This paper reviews the current status of laser fuse processing and discusses the challenges ahead. Various processing parameters are discussed including laser wavelengths, laser pulse duration and shaping, laser beam polarization, laser beam positioning accuracy, and thermal effect, and their impacts on fuse size limitation. We also examine the interrelationships among these parameters. Experiments and evaluations have been conducted, including the use of new laser sources and processing techniques, improved optical system, and beam positioning system. The results, together with several potential solutions for the next generation of laser memory repair system, are presented. Introduction Laser fuse processing has been widely used in semiconductor memory repair. DRAMs, and memories in general, are quite susceptible to process defects. Redundant elements are switched in to replace the defected elements by means of laser blown fuses, thus, enhancing the effective yield. Memory technology has progressed at a rapid pace since the invention of the one-transistor/one-capacitor DRAM cell in the late 1960s. In recent years, there has been a shift from a technology generation strategy (4 Mb/0.7 µm, 16 Mb/0.5 µm, etc.) to a shrink strategy (64 Mb/0.35 µm/0.25 µm/0.2 µm, etc.) with shorter development cycles. The fuse pitch for 0.14-µm node technology is 2.8 µm or less depending on the materials used. As the industry moves to 0.09-µm node technology and beyond, smaller fuse dimensions are expected. The fuse pitch will be less than what the current laser systems are capable of, which is 2 microns or less. This puts significant challenges to the equipment manufacturers. The capability of laser processing of finer pitch fuses has become the bottleneck to further miniaturization of redundant fuses and one of the major considerations in memory fuse design. We will first review the current status of laser processing of memory links. Solutions for effective processing of aluminum and copper links are developed and presented here. Finally, future direction for fine pitch memory link processing is discussed
[1]
Joseph B. Bernstein,et al.
Analysis of laser metal-cut energy process window
,
2000
.
[2]
J Cordingley,et al.
Application of a binary diffractive optic for beam shaping in semiconductor processing by lasers.
,
1993,
Applied optics.
[3]
R. G. Nelson,et al.
Laser programmable redundancy and yield improvement in a 64K DRAM
,
1981
.
[4]
Edward J. Swenson,et al.
Optimization of memory redundancy laser link processing
,
1995,
Advanced Lithography.