Introduction to Chemically Amplified Photoresists
In the transition to 248 nm lithography, the reduced intensity of a mercury lamp at 248 nm (see Figure 1) forced research to look at means of improving the sensitivity of photoresist materials.
Figure 1: Lamp Intensity spectrum for Mercury exposure tool. Mercury lamps are the UV light source for exposure tools having wavelengths greater than 193 nm.
From this research, it found that using a positive chemical amplification scheme (shown in Figure 2) dramatically improved the photosensitivity of these materials. The beauty of chemical amplification is that one photon can trigger the deprotection of hundreds (if not thousands) of acid-catalyzed deprotection reactions, rather than one proton to decompose one photoactive compound molecule system used for DNQ/Novolak systems.
Figure 2: Mechamism of Chemically Amplified Photoresists. In this diagram, a 248 nm photoresist film containing polymer (PBOCST or T-BoC protected polyhydroxystrene) and photoacid generator (Triphenylsulfonium aniomate) are present.
A chemically amplified resist (CAR) contains four components: a polymer resins that provides most of the properties of the photoresist film, a photoacid generator to provide sensitivity to ultraviolet light, and a dissolution inhibitor to provide a solubility switch before and after exposure. As mentioned earlier, dissolution inhibitors are oligomers of acid-labile protected monomers, or acid-labile protecting groups that are attached to the polymer resin. In the unexposed state, the acid-labile protecting group completely inhibits the dissolution rate of the photoresist by replacing the base-soluble hydroxide with an insoluble group. After exposure to ultraviolet light, the photoacid generator decomposes, generating small amounts of generated photoacid within the photoresist film. This generated acid attacks the acid-labile protecting group (usually an ester or anhydride), causing an acid-catalyzed deprotection reaction.
As with any other change in lithographic technology, the old resist and optical materials exhibit lower transparencies at the new exposure wavelength. The DNQ-Novolak resins exhibited high absorbance at 248 nm (see Figure 3), so new resist materials had to be designed for KrF lithogrpahy.
Figure 3: Absorbance spectra of PHOST and Novolac.
It was found that slight alterations to the novolak (removal of benzene methyl) monomer dramatically increased the optical transparency of these materials. Around 1995, the DNQ-Novolak system was replaced by PBOCST (TBoC protected PHOST, see Figure 4) as the primary resist technology for high-end lithographic applications.
Figure 4: TBOSCT, a DUV photoresist platform.
Polyhydroxystyrene (PHOST) was an analog to novolak that was found to have similar etch resistance, film forming properties, but a lower absorbance at the deep UV wavelengths than novolak (see Figure 3).1 However, the dissolution rate of these materials cannot be inhibited by resist additives, due to the intensive intermolecular hydrogen bonding found within these materials. , , By capping the hydroxyl group in poly(hydroxystyrene) with a tBoc protecting group, this resist platform became the first commercial chemically amplified resist.
One can protect the photoresist resin (the synthesis of PBOCST, tBOC protected PHOST is shown in Figure 5) through a simple reaction with an alkyl halide. One can easily manipulate the added protecting group by simply modifying this alkyl halide.
Figure 5: Synthesis of PBOCST. First, the oxygen is replaced with a bromide using NaBH4. Afterwards, a grinyard is synthesized to increase the monomer reactivity. Finally, the protecting group is added through a halide reaction.
Unfortunately, the first images of 248 nm photoresists exhibited dramatic t-topping (shown in Figure 6). Instead of linear sidewall profiles, t-topping is the formation of a large insoluble head to the feature, then an increase in CD size at deeper depths. It was found that base contamination in the room atmosphere neutralized the formed photoacid molecules during the exposure of the resist to ultraviolet light. After neutralization of these acid molecules, the top region of the photoresist is less soluble in aqueous base developers, reducing the dissolution rate at the top of these features. As one goes deeper into the resist profile, the amount of base decreases, and the expected feature profile begins to appear. T-Topping can be controlled by two mechanisms: base filtration of room atmosphere, and loading of base quenchers into the resist film. By eliminating the base in the room atmosphere, there are fewer base molecules available to neutralize the formed photoacid molecules. Base quenchers also eliminate t-topping by reducing the potential changes in the concentration of base within the resist film, because there is considerable base already present within the material.
Figure 6: T-Topping in 193 nm photoresist features after a 30 minute delay. This phenomenon is caused by base poisoning of the photoresist by the ambient atmosphere.
After nearly a decade of research, 193 nm lithography was fully implemented in semiconductor fabs for the 130 nm node. It was found that the double bonds strongly absorb at 193 nm, eliminating PHOST as a potential resin for 193 nm resist technology. Two dominant resist technologies have been developed for 193 nm lithography (see Figure 7): Acrylate based polymer and poly(norbornene)-co-malaic anhydride (COMA).
Both of these systems have shown the necessary transparency (~0.2 1/um) at 193 nm, solubility in aqueous base developers, and relative etch resistance to plasma chemistries to be used as effective photoresists. Initially, the acrylate resists exhibited a higher resolution with a lower etch resistance. Work with the carbon-rich acid-labile protecting groups has dramatically improved the etch resistance of acrylate based photoresists (such as adamantyl or lactone groups, see Figure 8). Currently, these materials have a relative etch resistance of 1.2-1.3 times the etch resistance of 248 nm resist materials, which is a significant improvement over PMMA.
Figure 8: Acrylate monomer with lactone protecting group.
COMA resists typically exhibit problems with shelf life due to the hydrolysis of maleic anhydride groups upon exposure to water. Though this problem can be reduced with a nitrogen blanket, they still exhibit much shorter shelf lives than acrylate based resists.
References:
- Thompson, L. F.; Willson, C. G.; Bowden, M. J. Introduction to microlithography, 2nd ed.; American Chemical Society: Washington, DC, 1994.
- Hanrahan, M. J.; Hollis, K. S. Proceedings of SPIE-The International Society for Optical Engineering 1987, 771, 128-135.
- Moskala, E. J.; Howe, S. E.; Painter, P. C.; Coleman, M. M. Macromolecules 1984, 17, 1671-1678.
- Coleman, M. M.; Lichkus, A. M.; Painter, P. C. Macromolecules 1989, 22, 586-595.
- Conlon, D. A.; Crivello, J. V.; Lee, J. L.; O'Brien, M. J. Macromolecules 1989, 22, 509-516.
- Kinkead, D. A.; Ercken, M. Proceedings of SPIE-The International Society for Optical Engineering 2000, 3999, 750-758.
- Przybilla, K.; Kinoshita, Y.; Kudo, T.; Masuda, S.; Okazaki, H.; Padmanaban, M.; Pawlowski, G.; Roeschert, H.; Spiess, W.; Suehiro, N. Proceedings of SPIE-The International Society for Optical Engineering 1993, 1925, 76-91.
- Yamana, M.; Hirano, M.; Nagahara, S.; Kasama, K.; Hada, H.; Miyairi, M.; Kohno, S.; Iwai, T. Proceedings of SPIE-The International Society for Optical Engineering 2003, 5039, 752-760.
- Allen, R. D.; Sooriyakumaran, R.; Opitz, J.; Wallraff, G. M.; Dipietro, R. A.; Breyta, G.; Hofer, D. C.; Kunz, R. R.; Jayaraman, S.; et al. Proceedings of SPIE-The International Society for Optical Engineering 1996, 2724, 334-343.
- Ronse, K. Microelectronic Engineering 2003, 67-68, 300-305.
