Achieving optimal parameter settings in epitaxial silicon carbide (Epi SiC) manufacturing is challenging due to the need to simultaneously meet conflicting multi-objective requirements, such as precise thickness control and uniform doping. This paper presents an extended Digital Twin (DT) framework that incorporates the Multi-Objective Optimization with Deep-Feature Gaussian Process (MOODFG) algorithm, explicitly extending our prior DT-in-the-loop framework (MRBORI) in [26] from single-objective tuning to spec-driven multi-objective recipe optimization. The framework enables the effective optimization of high-dimensional and interdependent process parameters, overcoming limitations of traditional methods. The key innovation of the MOODFG algorithm lies in its integration of deep feature extraction and Gaussian Process Regression, which allows dynamic refinement of feature representations and surrogate models through iterative learning. Unlike MRBORI [26] (one objective per run), MOODFG minimizes a weighted distance to a user-specified target vector (e.g., thickness and doping) and provides an uncertainty-aware target-attainment certificate with a practical stopping rule for deployment. This approach ensures robust convergence to optimal parameter values while efficiently balancing multiple objectives. Experimental validation using real-world Epi SiC manufacturing data demonstrates significant improvements in yield, parameter stability, and process adaptability, highlighting the framework’s transformative potential for semiconductor manufacturing.

High-precision wafer metrology poses significant cost and throughput challenges in modern semiconductor manufacturing, where frequent process changes and recipe variations demand highly adaptive and scalable solutions. In this paper, we present a Generative-FewShot-Active Virtual Metrology (GFA-VM) framework that unifies large-scale generative modeling, few-shot fine-tuning, and uncertainty-driven active sampling into a single, data-centric system. A foundational generative model, built on a hybrid architecture of Transformer networks and Variational Autoencoders (VAEs), learns diverse sensor characteristics in an offline stage without relying on extensive labeled data. During online inference, the model produces both wafer quality predictions and predictive uncertainties; samples exceeding a dynamic uncertainty threshold are selected for physical measurement and few-shot model recalibration. This selective sampling both reduces measurement costs and adapts rapidly to new process conditions (e.g., novel recipes or equipment upgrades), requiring only a handful of freshly labeled wafers. The paper further addresses the long-term stability of the system through a self-updating mechanism that adjusts the uncertainty threshold when distributional shifts occur. Empirical evaluations confirm that our GFA-VM approach achieves state-of-the-art accuracy while significantly reducing metrology overhead compared to conventional virtual metrology methods. Additionally, rigorous theoretical analyses—including proofs of convergence and label cost bounds—demonstrate the reliability of using a generative foundation plus meta-learning technique. By fostering on-demand adaptation within a closedloop framework, GFA-VM offers a comprehensive, scalable strategy for next-generation semiconductor process control.

The concepts of Industry 4.1 for achieving Zero-Defect (ZD) manufacturing were disclosed inIEEE Robotics and Automation Lettersin January 2016. ZD of all the deliverables can be achieved by discarding the defective products via a real-time and online total inspection technology, such as Automatic Virtual Metrology (AVM). Further, the Key-variable Search Algorithm (KSA) of the Intelligent Yield Management (IYM) system developed by our research team can be utilized to find out the root causes of the defects for continuous improvement on those defective products. As such, nearly ZD of all products may be achieved. However, in a multistage manufacturing process (MMP) environment, a workpiece may randomly pass through one of the manufacturing devices with the same function in each stage. Different devices of the same type perform differently in each stage, where the performances will be accumulated through the designated manufacturing process and affect the final yield. KSA can only identify the influence of univariate variables (i.e., single devices) on the yield, yet it cannot detect the manufacturing paths that have significant influence on the yield. In order to cope with this deficiency such that the golden path with a better yield amongst all the MMP paths can be found, this research proposes the Golden Path Search Algorithm (GPSA), which can plan golden paths with high yield under the condition of the number of variables being much larger than that of samples. As a result, it makes the improvement of manufacturing yield be more comprehensive.Note to Practitioners—Traditional scheduling only considers the capacity of the manufacturing devices for allocation; while, the impact on the yield is rarely considered. In fact, in a production process, the production deviations will be gradually accumulated and affect the product quality along with the processing influence of each device. Therefore, the purpose of this paper is to propose the GPSA scheme to quickly search for high-yield manufacturing paths before the production. Manufacturers can then configure production devices based on these paths. According to the experimental results of real manufacturers’ data, GPSA can not only quickly nail down the high-yield paths from a large amount of historical data, but also alert the users to avoid paths that are prone to defective rates for their reference.