Limits...
Beyond EUV lithography: a comparative study of efficient photoresists' performance.

Mojarad N, Gobrecht J, Ekinci Y - Sci Rep (2015)

Bottom Line: We observe high sensitivity of the photoresist performance on its chemical content and compare their overall performance using the Z-parameter criterion.Interestingly, inorganic photoresists have much better performance at BEUV, while organic chemically-amplified photoresists would need serious adaptations for being used at such wavelength.Our results have immediate implications for deeper understanding of the radiation chemistry of novel photoresists at the EUV and soft X-ray spectra.

View Article: PubMed Central - PubMed

Affiliation: Laboratory for Micro- and Nanotechnology, Paul Scherrer Institute, 5232 Villigen, Switzerland.

ABSTRACT
Extreme ultraviolet (EUV) lithography at 13.5 nm is the main candidate for patterning integrated circuits and reaching sub-10-nm resolution within the next decade. Should photon-based lithography still be used for patterning smaller feature sizes, beyond EUV (BEUV) lithography at 6.x nm wavelength is an option that could potentially meet the rigid demands of the semiconductor industry. We demonstrate simultaneous characterization of the resolution, line-edge roughness, and sensitivity of distinct photoresists at BEUV and compare their properties when exposed to EUV under the same conditions. By using interference lithography at these wavelengths, we show the possibility for patterning beyond 22 nm resolution and characterize the impact of using higher energy photons on the line-edge roughness and exposure latitude. We observe high sensitivity of the photoresist performance on its chemical content and compare their overall performance using the Z-parameter criterion. Interestingly, inorganic photoresists have much better performance at BEUV, while organic chemically-amplified photoresists would need serious adaptations for being used at such wavelength. Our results have immediate implications for deeper understanding of the radiation chemistry of novel photoresists at the EUV and soft X-ray spectra.

No MeSH data available.


Related in: MedlinePlus

Line-edge roughness and critical dimension of Inpria IB.(a) LER and (b) critical dimension as a function of dose-on-wafer for different half-pitch values, exposed at EUV (squares) and BEUV (triangles).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4363827&req=5

f2: Line-edge roughness and critical dimension of Inpria IB.(a) LER and (b) critical dimension as a function of dose-on-wafer for different half-pitch values, exposed at EUV (squares) and BEUV (triangles).

Mentions: The critical dimension (CD) and LER dose-dependence is illustrated in Fig. 2. The analysis of the CD and LER was done using a commercial software (SuMMIT®) and LER values correspond to 3σ deviation averaged over both walls of 20 lines. The horizontal axis has been calibrated to represent dose-on-wafer, Dw and not the commonly used dose-on-mask Dm. This calibration allows comparison of the exact dose at the wafer and takes into account the mask diffraction efficiency difference at the two wavelengths. Details of the procedure are described in Methods. As it can be seen, since Inpria IB is a negative-tone resist, for all HPs CD monotonically increases by increasing Dw, and remarkably, the exposure latitude (EL) drops at all HPs when moving from EUV to BEUV. The ratio of the ELs at BEUV to EUV, QEL is 0.25 for HP = 35 nm and increases to QEL ≈ 0.4 for HP = 20 and 22 nm, yet at the HP = 18 nm illuminated by BEUV, only patterns could be formed at one dose value. Within the same functional dose range, the dependence of the LER on Dw is illustrated in Fig. 2b. Clearly, the LER is lower at all HP values when exposed with EUV in comparison to BEUV. Whether rougher edges are solely because of the chemical formulation of the resist, or because of an intrinsic property of exposure at higher photon energies can be examined by comparing patterning performance of other photoresists. Figure 3a demonstrates the average LER as a function of HP for all photoresists. In terms of LER, HSQ has a good performance at both wavelengths, and provides LER < 2 nm, whereas for Inpria IB and the CAR, patterning at BEUV results in rougher edges. For HSQ, BEUV exposures provided HPmin = 22 nm, while with EUV HPmin = 18 nm was achieved, which confirms similar measurements that show HPmin ≈ 15 nm26. In principle this resolution limit for HSQ might be improved by using another developer based on NaOH solution, commonly used for developing patterns with HP < 15 nm2426. However, this improvement comes with the cost of losing sensitivity by a factor of 3 at EUV wavelength. The CAR we studied is among the highest-resolution resists and revealed down to HPmin = 18 nm at EUV. However, the smallest feature size observed at BEUV was only HPmin = 35 nm. In a previous study we demonstrated that another CAR, which has the EUV resolution limit of HPmin = 22 nm, can also be patterned with the same resolution at BEUV20. It could therefore be concluded that 35 nm is not the patterning resolution limit of organic CARs and that their performance is strongly dependent on their chemical composition and the incident photon energy.


Beyond EUV lithography: a comparative study of efficient photoresists' performance.

Mojarad N, Gobrecht J, Ekinci Y - Sci Rep (2015)

Line-edge roughness and critical dimension of Inpria IB.(a) LER and (b) critical dimension as a function of dose-on-wafer for different half-pitch values, exposed at EUV (squares) and BEUV (triangles).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4363827&req=5

f2: Line-edge roughness and critical dimension of Inpria IB.(a) LER and (b) critical dimension as a function of dose-on-wafer for different half-pitch values, exposed at EUV (squares) and BEUV (triangles).
Mentions: The critical dimension (CD) and LER dose-dependence is illustrated in Fig. 2. The analysis of the CD and LER was done using a commercial software (SuMMIT®) and LER values correspond to 3σ deviation averaged over both walls of 20 lines. The horizontal axis has been calibrated to represent dose-on-wafer, Dw and not the commonly used dose-on-mask Dm. This calibration allows comparison of the exact dose at the wafer and takes into account the mask diffraction efficiency difference at the two wavelengths. Details of the procedure are described in Methods. As it can be seen, since Inpria IB is a negative-tone resist, for all HPs CD monotonically increases by increasing Dw, and remarkably, the exposure latitude (EL) drops at all HPs when moving from EUV to BEUV. The ratio of the ELs at BEUV to EUV, QEL is 0.25 for HP = 35 nm and increases to QEL ≈ 0.4 for HP = 20 and 22 nm, yet at the HP = 18 nm illuminated by BEUV, only patterns could be formed at one dose value. Within the same functional dose range, the dependence of the LER on Dw is illustrated in Fig. 2b. Clearly, the LER is lower at all HP values when exposed with EUV in comparison to BEUV. Whether rougher edges are solely because of the chemical formulation of the resist, or because of an intrinsic property of exposure at higher photon energies can be examined by comparing patterning performance of other photoresists. Figure 3a demonstrates the average LER as a function of HP for all photoresists. In terms of LER, HSQ has a good performance at both wavelengths, and provides LER < 2 nm, whereas for Inpria IB and the CAR, patterning at BEUV results in rougher edges. For HSQ, BEUV exposures provided HPmin = 22 nm, while with EUV HPmin = 18 nm was achieved, which confirms similar measurements that show HPmin ≈ 15 nm26. In principle this resolution limit for HSQ might be improved by using another developer based on NaOH solution, commonly used for developing patterns with HP < 15 nm2426. However, this improvement comes with the cost of losing sensitivity by a factor of 3 at EUV wavelength. The CAR we studied is among the highest-resolution resists and revealed down to HPmin = 18 nm at EUV. However, the smallest feature size observed at BEUV was only HPmin = 35 nm. In a previous study we demonstrated that another CAR, which has the EUV resolution limit of HPmin = 22 nm, can also be patterned with the same resolution at BEUV20. It could therefore be concluded that 35 nm is not the patterning resolution limit of organic CARs and that their performance is strongly dependent on their chemical composition and the incident photon energy.

Bottom Line: We observe high sensitivity of the photoresist performance on its chemical content and compare their overall performance using the Z-parameter criterion.Interestingly, inorganic photoresists have much better performance at BEUV, while organic chemically-amplified photoresists would need serious adaptations for being used at such wavelength.Our results have immediate implications for deeper understanding of the radiation chemistry of novel photoresists at the EUV and soft X-ray spectra.

View Article: PubMed Central - PubMed

Affiliation: Laboratory for Micro- and Nanotechnology, Paul Scherrer Institute, 5232 Villigen, Switzerland.

ABSTRACT
Extreme ultraviolet (EUV) lithography at 13.5 nm is the main candidate for patterning integrated circuits and reaching sub-10-nm resolution within the next decade. Should photon-based lithography still be used for patterning smaller feature sizes, beyond EUV (BEUV) lithography at 6.x nm wavelength is an option that could potentially meet the rigid demands of the semiconductor industry. We demonstrate simultaneous characterization of the resolution, line-edge roughness, and sensitivity of distinct photoresists at BEUV and compare their properties when exposed to EUV under the same conditions. By using interference lithography at these wavelengths, we show the possibility for patterning beyond 22 nm resolution and characterize the impact of using higher energy photons on the line-edge roughness and exposure latitude. We observe high sensitivity of the photoresist performance on its chemical content and compare their overall performance using the Z-parameter criterion. Interestingly, inorganic photoresists have much better performance at BEUV, while organic chemically-amplified photoresists would need serious adaptations for being used at such wavelength. Our results have immediate implications for deeper understanding of the radiation chemistry of novel photoresists at the EUV and soft X-ray spectra.

No MeSH data available.


Related in: MedlinePlus