Issue



Developments in materials for 157nm photoresists


10/01/2001







International Sematech Universities 157nm Photoresist Research Project

overview
Within the past year, the extension of optical lithography to 157nm has received widespread support because it offers the prospect of improved resolution based on decades of optical lithography experience. This article provides an update on the progress in 157nm, from photoresists to pellicles and exposure tools.


Figure 1. A modular approach in the design of new photoresist materials provides flexibility in design, but transparency among all subunits must be maintained.
Click here to enlarge image

Existing deep UV lithographic technologies and associated optical extensions combined with early 193nm-based systems are expected to provide sufficient resolution to enable the 130nm device cycle outlined on the 1999 SIA Roadmap [1]. Recently, developments in higher numerical aperture systems (>0.8) have been discussed to support 100nm imaging at 193nm. There is no mature lithographic technology, however, to support the manufacturing of the device generations at 70nm.

Several different "NGL" technologies are vying to be the successor to optical lithography below the 130nm technology node. Currently, extreme ultraviolet (EUV) lithography has the most support. Though a transition to a next-generation lithographic technique will eventually be necessary, each poses an entirely new set of engineering and infrastructure challenges. Extension of optical lithography to 157nm has therefore received widespread support within the past year as it offers the prospect of improved resolution based on decades of optical lithography experience.

According to the 1999 SIA Roadmap, early pilot line fabrication for the 100nm node should begin during 1Q02, and volume production capability should be in place by 1H05. Considering that the development cycle for DUV photoresist materials spanned at least 10 years (1980-1990) and the 193nm resist development program approximately eight (1991-1999), 157nm photoresist development efforts are faced with an extremely abbreviated and challenging development schedule [2].

Associated areas report progress
The development of 157nm lithography has been challenging due to the high energy of the photons (7.9eV). Most common lithographic materials strongly absorb at this wavelength [3]. This has required research and development of all materials that come in contact with the beam line in an exposure tool — the mirrors, mask material, pellicles, optics, coatings, and resist. The environment of the tool has to be modified; the beam line must be nitrogen purged to reduce the amount of oxygen and water to the parts per million level since both strongly absorb at 157nm [4].

Research in critical areas other than photoresists is beginning to show significant progress. The lenses and mask substrates used in previous generations of photolithography are primarily comprised of fused silica due to its extremely low absorbance throughout the ultraviolet spectral region [5]. Unfortunately, the absorbance of fused silica increases dramatically at wavelengths below 180nm due to traces of OH in the bulk material. Recently, several companies have reported significant developments in fluorinated fused silica and have fabricated mask substrates that appear to have the requisite properties for use at 157nm. The transmission of the modified fused silica material is adequate (~85% for a 6mm-thick sample with target specification of 85%) and the lifetime appears to be stable up to ~300 million pulses [3].

Calcium fluoride (CaF2) will be the material to replace fused silica for the lenses, and the quality of CaF2 samples now available for optical elements has been improving at a rapid rate. 193nm production tools contain some CaF2 elements and, with increases in numerical aperture and further extension of this wavelength, more CaF2 will be needed.

This is the material of choice for transmissive optical elements. MIT Lincoln Labs has measured the initial transmission and lifetimes of samples from different suppliers [2]. The lifetime targets for initial and final bulk absorbance are not to exceed 0.002 cm-1 (base 10); the total lifetime dose target 1011 pulses at 0.5mJ/cm2/pulse. The majority of the samples measured have already met or are close to meeting target bulk absorption values. No change in sample transmission was observed for laser irradiation up to 250 x 106 pulses at fluences up to 4mJ/cm2/pulse, corresponding to a total dose of 1mJ/cm2, which exceeds the target [5].

While material quality looks promising, the supply of CaF2 is limited due to low yields in production. In several recent industry meetings, CaF2 suppliers have assured the semiconductor industry that there are furnaces available, so after yields are improved, the supply should be less problematic [6].


Figure 2. The SEM demonstrates 110nm 1:1 features in 68nm-thick XP-2332C photoresist material used as a baseline for 157nm exposures.
Click here to enlarge image

Industry goals for the pellicle material are initial transmission of 99%, an end-of-life transmission of 90%, and a lifetime dose of 7.5 kJoule with mechanical characteristics of good adhesion, strength (air jet test), thickness uniformity, and low surface roughness (scatter). Meeting these requirements is still such a big challenge that some companies are considering solutions without pellicles. Work in progress is investigating thermophoresis protection for pellicle-less masks. Still the development of new organic pellicle materials offering sufficient transparency at 157nm is underway, some of them offering sufficient mechanical properties with absorbance values <0.001µm-1 at 157nm. Some progress has been reported on polymer pellicles by DuPont, Asahi, and Shin-Etsu. Many of the polymers meet the transmission requirements, but have a serious durability issue.

At least two companies are working on hard pellicle solutions: Shin-Etsu and Asahi. Having been fabricated and tested, the hard pellicles meet transmission and durability requirements, but transmission uniformity needs improvement. The goal is 0.2%, whereas ~1.5% has been observed. The tool suppliers have indicated that they can accommodate hard surface pellicles. Several programs are in place to address tool considerations.

An infrastructure for developmental studies is also coming into place. At the end of 1Q00, a 157nm microstepper built by Exitech Ltd. and based on a 0.6 NA, catadioptric, Tropel lens was installed and available for use at International Sematech. Offering a relatively small 1.5 x 1.5 mm2 field size, the microstepper is designed for early material testing and photoresist evaluations. It also allows for testing of important process-related issues such as nitrogen purging of the beam path and wafer plane, a technique that is required to prevent parasitic absorbance from airborne contaminants, such as water vapor, that absorb strongly at 157nm [4].

Single layer photoresists
A single layer photoresist process for imaging at 157nm is the clear industry favorite and important details are beginning to emerge to guide their successful design. Fundamentally, the design of materials for 157nm photoresists is identical to that used in the previous lithographic generations [2]. The materials must adhere to the set of requirements common to all chemically-amplified photoresist materials. For use in a single layer photoresist, potential materials must meet three fundamental requirements: 1) high transparency at the exposure wavelength, 2) sufficient resistance to plasma etching processes, and 3) functional units that are capable of undergoing efficient photochemical transformations that change their solubility in developer solutions. For positive-tone photoresists, this is typically accomplished by designing the matrix resin with an acidic functionality that is masked by an acid labile protecting group. Strong acids generated within the exposed regions of the photoresist film catalyze the thermolysis of the protecting groups, rendering these regions soluble in the aqueous base developer solutions [7]. The new photoresist materials must also have accessibility through reasonably simple synthetic schemes, glass transition temperatures above typical processing temperatures, acceptable shelf storage lifetime, and minimum toxicological risk.


Figure 3. Using 157nm exposures, phase shift masks, and 148nm-thick resist on a 70nm AR layer, various resist recipes resolve at linewidths of: a) and b) 80nm; c) 100nm; and d) 70nm.
Click here to enlarge image

It is often convenient to adopt a modular approach in the design of new photoresist materials, allowing the polymer design to be broken down into several functional subunits that may be addressed in a somewhat independent fashion (Fig. 1). It is important to note that transparency requirements must be considered in the design of all subunits. The backbone subunit is responsible for tethering the monomer units together and providing the basic mechanical properties for the bulk polymer. The etch-resistant subunit is responsible for decreasing the rate at which the polymer is destroyed by a plasma. Solubility in tetramethylammonium hydroxide (TMAH) developer solutions is achieved by designing matrix resins that include acidic functional groups. Most common protecting groups from earlier photoresists, such as tert-butyl carbonates, and t-butyl esters cannot be used in 157nm photoresist materials since they contain carbonyl groups. Acetal protecting groups contain no carbonyls and are the best choice from a transparency standpoint; however, there are other issues.

Imaging materials
As previously discussed, resists designed for 248 and 193nm lithography are too absorbing at 157nm, which has required a two-pronged approach. The first is to use existing resists, but at a greatly reduced thickness (60-70nm). This provided a pathway for tool testing and early baseline processes. Figure 2 is a cross-section photo of XP-2332C from Shipley on bare silicon with 110nm (1:1) features and a resist thickness of only 68nm. International Sematech is using this as the baseline material for its 157nm exposure tool. This resist is based on a 248nm polymer platform, but is used at a thickness of 68nm [8].

The main focus within the majority of the photoresist research community is on imaging materials based on fluoropolymers. Over the past year, several papers from research groups discussed various new and revitalized approaches [9]. The Universities Photoresist Research Project sponsored by International Sematech works with an alliance of researchers and their students at five universities: U. of Texas, Cornell, Clemson, U. of California at Berkeley, and California Institute of Technology [10]. This project has provided a vast amount of information currently published on 157nm materials imaging and has mainly focused on fluoropolymers based on norbornane. There are other approaches based on fluorinated variations of the resins used in 248nm resists, for example F-ESCAP at MIT/Lincoln Labs and PF-ESCAP at IBM-Almaden Research Center [11, 12].

Lithographic evaluation
The Universities project has been optimizing the formulations and imaging conditions required to print sub-100nm features. This is a tremendous task since it is necessary to explore a variable space that includes photoacid generator, base, dissolution inhibitor, base polymer, exposure, bake, and development. Figure 3 shows some of the images that have been printed to this point in the program. The photoresist system is based on poly norbornane hexafluoroalcohol-co-norbornane t-butyl ester (NBHFA-co-NBTBE) and is blended with an additional polymer that acts as a dissolution inhibitor.


International Sematech Universities 157nm Photoresist Research Project authors.
Click here to enlarge image

Imaging is performed at 0.6NA, 0.3s, 50mJ/cm2, with an alternating PSM, using 148nm-thick resist on 70nm antireflective layer (DUV30) followed by a 60-sec 140°C PAB, 90-sec 130°C PEB, and 20-sec 0.26N TMAH developer. The image quality is good with sidewall angles between 85-88°C. These imaging results represent low k1 factors which will be required if 157nm is used at the 70nm or even 50nm node. The film thickness of nearly 150nm is a significant improvement over thicknesses generally required for pattern transfer. The etch properties are of extreme interest since this material is essentially a first-generation 157nm photoresist. Oxide etch resistance improved significantly over first-generation 193nm materials and performed slightly better than APEXE (2nd-generation KrF resist), but not as good as UV6 (4th-generation KrF resist).

Acknowledgments
The authors of this article are: Will Conley (Motorola assignee to International Sematech and reachable at ph 512/356-3669, email [email protected]); Kim Dean, Daniel Miller, Georgia Rich, Vicki Graffenberg, Shashi Patel, Shang-Ho Lin, and Andrew Jamieson (International Sematech); Raymond Hung, Shintaro Yamada (currently at Shipley Co., Marlboro, MA); Matt Pinnow, Scott MacDonald (IBM, San Jose, CA); Charles Chambers, Brian Osborne, Kyle Patterson (currently at Motorola, Austin, TX); Mark Somervell (currently at Texas Instruments, Dallas, TX); Brian Trinque, H.V. Tran, Sungseo Cho (currently at Shipley Co., Marlboro, MA); Takashi Chiba and Jeff Byers (currently at KLA-Tencor, Austin, TX) from the U. of Texas, departments of chemistry and chemical engineering; Brian Thomas, Greg Shafer, and Darryl DesMarteau (Clemson University, Clemson, SC); John Klopp and Jean Frechet (California Institute of Technology, Pasadena, CA); Dan Sanders and Robert Grubbs (U. of California at Berkeley); Chris Ober and Hilmar Körner (Cornell University, Ithaca, NY); and C. Grant Willson (U. of Texas at Austin).

References

  1. Semiconductor Industry Association Roadmap, 1999
  2. K. Patterson, M. Somervell, C.G. Willson, "The Challenges in Materials Design for 157nm Photoresists, Solid State Technology, 43(3), p. 41, 2000.
  3. R.R. Kunz, et al., "Outlook for 157-nm Resist Design," Proc. of the SPIE, 3678, 13, 1999.
  4. S. Bassett, M. McCallum, H.S. Kim, J. Cashmore, D. Ashworth, Future Fab, 10, 165, 2001.
  5. T.M. Bloomstein, et al., "Lithography with 157nm Lasers," J. Vac. Sci. Technol. B, 13(6), 2112, 1997.
  6. Sematech's "Critical Review on CaF2 Supply for 157nm Tools," Dallas, TX, April 2001.
  7. L.F. Thompson, C.G. Willson, M.J. Bowden, ed., Introduction to Microlithography, Second Ed., American Chemical Society, Washington, DC, pp. 212-232, 1994.
  8. K. Dean, J.Vac. Sci. Technol., 2001, in press.
  9. R.J. Hung, et al., Proceedings of the SPIE, (4345), 2001, in press.
  10. The International Sematech Universities Photoresist Research Project consists of: Prof. C. Grant Willson (U. of Texas), Prof. Chris Ober (Cornell), Prof. Darryl DesMarteau (Clemson), Prof. Jean Frechet (U. of California at Berkeley), and Prof. Robert Grubbs (California Institute of Technology).
  11. T. Fedynyshyn, et al., Proceedings of the SPIE, (4345), 2000, in press.
  12. H. Ito, et al., ibid.