In recent years, with the rapid development of the global information technology industry and the increasingly prominent core position of semiconductors in national strategy, China’s integrated circuit industry has achieved leapfrog growth. The industry scale has expanded from less than 10 billion USD in 2005 to several hundred billion USD, making it one of the largest semiconductor consumer markets in the world. Against this backdrop, the construction quality of semiconductor manufacturing facilities, especially cleanrooms, is directly related to the yield, performance, and reliability of chip production. Therefore, the engineering design, construction, and overall quality management throughout the process have received significant attention from the industry. As the core environment for semiconductor manufacturing, cleanrooms have extremely stringent requirements for the control precision of factors such as particle count, microbial content, temperature and humidity, air pressure, airflow organization, electrostatic discharge, and micro-vibration. Traditional cleanliness levels of 10,000 and 1,000 are no longer sufficient to meet the needs of advanced processes, with integrated circuit manufacturing typically requiring ISO Class 1 or higher standards. This results in cleanroom engineering exhibiting significant characteristics such as enormous initial investment, complex technology integration, deepening of multiple professional intersections, tight construction cycles, and high difficulty in risk management. Additionally, the need to handle high-purity chemicals, special gases, precision electromechanical systems, intelligent monitoring, and anti-vibration technologies during construction further complicates quality control and introduces uncertainty. If not managed properly, this can lead to project delays, increased costs, reduced product yield, frequent equipment failures, or even project failure, resulting in irretrievable economic and reputational losses. Therefore, based on traditional quality management theories, integrating lean management concepts, and systematically constructing a refined, standardized, and dynamically optimized quality management system covering the entire lifecycle of semiconductor cleanroom EPC projects is not only an intrinsic requirement for improving engineering quality but also an urgent requirement for promoting the healthy and sustainable development of the national integrated circuit industry.
This study focuses on the “EPC engineering of semiconductor cleanrooms” as the specific research object, emphasizing the construction and practical application of a lean quality management system during the construction process. The paper first comprehensively reviews the theoretical evolution and practical status of engineering quality management both domestically and internationally, from Taylor’s scientific management movement, statistical quality control, total quality management (TQM), to lean construction theory, systematically tracing the development of quality management thought and analyzing common issues in industrial plant and cleanroom engineering quality management, such as inadequate implementation of systems, non-standard acceptance, poor cross-stage collaboration, and insufficient specialized capabilities. Based on this, the study combines modern project quality management theories with the characteristics of the EPC general contracting model to propose a lean quality management system framework for the entire construction process of high-tech cleanrooms. This system emphasizes the core concept that “quality is built, not inspected,” focuses on early planning, preventive control, and continuous improvement, and pursues a zero-defect goal. It relies on practical engineering cases—Project G (GK Semiconductor’s 12-inch CIS integrated circuit R&D and industrialization project in Shanghai)—for empirical analysis and validation, thereby forming a comprehensive quality management solution from theory to practice, from overall planning to detailed operations.
In the planning phase, the study clarifies the primary position of lean quality planning, stating that quality policies and objectives must be scientific, measurable, achievable, and challenging, and must be systematically broken down to each department and even individuals. By using Fault Tree Analysis (FTA), the study decomposes and calculates the structural importance of the item “poor operation of the cleanroom engineering quality management system,” identifying key factors affecting quality in the following order: disconnection between quality policies and objectives and actual conditions, failure to establish an effective management system, and top-level design issues of the quality control system that are not systematic or difficult to implement. Furthermore, the study proposes refined control strategies targeting the five elements of “people, machines, materials, methods, and environment”: in terms of personnel, it emphasizes certified operation, systematic training, and behavior management; for machinery, it focuses on equipment status identification, regular inspection, and maintenance systems; material management emphasizes purification process control, storage protection, and incoming inspection; methodologically, it strictly implements special construction plans and technical disclosures; and environmental management centers around clean environment assurance measures such as dust prevention, water protection, vibration prevention, and electrostatic discharge prevention. At the same time, the study establishes a four-level quality management organizational structure in Project G, clarifying the responsibilities of each position from project manager to quality engineer, and formulates twelve quality management systems, including technical disclosures, drawing reviews, pre-planning, three-inspection system, quality meetings, sample guidance, concealed acceptance, measurement management, and document control, providing systematic and organizational guarantees for achieving project quality objectives.
Quality management during the design phase is the source of engineering quality. This study points out that design outcomes must not only comply with national standards and contract requirements but also meet the functional positioning of production processes, operational reliability, and ease of maintenance. The study proposes a multi-dimensional evaluation standard for design quality covering functionality, reliability, safety, economy, and feasibility, and aims to construct a quality control system covering the entire design process. Given the complexity of semiconductor plant processes and the numerous professional interfaces, it emphasizes the importance of design interface management and cross-professional collaboration, proposing a three-dimensional design coordination and conflict detection method based on BIM technology to reduce design changes during the construction phase. In Project G, by establishing a unified design standard library, implementing a graded design review process (including regular and non-regular reviews), and strengthening information management of the design process (relying on the DMCS system for document control and process tracking), the accuracy and completeness of design outputs have been effectively improved. Additionally, in addressing technical challenges such as AMC (gaseous molecular contaminants) control, external pollution source prevention, micro-vibration sensitive area design, and energy optimization (e.g., large temperature difference water systems, heat recovery technology), the study implements special technical reviews and CFD simulation-assisted design strategies to ensure the technical advancement and construction feasibility of design plans. Furthermore, through a strict design change management process (involving change proposals, technical reviews, cost assessments, approval implementation, and document update closed-loop management), it minimizes quality risks and cost deviations arising from design changes.
Quality control during the procurement phase is a core link to ensure that engineering materials and equipment meet quality requirements. The study emphasizes the standardization, compliance, comprehensiveness, and executability of the procurement process under the EPC model, proposing a supplier lifecycle management system centered on quality standards. This system includes supplier development (qualification review, on-site inspection, comprehensive evaluation), performance evaluation (scoring based on quality, delivery time, price, technical capability, service response, etc.), and classification management strategies (drawing on the Kraljic matrix to categorize suppliers into strategic, leverage, bottleneck, and routine types for differentiated management). In Project G, by establishing detailed supplier evaluation standards and processes, a cross-departmental collaborative supplier review mechanism has been formed, ensuring high-quality supply of key equipment (such as MAU, FFU, AMC filtration devices) and materials (such as PVDF pipes, epoxy flooring coatings). On the other hand, the study combines trends in information technology development to propose a construction plan for a refined procurement information management platform, which integrates core functions such as system management, procurement planning, supplier management, online bidding, contract management, and material ledgers, achieving full online management and traceability of procurement needs, bidding, orders, acceptance, payment, and supplier data, significantly improving the transparency, efficiency, and controllability of the procurement process, thereby providing supply chain-level guarantees for overall project quality.
The construction phase is the main battlefield for quality management. This study integrates lean construction and PDCA cycle concepts to construct a construction quality control model based on the “five elements” control, with the Inspection and Test Plan (ITP) as the execution mainline and continuous improvement as the goal. In terms of “people,” Project G ensures that the capabilities and awareness of the workforce meet the requirements for high-clean environment construction through strict training by trade (including quality awareness, standards, and special skills training), implementing dynamic labor planning and behavior management; in terms of “machines,” it establishes systems for equipment entry inspection, status identification (qualified/in repair/archived/scrapped), and regular maintenance to ensure the precision and reliability of construction tools; in terms of “materials,” it strengthens incoming material inspection, storage protection, and purification treatment, strictly executing material approval and traceability processes to eliminate the use of unqualified materials; in terms of “methods,” it focuses on developing special plans and quality control points for key processes such as cleanroom ceiling installation precision control, raised floor flatness assurance, epoxy flooring construction, and duct fabrication and installation, ensuring that process execution is in place through sample guidance, process handover inspections, and on-site supervision; in terms of “environment,” it implements clean control measures for the construction environment, including personnel and material purification, dust control, temperature and humidity monitoring, and finished product protection, ensuring that the on-site construction environment consistently meets cleanliness requirements. At the same time, the study emphasizes the core role of ITP in construction quality control, detailing the entire process inspection nodes, standards, and methods from material inspection, process checks, system testing to final acceptance, and establishing processes for controlling non-conforming products and correcting deviations (including classification, cause analysis, measure formulation, effect verification, and record archiving), achieving closed-loop management of quality issues and ensuring that engineering quality remains controlled and approaches the zero-defect goal.
The completion acceptance and handover phase is the final checkpoint for engineering quality. This study proposes a three-level acceptance system covering normative acceptance, user infrastructure functional acceptance, and user production functional acceptance, clarifying the hierarchical acceptance process and participation responsibilities from inspection batches, sub-projects, and component projects to unit projects. In Project G, by formulating detailed pre-acceptance and formal acceptance plans and organizing participation from construction, design, construction supervision, and other parties, a comprehensive verification and testing of the quality of the engineering entity, functional performance, and completion documentation is conducted, focusing on completing pressure tests, equipment debugging, comprehensive performance testing of the cleanroom (such as cleanliness, airflow organization, temperature and humidity, illumination, noise, electrostatic discharge, etc.), and system linkage trial runs. For defects identified during the acceptance process, the study introduces a defect management system, classifying issues by urgency (Class A critical items/Class B general items) and tracking closed-loop rectification to ensure that all issues are effectively resolved before handover. In terms of handover management, the study emphasizes the completeness and accuracy requirements of technical document handover, including as-built drawings, operation and maintenance manuals, test reports, certificates of conformity, and spare parts lists, and establishes a handover process based on rooms (including three phases: construction integrity inspection, pre-inspection, and final inspection) to ensure that each area meets design and usage requirements. Finally, the study also formulates a work plan for the project maintenance phase, clarifying the composition of the maintenance team, division of responsibilities, service processes, and response mechanisms, achieving seamless connection from project construction to operational maintenance, ensuring the long-term stable operation of cleanroom facilities.
This study constructs a complete, systematic, and operable lean quality management system for semiconductor cleanroom EPC engineering through systematic theoretical research and in-depth empirical analysis. This system covers the entire process of project planning, design, procurement, construction, and acceptance handover, integrating multidisciplinary knowledge such as lean management, TQM, risk management, and information technology, emphasizing goal orientation, process standardization, dynamic monitoring, and continuous improvement, providing important theoretical support and practical guidance for quality management of high-tech plant engineering. In the successful application of Project G, this system effectively improved the first-pass acceptance rate of engineering quality, reduced quality risks and cost waste, ensured timely and quality project delivery, and achieved good economic and social benefits. At the same time, the study also points out that the current system still leans towards quality control during the construction phase, and in the future, lean quality management can be further extended to the design planning phase and down to the operation and maintenance phase, achieving truly integrated quality management across the entire lifecycle, providing more comprehensive and in-depth support for the high-quality construction of semiconductor industry infrastructure in China.
This article is a condensed version of Ms. Zhang’s master’s thesis from Tongji University, and all rights belong to her.