Lithography: History

Leading the Charge on Moore’s Law (1988–2004)

The Chemical, Microsystem, and Nanoscale Technologies Group evolved from the Submicrometer Technology Group, which was formed in 1988. At that time, the smallest features patterned with optical lithography (at 248 nm) in commercial devices were 0.5 μm in size. By 2013, the smallest half-pitch used in the semiconductor industry was still patterned with optical lithography and was ~20 times smaller, 0.025 μm. Much of this remarkable extension of optical lithography can be traced back to the Submicrometer Technology Group's early work, which helped to sustain the worldwide trend known as "Moore’s Law," which saw the doubling of electronic circuit density every three years.

Lithography patterning of thin films used in the fabrication of semiconductor devices has been the key enabler of the continuous reduction in size of microelectronic circuits for over three decades. Among the various lithography methods, optical projection lithography became the dominant technology in the 1980s. It was well known at the time that one approach to reducing the patterned feature size was a shift to shorter wavelengths of the radiation used in lithographic systems. In the 1980s, the dominant lithography employing the blue spectral region (436 nm) had transitioned first to the near ultraviolet (365 nm) and then to deep ultraviolet (248 nm).

The accepted wisdom in the lithography community was that 248-nm lithography was the end of optical lithography: no further reduction in lithographic wavelength was possible because neither high-quality lens materials nor suitable photoresists were feasible at wavelengths shorter than 248 nm. If the lithographic dimensional shrinking was to continue, so it was argued, radically different technologies would have to be developed.

With these trends as a backdrop, in 1988, the Defense Advanced Research Projects Agency (DARPA) started a project at Lincoln Laboratory to explore the feasibility of optical projection lithography at the even deeper ultraviolet wavelength of 193 nm, which corresponded to another wavelength at which excimer lasers could operate. Compared to the investments at the time in nonoptical lithography, the 193-nm effort at Lincoln Laboratory was quite small. Still, it was significant enough to warrant the establishment of the Submicrometer Technology Group.

Over the next few years, the new group attacked all the issues raised by skeptics of 193-nm lithography. It identified failure mechanisms of amorphous fused silica and crystalline calcium fluoride, the two most promising lens materials, and then it collaborated with optical materials companies to improve the quality and laser damage resistance of the lens materials. The group also explored a wide range of photoresists that would have the right transparency and photosensitivity at 193 nm, and collaborated with other research groups in demonstrating the first 193-nm photoresists. It also demonstrated the feasibility at 193 nm of photomasks and their protective pellicle membranes.

Significantly, the Lincoln Laboratory program also included the construction by Silicon Valley Group Lithography Systems (SVGL) of a prototype large-field 193-nm step-and-scan projection system. When it was completed in 1993, this first-ever 193-nm system was installed in the then recently completed Microelectronics Laboratory at Lincoln Laboratory, and it was subsequently used to develop photoresists and related microfabrication processes, to explore the resolution limits of 193-nm lithography, and to fabricate microelectronic devices.

By 1994, the progress made under the Lincoln Laboratory program in 193-nm lithography had convinced the rest of the lithography community to examine this technology more closely. The Semiconductor Manufacturing Technology (SEMATECH) consortium started a series of workshops on 193-nm lithography, and most companies supplying projection systems, photomasks, and photoresists expanded their respective internal development efforts. From 1993 on, several of these companies also entered into Cooperative Research and Development Agreements (CRDA) with the Submicrometer Technology Group in order to transfer and expand the technology and know-how developed in this field by the Lincoln Laboratory team. Multiyear CRDAs in optical lithography were established with SEMATECH, Intel, IBM, Shipley, SVGL, DuPont, KLA-Tencor, and others. Today 193-nm lithography is the mainstream lithography used by the semiconductor industry. Commercial advanced microelectronic circuits have been fabricated using 193-nm lithography since at least as early as 2002. Thus, despite widespread initial doubts, the pioneering work at Lincoln Laboratory seeded a whole new era in lithography.

As soon as the paradigm change to 193 nm was accepted, the question arose as to how to extend optical lithography even further to enable patterning at even smaller dimensions. Again, the Lincoln Laboratory group had recently proven feasible an option that did not require abandoning the 193-nm wavelength, yet enabled higher resolution. This option was liquid-immersion lithography.

Liquid-immersion lithography is a method whereby a transparent liquid is introduced between the last optical element of the projection system and the photoresist surface. By adding this liquid, the effective wavelength at the photoresist surface is reduced by the refractive index of the liquid without changing the wavelength throughout the rest of the lithographic system. Although the underlying concept of improving lithographic resolution by liquid immersion had been known for at least two decades, it was once again the Lincoln Laboratory group that, in 2001, showed liquid immersion at 193 nm was a practical option. Liquid-immersion 193-nm lithography has been used in small-volume manufacturing since 2007 and is now the dominant lithography for large-scale device production.

 

top of page