Dow Corning uses a fully automated laser system to produce silicon solar cells for photovoltaic applications. Housed in a remarkably small space, the pilot-scale research plant carries out processes that would normally require a multi-stage laser station.
Anyone interested in seeing how tomorrow' solar cells might be produced should head to the Belgian municipality of Seneffe, where the multinational specialty chemicals company Dow Corning is busy conducting research into crystalline silicon solar cell manufacturing processes for the photovoltaic industry. A key tool for its research is a laser workstation that combines all the relevant production steps within a single highly compact system. It removes functional layers from the wafer selectively and precisely, drills thousands of holes through the silicon, machines the edges, and then marks the wafer with a Data Matrix code in a final step.
The various process steps would normally require multiple manufacturing stages with different lasers - but Dow Corning had other ideas, as Guy Beaucarne, head of the Solar Cell Department, explains: "Modern solar cell manufacturing lines make extensive use of laser processes. For our R & D activities, we wanted a compact laser workstation that not only could execute some laser processes that have become common in the industry, but also could provide a wide research flexibility, enabling us to apply more advanced laser processes required for emerging solar cell technologies." With this concept in mind, the project manager decided to call up TRUMPF.
Project Engineer Jörg Smolenski recalls what happened next: "We were fascinated and fairly sure we could come up with a solution for the laser components. But we needed a systems partner capable of designing the complex automation package that the lasers would require."
To ensure the system would be able to handle all the processes involved in the laser manufacturing of solar cells as well as have additional built-in flexibility for upcoming R & D projects, Dow Corning asked for three wavelengths emitted by non more than two beam sources: 1,030 nanometers (infrared), 515 nanometers (green) and 343 nanometers (ultraviolet).
In addition, it was essential for Guy Beaucarne to have the option of carrying out additional modulation of the light and working with pulse lengths in the nano and picosecond range. That's the only way in which the beams can ablate the various substrates while still supplying enough energy to cut the material. A key goal of Dow Corning's design concept was to maintain flexibility with regard to the sequence of the process steps.
"It wouldn't have made sense to choose anything else"
The TRUMPF application engineers proposed using two beam sources, one of which would be a marking laser: "The TruMark lasers are reliable, compact systems, so it wouldn't have made sense to choose anything else," says Smolenski. With its green 532 nanometer light and a pulse repetition frequency of between 25 and 100 kilohertz, the TruMark 6230 proved to be the perfect tool for marking silicon.
But the main workhorse in this miniature fab is a TruMicro Series 5000 ultrashort pulse laser. Depending on the application, it delivers pulses as short as 10 picoseconds, pulse energies up to 250 microjoules and an average laser output of up to 100 watts. Parameters, which would also allow for a conversion of its light into the green and ultraviolet spectral ranges. TRUMPF used it as a basis for creating a triple frequency solution.
A special "box" developed by the company Xiton Photonics, based in Kaiserslautern in Germany, in collaboration with TRUMPF plays the role of the frequency conversion module. It either lets the infrared light through or converts the laser pulses into either green or ultraviolet laser light as required, each of which is sent through its own individual scanner optics.
High-precision system engineering
To control the box, scanner and overall system, IPTE developed a special automation solution: "The biggest challenge for us was the level of precision required in positioning the wafers under the lasers," says Kris Smeers. In some cases, the all-in-one system has to position the wafers with an accuracy surpassing 10 microns while still being able to handle a broad range of different formats.
Kris Smeers and his team decided to implement a high-precision image capture system: "As long as the machine can see what workpiece is coming and how it is positioned, then the exact shape doesn't matter," says Smeers, explaining how they reached their decision.
The camera used for image acquisition offers a resolution of 12 x 12 megapixels and the positioning drives utilize fully adjustable motors and high-precision encoders. To ensure that no external factors interfere with this precision work, IPTE limits human interaction to the programming of the process steps through the user-friendly interface - plus of course the constant provisioning of fresh wafers to the machines. This is one area where the miniature fab clearly differs from its larger counterparts: In bigger machines, IPTE also automates the process of loading the silicon wafers.
The solar industry's future tool
The system has been up and running in the research center since fall 2012. In response to the question as to why it specifically needed to be a laser system, Guy Beaucarne explains that he sees the laser as one essential tool of the future for the solar industry: "Lasers are very flexible, fast and reliable. They are easy to control and they ensure a reproducible process. At the same time, they can offer a smaller footprint and lowertotal cost of ownership than alternative patterning and machining methods," he says. "Many in the solar industry see the laser as a necessary tool for advanced solar cells".
This applies not only to the miniature fab that Dow Corning is using for its research, but also to these same processes on an industrial scale: "IPTE and TRUMPF have managed to create an extremely flexible tool that we can use for multiple different process steps," says Guy Beaucarne, emphasizing Dow Corning's satisfaction with the miniature fab.