Deep-learning-based SEM image denoiser (Conference Presentation)

Implementation of extreme ultra violet lithography (EUVL) for printing features of dimensions around 32nm and below has faced multiple challenges that have been overcome mostly through tool and process engineering. Currently, the main concern for implementation of EUVL in high volume manufacturing is related to stochastic defects, such as line bridges or breaks, which may occur randomly and have an impact on device performance.1 Photon shot noise, in addition to the inherent process variability and material non-homogeneities can lead to defects on the final patterned structures. As the pitch and critical dimension (CD) approach 32nm pitch and below, the dependency of stochastics defects increases exponentially. This paper fits within imec’s EUV stochastics and defectivity program, which investigates the key parameters for the generation and mitigation of defects from both a theoretical and practical approach. Initially, a reference process has been established and a methodology has been developed to capture and characterize bridge-type defects after pattern transfer into relevant layers for metrology and applications. Optical inspection and scanning electron microscopy (SEM) review are used to collect information on the number and types of defects present over large patterned areas (in the order of cm2). With a well-defined inspection technique, experiments related to fundamentals as well as materials and processes are continuously run. In the past years, imec has been leading the fundamental work on stochastic defects.2,3 De Bisschop et al. identified the “defect cliff”, or the exponential dependency of defect density as a function of space CD for multiple parameters, such as resist formulation and exposure/imaging conditions. More recently, the focus has changed to the investigation of the “defect floor”, which as the name suggests, is a flat response of defects to CD. The defect floor suggests that defectivity cannot be further improved, making one of the most relevant problems to solve for implementation of EUV in high volume manufacturing (HVM). Our results show a decomposition of the potential contributors to explore the effect of processing conditions related to scanner, track and pattern transfer on final defect density for a chemically amplified resist and attempt a comparison to a metal-oxide resist. Goal is to understand the contributors to the defect floor and identify an effective route to avoid/minimize defect formation.

In 2019, finally, extreme ultraviolet (EUV) lithography has been applied to high volume manufacturing (HVM) for preparing advanced semiconductor devices. That was very important year for EUV enthusiasts and semiconductor industry. Because it takes for a long time, more than 30 years, to study EUV lithography for realizing HVM. With recent rapid progress on the source power improvement, EUV lithography development including photoresist materials has been achieved HVM requirements. However, the performance of EUV resist materials are still not enough for the expected HVM requirements, even by using the latest qualifying EUV resist materials. One of the critical issues is the stochastic issues, which will be become ‘defectivity’, like a nano-bridge or a nano-pinching. The analyzing summary of the stochastic factors in EUV lithography, and their improvement status are described in this paper. 2 (two) major stochastic issues, which are photon stochastic and chemical stochastic, were observed in the lithography steps. And the improvement status of each stochastic issues including lithographic results are also described.

This work explores the application of alternative developer solutions (“developers”) with the aim of understanding their potential effectivity in the reduction of resist-based stochastic pattern defects (or “stochastic defects”) in extreme ultraviolet lithography (EUVL). Specifically, the application of a quaternary ammonium hydroxide type aqueous developer; ethyltrimethylammonium hydroxide (ETMAH) in comparison to the industry de facto standard aqueous alkali tetramethylammonium hydroxide (TMAH) developer was investigated. Focusing on EUV exposed contact hole (CH) patterns on a typical chemically amplified resist (CAR), the effect of these developers on stochastic defects were assessed. As a result, patterning investigations showed that the lithographic performance of the CAR developed in ETMAH is comparable to the those obtained with TMAH. In situ resist dissolution analysis using the high-speed atomic force microscope (HS-AFM) confirms this as it showed that the rate of CH formation during resist dissolution in both developers are relatively the same. Moreover, it was also understood that compared to the commonly used alkali developer concentration of 0.26N, a lower ETMAH developer concentration of 0.20N resulted in stochastic defect margin improvement, while maintaining lithographic performance. In situ resist dissolution analysis showed an obvious slowing down of CH pattern formation rate at 0.20N concentration, suggesting the possibility that of over-dissolution at the higher concentration condition, translating to an increase in merging CH defects. The results from this study show the advantages of further pursuing optimal developers for EUVL. This is especially significant as these findings indicate how optimal developers mitigate resist-based stochastic defects while maintaining lithographic performance.

Extreme ultraviolet (EUV) lithography is crucial to enabling technology scaling in pitch and critical dimension (CD). Currently, one of the key challenges of introducing EUV lithography to high volume manufacturing (HVM) is throughput, which requires high source power and high sensitivity chemically amplified photoresists. Important limiters of high sensitivity chemically amplified resists (CAR) are the effects of photon shot noise and resist blur on the number of photons received and of photoacids generated per feature, especially at the pitches required for 7 nm and 5 nm advanced technology nodes. These stochastic effects are reflected in via structures as hole-to-hole CD variation or local CD uniformity (LCDU). Here, we demonstrate a synergy of film stack deposition, EUV lithography, and plasma etch techniques to improve LCDU, which allows the use of high sensitivity resists required for the introduction of EUV HVM. Thus, to improve LCDU to a level required by 5 nm node and beyond, film stack deposition, EUV lithography, and plasma etch processes were combined and co-optimized to enhance LCDU reduction from synergies. Test wafers were created by depositing a pattern transfer stack on a substrate representative of a 5 nm node target layer. The pattern transfer stack consisted of an atomically smooth adhesion layer and two hardmasks and was deposited using the Lam VECTOR PECVD product family. These layers were designed to mitigate hole roughness, absorb out-of-band radiation, and provide additional outlets for etch to improve LCDU and control hole CD. These wafers were then exposed through an ASML NXE3350B EUV scanner using a variety of advanced positive tone EUV CAR. They were finally etched to the target substrate using Lam Flex dielectric etch and Kiyo conductor etch systems. Metrology methodologies to assess dimensional metrics as well as chip performance and defectivity were investigated to enable repeatable patterning process development. Illumination conditions in EUV lithography were optimized to improve normalized image log slope (NILS), which is expected to reduce shot noise related effects. It can be seen that the EUV imaging contrast improvement can further reduce post-develop LCDU from 4.1 nm to 3.9 nm and from 2.8 nm to 2.6 nm. In parallel, etch processes were developed to further reduce LCDU, to control CD, and to transfer these improvements into the final target substrate. We also demonstrate that increasing post-develop CD through dose adjustment can enhance the LCDU reduction from etch. Similar trends were also observed in different pitches down to 40 nm. The solutions demonstrated here are critical to the introduction of EUV lithography in high volume manufacturing. It can be seen that through a synergistic deposition, lithography, and etch optimization, LCDU at a 40 nm pitch can be improved to 1.6 nm (3-sigma) in a target oxide layer and to 1.4 nm (3-sigma) at the photoresist layer.

The rapid advancement of semiconductor manufacturing has driven the need for increasingly precise metrology and defect detection techniques. Extreme Ultraviolet (EUV) lithography, a key technology enabling the production of smaller and more efficient integrated circuits, introduces complex challenges in process control and defect inspection. Traditional methods struggle to keep pace with the heightened resolution and precision required for EUV-based semiconductor production. This paper explores the integration of machine learning (ML) techniques into EUV metrology and defect detection, offering a transformative approach to address these challenges. By leveraging advanced algorithms such as deep learning, neural networks, and data-driven models, we propose a new paradigm that enhances the detection of process-related defects, improves the accuracy of dimensional measurements, and provides real-time feedback for optimized manufacturing processes. The application of ML in this context not only enables more efficient defect classification and reduction but also offers the potential for predictive analytics that can proactively address emerging issues in EUV lithography. This shift towards machine learning-driven metrology represents a significant leap in semiconductor manufacturing, promising to enhance yield, reliability, and performance in next-generation integrated circuits.

In applying EUV lithography to 5-nm technology node and beyond, stochastic defect issues have arisen, seriously affecting manufacturing yield and/or limiting applicable minimum device sizes. To develop materials/processes with suppressing such defects and to bring them under control, we discuss a probabilistic model for stochastic defect generation processes. To deal with extremely low probability (typically 10^-4~10-¹²⁾ while being based on physics and chemistry of resist exposure, our model combines Monte Carlo method with probabilistic models. We focus on two mechanisms as causes of stochastic defects, (A) accidental connections of photon shot noises enhanced by densely localized secondary electron (SE) generation and (B) cascading SE generations along photoelectron trajectories. Our analyses show significantly higher defect probabilities in EUV than in ArF and their strong dependences on patterns sizes and process conditions, which is attributed to a wider spatial inhomogeneity in SE generation. Material parameter optimization by combining the model with the multi-objective genetic algorithm shows a trilemma among defect probability, target size, and required exposure dosage to size. It also shows necessity of scaling material parameters with shrinking target design rules. Guidelines for defect suppression are also suggested.

Extremely low probability stochastic defects are becoming a major concern in extreme-ultraviolet (EUV) lithography due to the discrete/probabilistic nature of photons, photo- and secondary electrons, reactions, and (macro-) molecular dissolutions. We introduce minimum complete models for stochasticity in EUV exposed resist films and predict the ultimate resolution of EUV lithography from a stochasticity viewpoint. Principal component analysis (PCA) is applied to three-dimensional chemo-physical event distributions in pattern-exposed resist films calculated by a fully coupled Monte Carlo simulation combined with discrete development/etching models. It expresses variabilities in the event distributions as linear combinations of the principal components (eigen-distributions), each of which reflects correlated reactions and clustering of (macro-) molecules. It successfully explains the fluctuations in local pattern edges and their material/process dependencies. Large size event distribution data are generated using the principal components, and stochastic defects are detected in the generated data. This virtual defect generation and inspection explains the observed probabilities and structures of stochastic defects. Thus, PCA provides general and universal representations for reaction stochasticity in resist films exposed by EUV lithography. It predicts that the ultimate resolution of EUV lithography in terms of stochasticity is limited by photoelectron/secondary electron scattering/diffusion or by photo-reaction density around the optical limit of 8 nm.

The following topics are dealt with: low-power electronics; Internet of Things; integrated circuit design; video coding; integrated circuit manufacture; CMOS integrated circuits; field programmable gate arrays; magnetic sensors; calibration; power consumption.

Intel's efforts to build a practical quantum computer are focused on developing a scalable spin-qubit platform leveraging industrial high-volume semiconductor manufacturing expertise and 300 mm fabrication infrastructure. Here, we provide an overview of the design, fabrication, and demonstration of a new customized quantum test chip, which contains 12-quantum-dot spin-qubit linear arrays, code named Tunnel Falls. These devices are fabricated using immersion and extreme ultraviolet lithography (EUV), along with other standard high-volume manufacturing (HVM) processes as well as production-level process control. We present key device features and fabrication details as well as qubit characterization results confirming device functionality. These results corroborate our fabrication methods and are a crucial step toward scaling of extensible 2D qubit array schemes.

The feasibility of using extreme ultra violet (EUV) lithography in high volume manufacturing makes the technology a very strong candidate for sub 20nm patterning. However defect control remains a major challenge even today. The aim of this paper is to understand propagation of the programmed defects present on the EUV mask to wafer to get an understanding of how stochastic defects may evolve through processes and how we can mitigate it. The evolution of the defects from mask to wafer post lithography and post etch has been studied both theoretically by calculating change in defect area from a script on images taken with e-beam inspection and other metrology techniques like atomic force microscopy. The end goal is to study the propagation of these programmed defects from post lithography to post etch on wafer through parameters like defect area, defect sizes and stack height information[1][2].

Extreme ultra-violet lithography lithography resolves features below 11 nm. However, photonic and atomic variations at these photon energies and dimensions lead to less than 1:109 potential stochastic defects causing device failures in stable manufacturing processes. This study investigates a methodology intended to identify root causes of stochastic defects with potential mitigation paths. Simulation techniques using pseudo random numbers are used to identify failing photonic and chemical event or distribution combinations. Failing combinations occurring in many photon-chemical configurations are thought to have potential mitigation methodologies. Photonic effects demonstrated significant impacts on stochastic defect formation with approximately 73% of the photon seeds resulting in a failure in at least 60% of the trials. The material results were mixed with large failure quantities that demonstrated low impacts. The photonic shot noise based failures were dominating in this study and these failures will not be mitigated by material enhancement alone.

Extreme ultraviolet lithography (EUVL) is the leading technology for patterning at the 32 nm technology node and be-yond. EUVL light at 13.5 nm is used to print circuits. This light is produced by heating fuel (Xe, Sn) in EUV sources to a very high temperature by using either magnetic compression or laser irradiation. Today EUV source power remains the number one concern for implementation of EUVL in high volume manufacturing. Over the last few years, much pro-gress has been made in EUV source performance and availability. Today, alpha level high power (~10 W) EUV sources have been integrated in alpha level EUVL scanners. Medium and low power EUV sources are used for in-house metrol-ogy and performance studies on EUV mask blanks, EUV masks, photoresists, and optical elements. These compact dis-charge sources with medium power in the range of 10-100 mW/sr/2% bandwidth and low power EUV tubes are being used by various RD today, these low power EUV sources are instrumental in allowing in-house R&D projects. In this paper, the latest status of high power EUV sources, low and medium power metrology sources, and some of their appli-cations are described.

Extreme-UV lithography (EUVL) is the most promising candidate of next generation technology for hp20nm node device manufacturing and beyond, however there are many critical issues to solve in the light source, tools, mask and photo resist. Regarding the development of a photo resist, it is necessary for high volume manufacturing (HVM) to improve LWR, resolution limit and sensitivity. Additionally, concerning about deterioration of a patterning performance by Out-of-Band (OoB) light existing in the EUV light, and contamination problem of exposure tool due to the photo resist outgassing are the key issues which have to be resolved toward HVM by EUVL. Especially, the outgassing problem can become a significant issue for fine patterning under high dose exposure condition. This paper proposes the novel solution for these critical issues with the application of a top coat material which is named OBPL (Outgassing & out-of-Band Protection Layer) on photo resist. The key characteristics of OBPL material are to have a role in protection against the OoB adverse effect to keep up the photo resist performance, to suppress the outgassing from photo resist as a barrier layer and to enhance the lithographic performance such as photo resist profile and process window. In designing the OBPL material, the optical property needs having not only the high absorbance of DUV (Deep-UV) light in OoB range but also high transmittance for 13.5nm wavelength to prevent the sensitivity loss. Furthermore, it is found that the polymer backbone affects the outgassing barrier capability in previous evaluation. Based on these investigations, a state-of-the-art OBPL achieves quite a positive lithographic result with sufficient OoB absorption and outgassing suppression. Moreover, this material has applicability to all types of photo resist including NTD (Negative-Tone Development) process. This paper describes the result of the feasibility study for OBPL and the lithography performance with EUV full field scanner.

The effects of photoresist dissolution on randomly occurring (“stochastic”) pattern defects in extreme ultraviolet (EUV) lithography were investigated. Specifically, the effects of the alkali concentration of two developers of different alkyl chain lengths; tetramethylammonium hydroxide and tetrabutylammonium hydroxide (TBAH) on stochastic defect generation were investigated for two typical EUV photoresists. In the case of the PHS-type photoresist, results show that stochastic defect generation was almost the same regardless of developer type, given the same developer concentration. For the hybrid-type photoresist, a decrease in stochastic defects was found with the application of the longer alkyl chain length TBAH developer. Results also suggest the existence of an optimum developer concentration. Such optimum developer concentration is considered to be affected by the balance between polarity and non-polarity of the photoresist and developer components.

For designing high-resolution integrated circuits, extreme ultraviolet (EUV) lithography using 13.5-nm light is indispensable; however, stochastic issues which can result in pattern roughness and stochastic defects have yet to be resolved. Many of these issues can relate to the low photon density when compared to DUV lithography. To resolve stochastic concerns in chemically amplified (CA) photoresists, it is important to use appropriate photoacid generator (PAG). EUV PAGs are designed to capture secondary electrons that originate from EUV irradiation to the photoresists. Herein, we focused on in silico analysis of C-S bond cleavage of sulfonium cation series, which is a well-known mechanism in production of photogenerated acids. Change in cation types or substituted functional groups directly affect electron affinities of PAG cations, as well as the efficiency of C-S bond cleavage reaction. Finally, the correlation between actual lithographic performance and calculated properties was carried out to understand the factors that influence efficiency of photoacid generation.