Picometer sensitivity metrology for EUV absorber phase

Stuart Sherwin, Isvar A. Cordova, Ryan Miyakawa, Markus P. Benk, Laura Waller, Andrew R. Neureuther, Patrick Naulleau2020 · DOI

TL;DR

This paper introduces a novel reflectometry-based metrology technique that can precisely measure the phase difference between EUV photomask absorber and multilayer layers down to picometer sensitivity. This advancement is crucial for improving the image quality of advanced EUV masks, enabling more precise manufacturing of next-generation semiconductor devices.

What Problem Does This Paper Address?

As semiconductor manufacturing pushes towards smaller feature sizes, the demand for advanced EUV (Extreme Ultraviolet) lithography masks, specifically attenuated phase-shift masks (aPSMs), has increased. These masks offer superior image quality by manipulating light phase, but accurately controlling and measuring this phase is critical. Traditional methods struggle with the required precision. This paper addresses the challenge of precisely measuring and controlling the relative phase between the absorber and multilayer reflections on these masks, particularly in the presence of subtle changes like surface contamination. The goal is to achieve picometer-level sensitivity to these changes, which directly impact mask performance and ultimately chip manufacturing.

Methods

The authors developed a computational reflectometry technique that models the physical properties of the mask layers. The core of their method involves:

  • Physical Modeling of Reflectometry Data: They use a multidimensional nonlinear least-squares regression to fit a detailed physical model of the mask's multilayer and absorber layers to experimental reflectometry data. This model incorporates layer thicknesses, refractive indices, and surface roughness.
  • Fresnel Coefficient Calculation: From the fitted physical model, they compute the Fresnel reflectance, which allows them to extract the relative phase between the absorber and multilayer.
  • Brute-Force Contamination Measurement: To assess changes in surface contamination (specifically carbon growth), they employ a simplified brute-force search. This involves freezing the physical model parameters and then evaluating a range of carbon thicknesses to find the one that minimizes the mean-squared error (MSE) between the model and the measured reflectivity. This approach is designed to be robust and guarantee convergence to a unique optimum for each measurement.

Datasets

The study utilized reflectometry data collected from a test mask at the Advanced Light Source Calibration, Optics Testing and Spectroscopy Beamline 6.3.2. This data consisted of reflectivity scans over a range of wavelengths (12.5 to 14.5 nm) and angles (4 to 8 degrees). Measurements were repeated over several months to observe changes due to storage and repeated experimental procedures.

Key Insights and Ideas

  • Picometer Sensitivity is Achievable: The most significant insight is the demonstration that reflectometry, when combined with sophisticated physical modeling, can achieve phase sensitivity at the picometer level, far exceeding previous capabilities.
  • Contamination is a Key Factor: The study highlights that even small amounts of surface contamination (like a thin carbon film) can significantly alter the relative phase between absorber and multilayer, impacting mask performance.
  • Interference Effects Dominate: The paper emphasizes that a simple "transmission-only" analysis is insufficient to understand phase changes. Destructive and constructive interference between reflected light waves from different layers plays a crucial role and must be accounted for by rigorous Fresnel analysis.
  • Storage vs. Exposure Contamination: The research revealed distinct contamination behaviors. Multilayers accumulated significantly more contamination during storage, while absorbers showed a more consistent growth during both storage and repeated measurements, likely due to EUV absorption.
  • Effective Propagation Distance: The analysis allowed for the extraction of an "effective propagation distance" for both multilayer and absorber reflections. This difference in distance has implications for mask 3D (M3D) effects and how patterns translate under different illumination conditions.

Key Results

  • Phase Precision: The method achieved a precision of 3σ < 0.1 degrees for phase measurement on the multilayer and 0.16 degrees for the absorber, after accounting for systematic trends. This translates to a wavefront precision of approximately 7.5 pm.
  • Thickness Precision: The precision for measuring contamination thickness was found to be 3σ < 6 pm, with specific estimates of 3.5 pm for the multilayer and 5.8 pm for the absorber after accounting for linear trends.
  • Contamination Growth: During storage, the multilayer showed a significant contamination growth of 1068 pm, while the absorber accumulated only 28 pm. During repeated measurements, multilayer contamination slightly decreased (-1.7 pm/measurement), while absorber contamination increased (12.3 pm/measurement).
  • Relative Phase Change: The total change in relative phase due to contamination was -0.86 degrees for the multilayer and -1.12 degrees for the absorber.
  • Model Accuracy: The initial physical model was capable of producing a highly accurate match to the experimental reflectivity data, demonstrating the validity of the approach.

Limitations and Future Work

  • Absolute Accuracy: The authors acknowledge that while their method demonstrates high repeatability and sensitivity, the absolute accuracy of the initial physical model is not mathematically guaranteed to find a unique optimal solution. Further studies are needed to assess the exact requirements for illumination conditions and prior knowledge to ensure absolute accuracy.
  • Need for Further Validation: Additional measurements, such as AFM for surface roughness and grazing incidence reflectometry for refractive index, may be necessary to fully guarantee the absolute accuracy of the method.
  • Exploration of M3D Effects: The paper suggests that the observed differences in effective propagation distance and phase variation across the source will have consequences for imaging and warrant further investigation and potential mitigation strategies.
  • Broader Application: While validated on a specific test mask, future work could explore the applicability of this technique to a wider range of EUV mask designs and materials.

Summary updated

Abstract

With growing interest in EUV attenuated phase shift masks due to their superior image quality for applications such as dense contact and pillar arrays, it is becoming critical to model, measure, and monitor the relative intensity and phase of multilayer and absorber reflections. We present a solution based on physical modeling of reflectometry data, which is capable of achieving single picometer phase precision. During repeated reflectometry measurements we observed a systematic change in absorber reflectivity which we attribute to the growth of a carbon film from 44-156pm, causing a change in the relative phase of 0.3°. This represents sensitivity to changes in the average film thickness to well below one atomic monolayer. After separating out systematic drift from random noise, we estimate our precision to be 3σ = 0.1°, corresponding to 3-4pm.
BibTeX
@misc{sherwin2020picometer,
  author = {Stuart Sherwin and Isvar A. Cordova and Ryan Miyakawa and Markus P. Benk and Laura Waller and Andrew R. Neureuther and Patrick Naulleau},
  title = {{Picometer sensitivity metrology for EUV absorber phase}},
  year = {2020},
  doi = {10.1117/12.2574759},
}

Loading related content...

Figure