Multiple electron beam maskless lithography for high-volume manufacturing

Though electron beam lithography has been long used for mask writing, it is yet very slow and typically takes almost a day to complete a high-end mask. Of course direct writing on a 300-mm wafer with around 100 fields would take much longer. To make it production-worthy, its throughput should achieve at least >10 wafers per hour (WPH) from a single chamber or >100 WPH by a cluster within a scanner footprint while maintaining high resolution. A more than 3-order of improvement in throughput is required. Increasing the beam current in the conventional single beam system is apparently not enough. Using the maturing MEMS technology and electronic control technology to make more than ten thousand electron beamlets write in parallel is the most possible and straightforward solution. Several groups have proposed different multiple electron beam maskless lithography (MEBML2) approaches, with either Gaussian beams, variable shape beams or cell projections, to increase the throughput. In this paper, the challenges and possible solutions to achieve production-worthy throughput of MEBML2 for the 32-nm HP node and beyond are discussed.