By Eddy Blokken
The industrial laser is one of the key enabling technologies for cheap and massive production of photovoltaic (PV) systems. It also offers considerable potential for developing new tools for PV system manufacturing.
The precision and flexibility of laser technology for depositing heat energy locally makes it ideal for melting, stimulating diffusion of dopants in a substrate, or separating zones by cutting through thin layers of active or conductive material. Laser technology can also mark individual wafers with unique identifiers for process traceability and control. Some of these potential applications have been more widely accepted by the PV industry than others, however.
One area of laser use that received increasing acceptance for PV is edge isolation. After initial creation of the active layer, a wafer has a short circuit between its front (active) side and back side. To overcome this short circuit, a laser is used to cut a trench several microns deep about 30 to 50 µm from the rim of the wafer. This trench is deep enough to cut through the active layer into the bulk substrate. The advantage here is that it is fast, flexible, and accurate, so that high throughput is possible, and the active surface of the wafer is maximized. Other advantages are that this technique does not put any significant stress on the wafer and that it can be easily integrated in an in-line system with the next step, namely contacting the active front of the wafer.
As such, the process is compliant with several key performance indicators in a PV production facility: it does not present a bottleneck process for throughput; it is compatible with in-line processing (a continuous production process, in contrast to a batch process); and it is compatible with the trend of using thinner and thinner wafers (today 200 µm thick, but future trends predict as low as 140 µm or even 70 µm). Another application of such trench isolation through the use of lasers is thin-film manufacturing. The extreme accuracy allows each layer in a multiple-layer stack to obtain a unique structure through different laser process steps for each layer. Consequently, sufficient accuracy is obtained without the need for expensive lithography steps.
Lasers are also being accepted in PV manufacturing to help improve electrical contacts on the active layer. The current technique (screen printing of a silver paste, and the sintering of this paste into the active layer by a short thermal step in an oven) is the most critical step in the manufacturing process, as it has a huge impact on the yield (percentage of cells that generate electricity) and the efficiency of the cell (defined as the conversion rate of light energy received by the cell, and the electric energy generated by the cell, typically between 15% and 19% for industrial manufacturing). The printing process also induces some stress on the wafer, making it a challenge to print extremely thin wafers.
At the Fraunhofer Institute for Solar Energy Systems (Freiberg, Germany) researchers are developing a process in which a laser beam is used to melt a metal powder on top of a wafer. Extreme mastery of this process is the key to obtaining a good melt of the metal for the right conductivity and low contact resistance with the active layer. It is also the key to avoiding any damage to the crystal structure of the wafer.
Marking controversy
An area of controversy for the PV industry in terms of lasers, however, as evidenced at the first SEMI PV Fab Managers Forum in Leipzig last March, is whether the PV industry should use laser marking to put a unique ID on each processed wafer. Such a practice would enable manufacturers to log each process variance and its effect on the performance of the final solar cell, enabling improvements in yield, as well as providing a tool for dealing with inevitable process variations. This is a common technique in the semiconductor industry and was one of the elements that improved yield significantly and reduced the start-up time for new technology nodes.
In PV, cell manufacturers are less convinced of the need: the number of wafers processed in one hour in a PV production facility is equivalent to the number of wafers processed in one month in a semiconductor fab. In addition, the number of steps to make a chip is an order of magnitude higher then the number of steps needed to make a solar cell.
Many leaders in the PV industry claim that the amount of data generated by following-up on each wafer would be too massive for practical use. Another concern is that the PV industry is not yet able to address each process step through an integrated manufacturing execution system (MES). Reacting to process variances can now only be done by an equipment operator. In the future this situation will change, however.
First, more steps will be added to the process to increase the efficiency of the solar cell. In addition, if production facilities move to 100 MW per production line, automated control of the manufacturing process will be indispensable to achieving targeted cost levels, and traceability will be mandated by the increased complexity. Whether this process will evolve on a wafer-by-wafer basis, or on a lot-by-lot basis, and how any irregularities (individual wafers that leave or enter the lot) will be handled is unclear. It is clear, however, that lasers will ultimately be used to provide the needed traceability.
EDDY BLOKKEN is director of technology and Standards at SEMI Europe, Avenue des Arts 40 Brussels B-1040, Belgium. He is also the global PV segment owner at SEMI; e-mail: [email protected]; www.semi.org.