi solar cell and module design together. This approach has demonstrated performance advantages that should also scale more easily for future requirements (larger and thinner cells, easier scale-up of the production tools, etc.) Additionally, a science-based reliability methodology was described to enable faster new product and process qualification.
Figure 3. Illustration of science-based reliability methodology. This illustration is applied to a microelectromechanical system (MEMS) device, but is applicable to any new product or technology.
Advent Solar is a registered trademark of Advent Solar Inc. Ventura is a trademark of Advent Solar Inc.
- P. Mints, ?Global Potential for Solar Electricity, Supply and Demand Perspective,? IEEE/AMAT Solar Symp., 2 October 2008 (San Jose, CA).
- A.W. Weeber, et al., ?How to Achieve 17% Cell Efficiencies on Large Back-contact Multicrystalline-silicon Solar Cells,? 4th World PV Solar Energy Conf., pp. 1048-1051, Waikoloa, HI (2006).
- C. R. Morris, C. H. Ferguson, ?How Architectures Win Technology Wars,? Harvard Business Review, pp. 86-96, March-April 1993.
- D. M. Tanner, et al., ?Science-based MEMS Reliability Methodology,? Microelec. Rel. 47, 1806-1811 (2007).
- E. Van Kershaver, G. Beaucarne, ?Back-contact Solar Cells: A Review,? Progress in Photovoltaics, 14, 107-123 (2005).
James Gee received his MS in electrical engineering from Stanford U. and is chief scientist at Advent Solar, Albuquerque, NM USA; email@example.com; www.adventsolar.com.
Emitter wrap-through back-contact solar cells
Back-contact silicon solar cells have received a lot of attention in the research community . The primary advantages include improved efficiency by reducing grid obscuration, simpler automation of the module assembly, and enhanced aesthetics due to a more uniform appearance. More recently, they have gained considerable commercial attention. SunPower offers a high-efficiency solar cell with n-type monocrystalline-Si substrates and a proprietary processing technology. Kyocera, PhotoVoltech, Sollhand, and Q-Cells offer, or have exhibited back-contact cells, using p-type multicrystalline-Si and a metallization wrap-through (MWT) design. MWT wraps metal from the front to the rear surface through a limited number of holes—so there are still some grid obscuration and resistance losses on the front surface and associated cost of the additional print step.
Illustration of an EWT cell.
A promising back-contact cell approach is the emitter wrap-through (EWT) solar cell. This solar cell wraps the n+ emitter from the front to the rear surface through laser-drilled holes to form conductive vias (see figure). Unlike the MWT cell, this design features complete elimination of grid obscuration losses for higher efficiency and improved aesthetics. The cell structure also inherently features double-sided collection due to the presence of an n-type emitter on both surfaces.
Double-sided collection means that photogenerated carriers in the bulk of the p-type Si substrate can be collected by either the front- or the rear-surface emitter—which effectively increases the interior collection efficiency. The improved interior collection efficiency means that EWT cells can use lower cost Si materials, such as upgraded metallurgical-grade silicon, with reduced performance penalty. EWT cells will also suffer less from light-induced degradation that is commonly observed in p-type monocrystalline-silicon cells due to an interaction between B dopant atoms and interstitial O impurities. Monocrystalline-Si EWT cells have been demonstrated with efficiencies >21%, and projected efficiencies for multicrystalline-Si EWT cells are ~19%.