Shining a New Light on Silicon PV Manufacturing
Groundbreaking furnace tests photovoltaic wafer viability.
"It's all very complicated," says Principal Scientist Bhushan Sopori, "but what matters is that we're controlling, shaping light." While the National Renewable Energy Laboratory's (NREL) light-shaping innovation is indeed complicated, it actually simplifies the process of fabricating silicon solar wafers. It does this by creating just the right conditions for something to go wrong... acting almost like a time machine that predicts if something will go wrong in the future.
Leaning against a metal furnace that, despite containing a high degree of thermal heat, remains cool to the touch, Sopori describes its efficiency. "Almost none of the light is wasted. It is all directed precisely to the surface of the moving wafer."
The furnace, which directs a stationary beam of light across the surface of the wafer as a conveyor belt carries it through an illumination zone, appears to be a means of inspecting the surface of these very thin wafers for defects. However, silicon wafers have rough surfaces, covered with irregularities that would hide most fatal defects, even when viewed under the most sophisticated lens.
The sound of a gentle snap comes from inside the illumination zone and a broken wafer moves out from beneath the metallic structure. For the sake of the demonstration, the wafers are supported by an extra surface—otherwise, the broken pieces would fall through the metal rods on the belt, which would automatically eliminate them from the line.
"This was the wafer we just scratched." Sopori notes, showing the damaged wafer. He then demonstrates how he can adjust the attached computer interface to apply all the different parameters of thermal stress.
Meanwhile, the conveyer belt carries a second wafer through the other side of the furnace, unscathed. "This wafer, on the other hand, had no damage. It will stay in the production line to undergo the next cell fabrication steps."
Sopori explains that the test is unique because the scratched wafer was broken, but the undamaged wafer was not—and not only that, the test doesn't alter or weaken an undamaged wafer in any way.
Tackling the Serious Issue of Wafer Breakage
The silicon photovoltaic (Si-PV) wafer screening system was designed to solve a serious problem in the silicon wafer manufacturing industry, which comprises 85% of the PV cell market. Typically, between 5% to 10% of wafers break during the process of cell production, a material loss that limits production yields and increases the final cost of PV modules.
Manufacturers are using increasingly thin wafers to achieve higher cell efficiency and to lower the wafer cost. Unlike the microelectronics industry, which polishes the surface of silicon cells to remove the surface damage and microcracks that can occur during wafer sawing, the PV industry does not prepare wafers in a way that maintains their mechanical strength.
Synergies Among Innovative Technologies Solve Big Industry Problems
It's quite a feat when a new technology can change an entire manufacturing process, but even more so when it can be coupled with other innovations to solve new and expansive problems. In 2011, NREL and AOS Solar earned an R&D 100 Award from R&D Magazine, recognizing the Optical Cavity Furnace (OCF) as one of the top innovations of the year. The system performs a full array of processing steps including junction formation, annealing, metallization, and oxidation.
When developing the Si-PV wafer screening system, Sopori and his team employed basic optical principles of the OCF to provide a light source for an optical concentrator and harnessed optical excitation in a completely new way to produce very high efficiency.
Because solar wafers have rough surfaces, the team couldn't use conventional optical software to calculate the thermal stresses required for testing silicon wafers in a moving system. Instead, they needed to build a detailed model to demonstrate essential physical processes such as light reflection and absorption as a function of temperature, surface conditions, and wafer thickness. To solve the R&D challenge, researchers derived the needed information using NREL's R&D 100 Award-winning PV Optics software. Using the software in conjunction with thermal modeling, they extended the software's capability to evaluate silicon wafers' emissivity, or heat-transmitting properties.
Deceptively simple in appearance, the system reflects decades of research in optical physics, materials research, and fracture mechanics. It has attracted a licensing partner, AOS Solar—a company that is working with solar cell and wafer manufacturers that have purchased test systems for evaluation. It has also earned support among leaders in the research community and industry.
"We have been selling furnaces to the semiconductor industry for the last 30 years," said Jim Smith, the business development manager of Tystar Corporation, "and we consider the wafer screening tool to be a very low-cost solution to the wafer breakage problem in the solar industry."
During its development, researchers engineered the system to replicate the stresses imposed that can occur during solar cell processing: oxidation, annealing, metallization, diffusion, and wafer handling, all of which can cause existing defects to propagate and lead to wafer breakage. By creating stress equal to what has been measured in cell and module production, the test is much more accurate than methods that depend on inspection or representative sampling.
The system works by applying a non-uniform thermal profile to the wafer as it moves through the illumination zone, producing a dynamic stress distribution across the wafer's surface. If any defects are present in the wafer, they serve as stress concentrators and crack nucleation spots. As the dynamic stress is applied, it tends to intersect and widen the cracks, causing the wafer to break. Because the stress is highest at the defects, or nucleation spots, the rest of the wafer is not subject to enough stress to impose new damage.
The technique is not only accurate, but also efficient. It requires less than 0.002 kilowatt-hours (kWh) of optical energy per wafer and costs less than $0.02/wafer—an expense that has a negligible effect on the levelized cost of energy. Beyond these benefits, one of the most exciting advantages is its flexibility.
"We have designed this system to be integrated easily into existing PV manufacturing processes," Sopori says. "It is compatible with existing conveyor belt technology and can be integrated into the manufacturing lines—and because the parameters are adjustable, it can accommodate manufacturing changes that come in the future."
Learn more about NREL's work with silicon materials.
Learn more about NREL's Spectrum of Clean Energy Innovation and how the laboratory's capabilities emulate the nature of the innovation process.