Invention Could Reduce Flat Panel Display Manufacturing Costs
|By Allan Chen, A_Chen@lbl.gov
October 9, 1996
BERKELEY, CA -- A scientist at the Ernest Orlando Lawrence Berkeley National Laboratory (Berkeley Lab) has developed a new "selective ion source" that can improve the semiconductor manufacturing process and decrease the cost of flat-panel displays.
The inventor of the new process is Ka-Ngo Leung, the leader of the Plasma and Ion Source Technology Group in Berkeley Lab's Accelerator and Fusion Research Division. "Although the new technology can help improve the manufacture of conventional silicon-based semiconductors," Leung says "its primary application will be the manufacture of flat-panel displays.
The winners of the race to bring down the cost and reduce the defect rate of these complicated semiconductor-based display devices will reap a huge economic reward, since FPDs are expected eventually to dominate the market for high-resolution displays. There is a considerable amount of research in the U.S., Japan, Germany and elsewhere aimed at developing the technology for inexpensively manufacturing high-quality FPDs.
"Flat-panel displays consist of many pixels. Each pixel turns on and off individually, and is controlled by transistors," says Leung." The transistors are formed from semiconductors consisting of silicon 'doped' with phosphorus or boron."
Semiconductors are the materials from which most integrated circuits such as microprocessors and dynamic random access memories -- the critical components of desktop computers and many other electronic devices -- are made.
One way to dope a layer of silicon with phosphorus, arsenic, or boron ions is to aim a stream of dopant ions into the silicon, a process called ion implantation.
The manufacturing process must ensure that the dopant is pure and as free as possible from unwanted ions. "Conventional implantation might start with a beam of phosphorus or boron scanning the silicon wafer to dope it uniformly. However, flat panel displays start with a panel of glass that is too big to use the conventional method effectively" says Leung.
"So, a technique called ion shower is used. You employ a large beam of phosphorus or boron. You immerse this big panel of glass in the shower. This implants the whole panel with the dopant. To do this you need a big ion source."
"The ion source generates a stream of dopant ions such as phosphorus by starting with a phosphorus-containing gas like phosphine, PH3. However, phosphine is so toxic that it is always sold heavily diluted, often in a mixture of 95 percent hydrogen gas and five percent phosphine. When you put this gas into a conventional ion source, what do you get?" asks Leung. "Most of the ions coming out are hydrogen."
This creates two problems. One is that the energy of the beam is so high that the silicon-coated glass blank is heated too much by the accelerated hydrogen ions (or other unwanted ions in the mixture) causing damage to the glass surface. A second problem is damage caused by the buildup of excessive electrical charges in the plate from the unwanted ions. The damage can be significant enough to warrant throwing out the plate -- an expense that substantially raises the cost of manufacture.
Leung's solution was to develop an ion source technology that uses just enough energy to selectively ionize only the phosphine (or diborine, B2H6, or other gas), but not enough to ionize the hydrogen gas. This eliminates most of the unwanted impurities from the ion stream because they are electrically neutral, and reduces both the excess heating and the charging problem.
The technology also solves another problem. "Conventional ion sources use a tungsten filament to inject the gas containing the dopant with electrons. However the tungsten is an impurity that is easy to ionize, a source of contamination. So we developed a technology to inject electrons without tungsten that instead uses a principle called radiofrequency (RF) induction discharge. We create plasma using RF induction discharge and then inject electrons into the main chamber to ionize the source gas," he says. "The device is called an RF plasma cathode."
Leung also has another invention which consists of an apparatus for creating a large-diameter ion beam from a small-diameter diverging beam. In conjunction with the selective ion source, this invention can be used to generate a beam appropriate for ion implantation of large areas of material such as those used in flat-panel displays.
Developmental work on the selective ion source is supported by a corporate sponsor under a cooperative research and development agreement with DOE.
The mission of the Plasma and Ion Source Technology Group is to develop plasma and ion sources for nuclear and basic energy sciences, (for example, for spallation neutron research), and medical applications such as proton therapy to treat cancer.
The University of California was awarded patent number 5,517,084 by the Patent Office for Leung's selective ionization invention. Patents are pending for his rf plasma cathode and large diameter ion beam apparatus. The three patents are available for licensing from Berkeley Lab's Technology Transfer Department.
Berkeley Lab is a U.S. Department of Energy national laboratory located in Berkeley, California. It conducts unclassified research and is managed by the University of California.