This application note describes a procedure for determining the optimal electron beam exposure dose for experimental Hydrogen Silsesquioxane (HSQ) photoresist. Using the SURFMERA PROXIMA Atomic Force Microscope (AFM) for high-resolution morphological analysis, the minimum dose for complete exposure and the maximum dose before overexposure were successfully identified.
The Proxima accelerates the development of advanced lithography by providing rapid, reliable measurement of key parameters from large samples.
Introduction
In the development of novel photoresist formulations, precise characterization of the exposure process is a crucial step. The final structure morphology, which dictates device performance, is highly dependent on exposure parameters.
Hydrogen Silsesquioxane (HSQ) is one of the most popular inorganic photoresist for negative electron beam lithography [1]. It became widely accepted by its sufficient etching resistivity and high-resolution capability. Due to its popularity, the properties of HSQ have been well explored. However, the search for the perfect photoresist continues, hence, there is still room for improvement.
This application note focuses on experimental HSQ, in particular, on dose test for a new material. An effective way to quantify the optimal exposure dose range for a photoresist is to correlate exposure dose with morphological changes of exposed area [2]. High spatial resolution, stability and repeatability are required in order to detect subtle topographical features specific for under- or overexposure. Meeting these requirements, the SURFMERA PROXIMA AFM becomes an ideal tool to investigate photoresist behavior under electron beam exposure. Further are presented the results on HSQ exposure, optimal exposure dose for the experimental material and method of its determination.
Experimental Setup and Methodology
AFM Setup
All topographic measurements were taken by the Surfmera Proxima AFM. System features noise cancelling, active vibration protection and temperature stabilization, which allow to achieve high-resolution imaging on non-ideal surfaces outside a dedicated laboratory environment.
The measurements leveraged the system’s automated navigation to efficiently locate and sequentially analyze multiple test structures across the sample. Data were acquired in both contact and semi-contact (tapping) modes to ensure comprehensive morphological characterization. The system’s default software was used for all initial data capture and subsequent quantitative image analysis, including segmentation and height distribution profiling.
AFM probe NSG01 suitable for both contact and semi-contact mode (resonant frequency: 170 kHz, stiffness: 1.5 N/m) was used.
Sample Preparation
Thin film of experimental HSQ was formed on silicon substrate with size of 15×15 mm. In order to quantify the optimal exposure dose range a test structure was designed. It was composed of two 100 x 100 µm matrices, each consisting squares with size of 3.5 x 3.5 µm. The electron beam dose was varied from a baseline of 100 µC/cm², increasing by 20 µC/cm² for each subsequent square. The dose gradient was applied row-by-row, from top to bottom and right to left, across the two matrices (Fig. 1). Exposure was followed by development in a 2.5% TMAH solution.
Location and Navigation
The prepared sample was mounted on a universal holder and loaded into the AFM measurement chamber (Surfmera Proxima) for further morphology investigation (Fig. 2).
In the absence of fiducial marks, an optical panoramic map of the sample was generated over a 150×150 mm area (22×14 frames of 1.2×0.8 mm with a 0.1 mm overlap) to locate the test structures. The target area was positioned relative to the AFM probe using interactive targeting on the video image.
Results and Analysis
Determination of Lower Dose Limit
The top two rows of the right matrix (lower doses) were analyzed to establish the minimum dose for complete resist exposure. AFM image, corresponding height profile and height distribution histogram are shown in Fig. 3.
From Fig. 3c it is clear that height distribution histogram consists of separate peaks. A multi-peak Gaussian fit (Fig. 4) revealed distinct six height levels:
- substrate level (1),
- 4 underexposed squares of test structure (2-5),
- 4 squares of the same height (6), even though the exposure dose increased for each subsequent square. This indicates that these areas were formed within the optimal exposure dose range.
The described method of data processing allows one to accurately determine the height of the structures of interest. That, in turn, should lead to precise calculation of exposure dose required for their formation. In this particular case, the analysis of the growth curve (Fig. 4b) indicated that a dose of 300 µC/cm² is the minimum sufficient level for complete exposure, defining the lower limit of the process window.
Determination of Upper Dose Limit
Exceeding the optimal dose leads to overexposure, characterized by “bulging” around the structures and the formation of “ridges,” which reduce the effective structural height and resolution. An AFM image of a characteristic overexposed area is shown in Fig. 5.
To objectively determine the upper dose threshold, an automatic image segmentation analysis was applied to the bottom nine rows of the test structure (higher doses). The results were filtered by area and height (Fig. 6). The analysis identified that area #53, formed with a dose of 1140 µC/cm², was the last area free from overexposure effects.
Area #54 (1160 µC/cm²), shown in Fig.7a, exhibits clear ridge formation, confirming the onset of overexposure. For comparison, Fig. 7b shows structures formed at the optimal exposure dose without any artifacts.
Conclusion
This application note has demonstrated a robust methodology for defining the process window of an experimental HSQ photoresist using AFM for morphological analysis. The procedure successfully established the optimal exposure dose range by identifying the lower threshold for complete exposure and the upper threshold before overexposure.
Based on the AFM analysis:
- The minimum effective exposure dose is 300 µC/cm².
- The maximum dose before overexposure is 1140 µC/cm².
Therefore, the recommended optimal exposure dose window for this photoresist under the specified conditions is 300 – 1140 µC/cm². Employing this range ensures the fabrication of well-defined, high-fidelity structures, providing a critical foundation for process optimization and production yield improvement in advanced lithographic applications.
References
- J. Shen, F. Aydinoglu, M. Soltani, Bo Cui; E-beam lithography using dry powder resist of hydrogen silsesquioxane having long shelf life. J. Vac. Sci. Technol. B, 2019; 37 (2): 021601. DOI: 10.1116/1.5079657
- C. Martin, G. Rius, A. Llobera, A. Voigt, G. Gruetzner, F. Pérez-Murano; Electron beam lithography at 10keV using an epoxy based high resolution negative resist. Microelectronic Engineering, 2007, 84 (5): 1096. DOI: 10.1016/j.mee.2007.01.035.
