George Tsirogiannis

We present EUV-OVL-SYN, an open and fully reproducible synthetic benchmark dataset for EUV lithography. Overlay error, the misalignment between consecutive patterning layers, is a primary yield-limiting factor at advanced nodes, yet the absence of publicly available, physically realistic datasets impedes systematic progress in data-driven correction research. EUV-OVL-SYN is built on documented scanner physics and replicates the signal hierarchy of real production data. The dataset captures three distinct levels of physical structure: a dominant inter-field Brink/van den Brink polynomial fingerprint (cross-wafer R2 ≥ 0.93); four lot-level temporal drift scenarios modeling linear scanner heating, reticle alignment drift, multi-parameter process drift, and chiller-cycling oscillation; and a structured nonlinear intra-field residual analytically verified to lie outside the Brink polynomial basis. The dataset consists of 4 lots of 50 wafers each, measured at 420 metrology targets per wafer (284 000 measurements in total), formatted identically to industrial automatic process control exports for direct compatibility with existing correction pipelines. The nonlinear residual (1σ ≈ 1.1–1.4 nm per axis) is well above the 0.15 nm measurement noise floor yet non-trivially masked by the dominant Brink fingerprint, providing a detectable but genuinely challenging correction target. Full ground-truth decomposition is provided for every signal component, permitting exact, method-agnostic evaluation of any correction approach. The complete dataset is generated by a single reproducible Python script and released under an open license.

The demand for protein‐fortified products has increased due to several factors, with a major one being the growing population of older adults worldwide. The aim of this study was to optimize the formulation of protein‐fortified tomato soup by adding milk protein concentrate (MPC) and pea protein isolate (PPI) individually or in combination. The combination was used to balance the nutritional limitations of plant proteins (such as the low levels of leucine) while enhancing the environmental sustainability of milk proteins. The key objective was to develop a tomato soup with at least 20% of the caloric content coming from protein, with similar physicochemical properties to a control soup, reduced CO₂ emissions, and high leucine content to meet the nutritional needs of older adults. The effects of various protein combinations on the physicochemical properties of the tomato soups were examined using a machine learning approach. The findings revealed that while it is challenging to simultaneously maximize leucine content and minimize carbon emissions, a formulation with double leucine content and a relatively low CO₂ footprint was achievable. The models demonstrated high accuracy for most physicochemical properties but encountered greater complexity for certain attributes, emphasizing the importance of considering both individual and combined protein effects. In conclusion, the study highlights the potential to use machine learning to blend proteins in a strategic manner with the aim to develop nutritious and sustainable food products, addressing the nutritional needs of older adults, while providing more environmentally friendly options in protein fortification.