Leveraging Novel MEMS Technologies for Next-Generation Photovoltaic Applications

The MEMS industry is characterized by highly diversified, non-standardized processes for heterogeneous applications going into a variety of markets. Thus, technology successfully introduced for this market must be very flexible, and can therefore be easily leveraged for cross-industry innovation. MEMS technology has already been used in other sectors of the microelectronics industry, such as advanced packaging and 3-D integration in semiconductor devices, and has proven to generate tremendous competitive advantages.

Photovoltaics (PV) technology is substantially different from MEMS, but both industries are similar in age and their stage in volume manufacturing. Also, next-generation high-efficiency PV cell approaches call for novel manufacturing technologies to significantly reduce the cost per watt. This article will describe the business methodology of this cross-industry innovation process, and will briefly introduce the technologies that were leveraged as well as the targeted applications.

Business methodology

The PV industry must deal with compromises in all aspects. Among the most crucial ones is the efficiency of energy conversion and the related cost for producing complex PV structures. Even though the global PV sector has been growing significantly for the past eight years — at a compound annual growth rate (CAGR) of ~40% — a recorded 5% annual price decrease of PV modules on average will hardly boost the sector toward a widely adopted energy source within the near future. Implementing cost-effective production capabilities along with an innovation strategy for high cell efficiency is crucial.

Suppliers to the MEMS industry have been experiencing similar challenges over the past couple of years. Today’s MEMS growth is fueled by applications in the consumer industry. Accelerometers and gyroscopes for gaming, notebook and mobile applications, silicon microphones and RF resonators for mobile phone applications, and digital mirror devices for home entertainment and projection applications have put pressure on those suppliers to reduce total cost of ownership (CoO) for production equipment.

While this fact has forced equipment suppliers for the MEMS market to come up with innovative ways to save costs, it has also led to technological breakthroughs — including backside alignment for double-sided lithography; spray coating for conformal resist coverage on challenging topographies; and wafer bonding to permanently, as well as temporarily, join two or more wafer substrates. These technologies have already been successfully leveraged for advanced packaging and 3-D integration applications, as well as for the compound semiconductor industry.

Like MEMS, the PV industry is migrating toward high-volume manufacturing, maintaining a high degree of vertical integration. Silicon and thin-film materials are the predominant substrates used, and device packaging is primarily done in-house. In addition, there is a heavy interdependence of process and equipment because of the lack of standardization and the significant amount of proprietary know-how that is involved in the manufacturing process. So, although PV is a substantially different market than the MEMS and semiconductor industries, MEMS and PV are surprisingly similar in terms of intrinsic industry characteristics. Furthermore, the PV industry is characterized by an even higher price elasticity of demand. This fact provides a good feeding ground for technology that is able to substantially reduce total cost of ownership (TCO) and thus finally contribute to reducing the cell cost per watt.

To have a well-established TCO strategy for all equipment and services in place (Fig.1), it is essential to not only focus on hard facts like machine throughput, uptime, footprint, investment costs and process yield, but also on soft facts such as reaction time for new inquiries, the resulting transaction speed, the flexibility of the tools to deliver investment protection, the reputation in the market, the reliability of delivery times, and other factors that silhouette the company against the competition. This significantly helps to reduce the risk when targeting new markets with existing technology as long as the company is able to effectively communicate this innovation strategy through effective strategic marketing channels.

1. The total cost of ownership puzzle for MEMS equipment.

But successful cross-industry innovation is more than market differentiation with existing technologies. It also involves the hunt for analogies in the source and target industry sector on a high level of abstraction, as well as the targeted adaptation of the technology. Ideally, cross-industry innovation is screening for technology and business potential outside the industry due to the probability for a higher degree of innovation, resulting in stronger competitive advantages. On the other hand, the higher risk associated with a cross-industry innovation strategy calls for a solid risk management process. Figure 2 illustrates the strategies that have been evaluated to reduce business risk within a cross-industry innovation process.

2. Risk management strategies within a cross-industry innovation process.

Technology and applications: MEMS for PV

The photovoltaics industry is actually borrowing most of its manufacturing technologies from the semiconductor industry. Likewise, there are strong interdependencies between MEMS manufacturing technologies and next-generation semiconductor devices and packages. To list just a few example: wafer bonding for 3-D interconnect and thin-wafer handling, double-sided mask alignment and spray coating for advanced semiconductor packaging, and nanoimprint lithography for wafer-level cameras.

This MEMS-semiconductor-photovoltaics chain boosts the success rate for transferring technologies between MEMS and PV. On the other hand, it is evident that the degree of innovation in cross-industry strategies created from a transferred technology is higher the more different the characteristics of the two markets are. When comparing MEMS and PV, the obvious differences from the point of an equipment manufacturer can be seen in the device volumes, the size of the substrates, and the added costs per substrate through the manufacturing process as well as the throughput numbers per investment costs for the tool. These facts impose challenges to the cross-industry innovation process and confine the target market mainly to high-efficiency cells and cells that require challenging process and handling steps.

Wafer bonding is a proven technology in MEMS that was introduced in the early 1990s to join two substrates together in a vacuum chamber with the help of high forces and temperatures up to 600°C. When optically aligned wafer bonding or layer transfer is required, the two wafers are aligned separately prior to transferring the pre-aligned and clamped wafers to the bond chamber. Today’s bond alignment technology is able to deliver post-bond alignment accuracies in the range of ±0.5µm (3σ) with non-UV and non-IR transparent wafers. Wafer bonders with modified heat and pressure transfer concepts (Fig. 3), which were optimized for thin-layer transfer processes, deliver added value to customers pursuing thin-film solar cells, such as CIGS, multi-junction and multi-band PV cells. The wafer bonding tools were adapted for higher bond forces up to 60 kN in order to address larger substrate sizes, new cluster handling concepts to accommodate atmospheric handling restrictions, and rapid heating concepts for the intrinsic physical reaction characteristic for thin-layer transfer.

3. The Gemini fully automated bonding cluster for wafer bonding and layer transfer processes has been optimized for thin-film solar cells.

III-V concentrator cells require lithographic micro-structuring at certain stages, and thus represent another area where MEMS equipment suppliers can contribute with a significant reduction in total cost of ownership. Here, current screen printing technologies cannot be used because of their limited resolution capability, which would result in shadow effects on the active layers. Current mask aligners (Fig. 4) can achieve minimum resolutions down to 1µm in proximity exposure modes with a throughput that easily exceeds that for inkjet or UV stepper technology.

4. The Hercules lithography track system for front and double-sided lithography can be optimized for polysilicon solar cells.

Mask alignment tools can also be used in novel concepts for polysilicon solar cells targeting an optimized use of the silicon substrate by increasing the active silicon surface. Innovative manufacturing concepts like Origin Energy and ANU’s Sliver technology, which uses a double-sided wet-etching process to create the thin silicon sliders and thus requires a lithography-structured etching mask on both sides of the substrate, are using mask alignment technology to create the etching mask.

Particularly for thin crystalline and polymer solar cells, spray coating can deliver an innovative solution for applying thin films of metallic particles suspended in a polymeric matrix. This technology has already been successfully transferred from MEMS into flexible display applications and next-generation wafer-level packaged semiconductor devices with electrical through-silicon via (TSV) interconnects (Fig. 5).

5. The EVG 150 large-area spray coating platform (left), whose technology has already been transferred to through-silicon via (TSV) technologies (right), could also be leveraged for photovoltaics.

Spray coating is different from conventional spin coating processes for the application of polymer-based resist materials. Instead of dispensing resists to the center of the wafer and spinning off the polymer at high speeds to create a uniform layer of resist on a flat substrate, spray coating uses a low-pressure ultrasonic atomizer to spray fine resist droplets onto the stationary wafer surface and is thus able to dispense an extremely uniform layer of resist over high-topography substrates at comparable throughputs. Additionally, it can save up to 80% of resist compared with spin coating and can generate highly uniform and thin resist layers below 100 nm. The EVG 100 series can be scaled for substrates up to 780 × 650 mm, and is uses high-speed spray nozzles for optimized throughput and coating uniformities.

Nanoimprinting for the creation of sub-100 nm structures and temporary bonding and de-bonding for ultrathin-substrate handling are technologies that have already been successfully transferred from MEMS to other industries. These technologies could potentially be directly leveraged with minimum risk to the PV industry as well. Nanoimprinting has high potential for future organic cells with vertical pn-junctions.

Thin-wafer handling concepts might even find their first volume applications earlier in the PV industry. One of the most important factors to reduce the cost per watt for silicon solar cells is to reduce the thickness of the silicon wafer material. Depending on the size of the substrate, the minimum thickness that can be handled with standardized equipment transfer and process concepts is in the range of 100-150 µm (Fig. 6), and this value is even higher for brittle non-silicon materials. Leading silicon cell manufacturers will hit this thickness limit within the next 1-3 years. A production-proven technology for ultrathin-wafer handling that is seen to be the industry standard in future is handling the wafer over rigid carriers (referred to as “temporary bonding” and “debonding”). The thin wafer is temporarily bonded to a reusable rigid carrier by means of special intermediate layers with a dedicated release mechanism. These intermediate layers can be dry-film laminates, spin-on polymers or special waxes. Thus the ultrathin wafer can be handled throughout the full process chain just like a standard wafer, and there is no additional investment necessary for adapting equipment to thin-wafer compatibility. At the end of the manufacturing process, the finished ultrathin wafer is debonded from its carrier through the activation of the corresponding release mechanism (usually either temperature or UV radiation).

6. Ultrathin substrates offer needed cost advantages for photovoltaics, but they come with handling challenges, as shown with this 100 µm thin 300 mm silicon wafer.

Conclusion

Cross-industry innovation on the technology level has proven to be successful between the MEMS and the PV industry. For the PV industry, which is characterized by a high price elasticity for demand, only technologies that are able to substantially reduce the total cost of ownership for cell manufacturing will be successfully leveraged. Permanent wafer bonding for thin-layer transfer processes, temporary bonding for ultrathin-substrate handling, lithography processes for defining micron-scale structures, and spray coating for applying ultrathin layers of conductive material are examples of successful cross-industry innovation between MEMS and the photovoltaic industry.