Abstract: To address the limitations of traditional diffusion/oxidation control systems in terms of temperature control accuracy, process control capability, and automation, an intelligent diffusion/oxidation process control system is proposed. This system uses a programmable logic controller (PLC) as its core, a fuzzy self-tuning PID algorithm as its controller, and the Modbus protocol for communication between the PLC and various devices. Under this system's control, the temperature process curve of the diffusion furnace and auxiliary processes can be effectively controlled, and the PID parameters can be self-tuned online, achieving automation of the semiconductor diffusion process and improving temperature control accuracy and efficiency. Application results demonstrate that this temperature control system exhibits good adaptability and robustness. Keywords: diffusion/oxidation; programmable logic controller; fuzzy self-tuning PID; control accuracy Abstract: Aiming at the current conditions of traditional diffusion / oxidize control system which precision, ability to produce the control ability of the craft and automatic operation is the getting lower relatively, it is difficult to realize the current situation of centralized management, diffusion / oxidize the control system of the craft after putting forward a kind of intelligence. This system regard programmable logic controller (PLC) as core, with fuzzy self-turning PID algorithm for the controller, adopt Modbus agreement to realize PLC and every apparatus communication. Under the control of this system, can realize to the control spreading the temperature craft curve of stove and auxiliary process effectively, and can online to make PID parameter exactly since, realize semiconductor automation to spread craft, have improved the precision and working efficiency of temperature control. Employ the result to indicate, this temperature-controlled system has well adaptively and stupid and excellent. Key words: Diffusion/oxidize, Programmable logic controller, Fuzzy self-turning PID, Control the precision 0 Introduction High-temperature diffusion/oxidation systems are thermal processing equipment used in semiconductor device and integrated circuit manufacturing processes for diffusion, oxidation, annealing, and alloying of wafers. They are also suitable for special temperature processes on other materials and are automated control devices designed for long-term continuous operation with high precision and stability. In semiconductor production, the accuracy of diffusion furnace temperature control and its operational stability have become decisive factors in semiconductor product quality. However, in traditional control systems, diffusion furnace control and management are performed by microcontroller-based instruments, which have low temperature control accuracy, production process control capabilities, and automation capabilities, making centralized management difficult and resulting in poor product quality and low production efficiency. This system is designed and developed based on the current state of semiconductor diffusion/oxidation technology in China. It uses an advanced and reliable PLC as the control core, a touch screen as the human-machine interface, a high-precision temperature detection circuit, and a fuzzy self-tuning PID controller to achieve complex process temperature control. It features high detection and control accuracy, good operational stability, self-tuning control parameters, and convenient parameter setting. 1. Semiconductor Diffusion/Oxidation Process Flow The high-temperature diffusion/oxidation system mainly consists of a diffusion furnace, a clean bench, a push-pull boat system, and a gas source cabinet. The diffusion/oxidation process flow in semiconductor device manufacturing is as follows: 1) The temperature of the diffusion furnace is raised to a specific temperature according to a certain temperature process, and the furnace is kept at a constant temperature; 2) The wafer to be diffused is placed on the push-pull boat tray by the operator and fed into the diffusion furnace through the push-pull device; 3) Under the specific constant temperature condition of the diffusion furnace, various gases to be doped are injected into the diffusion furnace. The entire doping process must ensure the formation of a specific constant temperature zone within the furnace to achieve uniform wafer diffusion. Therefore, temperature control is the most important link in the diffusion process control system, and the quality of the control directly determines the quality of semiconductor diffusion. 2. System Hardware Design Based on the requirements of the semiconductor diffusion/oxidation process, namely, to automate the entire process control and meet the requirements of the semiconductor diffusion/oxidation process, the system uses a PLC as its core, an LCD touch screen as the human-machine interface, and a self-developed JC-9 temperature detection module to collect temperature data. The system hardware structure is shown in Figure 1. [align=center] Fig.1 System Hardware Structure Diagram[/align] As shown in Fig.1, the system hardware includes three closed-loop controls: ① Temperature Control: Temperature detection is achieved using a JC-9 temperature detection module. Temperature control is achieved by using a self-developed three-phase high-power voltage and power regulating unit controlled by a PLC using a PID control algorithm, thus realizing closed-loop control. Temperature control is the most important link in the diffusion process control system, and the quality of control directly determines the quality of semiconductor diffusion. To ensure the accuracy of temperature control, the accuracy of temperature detection must be guaranteed first, and a corresponding control algorithm is required. The detection accuracy of the JC-9 temperature detection module used in the system is better than 0.5‰, and a fuzzy self-tuning PID control algorithm is written to achieve a control accuracy better than 1‰. ② Gas Flow Control: The PLC directly reads the actual gas flow value from the digital interface of the gas mass flow meter and controls the flow output through the digital interface according to the gas flow requirements in the diffusion process, thus realizing closed-loop control. ③ Push-pull boat control: The PLC directly reads the data of the push-pull boat position encoder to determine the position of the push-pull boat, controls the stepper motor driver through the high-speed pulse output port to control the operation of the stepper motor, controls the position of the push-pull boat by sending the number of pulses, and controls the speed of the push-pull boat by the frequency of the pulse. Realize the position/speed closed-loop control of the push-pull boat. 3 System software design System software design includes touch screen operation interface design, fuzzy self-tuning PID control algorithm design and Modbus communication program design. 3.1 Touch screen operation interface[3] The touch screen is a human-machine dialogue interface with rich screen, large amount of information, flexible and intuitive operation. In view of the semiconductor diffusion process and system function requirements, the user interface is designed with the following main interfaces: automatic operation interface, manual operation interface, process parameter setting interface, process curve drawing interface, PID parameter self-tuning interface, etc. (1) Automatic operation interface. Realize the real-time display of temperature detection value, set value, various gas flow value, status of each valve, push-pull boat operation status and speed of each measuring point of the diffusion furnace, and complete the supervision function of the entire process. (2) Manual operation interface. Manual operation is suitable for system debugging, status testing and other applications. In addition to displaying the three-point temperature of the furnace body, other parts can be manually operated by buttons. The main contents include: the "forward" and "backward" operation control of the push-pull boat; the control of the four valves of the gas path; the control of hydrogen flow and nitrogen flow. In the "manual" screen, there are buttons for setting the forward and backward speed of the push-pull boat and setting the flow of various gases. Users can set according to requirements. (3) Process parameter setting interface. The system can store 20 sets of process formulas. Each process formula consists of 20 process segments. Each process segment can be divided into two steps: heating (cooling) and constant temperature. The process parameter setting interface can set the temperature value, duration, gas flow, push-pull boat position and other parameters for each step. (4) Process curve drawing interface. The parameters of the control process are displayed in digital form, and the three-point temperature of the furnace body and the set temperature value are displayed in the screen in the form of curves. (5) PID parameter self-tuning [4] interface. According to the temperature requirements of the process, the set temperature is selected, the system starts to heat the furnace body and starts the self-tuning process to realize the tuning of PID parameters. 3.2 Design of Fuzzy Self-tuning PID Control Algorithm [5][6] As the core of the system, the PLC completes the detection and processing of various information, the implementation of the control algorithm, and the output of control quantity. The system implements control including: process control of the nine-point temperature of the diffusion furnace; control of the gas flow process; and automated process control of the diffusion furnace production process. Among them, temperature control is crucial, and the quality of control directly determines the quality of semiconductor diffusion. This system uses fuzzy self-tuning PID control algorithm to realize temperature control. The flowchart of the fuzzy inference control algorithm program is shown in Figure 2. The calculation has the following steps: (1) Determine the error and the rate of change of error and the domain of the control quantity, divide the actual change range of the and into 7 levels (considering that the control system requires high control accuracy), so that each level corresponds to a certain element in its domain. In this way, each actual measurement of the system can be quantified as a certain element in the domain. (2) The system calculates the fuzzy control table offline. When running in actual operation, the PLC main program interrupts and executes the table lookup subprogram to obtain the control quantity. [align=center] Fig.2 Fuzzy reasoning subprogram[/align] In the field control, the fuzzy self-tuning PID controller has strong robustness, the parameter self-tuning is easy to implement, and the control accuracy is high. It has faster dynamic response characteristics than the simple PID controller, and its temperature control effect is better. 3.3 Modbus communication program design The system uses Modbus fieldbus to realize the communication between the PLC master station and the temperature detection module, gas mass flow meter and other slave stations. 1. Communication between master and slave stations with OMRON PLC as the master station[6] The system selects OMRON CP1H series PLC as the communication master station. CP1H is a powerful integrated small PLC launched by Omron. Its two serial ports have built-in Modbus-RTU master station functions, and it is also equipped with two serial communication option boards RS422/485 serial port option board and RS232 serial port option board. The system selects the JC-9 temperature acquisition module, gas mass flow meter, and push-pull boat as slave stations, all of which have RS485 interfaces and built-in Modbus protocol, providing Modbus-RTU slave station functionality. Master station communication can be achieved simply by following the programming manual's settings. 2. Slave Station Communication Flow: The system slave stations include the JC-9 temperature acquisition module, gas mass flow meter, and push-pull boat device. The gas mass flow meter and push-pull boat device are purchased with built-in Modbus protocol (not described here). The JC-9 temperature acquisition module is a self-developed nine-channel temperature acquisition module with an RS485 interface. Since the system communication uses Modbus communication mode, a custom Modbus slave station protocol needs to be defined on the module. The slave station receives commands from the master station and responds according to the master station's requirements. The communication interruption handling procedure between the slave and master stations is shown in Figure 3. [align=center]Figure 3. Modbus communication procedure of the sub-site[/align] 5. Conclusion The system is designed based on the diffusion/oxidation process requirements in semiconductor manufacturing. It features a simple structure, easy operation, low cost, and high reliability, significantly improving temperature control accuracy and process automation. It is now widely used in semiconductor manufacturing enterprises, yielding substantial economic and social benefits. References [1] Michael Quirk, Julian Serda. Semiconductor Manufacturing Technology [M]. Translated by Han Zhengsheng et al. Beijing: Electronic Industry Press, 2004, 445-447 [2] Liu Tao, Zhao Jiabao. Application of microcomputer measurement and control technology in semiconductor diffusion oxidation process equipment [J]. China Instrument and Control, 2005, (12): 94-96 [3] Zhong Zhaoxin. Programmable control: its principles and applications [M]. South China University of Technology Press, 2004 [4] Liu Yan, Zhang Tiankai, Huang Dong. Design and application of intelligent temperature control system for multi-station sintering furnace. Automation Instrumentation [J], 2006, (11): 48-50 [5] Wang Xianchun, Cai Jianhua, Hu Weiwen. Implementation of PID algorithm and parameter self-tuning in temperature control system. Automation Instrumentation [J], 2007, (2): 16-18 [6] Liu Guoguang. Adaptive temperature controller based on Fuzzy-PID. China Instrument and Meter, 2002, (5): 19-22 [7] Guo Jun, Liu Chang, Yao Fenglin. Implementation of PID controller based on PLC temperature control system. Mechanical Engineering and Automation [J], 2005, (3): 31-33 The first author, Zhou Hongli, female, born in 1982, is currently a master's student in the School of Computer Engineering of Qingdao University of Technology. Her main research direction is the design of computer control and detection systems.