ArF and F2 lithography using bilayer resists
07/01/2003
Overview
Contamination of the exposure system optics with silicon has been identified as a major issue for the adoption of bilayer technology across all wavelengths. A laser outgassing system was developed and calibrated using model silicon compounds to better understand the effect of polymer architecture on silicon outgassing. It was observed that the placement of silicon into the polymer backbone as a poly(silsesquioxane) allows the incorporation of high-silicon content for improved etch resistance, with no detectable outgassing of silicon during the exposure step. The design concepts used for these ultra-thin silicon-imaging systems enabled the resolution of sub-100nm dense patterns.
To achieve production at sub-100nm device targets, a thinner (<2000Å) imaging layer with improved etch resistance will be required to obtain robust manufacturing processes. Recent developments in 193nm single-layer technology have demonstrated that a reduction in the film thickness of 193 single-layer resists adversely affects the fidelity of the dry etch transfer process from photoresist to substrate, and as a result, the use of ultra thin resists will require alternate etch processes, such as the use of hardmask technologies, bilayer or trilayer schemes.
Thin film imaging techniques such as a bilayer photoresist scheme have long been considered as an alternative to single-layer resist technology [1–12]. The bilayer lithographic process involves imaging a thin silicon top layer (1000–2000Å) over an etch-resistant underlayer (3000–6000Å). The top thin silicon layer is imaged and developed to generate a pattern and this image is transferred to the bottom layer using oxygen reactive ion etch (RIE). During the pattern transfer process, the silicon resist in the masked region is oxidized in the plasma to form a refractory oxide which acts as an etch barrier. The use of a bilayer resist has an advantage over conventional single-layer photoresists in that the underlayer can be tailored for a variety of applications through superior etch resistance, planarization, and reflection control. Further, the underlayer can be modified for damascene applications, such as imaging over complex topography, and minimizing resist poisoning from low-k dielectric materials.
One of the main concerns for the implementation of a bilayer lithographic process in production is the potential contamination of the optics by silicon species during the lithographic process. Kunz and coworkers [13] from MIT Lincoln Laboratory, and Irie and coworkers [14] from the Association of Super-Advanced
Electronics Technologies (ASET), have shown that both organic and silicon fragments can be outgassed from a photoresist during the exposure step. Consequently, these outgassed species can deposit on the lens and potentially degrade transmission and uniformity. Further, it has been shown that the deposition of a silicon species on the lens is irreversible using standard cleaning procedures.
Previous work on outgassing of silicon containing polymers upon exposure to 157nm light illustrated that the method of incorporating silicon into the imaging layer is important. Hien and coworkers from Sematech reported that the location of the Si atom in the polymer is critical to the outgassing of silicon during the exposure step [15]. The Sematech investigation found that polymers containing silicon in side chain groups tend to outgas at higher rates than polymers in which the silicon is incorporated into the backbone. The development of a silicon polymer with the following attributes is therefore essential for the adoption of bilayer as a feasible photoresist technology for the 65 and 45nm technology nodes: 1) minimal or no outgassing of silicon, 2) functionality with high transparency at 193 and 157nm, 3) good etch resistance, and 4) uniform aqueous base dissolution behavior.
 |
null
Several research groups have focused on the development of silicon containing bilayer resists. Numerous materials approaches to silicon-containing imaging systems have been investigated including acrylate polymers with silicon incorporated in groups pendant to the polymer backbone, to silsesquioxane polymers in which the silicon is incorporated in the polymer backbone [1–12]. One approach is to incorporate silicon pendant groups in conventional 193nm single-layer acrylic polymers. Kwong and coworkers developed a high-resolution, 193nm bilayer resist based on norbornene-maleic anhydride alternating a copolymer with a pendant silicon group [12]. Although this copolymer showed excellent lithographic performance and etch stability, high outgassing of silicon species at a 193nm exposure was observed. Silicon outgassing rates for these polymers were on the order of 1013 molecules/cm2, as tested at MIT Lincoln Laboratory.
 Figure 1. Detection of silicon model compounds using a porous polymer GC column. |
null
Even though further chemical modifications to this 193nm bilayer resist led to significant reduction in the outgassing of silicon species (108–1010 molecules/cm2), outgassing of silicon from polymers containing small pendant silicon groups is very likely. Therefore, it is necessary to understand the effect of silicon polymer architecture on the outgassing of silicon species. Model polymers were prepared in which silicon was incorporated in a number of different ways pendant to the polymer backbone and in the polymer backbone. It was observed that the placement of silicon into the polymer backbone as a poly(silsesquioxane) allows the incorporation of high silicon content for superior etch resistance, with no detectable outgassing of silicon during the exposure step. The design concepts used for these ultra thin silicon imaging systems have resulted in excellent imaging capability, resolving sub-100nm dense patterns.
Materials
The various silicon model compounds used to calibrate the outgassing equipment (see table) were obtained from Gelest Inc. Silsesquioxanes were prepared by acid catalyzed condensation polymerization according to standard literature procedures. The average molecular weight (Mn) of the silsesquioxanes, siloxanes and pendent acrylate polymers is in the range of 3000–10,000, as determined by gel permeation chromatography calibrated with polystyrene standards. All polymers used in this study were characterized by 1H, 13C, 29Si NMR, HPLC, FT-IR, thermal analysis, elemental analysis and MALDI-TOF MS spectrometry.
Experimental set-up for outgassing study
Outgassing studies on small molecule silicon species and silicon polymers were performed at Rohm and Haas Co. The detection equipment consists of an atmospheric pressure stainless steel chamber equipped with a wafer stage, a laser input window, and the appropriate optical elements for exposure. The laser linked to this chamber can be tuned for 248, 193, and 157nm light. The chamber is connected to a sample concentrator that is coupled to a trace gas chromatograph equipped with an ion trap mass spectrometer detector. A range of chromatography columns was investigated ranging from porous polymer columns to wall-coated open tubular columns (WCOT).
Detection limit for silicon compounds
Prior to evaluating the effect of polymer architecture on silicon outgassing, the analytical equipment was validated for the detection of silicon species. The detection of a range of silicon-containing model compounds with different size and silicon content were studied in order to estimate the detection limit for the compounds. The type of gas chromatography (GC) column played an important role in the detection of these silicon compounds. Two general classes of GC columns were investigated: porous polymer coated capillary columns, in which the retention mechanism is based upon adsorption chromatography, and WCOT. The porous polymer column is excellent for low boiling hydrocarbons and silicon compounds but it strongly adsorbs tetraethyl-orthosilicate.
 Figure 2. Silicon species outgassed from model polymers. |
null
Alternatively, WCOT columns provide much better detection for the higher boiling silicon compounds and tetraethylorthosilicate is not adsorbed, in contrast to porous polymer columns. Figure 1 illustrates the detection of all six model compounds using a porous polymer column. As the size and silicon content of the silicon compound is increased, the detection limit is reduced.
For chromatography, the method used to determine the detection limit consists of estimating the concentration for which the chromatographic peak of the analyte of interest is 6x higher than the baseline noise standard deviation.
This is usually done by injecting a standard of known concentration measuring the peak height over the baseline and calculating the signal/noise ratio (S/N). The detection limit can be calculated by dividing the concentration by (S/N) and multiplying the result by six. An estimated detection limit of the six silicon compounds has been calculated using the peak height method for the two best GC columns; results are presented in the table on p. 73.
Outgassing results of bilayer resists
To study the effect of polymer architecture on the outgassing of silicon species, model polymers were prepared in which the silicon was incorporated in a number of different ways pendant to the polymer backbone and in the polymer backbone. These polymers consisted of acrylate polymers containing pendant silicon units, siloxane, and silsesquioxanes (Fig. 2).
Figure 2 tabulates the silicon species outgassed from the various silicon containing polymers — all of the side chain and siloxane-based silicon materials exhibited outgassing of silicon species during exposure. In contrast, under the exact same exposure conditions and formulation, all of the silsesquioxanes-based materials exhibited no detectable outgassing of silicon species. The acrylic polymers containing silicon pendant groups produced dimethylsiloxane on the order of 1012 molecules/cm2 upon 193nm exposure. A similar outgassing trend was also observed at MIT Lincoln Laboratory. Hien et al. also observed outgassing of silicon from siloxanes [15]. In the siloxane systems investigated for the present study, the rate of outgassing for siloxane polymers is much higher than in the previous Sematech study [15]. The difference in the silicon outgassing rate for siloxanes between the Sematech and Shipley reports is under further investigation and may be related to the chemical functionality pendant to the siloxane.
 Figure 3. Dissolution contrast of 193nm silsesquioxane-based photoresist. |
null
A possible mechanism for the outgassing of silicon from the side chain polymers involves the acid generated from the photoacid generator attacking the oxygen of Si-O-Si leading to the formation of silanol and a cation. The silanol generated will then be converted to hexamethylsiloxane in the presence of acid. Mechanistic studies into the acidolysis reactions of silicon containing photoresists has been reported by Zharov et al. [16].
As a result of these outgassing studies, silsesquioxanes have been identified as the preferred silicon platform for the development of 248, 193, and 157nm bilayer photoresists at Shipley Co.
193nm and 157nm silsesquioxane design
In addition to reduced outgassing, the silsesquioxane approach allows the incorporation of high-silicon content, ranging from 10–20wt% silicon depending upon polymer composition. The organic groups (R) pendant to the silsesquioxane have been optimized for transparency at the appropriate wavelength (193nm or 157nm), control of the dissolution behavior, and contrast of the photoresist. Figure 3 shows the dissolution rate contrast of a silsesquioxane-based ArF bilayer photoresist optimized for high-contrast applications.
 Figure 4. 193 bilayer 90nm and 110nm contact holes at 0.75NA. |
null
Lithographic performance
For sub-100nm device rules, both 193 and 157 bilayer photoresists based upon silsesquioxanes have demonstrated excellent resolution capability and performance. The design concepts used for these ultra-thin silicon imaging systems have resulted in good imaging capability, resolving sub-100nm dense patterns using an ASML 5500/1100 ArF exposure tool at 0.75 NA. Figure 4 shows dense 90nm contact holes (at a 220nm pitch) imaged using a functionalized silsesquioxane 193nm bilayer resist system. Also shown are 110nm dense contact holes imaged at 193nm and subsequent etch transfer of the image into the underlayer via O2-RIE.
F2 bilayer photoresists have also been developed using this silsesquioxane approach in which the pendent R groups have been modified to increase transparency at 157nm. Using this design, F2 bilayer resists have been developed that show great promise, resolving 70nm dense line space patterns (Fig. 5).
 Figure 5. 157 bilayer 70nm dense line and space pattern. |
null
Conclusion
Bilayer photoresists that show promise for sub-100nm lithography have been developed. Both ArF and F2 bilayer photoresist systems have been developed based upon a silsesquioxane approach. It was observed that incorporating silicon into the backbone of the silsesquioxane resulted in no detectable outgassing of silicon species during the exposure step at either the 193nm or 157nm wavelengths. Sub-100nm resolution has been achieved using the thin silicon imaging systems developed for both ArF and F2 lithography. Further, complementary underlayers have been developed for both ArF and F2 systems and have been optimized for etch resistance, planarization over complex topographies and tunable reflection control.
George Barclay, Subbareddy Kanagasabapathy, Gerd Pohlers, Shipley Company, Marlboro, Massachusetts, Francois Huby, Kenneth Wiley, Rohm and Haas Co., Spring House, Pennsylvania
Acknowledgments
Co-authors on this paper are: Joe Mattia, Kao Xiong, Sheri Ablaza, Jim Cameron, Shintaro Yamada, Sungseo Cho and Anthony Zampini of Shipley Co.
References
1. Schaedeli et al., J. Photopolym. Sci. Technol., 9(3), 435–446, 1996.
2. Sooriyakumaran, et al. Proc. SPIE, 3333, 219, 1998.
3. Wallraf, et al., J. Photopolym. Sci & Technol., Vol. 11, No. 4, 673, 1998.
4. Goethals, et al., J. Photopolym. Sci & Technol., Vol. 11, No. 3, 513, 1998.
5. Lin, Proc. SPIE, 3678, 241, 1999.
6. Namba, H. Takahashi, J. Photopolym. Sci & Technol., Vol. 11, No. 4, 663, 1998.
7. Morisawa, et al., J. Photopolym. Sci & Technol., Vol. 11, No. 4, 667, 1998.
8. Hatakeyama, et al., Proc. SPIE, 3333, 62, 1998.
9. Hien et al., Proc. SPIE, 3333, 154–164, 1998.
10. Uetani et al., Proc. SPIE, 3678, 510–513, 1999.
11. Kwong et al., Proc. SPIE, 4345, 50–57 2001.
12. Kwong et al., Proc. SPIE, 4690, 403–409, 2002.
13. Kunz et al., J. Vac. Sci. Technol. B, 18(3), 1306–1313, 2000.
14. Irie et al., J. Photopolym. Sci. Technol., 14(4), 561–566, 2001.
15. Hien et al., Proc. SPIE, 4345, 439–447, 2001.
16. Zharov et al., Chem. Mater., 656–663, 2002.
For more information, contact George Barclay, principal scientist, Shipley Co., 455 Forest St., Marlborough, MA 01752; ph 508/229-7262, fax 508/229-2473, email [email protected].