Technological advances have led to the expansion of the installations of solar photovoltaic (PV) systems as well as the lowering of the prices of solar electricity in recent years. With solar energy becoming a highly efficient source of power generation worldwide, corporate and academic research organizations are also becoming more interested in applying various fields of science and technology to raising the quality of PV products. The latest example of this is a robot developed by Norway’s private research organization SINTEF for quality assurance during the production of monocrystalline silicon (mono-Si) ingots.
SINTEF has long been studying the optimization of the manufacturing of mono-Si end products, and the organization is currently focusing on quartz crucibles, which make up a vital group of consumable items in the entire production process. These containers are designed to withstand the heat and weight of the melted silicon material that they hold within the furnace. They have to remain intact during the crystal formation stage, during which a mono-Si ingot emerges out of the molten silicon in the crucible. Afterwards, the crucibles have to become sufficiently brittle in the cool-down stage so that they can be easily separated from the ingots. Structurally, a crucible is made of different layers of quartz materials that serve specific functions at different stages of the ingot production. Defects within these layers can lead to ingots falling out of their containers or ingots tainted with impurities related to the quartz materials.
John Alte Bones, a PV cell researcher at SINTEF, said that the efficiency rate of a PV cell correlates to the quality of the raw materials. Furthermore, reliable quartz crucibles ensure a good yield of mono-Si ingots and thus reduce the environmental impact of the production process. While quartz crucibles play a crucial role in the quality of PV cells, there have been no significant improvements in the quality assurance systems for the crucibles. The human eyes are still mainly responsible for checking these containers for defects.
Researchers at SINTEF therefore have collaborated with students from the Norwegian University of Science and Technology (NTNU) to build a robot that would replace human eyes in performing in-depth inspection on quartz crucibles. Because a crucible is composed of different quartz material layers featuring their own reflectivity and transparency values, the robot uses optical instruments and a machine vision technology to identify different types of faults and defects. Furthermore, the special sensor package carried by the robot allows a “super sight” capability that specifically deals with this multi-layer challenge.
Bones noted that the SINTEF researchers employed “destructive methods” for determining the relations between the quality of crucibles and the quality of mono-Si end products. According to him, researchers had crushed and dissolved crucible materials in chemicals to analyze the properties of different quartz material layers.
At present, the SINTEF-NTNU team has already successfully deployed their robot in trials. The robot can quickly and accurately scan crucibles and prevent those with defects from entering the furnace.
To create a sensor module that can do everything during the inspection was an impossible hurdle for the researchers, so they instead devised a package of sensors that are linked together and can communicate with each other. Bones, however, pointed out that the confocal white light sensor is one of the most important sensors in the package because it responds to different colors and their corresponding wavelengths within the white light spectrum.
Additionally, the robot is equipped with a high-resolution digital CCD camera that can focus on miniscule details of a product. The camera is connected to a machine vision system to further allow the robot to recognize material variations and defects. Bones said that the robot also examines the outward appearance of a quartz crucible, ensuring it meets the curvature and thickness requirements. Therefore, the robot uses numerous distance sensors to measure the dimensions of a product. At the same time, the robot does self-correcting calculations to make sure that it is always in the right position to do these measurements as it moves around on a track.
(The above article is an English translation of a Chinese article written by Daisy Chuang. The credit of the photo at the top of the article goes to Jose Mesa via Flickr and falls under the license of CC BY-SA 2.0.)