Photoresists
S. C. FU,1 H. W. LIN,2 W. Y. CHOU,1 L. A. WANG,3 K. H. HSIEH1 1Department of Chemical Engineering, National Taiwan University, Taipei 10617, Taiwan
2
Graduate Institute of Material Science and Engineering, National Taiwan University, Taipei 10617, Taiwan
3
Graduate Institute of Electro-Optical Engineering, National Taiwan University, Taipei 10617, Taiwan
Received 28 March 2000; accepted 9 February 2001 Published online 18 December 2001
ABSTRACT: Polymethacrylate copolymers used as single-layer resists (SLR) were in-vestigated in this study. A photoacid generator (PAG) in the chemically amplified resist (CAR) can generate a Bro¨nsted proton after light stimulation, which can catalyze the deprotection reaction of t-butyl methacrylate during the postexposure baking (PEB) process. Excimer lasers (KrF and ArF) were used as light source to evaluate the lithographic performance of resists. The hydrophobic property of the resist changed to hydrophilic property after the deprotection reaction by light exposure, which made the exposed resist soluble in the base developer. The deprotection reaction was analyzed by using FTIR and the reaction kinetics of the resist is discussed. The effects of the t-butyl methacrylate and the isobornyl methacrylate in the resists on the light sensitivity, con-trast, reactive ion etching (RIE) resistance, and lithographic performance were systemat-ically investigated.© 2002 John Wiley & Sons, Inc. J Appl Polym Sci 83: 1860 –1869, 2002 Key words: methacrylate resists; chemical amplification; deprotection reaction; con-trast; lithography
INTRODUCTION
The increasing progress in integrated circuit (IC) manufacture has brought great convenience and efficiency to human life. The increased complexity and enhanced reliability of the advanced IC re-quires the producer to manufacture the circuit with more and more stringent line width. How-ever, to construct a thinner critical dimension (CD) in the circuit, both the resists and the light source must be modified or changed. For light interference limitation, the light source required to expose the resist has to move from visible light
(g-line, i-line) toward KrF (248 nm) and ArF (193 nm). Because of strong phenolic absorption in the phenolic-containing resists at 193 nm, the No-volac and poly(4-hydroxystyrene) (PHS) resists, which are conventionally used in i-line and KrF (248 nm), respectively, cannot be used as ArF (193 nm) photoresists.
To research resolution for this problem, IBM and Lincoln researchers1,2 first developed the
methacrylate copolymers for a chemically ampli-fied resist (CAR) used in 193-nm lithography. The
t-butyl group in methacrylate copolymer can
change the polarity after the postexposure baking (PEB) process. The exposed resist becomes solu-ble in the basic developer and exhibits the posi-tive-tone pattern transferring. The photoacid gen-erator (PAG) in the resist needs energy from light
Correspondence to: K. H. Hsieh (khhsieh@ccms. ntu.edu.tw).
Journal of Applied Polymer Science, Vol. 83, 1860 –1869 (2002) © 2002 John Wiley & Sons, Inc.
DOI 10.1002/app.2313 1860
exposure to activate the deprotection reaction, and can magnify the reaction concentration by repeatedly catalyzing the deprotection reaction of the photoresist.3– 6
Investigations on the methacrylate CAR have mainly focused on improving the etching resis-tance by incorporating the high alicyclic groups,10,11 such as isobornyl,5 adamantyl,11 or
the tricyclodecyl12groups, or on developing a
bet-ter latitude by changing the protective groups, such as tetrahydrofuranyl,13ethoxyethyl,14or the
mevalonic lactone15groups. Moreover, the rate of
acid diffusion, about 10⫺6m2/s in resists, is
be-lieved to be an important factor that influences the line width in lithography.16,17The diffusion–
reaction models proposed by researchers have been integrated into the computer package Pro-lith,18which has been widely used to predict the
effect of processing conditions on the lithographic performance of resists.19,20However, the effect of
alicyclic groups in isobornyl methacrylate (IBMA) on lithography cannot be clearly understood from previous investigations,4,21–23 and the effect of
deprotection groups in resists has not been clearly investigated. This work aims to investigate the lithographic performance of the copolymers from
t-butyl methacrylate (TBMA), IBMA, methacrylic
acid (MAA), and methyl methacrylate (MMA). Co-polymers were synthesized with varying compo-sitions of IBMA and TBMA, whose effects on light sensitivity, contrast, reactive ion etching (RIE) resistance, and lithographic performance are sys-tematically discussed in the study.
EXPERIMENTAL
Materials
The monomers MMA, MAA, TBMA, and IBMA were purified by vacuum distillation to remove
the inhibitor. Methacrylate copolymers of varying compositions were prepared by adding an appro-priate equivalent amount of monomers (MAA, TBMA, MMA, MAA) in dry THF, with 2,2 ⬘-azobi-sisobutyronitrile (AIBN) as initiator (Table I). The reaction was carried out at 65°C under a nitrogen atmosphere over 24 h. The copolymers were precipitated by using nonsolvent hexane and then dried in vacuum.
Measurement
The Fourier transform infrared (FTIR) analysis was carried out on a Bio-Rad FTS-40 (Bio-Rad, Hercules, CA). Thermal properties were obtained by using Perkin–Elmer DSC-7 and TGA-7 (Perkin Elmer Cetus Instruments, Norwalk, CT). Gel per-meation chromatography (GPC) was performed on a Showa Shodex RI-71 (Showa, Japan) using THF as solvent. Hitachi S-800 SEM (Hitachi, To-kyo, Japan) was used to investigate the litho-graphic performance of the resist. Lambda-physik excimer lasers or Oriel 6810 mercury-arc lamp (Oriel, Stratford, CT) equipped with 254-nm band-passing filter was used as the exposure tool for the investigation. RIE resistance was con-ducted using a Roth and RAU Unilab RIE instru-ment. Film thickness was measured by using Dektak 3030ST surface profiler.
Lithographic Evaluation
Methacrylate copolymers mixed with 5 wt % PAG (triarylsulfonium hexafluoro-antimonate) were dissolved in 2-(1-methoxy)propyl acetate (PG-MEA) and were filtered through a 0.2-m filter. Wafers used in the experiment were primed with hexamethyldisilazane (HMDS) for better adhe-sion properties. Resists were spin-coated on the wafers and then prebaked at 100°C for 1 min. The thickness was controlled by varying the solution
Table I Composition and Properties of Photoresist Copolymers
Resist MAA MMA TBMA IBMA
AIBN (mol %) Mw Tg (°C) ␥ E0 (mJ/cm2) A 20 50 30 0 0.5 70,600 135.6 6.0 9 B 20 40 30 10 0.5 64,300 147.3 5.6 13 C 20 30 30 20 0.5 79,000 156.2 5.3 27 D 20 20 30 30 0.5 68,000 167.3 4.6 30 E 20 10 30 40 0.5 79,000 173.7 3.6 58 F 20 0 30 50 0.5 74,000 175.8 — — G 20 40 40 0 0.5 72,000 141.8 14.9 22 H 20 30 50 0 0.5 76,000 131.5 17.2 26
concentration or by changing the spin rate. Sam-ples were exposed to lambda-physik KrF and ArF excimer lasers through a 1-m line and space (L/S) contact mask, and were then subjected to a postexposure bake (PEB) at 120°C for 1 min. Re-sists were developed on tetramethylammonium hydroxide (TMAH) developer and then rinsed in water for 10 s.
RESULTS AND DISCUSSION
Characteristics of Photoresists
The monomer feed ratio and physical properties of the photoresist copolymers A to H are listed in Ta-ble I. The MAA content was kept at 20 mol % for all copolymers because it was previously reported to have optimum sensitivity to polarity change.4,5,7–10
The copolymers were prepared through free-radical polymerization with the same amount of AIBN ini-tiator. Therefore, the molecular weights of the co-polymers are similar in spite of the composition variation. The Tg of copolymers increased as the IBMA monomer was introduced into the system. Variations of their Tg’s were correlated and followed the mixing rule of their components. The IBMA incorporated in the resist was chosen as a modifying segment to increase the RIE resistance2,4 of the
copolymer. The Tg’s of the resultant resists
in-creased with an increase of IBMA content resulting from the bulky side group of IBMA. The thermal stability of the resists, determined by thermo-gravimetric analysis (TGA), demonstrated a two-stage degradation. The first two-stage of degradation, in which the t-butyl group degrades into t-buty-lene species, shifted slightly with varying compo-sitions. The onset degradation temperature (Td) was around 200°C.
The resists were spin-coated on wafers and the effects of solid contents and spin rate on the film thickness are shown in Figure 1. The film thick-ness of the copolymers varies with the 1 power of solid contents and the⫺12power of the spin rate.27
The spin-coated film displayed poor adhesion to the wafer and resulted in a rugged surface when the IBMA was above 40 mol %. This result indi-cates that further increasing IBMA content in copolymer is not beneficial for the resist. To ob-tain a 1-m-thick spin-coated film for the follow-ing investigation, the solution of 2 g solid/10 g PGMEA with the spin rate of 3000 rpm was fol-lowed for this system.
Deprotection Kinetics
The broadening of CAO around 1700 cm⫺1, the blue shift of COOOC around 1150 cm⫺1, and the decrease of CO(CH3)3 around 1368 cm⫺1in the FTIR spectra is characterized as the deprotection
Figure 1 Dependence of film thickness of spin-coated resist A on spinning rate at various solid contents of polymer in 10 g PGMEA: f, 2 g; F, 1.8 g; Œ, 1.5 g; , 1.2 g;䉬, 1 g.
of the t-butyl group in the methacrylate re-sist.7,24,25The deprotection level is defined as
Deprotection level X⫽[COOR]0⫺[COOR] [COOR]0
The mercury lamp was used for exposure and the concentration of stimulated photoacid was
as-sumed to be constant with a long-time exposure under verification by phototitration techniques.26 First-order approximation of the deprotection re-action rate to the t-butyl content is expressed as
⫺ d[COOR]
dt ⫽k[H
⫹]a[COOR]⫽k
1[COOR] Figure 2 Deprotection level versus time at different temperatures: f, 100°C; F,
105°C; Œ, 115°C; , 120°C.
Figure 3 Arrhenius plot of deprotection rate constant versus temperature.
where [COOR] and [H⫹] represent the relative concentration of the t-butyl group and photoacid, respectively.
The fitting curves of the deprotection level X, as shown in Figure 2, are found to correlate well
with the experimental data. The Arrhenius plot of the kinetic constant k1 to the reciprocal of tem-perature is illustrated in Figure 3. The activation energy for deprotection of the t-butyl group of methacrylate was obtained as 31.1 kcal/mol in the
Figure 4 Contrast curves of f resist A, F resist B, Œ resist C, resist D, and䉬 resist
E exposed to 248-nm light and developed with 0.065N TMAH for 10 s.
Figure 5 Contrast curves of f resist A, F resist B, and Πresist D exposed to 193-nm light and developed with 0.065N TMAH for 10 s.
film. The saturation of deprotection was found in Figure 2 for 60 s at 120°C. The condition was used to characterize the performance of the resists.
Lithographic Performance
The conventional developer of 0.26N TMAH is usually applied for Novolac and PHS photore-sists. However, the developing rate for this sys-tem was too rapid to measure the remaining film thickness and even unexposed resist was soluble in the developer. A diluted developer was used and proved to work well in our previous investi-gation.7The contrast of the resists was measured
from the slope of the remaining film thickness after being developed by 0.065N TMAH developer for 10 s at the various exposure doses of 248-nm light source. The sensitivity E0was measured as
the minimum energy required for the exposed film beginning to be dissolved in the developer, the results of which are listed in Table I.
The effects of the IBMA content on the contrast curve and light sensitivity of the resists (A to E) are displayed in Figure 4. The resist with IBMA content above 40 mol % cannot give a smooth film and, therefore, the pattern of the resist F cannot be resolved in the experiment. The increase of IBMA content in the copolymers reveals a de-crease of contrast and sensitivity of the resist. This result can be attributed to the bulky isobornyl group in IBMA, which resulted in ab-sorbing the incident light and inhibiting the dif-fusion–reaction mechanism. Use of a 193-nm light source instead of 248 nm for the resists was also investigated and the resultant contrast
Figure 6 Lithographic pattern of resists exposed with various energy doses and developed with 0.0615N TMAH for 20 s.
curves are shown in Figure 5. Comparing the contrast curves in Figures 4 and 5, contrasts were significantly decreased by using a 193-nm light source, and this finding may be attributed to the strong absorption of the PAG under exposure with a 193-nm light source. The fast decay of photons in the film would result in the decrease of the contrast. Therefore, a clearer pattern can be transferred by exposing the resists with a 248-nm light source than by exposing with a 193-nm light source.
The resists were exposed with different doses and the lithographic patterns are displayed by SEM in Figure 6. Resists A and B can be devel-oped well, using the exposure dose of 25 mJ/cm2.
However, the resist requires a further dose of about 25 mJ/cm2 for a further increase of 10%
IBMA content in the resist. The introduction of the isobornyl group of IBMA into the resist will increase the Tg of the copolymer resist. The free volume of the resist available for photoacid diffu-sion is decreased16; thus, the exposure energy
required for the photoacid is increased.
The effect of the TBMA content in the resist on the contrast curves is shown in Figure 7. The greater the TBMA content in the resist, the more energy is needed to stimulate more photoacid to accelerate the deprotection reaction. The contrast value elevates with the increase of TBMA content
in the resist. Figure 8 shows the lithographic pat-terns by SEM for resists A, G, and H. The pattern of resist H, which has the lowest sensitivity and the highest contrast value among the three re-sists, was easily stripped off the wafer in just 20 s. The smallest processing window among the three resists is found for resist H. The patterns shown in Figure 8 imply that the greater the TBMA content in the resist, the easier the activated pho-toacid diffuses from the exposed area to the un-exposed area. It will subsequently change the po-larity of the unexposed region during the postbak-ing. Thus, the diffusion of photoacid from the exposed area to the unexposed area should also play an important role in lithography.
RIE is a process to etch the substrate by plasma after the exposed resist has been devel-oped. The unexposed resist may also be etched at the same time. The IBMA incorporated in the resist is expected to increase the RIE resistance. The effect of the IBMA content on the physical (Ar) and chemical (CHF3) resistance is shown in Figure 9. As we expected, as the IBMA content increases in the resist, the RIE rate decreases in argon and CHF3atmospheres.
The sensitivity of the resists measured by the plane exposure method can be obtained from the contrast curve. However, more information is needed to evaluate lithographic performance,
Figure 7 Contrast curves of f resist A, F resist G, and Πresist H exposed to 248-nm light and developed with 0.065N TMAH for 10 s.
such as Tg, developing rate, and stability of the resist. Photoacid diffusion plays an important role in the resist chemistry. The diffusion–reac-tion process is used to resolve the depth of focus (DOF) and standing wave effect. However, the photoacid diffusion may vary with the extent of reaction and is hard to measure in the experi-ment. We cannot judge the performance simply from the contrast curve because it is an oversim-plified experiment that the photoacid diffusion is not considered. Although the contrast curves and lithography performance in the research demon-strate that the processing feasibility decreases with the higher IBMA content in the resist, the RIE process shows that the resist of higher IBMA content can withstand the etching in the process. Therefore the trade-off technique or modification of recipes (such as adding some dissolution inhib-itor or crosslinking agent) is required to
formu-late a suitable resist. From this experiment, the resist with 10 to 20% IBMA content and 30% TBMA content is evaluated to be the most suit-able composition for the best lithographic perfor-mance.
CONCLUSIONS
The methacrylate resists with TBMA as acid-la-bile group and IBMA as anti-RIE group were found a suitable composition for the lithographic performance of resists. The Tg’s of copolymers were shown to meet the rule of mixture. The film thickness of a spin-coated resist is observed to be reciprocally proportional to the square root of the spin rate and proportional to the polymer concen-tration. The positive-tone images are observed in the experiments. The deprotection kinetics of the
Figure 8 Lithographic pattern of resists of various TBMA contents at various devel-oped times. Developer: 0.0615N TMAH; exposure energy: 25 mJ/cm2.
t-butyl group of TBMA in the resist using FTIR
analysis was proved to proceed as a first-order reaction to the TBMA concentration, with an ac-tivation energy of 31.1 kcal/mol. The increase of IBMA content in the resist increases the Tgof the
resist and improves its reactive ion etching (RIE) resistance, although the sensitivity and contrast of the resist decrease at the same time. In addi-tion, the increase of TBMA content in the resist will decrease the sensitivity and increase the con-trast but result in a smaller processing window. Only with critical tailoring of the resist composi-tion can the perfect lithography be achieved.
REFERENCES
1. Allen, R.; Wallraff, G.; Dipietro, R.; Hofer, D.; Kunz, R. J Photopolym Sci Technol 1994, 7, 507. 2. Rothschild, M.; Forte, A. R.; Kunz, R. R.;
Palma-teer, S. C.; Sedlacek, J. H. C. IBM J Res Develop 1997, 41, 49.
3. Hiro, M.; Akahori, T.; Tachiki, S. Polym Prepr (Am Chem Soc Div Polym Chem) 1997, 38, 137. 4. Allen, R. D.; Wallraff, G. M.; Hofer, D. C.; Kunz,
R. R. IBM J Res Dev 1997, 41, 95.
5. Jung, J. C.; Bok, C. K.; Bail, K. H. Proc SPIE–Int Soc Opt Eng 1998, 3333, 11.
6. Ito, H. Proc SPIE–Int Soc Opt Eng 1999, 3678, 2. 7. Fu, S. C.; Hsieh, K. H.; Wang, L. A. J Polym Res
1999, 6, 2.
8. Nakano, K.; Iwasa, S.; Maeda, K.; Hasegawa, E. Proc SPIE–Int Soc Opt Eng 1998, 3333, 43. 9. Allen, R. D.; Opitz, J.; Ito, H.; Wallow, T. I.;
Cas-mier, D. V.; Dipietro, R. A.; Brock, P. J.; Breyta, G.; Sooriyakumaran, R.; Larson, C. E.; Hofer, D. C. Proc SPIE–Int Soc Opt Eng 1999, 3678, 66. 10. Gokan, H.; Esho, S.; Ohnishi, Y. J Electrochem Soc
1983, 130, 143.
11. Takahashi, M.; Takechi, S.; Kaimoto, Y.; Hanyu, I.; Abe, N.; Nozaki, K. Proc SPIE–Int Soc Opt Eng 1995, 2438,422.
12. Nakano, K.; Maeda, K.; Iwasa, S.; Ohfuji, T.; Ha-segawa, E. Proc SPIE–Int Soc Opt Eng 1995, 2438, 433.
13. Allen, R. D.; Sooriyakumaran, R.; Opitz, J.; Wall-raff, G. M.; DiPietro, R. A.; Breyta, G.; Hofer, D. C.; Kunz, R. R.; Jayaraman, S.; Shick, R.; Goodall, B.; Okoroanyanwu, U.; Wilson, C. G. Proc SPIE–Int Soc Opt Eng 1996, 2724 334.
14. Maeda, K.; Nakano, K.; Iwasa, S.; Hasegawa, E. Proc SPIE–Int Soc Opt Eng 1997, 3049, 55. 15. Dammel, R. R.; Ficner, S.; Oberlander, J.; Jacobs,
A. K.; Padmanaban, M.; Khanna, D. N.; Durham, D. L. Proc SPIE–Int Soc Opt Eng 1998, 3333, 144.
16. Toriumi, M.; Ohfuji, T.; Endo, M.; Morioto, H. Proc SPIE–Int Soc Opt Eng 1999, 3678, 368.
17. Itani, T.; Yoshino, H.; Hashimoto, S.; Yamana, M.; Smoto, N.; Kasama, K. J Vac Sci Technol B 1996, 14, 4226.
18. Mack, C. A. Inside Prolith; Finle Technologies, Inc., 1997.
Figure 9 Effect of IBMA content on the RIE resistance under CHF3 f and Ar F
19. Zuniga, M.; Neureuther, A. R. J Vac Sci Technol B 1995, 13, 2957.
20. Sekiguchi, A.; Mack, C. A.; Isono, M.; Matsuzawa, T. Proc SPIE–Int Soc Opt Eng 1999, 3678, 985. 21. Allen, R. D.; Wallraff, G. M.; Dipietro, R. A.; Hofer,
D. C.; Kunz, R. R. Proc Am Chem Soc Natl Symp 1995, 72, 100.
22. Kunz, R. R.; Allen, R. D.; Hinsberg, W. D.; Wallraff, G. M. Proc SPIE–Int Soc Opt Eng 1993, 1925, 167.
23. Allen, R. D.; Wallraff, G. M.; Hinsberg, W. D.; Simpson, L. L.; Kunz, R. R. in Polymers for Micro-electronics; Thompson, L. F.; Willson, C. G.;
Ta--gawa, S., Eds.; ACS Symposium Series 000; Amer-ican Chemical Society: Washington, DC, 1994. 24. Hutchinson, J. M.; Wallraff, G. M.; Hinsberg,
W. D.; Opitz, J.; Oldham, W. G. Proc SPIE–Int Soc Opt Eng 1995, 2438, 486.
25. Petersen, J. S.; Mack, C. A.; Thackeray, J. W.; Sinta, R.; Fedynshyn, T. H.; Mori, J. M.; Byers, J. D.; Miller, D. A. Proc SPIE–Int Soc Opt Eng 1995, 2438, 153. 26. Buhr, G.; Dammel, R.; Lindley, C. R. Polym Mater
Sci Eng 1989, 61, 269.
27. Willson, C. G. in Introduction to Microlithography; Thompson, L. F.; Willson, C. G.; Bowden, M. J., Eds.; American Chemical Society: Washington, DC, 1994.
