Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2026 Feb 3;18(6):10261–10269. doi: 10.1021/acsami.5c18774

Integrated Strategies for Overcoming Resolution Limits in Electron Beam Lithography of Chemically Amplified Resists

Xinpei Wu †, Nikhil Tiwale †,*, Aaron Stein †,*, Chang-Yong Nam †,‡,*
PMCID: PMC12926953  PMID: 41633557

Abstract

Electron beam lithography (EBL) of chemically amplified resists (CARs) faces fundamental challenges, including stochastic electron scattering and acid diffusion, that limit resolution and reproducibility. Using SU-8 as a model CAR, this study systematically investigated complementary strategies to address these challenges, combining multipass exposure, proximity effect correction (PEC) with midrange correction factors, base quencher incorporation, and post-exposure bake (PEB) suppression. Monte Carlo simulations and calibrated PEC modeling revealed that extending the point spread function to include a midrange scattering component significantly improved critical dimension (CD) control across varying pattern densities, correcting deviations that conventional two-term PEC failed to capture. Multipass exposure, particularly 4-pass writing with a 25% offset, redistributed the dose to average stochastic beam and scattering fluctuations, reducing line-width roughness by more than 50% and yielding more uniform nanoscale features. Photoacid confinement was investigated by adding urea as a base quencher, which successfully reduced acid diffusion but introduced substantial sensitivity penalties without improving ultimate resolution or Z-factor performance, underscoring the trade-offs of chemical versus physical confinement. Suppressing PEB most directly minimized acid diffusion, resulting in improved Z-factors and reproducible 30 nm half-pitch dense line/space patterns. Overall, these results demonstrated that PEC with midrange correction, multipass strategies, quencher additives, and PEB-free processing addresses different aspects of the EBL process window and that their integration provides a comprehensive framework for managing stochastic scattering, diffusion, and chemical amplification effects. This framework advances dense nanoscale patterning in CARs and establishes guiding principles for optimizing resist design and process strategies in high-resolution EBL and potentially other advanced lithographies, such as extreme ultraviolet (EUV) lithography.

Keywords: electron beam lithography, chemically amplified resists (CARs), proximity effect correction, high-resolution patterning, photoacid diffusion, SU-8

graphic file with name am5c18774_0010.jpg

graphic file with name am5c18774_0008.jpg

1. Introduction

Electron beam lithography (EBL) is a cornerstone technique in nanofabrication, enabling direct-write patterning with sub-10 nm resolution and exceptional flexibility in feature geometry. − Unlike optical lithography, which relies on mask projection, EBL employs a focused beam of high-energy electrons to locally modify the solubility of a resist material. This capability has made EBL indispensable for applications such as nanoelectronics, biotechnology, and advanced materials research, where rapid prototyping and high-resolution pattern transfer are essential.

Among the various resist systems used in EBL, chemically amplified resists (CARs) have attracted significant attention due to their high sensitivity and compatibility with both electron beam and extreme ultraviolet (EUV) lithography. ,− In CARs, pattern formation is mediated by a catalytic chain reaction initiated by photoacid generators (PAGs) upon electron or photon exposure. For EBL, incident electronsthrough primary interactions and cascades of secondary and backscattered electronsgenerate acids that catalyze a deprotection or cross-linking reaction during the post-exposure bake (PEB). For EUV lithography, <90 eV low-energy secondary electrons generated by incident EUV photons are responsible for activating PAGs. This chemical amplification enables high sensitivity, reducing the required exposure dose and thus increasing throughput. Furthermore, CARs like SU-814, 15 and other epoxy-based systems offer high etch resistance, making them suitable for pattern transfer into underlying substrates.

However, the use of CARs in high-resolution EBL presents several challenges. − First, as shown in Figure a,b, Monte Carlo simulations , reveal that even with highly energetic 100 kV incident electrons, extensive electron scattering occurs within the resist and substrate. Forward-scattered primary electrons and backscattered electrons generated from the substrate produce significant mid- and long-range scattering tails in the deposited energy profile. This scattering broadens the spatial distribution of deposited energy beyond the beam footprint, leading to a non-negligible proximity effect. Second, as illustrated in Figure c, the acids generated in these partially exposed zones can further migrate during PEB, driven by thermal activation and acid diffusion kinetics. This lateral acid diffusion can broaden patterned features and simultaneously limit ultimate resolution, particularly in dense patterns or under elevated PEB conditions. − The combined effects of extended-range electron scattering and post-exposure acid diffusion, thus, impose fundamental limits on resolution and pattern fidelity in CAR-based EBL processes.

1.

1

Open in a new tab

Electron scattering and acid diffusion in CARs: (a) Monte Carlo simulation of primary (black) and backscattered (red) electron trajectories for two isolated 100-electron shots at 100 kV on Si. (b) Simulated energy deposition map showing localized exposure and proximity effects. (c) Schematic of EBL on CARs, illustrating limited electron scattering and enhanced lateral acid diffusion during PEB.

To address these challenges, we implemented a multipass writing strategy combined with advanced proximity effect correction (PEC) that incorporates midrange correction. This approach compensates for intermediate-range electron scattering along with acid diffusion and development blur and is particularly effective for maintaining dose accuracy and feature fidelity in high-density patterning. Furthermore, to mitigate acid diffusion without compromising cross-linking efficiency, we investigated two complementary process modifications: (i) incorporating urea as a base quencher into the resist formulation, which reduced the effective acid concentration, moderating the rate of cross-linking and further limiting unwanted acid spread, and (ii) adjusting PEB conditions. In parallel, eliminating PEB further suppressed acid diffusion, enabling us to study the intrinsic resolution limits determined by electron scattering alone. Our combined approachPEC with midrange correction, optimized PEB conditions, and urea-based diffusion controlsignificantly enhanced the high-density resolution and line edge roughness of EBL-patterned CARs.

2. Results and Discussion

2.1. Multipass EBL Patterning to Smoothen Energy Deposition

50 nm thin films of SU-8 were chosen as the model CAR system for its mechanical robustness, high sensitivity and etch resistance, and widespread use in diverse EBL applications. ,,− Figure compares single-pass and 4-pass writing strategies in SU-8. In the single-pass approach (Figure a), the electron beam delivers entire resist exposure dose in one exposure shot, and Monte Carlo simulations , of energy deposition (Figure c) for 20 nm wide line patterns reveal pronounced local dose variation caused by stochastic electron scattering. This variation led to nonuniform feature development and increased LWR. Details of the Monte Carlo simulation procedure are provided as the Supporting Information.

2.

2

Open in a new tab

Comparison of single-pass and 4-pass EBL writing in SU-8: single-pass writing (top row) is illustrated schematically (a), with the corresponding simulated energy deposition map (c) showing higher local dose variation. In contrast, 4-pass writing with 25% x-shift (bottom row) (b) produces a more uniform energy deposition profile (d). The corresponding SEM images (e, f) confirmed that 4-pass writing has improved pattern fidelity.

In contrast, the multipass writing strategy has been widely explored for improving EBL performance. , In this approach, instead of exposing the entire dose in a single exposure shot, the total dose is divided into multiple spatially offset exposure shots. In this work, we implemented a 4-pass exposure with a 25% x-shift (Figure b), dividing the total dose evenly across four spatially offset exposures shotseach pass delivering one-quarter of the total required dose. A lateral (x-direction) offset was used because the critical dimension (CD) and LWR in 1D line/space patterns are governed primarily by variations across the line width. Lateral shifting therefore provides the dominant averaging of stochastic dose and scattering variations, whereas orthogonal shifts offer negligible additional benefits for 1D features. Other offsets (10–50%) were tested experimentally and in simulation, with 25% providing the best balance between smoothing, pattern stability, and write time. This approach averages stochastic variations in beam position, beam blur, and electron scattering. By redistributing the dose, local fluctuations are minimized, producing a more uniform energy deposition profile (Figure d). Additionally, 4-pass writing mitigates thermal and charging effects that can accumulate during extended single-pass exposures, thereby further enhancing pattern fidelity.

The scanning electron microscopy (SEM) images of 20 nm wide isolated lines (Figure e,f) verify the advantage of the 4-pass approach. The measured CD decreases from 39.2 to 33.2 nm, and LWR is reduced from 11.1 to 5.1 nm, which shows more uniform and well-defined edges. These improvements agree with previous high-resolution EBL studies, , where multipass writing was shown to suppress local dose inhomogeneities, mitigate beam-induced distortions, and enhance CD controlparticularly for dense line/space patterns.

2.2. PEC with Midrange Correction

Figure illustrates the experimentally measured contrast and CD data and PEC point spread function (PSF) calibration models for 50 nm thick SU-8. The contrast curve in Figure a shows the normalized remaining height as a function of exposure dose, yielding the resist’s sensitivity (D 100 = 54.39 μC/cm2) and contrast (γ = 1.22), which are critical measures of resist performance. Figure b plots the measured CD versus dose for line-space patterns with densities ranging from 0 to 100%, all written using the 4-pass strategy. The schematic of the patterns can be found in the Supporting Information (Figure S1). Similar to our earlier work on the 100 nm thick SU-8, by writing line/space patterns at densities ranging from isolated features (0%) to fully dense arrays (100%), we can evaluate how electron scattering and acid diffusion influence critical dimension control across varying layout conditions. This approach allows us to quantify density-dependent proximity effects and calibrate the point spread function accordingly.

3.

3

Open in a new tab

(a) Contrast curve of 50 nm thick SU-8. (b) Measured CD as a function of dose for different pattern densities using 4-pass writing. (c) Initial PSFs of 50 nm thick SU-8 obtained from MC simulation. (d) Calibrated PSFs incorporating midrange correction by tuning the midrange weight factor Nue = 30%, derived from the contrast curve and CD data in (b).

The initial PSF for 50 nm SU-8 (Figure c) was generated using Monte Carlo simulations in TRACER software, capturing the short- and long-range scattering components. While this two-term PSF adequately represents forward scattering and long-range backscattering, for CARs, this correction is insufficient to account for acid diffusion that can significantly impact the fidelity of dense patterns. To address this, we introduced a midrange term as shown in yellow, into the PSF model, using calibration features in TRACER software, producing the calibrated three-term PSF shown in Figure d. This calibration was achieved by iteratively adjusting the midrange parameters as the midrange weight factor (Nue), which is the ratio of the energy in the midrange region to the short-range energy. We refined the model by tuning Nue to find the optimal additional midrange length during data fitting, which allowed the model to balance the electron-scattering correction with acid diffusion control.

Incorporating the midrange correction into PEC allows the algorithm to redistribute the dose and process blur more accurately, compensating for scattering over intermediate distances that are not captured in standard two-term models. The process blur is defined as a combination of resist processing and electron beam artifacts. This refinement reduces CD variation across different pattern densities and improves uniformity in both isolated and dense features.

2.3. Dense Line-Space Patterning of SU-8

Following the PEC calibration described in Figure , we used the “Z-factor” as a metric to assess the overall patterning performance, including resolution, pattern roughness, and dose sensitivity, to find the optimal Nue for SU-8 patterning. The Z-factor was calculated based on the half-pitch data using the following expression: ,

Z‐factor=CD3·LWR2·DtS 1

where CD is the critical dimension/half-pitch, LWR is the line-width roughness, and DtS is the dose-to-size (exposure dose that yields pattern dimension equal to CD). We performed L/S patterning at different half-pitches while systematically varying Nue to represent different midrange corrections. By comparing the resulting Z-factor values (Figure S2), we achieved an optimized PEC model for SU-8 with Nue = 30% that better reflects the actual electron energy distribution behavior in high-resolution EBL, enabling more precise pattern writing and reduced proximity effects.

Additionally, we evaluated the combined effects of midrange correction and multipass writing on dense SU-8 line-space patterns, as shown in Figure . Three exposure strategies were compared at a fixed dose of 30 μC/cm2: (a) 4-pass without midrange correction, (b) single-pass with midrange correction at Nue = 30%, and (c) 4-pass with midrange correction at Nue = 30%. The SEM images reveal that midrange correction alone improves the resolution from 50 nm (Figure a) to 35 nm (Figure b), but the best result is achieved when midrange correction is combined with four-pass writing (Figure c), producing the lowest line-width roughness (LWR = 5.8 nm) and the most uniform, sharply defined lines.

4.

4

Open in a new tab

(a–c) SEM images of dense SU-8 line-space patterns at 30 μC/cm2: (a) 4-pass without midrange correction, (b) single-pass with correction, and (c) 4-pass with correction. (c) Lowest LWR and best fidelity. (d) Z-factor evaluation at Nue = 30% for single-pass and 4-pass writing. (e) Line-width roughness (LWR) as a function of half-pitch for single-pass and 4-pass writing.

Quantitative analysis further supports these observations. The Z-factor evaluation at Nue = 30% (Figure d) shows that, across half-pitches from 30 to 100 nm, the four-pass strategy consistently achieves better Z-factor performance than single-pass writing, indicating improved dose latitude and process stability. Likewise, LWR measurements as a function of half-pitch (Figure e) confirm that four-pass writing reduces roughness across all tested dimensions, with the most pronounced improvements in the sub-50 nm regime. The midrange PEC term represents the integrated blur from intermediate-range electron scattering and acid diffusion. While the calibrated value used here (Nue = 30%) is specific to 50 nm SU-8 at 100 kV, the same three-term framework is applicable to thicker resists. Thus, the approach is general, and only the calibration parameter requires adjustment for other thickness/energy regimes. Notably, this approach is highly effective for dense line/space patterns near the resist resolution limit, where proximity scattering and stochastic fluctuations become dominant contributors to pattern variability.

2.4. Mitigating Photoacid Diffusion by Base Quenchers

In addition to PEC calibration and multipass writing strategies, we explored the role of base additives in modulating SU-8 resist performance. Quench bases (QBs) are known to suppress acid diffusion by scavenging a fraction of the photoacid, thereby limiting unwanted cross-linking beyond the intended exposure region. ,, To test this effect, urea was incorporated into SU-8 at varying concentrations (0–15 mol %), taking advantage of its small, simple molecular structure that facilitates uniform incorporation and efficient quenching.

The contrast curves shown in Figure a reveal a systematic shift in dose-to-clear with an increasing QB concentration. Neat SU-8 exhibits high sensitivity, cross-linking at ∼50 μC/cm2, whereas SU-8 with 15 mol % QB requires nearly double the dose, cross-linking at ∼90 μC/cm2. This behavior reflects reduced effective acid concentration and slower cross-linking kinetics, consistent with prior reports on CARs where base quenchers reduce sensitivity.

5.

5

Open in a new tab

Effect of urea as a QB on SU-8 contrast and line resolution. (a) Contrast curves for SU-8 with 0, 5, 10, and 15 mol % urea (relative to photoacid groups). (b–d) SEM images of 20 nm isolated lines patterned in SU-8 with varying QB concentrations.

The SEM images of 20 nm isolated lines further highlight this trade-off. At 5 mol % QBs, features with a CD of 31.8 nm are achieved at 50 μC/cm2 (Figure b). Increasing the QB content to 10 and 15 mol % yields finer features of 29.3 and 27.4 nm, respectively (Figures c,d). The narrowing of line width demonstrates the ability of QBs to localize photoacid-driven reactions, minimizing lateral blur and enhancing pattern fidelity. However, this improvement comes at the expense of sensitivity, as higher exposure doses are required to compensate for the quenched acid concentration.

These results underscore a central trade-off in resist optimization: PEC with midrange correction and multipass writing mitigates electron scattering and local dose variations, whereas base quenchers act directly on the chemical amplification step to confine acid diffusion, but at the expense of requiring higher exposure doses. The increased dose, however, amplifies electron scattering and, if increased too excessively, counteracts the benefits of reduced acid diffusion. Importantly, while PEB strongly impacts neat SU-8 by broadening features through acid diffusion, its influence on urea-containing samples is negligible since quenchers already dominate the chemical kinetics of acid confinement. As seen in the CD–dose trends across different densities (Figure S4), increasing urea concentration shifts sensitivity to higher doses without significant PEB dependence. Although urea suppresses acid diffusion, the higher exposure dose required increases stochastic energy deposition fluctuations and electron-scattering blur. Consequently, as shown in the high-resolution patterning data (Supplemental Figure S5), the addition of quenchers yields no significant improvement in resolution or Z-factor performance.

2.5. Minimized Photoacid Diffusion by Omitting PEB

In the early step of radiation chemistry, a pair of electrons and a radical cation are formed by ionization. The electrons ejected from molecules lose their kinetic energy through thermalization, and these thermalized electrons can still drive photoacid generation and subsequent chemical reactions. Without a PEB step, EBL can still pattern SU-8 but requires a higher exposure dose due to the absence of thermal driven cross-linking. We hypothesize that the omission of PEB in principle can improve patterning resolution and reduce LWR despite loss in sensitivity, and if combined optimally, it should be able to also improve the overall Z-factor.

Indeed, Figure a demonstrates the successful patterning of 20 nm isolated lines in SU-8 without PEB at 65 μC/cm2, confirming that target CDs can be achieved when acid diffusion driven by externally elevated temperature is suppressed by intentionally forgoing PEB. Even with added urea QBs, the same applies; as shown in Figure b, the minimum CD decreases with increasing urea concentration under PEB, whereas no significant CD reduction occurs without PEB due to the absence of active acid diffusion by external heating. Correspondingly, Z-factor analysis (Figure c) shows superior performance for SU-8 without PEB due to the reduced variability in acid reaction volume.

6.

6

Open in a new tab

SU-8 patterning without PEB: (a) SEM image of a 20 nm isolated line without PEB at 65 μC/cm2 demonstrates the achievable resolution when acid diffusion is suppressed. (b) CD decreases with increasing QB concentration for both processes, with PEB yielding wider lines. (c) Z-factor evaluations indicate superior performance by SU-8 without PEB. (d) Contrast curves show decreased sensitivity for SU-8 without PEB. (e) CD versus dose for different pattern densities using 4-pass writing without PEB. (f) Calibrated PSFs indicate reduced process blur without PEB (red) compared to with PEB.

The sensitivity and contrast values for SU-8 with and without PEB and QB can be found in Table . Omission of PEB increases the required dose but slightly improves contrast, consistent with reduced acid diffusion. Adding urea further shifts sensitivity to higher doses, while contrast remains nearly unchanged (Figure S3). Thus, both strategies restrict acid activity but at the cost of reduced sensitivity, underscoring the trade-off between dose efficiency and resolution control.

1. Dose Behaviors for SU-8 and SU-8 with QB.

  PEB
 non-PEB
samples sensitivity (μC/cm2) contrast sensitivity (μC/cm2) contrast
SU-8 54.39 1.22 63.61 1.35
SU-8 + 5 mol % QB 69.99 1.15 72.14 1.18
SU-8 + 10 mol % QB 84.55 1.17 96.06 1.03
SU-8 + 15 mol % QB 86.61 1.11 98.07 1.10
Open in a new tab

For dense line-space patterns, CD versus dose data (Figure e) reveal a much smaller CD increase in the absence of PEB compared to conventional processing (Figure b). Derived from the contrast curve data (Figure d) and CD-versus-density trends (Figure e), Figure f compares the calibrated PSFs for PEB and non-PEB conditions. The transparent curve is the PEB PSF (identical to Figure d and overlaid here for reference), while the solid curve represents the non-PEB PSF. The non-PEB PSF exhibits a steeper midrange decay, indicating reduced process blur due to suppressed acid diffusion. These results confirm that eliminating PEB sharpens the effective PSF relative to the diffusion-broadened profile under PEB conditions.

Together, these results demonstrate that omitting PEB effectively suppresses acid diffusion and sharpens feature definition, though at the expense of a higher dose. This trade-off highlights the coupled roles of electron scattering and acid diffusion in setting SU-8’s resolution limit and establishes PEB-free processing as a valuable strategy for advancing high-resolution EBL on CARs.

2.6. Ultimate High-Resolution Patterning of SU-8 without PEB

By combining strategies such as 4-pass writing, PEC with midrange correction, and PEB omission, the ultimate resolution limits of SU-8 can be evaluated. As shown in Figure a, omitting PEB enables the successful fabrication of 30 nm half-pitch line/space patterns with improved line continuity, reduced LWR (4.9 nm), and a smaller CD (29.3 nm)one of the best SU-8 patterning resolutions reported for high pattern density to date. These results confirm that suppressing acid diffusion is the most effective route to sharpen features and broaden the process windows (Figure c).

7.

7

Open in a new tab

Resolution limits for dense line and space patterns in SU-8. (a) and (b) SEM images of line/space patterns with target half-pitches (HPs) of 30 and 25 nm, showing improved line continuity and edge definition. (c) Process windows for HP 25 and 30 nm line/space patterns, with dotted lines indicating target CDs.

Meanwhile, attempts to pattern 25 nm half-pitch structures without PEB reveal partial line collapse and discontinuities (Figure b), indicating that electron scattering combined with residual acid migration still imposes a practical limit. At 20 nm half-pitch, no continuous features were obtained, reflecting the point where stochastic effects, acid diffusion, and resist blur dominate the PEC correction. This observation illustrates that, even with PEB suppression, SU-8 cannot fully confine acid reactions at such dimensions, defining a resolution floor of ∼ 25–30 nm for dense periodic features.

The Z-factor values summarized in Table provide further insight into the resist’s performance near its resolution limit. Omitting PEB reduces the Z-factor from 0.054 (PEB) to 0.031 (non-PEB), reflecting improved patterning performance balance (better resolution and LWR despite small sensitivity loss). In contrast, SU-8 with QB additives shows significantly larger Z-factors (≥0.182 with PEB and up to 0.569 without PEB), indicating poorer pattern fidelity and degraded process performance.

2. Summary of Resolution Limits for SU-8 Samples.

  PEB
 non-PEB
samples CDmin (nm) LWR (nm) Dts (μC/cm2) Z-factor (mC·nm3) CDmin (nm) LWR (nm) Dts (μC/cm2) Z-factor (mC·nm3)
SU-8 35.8 5.8 35 0.054 29.3 4.9 55 0.031
SU-8 + 5 mol % QB 35.2 10.2 40 0.182 34.4 10.5 60 0.272
SU-8 + 10 mol % QB 40.3 9.4 40 0.231 34.8 11.4 60 0.329
SU-8 + 15 mol % QB 41.2 10.5 40 0.308 39.2 12.6 60 0.569
Open in a new tab

3. Conclusions

We investigated multiple strategies to overcome the persistent challenges of high-resolution patterning in CARs for EBL, using SU-8 as a model system. Through a systematic evaluation of multipass exposure, PEC with midrange correction, base quencher, and PEB-free processing, we were able to separately probe the contributions of electron scattering, acid diffusion, and resist chemistry to the ultimate resolution limits of SU-8. Multipass (4-pass) writing emerged as a straightforward and highly effective strategy for reducing stochastic variability in dose delivery. The result is a measurable reduction in LWR and an improvement in line edge fidelity compared with single-pass exposures. SEM imaging and Monte Carlo simulations confirmed that multipass writing produces a more uniform energy deposition profile, leading to sharper and more reproducible features.

Building on this, PEC with an additional midrange correction term was introduced to account for midrange electron scattering not adequately described by conventional two-term point spread function (PSF) models. This refinement improved CD control across a range of pattern densities by compensating for scattering contributions that dominate at sub-50 nm scales. Together, the combination of PEC with midrange correction and multipass writing was shown to mitigate scattering-induced proximity effects in SU-8, enabling improved dose placement and feature uniformity in both isolated and dense line/space structures. The role of photoacid confinement was explored by incorporating urea as a quench base. While quenchers partially suppress acid diffusion and help restrict cross-linking to the intended regions, they also reduced resist sensitivity by significantly increasing the required exposure dose. Importantly, the addition of quenchers did not yield measurable improvements in resolution or Z-factor performance.

In contrast, omitting PEB provides the most direct and effective suppression of acid diffusion. Although higher doses were required to achieve comparable cross-linking, SU-8 without PEB exhibited sharper calibrated PSFs, smaller CDs, and improved Z-factors relative to conventional PEB processing. Crucially, high-resolution patterning experiments demonstrated that PEB-free SU-8 EBL could reproducibly achieve 30 nm half-pitch dense line/space patterns with a low LWR (∼4.9 nm). However, attempts at 25 and 20 nm half-pitch features resulted in line discontinuities and collapse, establishing that the intrinsic resolution limit of SU-8 lies between 25 and 30 nm for dense periodic patterns. We note that the PEB-free approach is most effective for thin resist films (≲100 nm) where cold cross-linking remains sufficiently efficient; for thicker CAR films or systems with slower acid-catalyzed kinetics, partial-PEB or hybrid thermal strategies may be required to maintain full cross-linking and pattern integrity.

Overall, these findings reveal the intricate interplay between electron scattering, acid diffusion, and resist chemistry in defining the resolution of CARs and establish a framework for pushing the boundaries of high-density nanoscale patterning in advanced EBL. The revealed insights into the external-heating-free PAG activation and minimized acid diffusion may also suggest new approaches for improving CAR patterning using EUV lithography. As EBL continues to serve as a critical research and prototyping tool for emerging nanotechnologies, the methodologies presented here will guide both the design of next-generation CARs and the development of process strategies to achieve reproducible sub-30 nm dense features.

4. Materials and Methods

4.1. Materials

SU-8 2002, SU-8 thinner, and SU-8 developer were purchased from KAYAKU. Urea (ACS reagent, 99.0–100.5%, Sigma-Aldrich) and SurPass 3000 (DisChem) were used as received.

4.2. Fabrication of SU-8 Thin Films

Prior to spin coating, silicon wafers were exposed to an oxygen plasma for 2 min, immersed in SurPass 3000 solutions for 1 min, and then rinsed in deionized (DI) water for 15 s. Thin resist films were prepared by dropping ∼50 μL of diluted SU-8 solution (1:15 by volume in SU-8 thinner) onto plasma-treated silicon substrates spinning at 4000 rpm to attain the thickness of ∼50 nm as measured by ellipsometry. For quencher-modified samples, urea was incorporated into the SU-8 formulation at defined molar ratios relative to the PAG. Specifically, 0, 5, 10, and 15 mol % urea were added by dissolving the appropriate amount of urea into the diluted SU-8 solution prior to spin coating. The solutions were stirred until homogeneous and then processed in a manner identical to those used for the neat SU-8 samples. All coated wafers were baked at 95 °C for 1 min before exposure.

4.3. EBL Patterning and Development

A JEOL JBX-6300FS EBL system (100 kV) was used to pattern the SU-8 resists. For exposure dose matrix tests, 60 μm squares were written at 100 pA with a 16 nm shot spacing, varying the dose from 3 to 120 μC/cm2. Line–space nanopatterns were exposed at 100 pA with a 5 nm shot pitch. PEB samples were baked at 90 °C for 1 min, while non-PEB samples proceeded directly to development. Development was performed in SU-8 developer for 1 min, followed by a 15 s isopropyl alcohol rinse and drying with a nitrogen gun.

4.4. Characterization

Post-development resist height was measured by using a Park NX20 atomic force microscope (AFM) equipped with PPP-NCHR tips. Scanning electron microscopy (SEM) was performed on a Hitachi S-4800 instrument operated at 2 kV. CD and unbiased LWR were extracted by using ProSEM software (GenISys GmbH) for high-precision feature analysis.

Supplementary Material

am5c18774_si_001.pdf (918KB, pdf)

Acknowledgments

This research is supported by the U.S. Department of Energy Office of Science Accelerate Initiative Award 2023-BNL-NC033-Fund. This research used the Nanofabrication and Materials Synthesis and Characterization Facilities of the Center for Functional Nanomaterials (CFN), which is a U.S. Department of Energy Office of Science User Facility, at Brookhaven National Laboratory under Contract No. DE-SC0012704.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.5c18774.

  • Detailed descriptions of the Monte Carlo simulation of electron energy deposition and high-resolution patterning of SU-8 with urea; schematics and corresponding SEM images of calibration patterns with varying densities; Z-factors for various Nue values; contrast curves of SU-8 with quenchers without PEB; the measured CD of SU-8 calibration patterns with quenchers; SEM images of high-resolution SU-8 patterns with quenchers; and detailed CD analysis values for half-pitch 30 and 25 nm SU-8 patterns (PDF)

The authors declare no competing financial interest.

References

  1. Chen Y.. Nanofabrication by electron beam lithography and its applications: A review. Microelectron. Eng. 2015;135:57–72. doi: 10.1016/j.mee.2015.02.042. [DOI] [Google Scholar]
  2. Gangnaik A. S., Georgiev Y. M., Holmes J. D.. New Generation Electron Beam Resists: A Review. Chem. Mater. 2017;29(5):1898–1917. doi: 10.1021/acs.chemmater.6b03483. [DOI] [Google Scholar]
  3. Saifullah M. S. M., Rajak A. K., Hofhuis K. A., Tiwale N., Mahfoud Z., Testino A., Karadan P., Vockenhuber M., Kazazis D., Valiyaveettil S., Ekinci Y.. Approaching Angstrom-Scale Resolution in Lithography Using Low-Molecular-Mass Resists (<500 Da) ACS Nano. 2024;18(35):24076–24094. doi: 10.1021/acsnano.4c03939. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Groves T. R., Pickard D., Rafferty B., Crosland N., Adam D., Schubert G.. Maskless electron beam lithography: prospects, progress, and challenges. Microelectron. Eng. 2002;61–62:285–293. doi: 10.1016/S0167-9317(02)00528-2. [DOI] [Google Scholar]
  5. Manfrinato V. R., Camino F. E., Stein A., Zhang L., Lu M., Stach E. A., Black C. T.. Patterning Si at the 1 nm Length Scale with Aberration-Corrected Electron-Beam Lithography: Tuning of Plasmonic Properties by Design. Adv. Funct. Mater. 2019;29(52):1903429. doi: 10.1002/adfm.201903429. [DOI] [Google Scholar]
  6. Kolodziej C. M., Maynard H. D.. Electron-beam lithography for patterning biomolecules at the micron and nanometer scale. Chem. Mater. 2012;24(5):774–780. doi: 10.1021/cm202669f. [DOI] [Google Scholar]
  7. Cui, Z. Electron Beam Lithography. In Nanofabrication: Principles, Capabilities and Limits; Springer International Publishing: 2024. DOI: 10.1007/978-3-031-62546-6. [DOI] [Google Scholar]
  8. Hassaan M., Saleem U., Singh A., Haque A. J., Wang K.. Recent Advances in Positive Photoresists: Mechanisms and Fabrication. Materials (Basel) 2024;17(11):2552. doi: 10.3390/ma17112552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Manouras T., Argitis P.. High Sensitivity Resists for EUV Lithography: A Review of Material Design Strategies and Performance Results. Nanomaterials (Basel) 2020;10(8):1593. doi: 10.3390/nano10081593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Sharma E., Rathi R., Misharwal J., Sinhmar B., Kumari S., Dalal J., Kumar A.. Evolution in Lithography Techniques: Microlithography to Nanolithography. Nanomaterials (Basel) 2022;12(16):2754. doi: 10.3390/nano12162754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Li L., Liu X., Pal S., Wang S., Ober C. K., Giannelis E. P.. Extreme ultraviolet resist materials for sub-7 nm patterning. Chem. Soc. Rev. 2017;46(16):4855–4866. doi: 10.1039/C7CS00080D. [DOI] [PubMed] [Google Scholar]
  12. Kozawa T., Yoshida Y., Tagawa M. U.. Radiation-Induced Acid Generation Reactions in Chemically Amplified Resists for Electron Beam and X-Ray Lithography. Jpn. J. Appl. Phys. 1992;31(12S):4301. doi: 10.1143/JJAP.31.4301. [DOI] [Google Scholar]
  13. Thackeray J. W.. Materials challenges for sub-20-nm lithography. J. Micro/Nanolithogr., MEMS, MOEMS. 2011;10(3):033009. doi: 10.1117/1.3616067. [DOI] [Google Scholar]
  14. Arun K. N., Sang Won P., Jeong Bong L.. Characterization of SU-8 as a resist for electron-beam lithography. Proc. SPIE. 2003;5116:414–423. doi: 10.1117/12.499112. [DOI] [Google Scholar]
  15. Aktary M., Jensen M. O., Westra K. L., Brett M. J., Freeman M. R.. High-resolution pattern generation using the epoxy novolak SU-8 2000 resist by electron beam lithography. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena. 2003;21(4):L5–L7. doi: 10.1116/1.1596216. [DOI] [Google Scholar]
  16. Mack, C. A. Lithographic Effects of Acid Diffusion in Chemically Amplified Resists. In Microelectronics Technology, ACS Symposium Series, Vol. 614; American Chemical Society: 1995; pp 56–68. [Google Scholar]
  17. Kozawa T., Saeki A., Nakano A., Yoshida Y., Tagawa S.. Relation between spatial resolution and reaction mechanism of chemically amplified resists for electron beam lithography. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena. 2003;21(6):3149–3152. doi: 10.1116/1.1619959. [DOI] [Google Scholar]
  18. Greul M., Shoshi A., Klikovits J., Martens S., Hofmann U., Barahona O., Shnirman B., Starz L., Wintrich P., Sailer H.. Universal approach for process optimization of chemically amplified photoresists in electron beam lithography. J. Micro/Nanopattern. Mater. Metrol. 2024;23(02):024601. doi: 10.1117/1.JMM.23.2.024601. [DOI] [Google Scholar]
  19. Shimizu R., Ding Z.-J.. Monte Carlo modelling of electron-solid interactions. Rep. Prog. Phys. 1992;55(4):487. doi: 10.1088/0034-4885/55/4/002. [DOI] [Google Scholar]
  20. Joy, D. C. Monte Carlo Modeling for Electron Microscopy and Microanalysis; Oxford University Press: 1995. [Google Scholar]
  21. Shin J., Ma Y., Cerrina F.. Depth dependence of resist line-edge roughness: Relation to photoacid diffusion length. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena. 2002;20(6):2927–2931. doi: 10.1116/1.1526638. [DOI] [Google Scholar]
  22. Lee C.-T., Lawson R. A., Henderson C. L.. Understanding the effects of photoacid distribution homogeneity and diffusivity on critical dimension control and line edge roughness in chemically amplified resists. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena. 2008;26(6):2276–2280. doi: 10.1116/1.2976601. [DOI] [Google Scholar]
  23. Koyama M., Imai K., Shirai M., Hirai Y., Yasuda M.. Effects of acid diffusion and resist molecular size on line edge roughness for chemically amplified resists in EUV lithography: computational study. Jpn. J. Appl. Phys. 2021;60(10):106505. doi: 10.35848/1347-4065/ac26e5. [DOI] [Google Scholar]
  24. Wu, X. ; Tiwale, N. M. ; Stein, A. ; Nam, C.-Y. ; Callahan, R. ; De Silva, A. . Modeling photoacid diffusion in chemically amplified resists via proximity effect correction in electron beam lithography. In Advances in Patterning Materials and Processes XLII, 2025. [Google Scholar]
  25. Wang M., Lee C.-T., Henderson C. L., Yueh W., Roberts J. M., Gonsalves K. E.. Incorporation of ionic photoacid generator (PAG) and base quencher into the resist polymer main chain for sub-50 nm resolution patterning. J. Mater. Chem. 2008;18(23):2704. doi: 10.1039/b801517a. [DOI] [Google Scholar]
  26. Pain L., Higgins C., Scarfoglière B., Tedesco S., Dal’Zotto B., Gourgon C., Ribeiro M., Kusumoto T., Suetsugu M., Hanawa R.. Resolution limit of negative tone chemically amplified resist used for hybrid lithography: Influence of the molecular weight. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena. 2000;18(6):3388–3395. doi: 10.1116/1.1321288. [DOI] [Google Scholar]; (acccessed 8/14/2025).
  27. Vinje J., Beckwith K. S., Sikorski P.. Electron Beam Lithography Fabrication of SU-8 Polymer Structures for Cell Studies. Journal of Microelectromechanical Systems. 2020;29(2):160–169. doi: 10.1109/JMEMS.2020.2967174. [DOI] [Google Scholar]
  28. Golvari P., Kuebler S. M.. Fabrication of Functional Microdevices in SU-8 by Multi-Photon Lithography. Micromachines (Basel) 2021;12(5):472. doi: 10.3390/mi12050472. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Pepin A.. Exploring the high sensitivity of SU-8 resist for high resolution electron beam patterning. Microelectron. Eng. 2004;73–74:233–237. doi: 10.1016/j.mee.2004.02.046. [DOI] [Google Scholar]
  30. Bilenberg B., Jacobsen S., Schmidt M. S., Skjolding L. H. D., Shi P., Bo̷ggild P., Tegenfeldt J. O., Kristensen A.. High resolution 100kV electron beam lithography in SU-8. Microelectron. Eng. 2006;83(4–9):1609–1612. doi: 10.1016/j.mee.2006.01.142. [DOI] [Google Scholar]
  31. Ivin V. V., Silakov M. V., Babushkin G. A., Lu B., Mangat P. J., Nordquist K. J., Resnick D. J.. Modeling and simulation issues in Monte Carlo calculation of electron interaction with solid targets. Microelectron. Eng. 2003;69(2–4):594–605. doi: 10.1016/S0167-9317(03)00351-4. [DOI] [Google Scholar]
  32. Herr, D. J. C. ; Martin, L. ; Manakli, S. ; Icard, B. ; Pradelles, J. ; Orobtchouk, R. ; Poncet, A. ; Pain, L. . Multiple pass exposure in e-beam lithography: application to the sub-22nm nodes. In Alternative Lithographic Technologies II, 2010. [Google Scholar]
  33. Hasan M. N., Lee S.-Y., Ahn B.-S., Choi J., Park J.-S.. Effectiveness of multipass and multirow writing methods for massively parallel e-beam systems. J. Vac. Sci. Technol. B. 2020;38(6):062602. doi: 10.1116/6.0000547. [DOI] [Google Scholar]
  34. GenISys; GmbH: 2018. https://www.genisys-gmbh.com/tracer.html. [Google Scholar]
  35. Lopez G. G., de Villafranca G., Azadi M., Metzler M. G., Lister K., Labella M., Eichfeld C., Belic N., Hofmann U.. On the trends and application of pattern density dependent isofocal dose of positive resists for 100 keV electron beam lithography. J. Vac. Sci. Technol. B. 2018;36(6):062601. doi: 10.1116/1.5048206. [DOI] [Google Scholar]
  36. Subramanian A., Tiwale N., Lee W. I., Kisslinger K., Lu M., Stein A., Kim J., Nam C. Y.. Vapor-Phase Infiltrated Organic–Inorganic Positive-Tone Hybrid Photoresist for Extreme UV Lithography. Adv. Mater. Interfaces. 2023;10(28):2300420. doi: 10.1002/admi.202300420. [DOI] [Google Scholar]
  37. Sturtevant, J. L. ; Michaelson, T. B. ; Jamieson, A. T. ; Pawloski, A. R. ; Byers, J. ; Acheta, A. ; Willson, C. G. . Understanding the role of base quenchers in photoresists. In Advances in Resist Technology and Processing XXI, 2004. [Google Scholar]
  38. Allen, R. D. ; Wang, C. W. ; Chang, C. Y. ; Ku, Y. . Photobase generator and photo decomposable quencher for high-resolution photoresist applications. In Advances in Resist Materials and Processing Technology XXVII, 2010. [Google Scholar]
  39. Sturtevant, J. L. ; Kozawa, T. ; Tagawa, S. . Basic aspects of acid generation processes in chemically amplified resists for electron beam lithography. In Advances in Resist Technology and Processing XXII, 2005. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

am5c18774_si_001.pdf (918KB, pdf)

Articles from ACS Applied Materials & Interfaces are provided here courtesy of American Chemical Society

RESOURCES