Advertiser and

Buyer's Guide
Buyers Guide

Chip Shots blog

Greatest Hits of 2005
Greatest Hits of 2005

Featured Series
Featured Series

Web Sightings

Media Kit

Comments? Suggestions? Send us your feedback.


Leaching, line-edge roughness, topcoats among key challenges facing immersion photoresist development


Immersion lithography has emerged as the industry's imaging technology of choice for staying on the 45- and 32-nm process-node roadmaps. Multiple research programs in universities on three continents as well as projects at TSMC, IBM, IMEC, Sematech, and the equipment and materials companies have propelled the wet lithography proposition forward as aggressively as any technological solution the industry has ever seen. One key to the ultimate success of immersion is the development of robust, high-performance photoresists.

While the early prospects for immersion resists look quite good, many issues remain unresolved. The leaching of resist components during their first contact with water, the refinement of protective topcoats, and line-edge roughness and other defect sources are among the concerns facing resist developers and users. This issue's Hot Button panel of experts shed some light on many of the challenges surrounding the development and implementation of advanced photoresists for 193-nm ArF immersion lithography.

WILL CONLEY (advanced lithography group, ARDL, Dan Noble Center, Freescale Semiconductor): Argon-fluoride (ArF) liquid immersion lithography (IML) has the potential to extend optical lithography to meet the needs of the 32-nm semiconductor device nodes and beyond. Water is the immersion fluid of primary interest because of its low intrinsic absorption at 193 nm. Research to date has established the feasibility of the technology, with no critical showstopper technical barriers.

The next phases of R&D will require the availability of prototype exposure tools operating at the 193-nm wavelength of an ArF excimer laser. Such tools will allow for the initial exploration into such issues as photoresist and materials interaction. Early efforts are attempting to understand the effects of resist components leaching, water absorption, and general image quality of current resist formulations. Other important issues under investigation include polarization, high-numerical-aperture (NA) effects, and general imaging requirements. The use of interferometric imaging has been a useful tool for understanding materials and optics issues.

The leaching of resist components could contaminate and even damage the lens element that comes in contact with the immersion fluid.
—Will Conley

Although the early prospects look very good, many IML-related resist issues are yet to be resolved. One issue is the leaching of the resist components during the initial contact with water. Researchers are developing analytical techniques from simple water extraction to complex micro-pH measurements. Water-soluble photoproducts such as acids can easily leach out of the surface of the film. Other resist additives, such as quenchers or contrast-enhancement materials, can also leach out of the film.

Resist researchers are paying particular attention to these issues because of their impact on the imaging system. The leaching of resist components could contaminate and even damage the lens element that comes in contact with the immersion fluid. Recent work focused on the measurement of "leached" products indicates that the photoacid generator and base quencher can be engineered to minimize or possibly even eliminate leaching.

Several resist manufacturers have introduced base-developable topcoats, which act as a protective layer during the immersion imaging process. The topcoat material is spin-applied on the resist layer and soft baked. The protective overcoat and resist do not intermix and the protective coat is insoluble in water. Once the wafer is exposed and postexposure baked, the protective coat is removed during the develop process and is designed not to affect the resist.

Defects remain the first and foremost IML issue. A number of full-field exposure tools are being built, and several have already been delivered. As the first full-field systems are installed, a deeper understanding of the impact on device yields will develop. Until then, many aspects can be addressed without a full-field system. Over the past year, the first >0.8-NA ArF dry systems have become available for sub-90-nm process development and manufacturing. These processes will be used for comparison of yield-contributing issues as well as for evaluation of the technology's general usefulness.

The outlook for immersion lithography is extremely promising. The technology is attractive because much of the infrastructure already exists, and the existing materials and tools have great potential for extendability. The progression of tool and material technology continues at a frantic pace. Tool manufacturers are designing >1.0 NA systems for availability in the 2006 time frame. Many in the industry are already proposing further extensions of ArF immersion with higher NAs (>1.2) and new higher-index immersion fluids and resists for the further extension of optical lithography.

KURT RONSE (director, lithography department, IMEC): The removal of 157-nm (F2) lithography from most of the industry's roadmaps and its replacement by 193-nm (ArF) immersion lithography has certainly caused relief among the photoresist suppliers, who did not know how to meet the requirements put forward in the International Technology Roadmap for Semiconductors (ITRS). Furthermore, the resist companies knew even less about whether they would ever get any returns on their 157-nm investments.

Nevertheless, consensus is growing among the industry that chemically amplified resists, which have become the mainstream products since the introduction of 248-nm (krypton fluoride, or KrF) lithography, start to face some serious challenges in further scaling down the dimensions toward the 45-nm technology node and beyond. Chemically amplified resists were introduced to meet the required sensitivity that ensured optical lithography was still cost-effective when 365-nm (i-line) lithography reached its limits. Especially for extreme ultraviolet (EUV) lithography (13.5-nm wavelength), where the source power may go down to protect the lifetime of the optics, it is important that this high sensitivity (5 mJ/cm2) is kept , although ArF lithography cannot afford a significant reduction of sensitivity either.

All chemically amplified resists work on the following principle: acid, generated by the exposure reaction, initiates a catalytic reaction during the subsequent bake step. During that period, the high to low concentrations of acid can diffuse through the resist structure and, as such, reduce the contrast obtained during the exposure step. Typical diffusion lengths are on the order of 40 nm or higher and do not cause too many worries for 90-nm technologies and above. However, it is quite clear that such diffusion lengths can no longer be tolerated for 45-nm and smaller technologies without complete loss of contrast.

A second issue is line-edge or line- width roughness (LER or LWR), to which variable sources contribute, such as resist polymer structure, image contrast, and diffusion length. The trends in LER reduction do not keep up with the required LER numbers listed in the ITRS guidelines. On the other hand, for 90-nm technologies, the electrical performance of the devices does not seem to suffer too much from LER. It is expected that this will change for the 45-nm and smaller technologies. Moreover, reducing the diffusion length and LER are typically conflicting requirements, just like increasing the sensitivity and reducing LER are generally competing trends.

In conclusion, even though lithographers can stick with ArF resist chemistry for a while, now that ArF immersion is steadily taking off, and EUV lithography is supposed to be able to use traditional KrF resists as a starting point, resist development in both cases faces some major challenges. Creative solutions are required to overcome those challenges, which are more wavelength related than technology node related and thus quite generic in nature.

MARK SLEZAK (technology manager, lithography group, JSR Microelectronics): As 65-nm-node technologies transition from R&D groups to early production, we are forced to make tough decisions regarding our choices for the next technology nodes. The 45-nm node is quickly approaching and our viable options are narrowing. It was thought that technologies such as 157 nm and EUV lithography would be ready; however, as we get closer to making capital investment decisions, it is clear that neither of these technologies will be ready in time. Immersion lithography is here, and it is time to get our feet wet.

Along with high-NA dry lithography, IML systems are becoming more realistic as we set our sights on solutions for the 45-nm node. However, there are still many questions. One prevailing question we hear as a resist supplier involves topcoats. Many customers ask if topcoats are necessary to move this technology from an R&D state to a production state. The fears associated with not using a topcoat are not only resist extractables damaging the lens, but also the degradation of the aerial image caused by changing the optical properties of the water. By omitting a topcoat, the lens protection resides solely with the water handling, which worries many tool vendors and end-users.

Initial results indicate that the use of a topcoat greatly reduces the amount of low-molecular-weight components, such as photoacid generators or quenchers, which can leave the resist and enter the water. Along with protecting the imaging system's lens, other positives associated with using a topcoat for immersion are the ability to tune surface tension characteristics between the water and the wafer (contact angle), postexposure delay protection from airborne contamination, and facilitation of defect-mitigation strategies.

Other questions associated with immersion lithography deal with the type of fluid used between the optical lens and wafer. Early work on first-generation immersion systems using water as the fluid, with an index of refraction of 1.44, have shown depth of focus (DOF) improvements from 1.5 to 2 times better than equivalent-NA dry systems. With this same premise in mind, higher-index fluids will not only improve the DOF relative to dry systems, but also allow for significant resolution enhancements. Early work at our company has shown imaging down to 32-nm half-pitch when using a high-refractive-index fluid—the first imaging below 35-nm half-pitch.

As we get our feet wet with immersion, it seems that the positives of topcoat technology outweigh the process complexity, and the introduction of new high-index fluid materials should allow the industry to use ArF immersion below the 45-nm node.

ROB CROWELL (product marketing manager, Clean Track business unit, Tokyo Electron [TEL]): It is now certain that 193-nm immersion lithography is the path to the 65- and 45-nm process nodes. Inevitably, resist processes will reach unprecedented complexity in order to fully realize the potential of this emerging technology. Some industry experts predict that this technology is extendable to the 32-nm node.

All initial resist processes for IML include some form of topcoating that serves as a protective barrier between the immersion fluid and the imaging resist film. The topcoat material has two primary functions: one, to prevent resist components from leaching out into the immersion fluid that could degrade pattern resolution; and two, to help protect the bottom scanner lens element from potentially detrimental resist effluents. Particular attention must be focused on ensuring that the topcoat process does not adversely affect the underlying imaging layer. Additionally, removal of the topcoat prior to development must be complete to ensure minimal impact to the final resist image. The inclusion of topcoat application and removal adds process modules to the lithography flow, which affects overall cluster throughput and cost of ownership.

Another challenge facing lithographers is line-edge roughness. Although it has been a component of past technology nodes, LER accounts for an increasingly larger portion of the overall CD budget at the 65- and 45-nm technology nodes. To compensate for measurement errors caused by LER, lithographers must explore increased sample rates or alternate metrology, including optical digital profilometry and scatterometry. Since there isn't a single factor that appears to cause LER, incremental relief is likely to come from several sources within the process and materials or perhaps from a postimaging treatment.

Not only will lithographers continue to whittle away at defect densities at increasingly smaller sizes, but they will also need to attack novel categories of anomalies. Defects are likely to come from new sources, such as processes associated with the application and removal of topcoats, interactions between immersion fluids and wafer surfaces, and the formation of bubbles. How water bubbles affect resolution locally and form defects remain to be explored in the early evaluation phases. Complex process and material interactions could result in unprecedented defect types and levels.

MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit

Questions/comments about MICRO Magazine? E-mail us at

© 2007 Tom Cheyney
All rights reserved.