Chemically amplified resists

Citation metadata

Authors: Angelo A. Lamola, Charles R. Szmanda and James W. Thackeray
Date: Aug. 1991
From: Solid State Technology(Vol. 34, Issue 8)
Publisher: PennWell Publishing Corp.
Document Type: Article
Length: 5,574 words

Main content

Article Preview :

Since the late 1950s most of the photoresists used in integrated circuit production have come from two basic chemical families: crosslinking rubber-based resists and the so-called PAC/novolak resists. The negative imaging resists comprised of bis-azides and cyclized polyisoprene offer high sensitivity and wide process latitude, but processing associated distortion of fine patterns limits the use of these materials to dimensions [is greater than or = to][micro]m. The early 1960s introduction [1] of positive-tone resists containing a diazonaphthoquinone PAC (photoactive component) and a novolak resin (PAC/novolak resist) initiated an era in which advances in microlithography were gated by improvements in exposure systems. Used with state-of-the-art, high numerical aperture, i-line steppers, PAC/novolak resists now achieve better than 0. 5 [micro]m line widths in production lithography. Some device manufacturers plan to use such systems for 64 Mbit DRAM processing [2]. A few years ago, when serious efforts were launched to develop deep ultraviolet (DUV) lithography as the technology for 16 Mbit devices, such longevity for PAC/novolak resist technology was not anticipated. Exactly when DUV lithography will be in production may be clarified in the coming year. Toshiba, for example, has already announced the use of DUV in the fabrication of a prototype 64 Mbit DRAM [3].

DUV single layer resists will employ chemistries quite different from those of the PAC/novolaks. In fact, two features of DUV lithography actually prevent the use of PAC/ novolak systems. First, the nonbleachable absorbance of PAC/novolaks is too high for proper imaging in the desired DUV wavelength region around 250 nm (248.3 nm for excimer laser based exposure systems). Second, the relatively low exposure fluxes delivered by available high resolution DUV exposure systems in acceptable exposure times are such that, to achieve exposure in reasonable time, resist sensitiVities must be < 50 mJ/cm[.sup.2]. This is much faster than any high resolution PAC/novolak resist. Further, while the optical absorbance of such resists is not an issue for e-beam or x-ray exposure systems, the low resist sensitivity still precludes the use of PAC/novolak resists for these exposure modes. In recent years, several resists have been introduced which satisfy the absorbance, speed, and resolution requirements of DUV lithography. Virtually all fall into the "chemically amplified" resist category, a term coined at IBM [4].

Figure 1 shows the basic mechanism of imaging in a chemically amplified resist. Such imaging may be defined as an operation in which exposure produces a latent image that does not immediately generate a concomitant change in dissolution rate, i.e., development immediately after exposure will not produce a relief image in the resist. Image formation first requires a resist activating thermal step (postexposure bake [PEB]). Resist speed is fast, but depends on the PEB temperature and duration as well as on other conditions of development.

One may also define a chemically amplified* resist as a material in which exposure results in the formation of a catalytic photoproduct, the three-dimensional distribution of which defines the latent image. During the PEB, the catalyst activates events that enable dissolution-modifying chemistry to...

Source Citation

Source Citation
Lamola, Angelo A., et al. "Chemically amplified resists." Solid State Technology, vol. 34, no. 8, Aug. 1991, pp. 53+. Accessed 23 Sept. 2021.

Gale Document Number: GALE|A11137024