Topcoat solution for outgassing and out-of-band light in extreme-UV lithography

Extreme-UV lithography (EUVL) is one of the next-generation lithography candidates for device manufacturing at the so-called 20nm half-pitch node and beyond. However, a number of critical issues remain to be solved with regard to the light source, exposure tools, and photoresist demands of the technology. For example, high-volume manufacturing by EUVL will require improvements in line-width roughness, resolution limit, and sensitivity of the resist to the light for patterning. Recent reports have focused on out-of-band (OoB) radiation— i.e., non-EUV light emitted by the EUV source— owing to concern that it may cause deterioration of lithographic patterning performance.^1–5 In addition, contamination through outgassing from photoresist is a significant problem for exposure tool optics.

We previously addressed these problems by developing an outgassing and OoB protection layer (OBPL),⁶ i.e., a topcoat material that absorbs OoB radiation and prevents outgassing from resist.⁷ However, we realized that it would be advantageous to expand the range of materials that can be used in resist design. In other words, we would like make the OBPL compatible with the various photoresists already available that are, however, hindered by outgassing.⁶Here, we describe experiments aimed at enhancing the design concept and material performance of the OBPL.

Figure 1. Resist outgassing rate through OBPLs (outgassing and out-of-band protection layers) of various types for benzene-derived species from photoacid generator (bottom) and all species (top). amu78: Molecular benzene.

To protect against optics degradation in EUV exposure tools, two important considerations for the OBPL are that it keep the outgassing level very low and that it block contamination from the resist.⁸Accordingly, candidate materials must have the characteristic that they do not mix with the resist. In addition, they must be removable by the development and rinse process during fabrication. We prepared two polymer-type OBPLs— one rigid, one flexible— of different thicknesses (10 and 30nm) by spin-coating them on 60nm-thick resist and baking them at 90°C for 60s. We then tested the samples to determine the impact of the OBPL polymer backbone on the outgassing barrier property of the layers.

We measured the resist outgassing rate by residual gas analysis (RGA) mass spectrometry. Figure 1 shows the resist outgassing rate through both thicknesses for rigid- and flexible-type OBPLs. The rigid-type OBPL showed good barrier performance against benzene-derived outgassing species from PAG (photoacid generator) regardless of the layer thickness. In contrast, the flexible OBPL (in particular, the 10nm-thick layer) did not sufficiently block contamination.

To investigate the outgassing barrier mechanism in greater detail, we performed a separate measurement using time-of-flight secondary-ion mass spectrometry (TOF-SIMS). We subjected rigid- and flexible-type OBPL/resist stacked wafers (30nm thickness of OBPL) to EUV flood exposure at 30mJ/cm². We then employed TOF-SIMS depth profiling to estimate the barrier properties based on changes in the concentration of sulfur atoms in the OBPLs. Figure 2 shows the sulfur atoms detected at the surface of the flexible OBPL, which exhibited a low barrier property by RGA. The rigid OBPL, in contrast, showed a dramatically increased change in sulfur concentration at the interface between the OBPL and the resist, suggesting its suitability as a barrier layer. We then measured the rigid OBPL/resist stacked film by witness sample (WS) testing. As we expected, the OBPL succeeded in blocking not only cleanable but also noncleanable contamination from the resist. We found the RGA/WS measurements to be in good agreement with those of TOF-SIMS.

Figure 2. Variation in sulfur (S) atom concentration after exposure of 30nm-thick flexible (left) and rigid (right) OBPL on resist. Sub.: Substrate.

We also sought to describe the effect of OoB (i.e., deep-UV) protection on the optical properties of the layers. We designed the OBPLs taking into account OoB light absorbance and EUV light transmittance. In particular, measurement of deep-UV light passing through the 30nm-thick rigid OBPL revealed ∼25% transmittance, which indicates a decreased risk of pattern degradation by OoB light. In addition, we calculated EUV transmittance of 88%, which we consider to be low loss. The 30nm film offers the best balance of OoB absorbance and EUV transmittance.

We further tested the efficacy of the OBPL by patterning. To determine the OoB contribution to pattern smoothness, we compared an untreated resist pattern against an OoB-exposed pattern prior to MET (Microfield Exposure Tool, Sematech) 26nm half-pitch patterning (see Figure 3). The amount of smooth patterning was dramatically decreased by OoB exposure. We then repeated the same test with the rigid OBPL (30nm sample), and found no significant variation in pattern margin, sensitivity to EUV, or line-width roughness. The OoB imaging studies were completed by pre-exposing resists coated on a 4-inch wafer, using an external broadband source of sufficient brightness and uniformity. The OoB exposure is 5% of the EUV dose required for patterning a given resist.

Figure 3. Pattern feature comparison of the rigid OBPL/resist, without and with exposure to out-of-band (OoB) light. LWR: Line-width roughness. hp26nm: 26nm half-pitch.

We evaluated the impact of the OBPL on lithographic performance using an ASML NXE:3100 at IMEC Belgium. A cross-sectional scanning electron microscopy image of the resist pattern reveals a rectangular shape at the top but severe footing, i.e., an inadequate vertical profile (see Figure 4). In contrast, the OBPL/resist stacked sample produced a straight pattern with slight top rounding. We believe that the OBPL enhances lithographic performance even under high OoB irradiation.

Figure 4. Lithographic performance comparison of rigid OBPL/resist. CD: Critical dimension. DOF: Depth of focus. X-SEM: Cross-sectional scanning electron microscopy.

In summary, we investigated OBPL materials with reference to their outgassing barrier properties, protection against OoB radiation, and lithographic performance. The rigid OBPL proved a sufficient barrier against outgassing as well as having effective OoB absorbance and high EUV transmittance. It also provided smooth patterning with no negative impact on the resist. Additional, state-of-the-art OBPLs are applicable to all types of resists, including NTD (negative-tone-development) processes. We thus conclude that OBPLs are essential materials for high-volume, next-generation device manufacturing. To improve EUV transmittance, we plan next to develop thinner OBPL materials possessing higher absorption of OoB irradiation without weakening the outgassing barrier properties.