There is something disconcerting about going through a speed-trap at 100mph and not breaking any laws. No flash, no flashing blue lights, just a high-speed procession of cars past an inactive radar camera. This is very typical of Germany, a country that will accept costly new technologies under certain circumstances, but only when it is deemed necessary and popular. The rapid adoption of solar photovoltaics (PV) is broadly seen as necessary in Germany (although not without spirited discussion), and it is for this reason they have been in the vanguard of the PV industry.
Reducing the Cost of Power
In the world of PV, cell manufacturers have an absolute need to reduce the cost of the power generated by photovoltaic installations, because even in Germany the patience to subsidize this technology is limited. A world class competitor in cell manufacturing must choose cost reduction strategies. One approach is one of capital intensive, high technology implementation that reduces cell variable costs through quality, utilization, process control, and efficiency optimization. Another approach is one of high utilization of a low-cost asset base, leveraging cheap factory costs and regional financial advantages.
The reality that a PV module is expected to be productive for at least 20 years means that quality and consistency of construction are important attributes for a supplier to establish. Even a low-cost supplier will make significant efforts to ensure that quality certifications are achieved and maintained, allowing the final product to be financed, or “bankable” in very large installations.
These considerations mean that in c-Si manufacturing technology, a constant battle rages between the implementation of new technologies with the promise of high efficiency, perhaps with added process steps, and a conservative approach that demands rigorous evaluation of change before it is accepted. Some of the most important manufacturing developments within c-Si technology currently under discussion include:
- Wafers. Despite the advances in wafer sawing and other process improvements, there is increasing consensus that the practical process limit for mc-Si cell manufacture is about 160µm. This will focus efforts on improvements in handling systems and reductions in micro-cracking.
- Texturization. There have been efforts to improve texturization formulations for better roughness control, improved uniformity, and easier process control. There are also efforts to go beyond wet texturization with imprint lithography and novel plasma processes.
- Emitter Formation. The many approaches to selective emitter (“SE”) published 12 to 18 months ago have proliferated and evolved. However, only a certain few SE based cell technologies are finding their way into commercialization. One competing pressure is the payoff of quality efforts in improving the efficiency of basic cells designs. Notable novel entrants into SE manufacture are screen-printable dopants, both glass- and silicon-based, and screen-printable etching pastes for precise etch control with reduced consumable chemicals.
- Metallization. The reliable standard of Ag and Al paste printing continues as a fundamental process, but a need to reduce the Ag content (and cost) is recognized, and developments in Ag augmentation and replacement through plating continue to be developed. Techniques of light-induced plating and variations on electroless plating are being improved in an effort to migrate from plating up on fired paste seed layers, to full direct plating on SE. Printing technologies are not giving up easily, and double-printing technology suppliers now claim well-over three GW of commercial capacity.
- Automation, Statistical Process Control (SPC), Advanced Process Control (APC). Approaches to implementation of automation vary by degree. Benefits of automated processing include the repeatability of positioning and actions, repeatable dispense and use of chemicals, and minimum unforeseen handling, and fewer process deviations.
Achieving Well-controlled Processes
Characteristically, there is no flat response of the c-Si PV process, but the optimum cell properties are a single combination of the multiple attributes of the finished product. Process variability within manufacturing is high, and to some extent due to the relatively simple design of the process tooling, resulting in a broad, distribution of individual cell properties. This makes assembling modules with like cells difficult, and lowers the average module performance. Any method of reducing the distribution of cell properties will result in more high-value cells, more flexibility in module production, improved pricing, and better financial returns.
Critical for implementation of an automated process is that all process elements are in good control (in a statistical sense) and that most or all of the input variables are well characterized and under active control. This requires a competent staff trained in implementation of SPC, and suitable investment in metrology hardware (i.e., sensors, etc.) and software to ensure control is implemented. All of this adds investment and direct cost while offsets of improving process yield and higher cell efficiencies are designed to lower total costs.
The next step in process automation is APC in which each in-line measurement feeds forward and backward in the process to adjust input parameters to achieve optimum processing. In a high-volume line this needs to be done automatically and continuously, which involves sophisticated line management and well engineered manufacturing execution software. Although these techniques are common in industrial manufacturing, they have not been widely implemented in PV cell manufacture.
The improvements in process technology discussed above all offer incremental improvements in wafer and cell processing, and will be augmented by further improvements. The conservatism of current manufacturers can only stall process change if manufacturers can be convinced that scale-economies alone can deliver grid parity. As evidenced by ongoing process development and the efforts to produce cell improvement roadmaps, few, including this author, believe this is the case. New process technologies will continue to be developed and commercialized in the future.
Most of the cell efficiency improvement approaches detailed here will be evaluated in mass production, and many will be adopted. Those methods adopted will have to meet high barriers of acceptance since improvements will have to deliver cost reductions against possible increased investment costs, increased materials costs, and increased process complexity – none of which a cell maker wants to hear!
Design plans for plant investment to make these improved cells require a choice between multiple low sophistication tools at low cost with a larger operator count versus integrated lines with high automation and the need for a small group of skilled engineers. Perhaps this is an artificially simple illustration, but it is somewhat representative of reality today in those regions where capital subsidies and industrial policy have a profound influence.
Arguably, many of the benefits of a well controlled manufacturing process accrue equal advantage to both a manual and automated production, but modern industrial technology tends to the latter with demonstrable results. Set against this reality, manufacturers establishing in China know that the expectation is primarily to employ large numbers to support a growing export business. China has shown the ability to outstrip other regions in high-tech manufacturing in other industries once its manufacturing base has been established, but perhaps a viable strategic option today is to attempt to get there first with investment in integrated advanced manufacturing capability.
Back in Germany, a severe winter is beginning to give way to spring. A new round of PV systems installations will accompany the warm weather. As the volume of new PV installations continues to grow, we expect large markets other than Germany to emerge to spur the innovation cycle further.
Mark Thirsk received his Honors BSc in metallurgy and materials science from Birmingham U., and an MBA from The Open Business School; he is a principal at Linx-AEI Consulting and is also a founder and managing partner at Linx Consulting LLC.