Judgment method for gas environment safety of general furnace chamber for semiconductor manufacturing

Abstract

The invention discloses a method for judging the gas environment safety of a general furnace chamber for semiconductor manufacturing. The method comprises the following specific steps: step 1, according to process design characteristic parameters of a target furnace chamber, definition for safety in a safety state judgment device and an algorithm model in a gas state inference engine; 2, characteristic parameters of a target furnace chamber are obtained in real time; 3, the process data obtained in the step 2 are analyzed through a gas state inference engine, and an analysis result is sent to a safety state judgment device; 4, comparing the result output in the step 3 with the safety definition by a safety state judging device to obtain a judging result that the target furnace chamber is in a safe state or an unsafe state; and 5, sending a judgment result obtained in the step 4 to a control system, and carrying out the next operation according to the judgment result.

Description

Technical Field

[0001] The present application relates to the field of semiconductor manufacturing technology, and in particular to a method for determining the safety of a general furnace chamber gas environment in semiconductor manufacturing. Background Art

[0002] In the semiconductor device manufacturing process, various types of process furnaces / chambers play a crucial role. These furnaces include, but are not limited to, diffusion furnaces, oxidation furnaces, annealing furnaces, chemical vapor deposition furnaces (CVDFurnaces, including LPCVD, PECVD, ALD, etc.), and etch chambers. These processes typically require a strictly controlled gas environment, involving the precise delivery, mixing, and reaction of multiple gases.

[0003] Among these gases, some are inert or relatively safe, such as nitrogen (N2) and argon (Ar), commonly used as carrier gases, protective atmospheres, or for purge cleaning. Others are reactive, flammable, explosive, toxic, or corrosive gases necessary for the process, such as hydrogen (H2), ammonia (NH3), silane (SiH4), dichlorosilane (DCS), methyltrichlorosilane (MTS), various organometallic precursors (such as TEOS), oxidants (such as O2, O3, N2O), etching gases (such as fluorocarbons such as Cl2, HBr, SF6, NF3, and CHF3), and doping gases (such as PH3, AsH3, and B2H6).

[0004] Ensuring the safety of the gas environment within these process chambers during all stages of operation (including processing, process switching, standby, maintenance, and exception handling) is a top priority in semiconductor manufacturing. Accidental incorporation of hazardous gases into the chamber, incomplete inert gas purges resulting in residual hazardous gases, or gas leaks can cause combustion, explosion, equipment damage, product contamination, and even pose serious safety threats to operators.

[0005] Currently, semiconductor process equipment is typically equipped with mass flow controllers (MFCs) to precisely control the flow of various gases, as well as pressure and temperature sensors to monitor the physical conditions within the furnace chamber, thereby ensuring a safe environment. Existing technologies typically design targeted monitoring systems for specific processes and applications. For example, temperature control systems for specific process chambers (such as US6015465A, which uses fluid film evaporation for temperature regulation) or process monitoring systems (such as US20090314205A1, which uses a vision system to monitor chamber components) monitor conditions within the furnace chamber. Leak detectors or concentration analyzers are also available for specific gases. In some specific applications, MFCs and exhaust gas analysis are used to control specific atmosphere compositions.

[0006] However, these existing technologies often have the following limitations: 1. Strong specificity and lack of versatility: Many safety monitoring or atmosphere control methods are designed for specific types of furnace chambers, specific processes, or specific gases. There is a lack of a universal gas safety judgment framework that can be widely applied to a variety of different semiconductor process furnace chambers; 2. Reliance on direct composition analysis: Some precise atmosphere control or safety monitoring methods may rely on expensive and complex-to-maintain online gas composition analyzers (such as mass spectrometers, gas chromatographs, etc.), which not only increases monitoring costs but also makes the monitoring system more complex. Summary of the Invention

[0007] In response to the above-mentioned deficiencies in the prior art, the purpose of this application is to provide a method for determining the safety of the gas environment in a general-purpose furnace chamber for semiconductor manufacturing, so as to solve the problems in the prior art of highly specialized and insufficiently versatile methods for monitoring the internal environment of the furnace chamber, the difficulty of the existing monitoring methods in monitoring a variety of different semiconductor process furnace chambers, and the reliance of the prior art on expensive and complex-to-maintain online gas composition analyzers for controlling the internal atmosphere of the furnace chamber, which increases the monitoring cost and makes the monitoring system more complex.

[0008] In order to solve the above technical problems, this application adopts the following technical solutions: In a first aspect of the present application, a method for determining the safety of a general furnace gas environment in semiconductor manufacturing is provided, wherein the specific steps are as follows: Step 1: Based on the process design characteristic parameters of the target furnace chamber, the definition of safety in the safety state determiner, and the algorithm model in the gas state inference engine; Step 2: Acquire the characteristic parameters of the target furnace chamber in real time; Step 3: The process data obtained in step 2 is analyzed by the gas state inference engine, and the analysis results are sent to the safety state determiner; Step 4: The safety state determiner compares the result output from step 3 with the definition of safety to determine whether the target furnace chamber is in a safe state or an unsafe state; Step 5: Send the judgment result obtained in step 4 to the control system, and perform the next operation based on the judgment result.

[0009] Preferably, the target furnace chamber is a process furnace chamber in a semiconductor device manufacturing process.

[0010] Preferably, the process furnace chamber includes a diffusion furnace, an oxidation furnace, an annealing furnace, a chemical vapor deposition furnace and an etching reaction chamber.

[0011] Preferably, in step 2, characteristic parameters are obtained in real time from the PLC or sensor network of the target furnace chamber; wherein the characteristic parameters include the target furnace chamber volume, the type of gas in the target furnace chamber, the exhaust gas characteristics and the data obtained by the mass flow controller.

[0012] Preferably, the definition of safety includes the types of safe gases and their corresponding concentration thresholds, and the types of unsafe gases and their corresponding concentration thresholds.

[0013] Preferably, the algorithm model includes mixing efficiency, permutation model selection and judgment rules.

[0014] Preferably, in step 3, the following operations are performed by the gas state inference engine: Step 1: Determine the initial gas concentration based on the target furnace chamber inlet gas accumulation calculator; Step 2: Call at least one model from the configurable gas displacement and concentration estimation model library to calculate the gas concentration change; Step 3: Correct the gas flow and volume parameters through the target furnace chamber pressure and temperature influence corrector; Step 4: Analyze abnormal gas fluctuations through flow balance and preliminary leakage diagnosis logic.

[0015] Preferably, in step 5, if the judgment result is a safe state, the next process is carried out; if the judgment result is an unsafe state, the control system displays that the target furnace chamber is in an unsafe state, and at the same time, determines whether to trigger an alarm signal based on the process characteristics of the target furnace chamber.

[0016] In a second aspect of the present application, a monitoring system for the gas environment safety of a general-purpose semiconductor manufacturing furnace chamber is provided, which implements the above-mentioned judgment method; the monitoring system includes a database, a real-time data acquisition module, a gas state inference module, a safety state determination module, and a result output and response module; The database is used to store and manage specific parameters, safety definitions, and algorithm models of the target furnace chamber, so that the gas state inference module can select the corresponding algorithm model from the database and the safety state determination module can obtain the safety definitions in the database; The real-time data acquisition module obtains real-time data from the PLC or sensor network of the target furnace chamber through the interface and sends the data to the gas state inference module; The gas state inference module receives the real-time data sent by the real-time data acquisition module, analyzes the real-time data according to the algorithm model obtained from the database, and sends the analysis results to the safety state determination module; The safety state determination module receives the analysis results sent by the gas state inference module, compares them with the safety definition in the database, determines whether the state in the target furnace chamber is safe or unsafe, and sends the judgment result to the result output and response module; The result output and response receives the judgment result sent by the safety status judgment module and transmits the judgment result to the upper control system; at the same time, when the judgment result is unsafe, it determines whether to trigger an alarm signal based on the process characteristics of the target furnace chamber.

[0017] Compared with the prior art, this application has the following beneficial effects: 1. The judgment method described in this application adopts a configurable framework design, so that the same set of core judgment methods can be applied to semiconductor manufacturing furnaces of various types, processes, and gas formulas, reducing the complexity and cost of developing and verifying safety monitoring logic for each device separately, and has strong versatility and adaptability.

[0018] 2. The judgment method described in this application can be implemented based on existing sensors, mainly relying on the performance of MFC, pressure, temperature sensors and exhaust gas flow monitoring (or estimation) commonly equipped in semiconductor furnaces. There is no need to forcibly equip each furnace with an expensive and complex-to-maintain dedicated online gas composition analyzer, and effective judgment of the safety of the gas environment can be achieved.

[0019] 3. The judgment method described in this application comprehensively analyzes real-time data from multiple different physical dimensions and makes inferences based on the gas dynamic model. Compared with simple alarms based on a single parameter threshold, it can more comprehensively and accurately evaluate the overall safety status of the complex gas environment in the furnace. Especially in dynamic processes such as gas switching and purging, the accuracy and reliability of the judgment method are significantly improved by making a comprehensive judgment after multi-parameter analysis.

[0020] 4. The judgment method described in this application can perform continuous or high-frequency real-time judgment on the furnace chamber gas environment. Once an unsafe state or potential risk is found, it can quickly output the results and trigger corresponding measures to effectively prevent the occurrence of safety accidents; at the same time, the judgment method described in this application allows users to flexibly define "safe gas", "unsafe / process gas" and their safety thresholds according to specific process requirements and safety standards, so that the judgment criteria are more in line with actual needs.

[0021] 5. The judgment method described in this application has the ability to assist in the diagnosis of abnormal conditions. Through flow balance analysis and abnormal pressure and temperature fluctuation monitoring, it can, to a certain extent, assist in determining whether there are equipment problems such as gas leakage, MFC failure, and exhaust abnormality, and provide clues for equipment maintenance. It can also be linked with other equipment and write to the PLC specified address to assist in safety judgment, and is very practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a flow chart of a method for determining the safety of a general furnace gas environment in semiconductor manufacturing. DETAILED DESCRIPTION

[0023] This application will provide a clear and complete description of the technical solutions in the embodiments of this application in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on this application are within the scope of protection of this application.

[0024] Unless otherwise indicated in specific cases in this application, the numerical ranges listed herein include the upper and lower limits, as well as all integers and fractions within the range, and are not limited to the specific values ​​listed when defining the range.

[0025] 1. A method for determining the safety of the gas environment in a general-purpose semiconductor manufacturing furnace Step 1: Based on the process design characteristic parameters of the target furnace chamber, the definition of safety in the safety state determiner, and the algorithm model parameters in the gas state inference engine; Step 2: Acquire the characteristic parameters of the target furnace chamber in real time; Step 3: The process data obtained in step 2 is analyzed by the gas state inference engine, and the analysis results are sent to the safety state determiner; Step 4: The safety state determiner compares the result output from step 3 with the definition of safety to determine whether the target furnace chamber is in a safe state or an unsafe state; Step 5: Send the judgment result obtained in step 4 to the control system, and perform the next operation based on the judgment result.

[0026] After studying the commonly used monitoring methods for existing semiconductor process equipment, the present applicant found that although current furnace chambers are generally equipped with mass flow controllers (MFCs), pressure, temperature and other sensors, when it is necessary to monitor the chemical safety of the gas environment inside the entire furnace chamber, the existing technology still relies on component analysis, and does not consider how to systematically and comprehensively utilize these real-time data from different physical quantities to dynamically evaluate the chemical safety of the gas environment inside the entire furnace chamber through an indirect, inferential method (for example, to determine whether it is a safe inert atmosphere or an atmosphere containing specific hazardous components), especially when it does not rely on direct component analysis. To this end, the present applicant conceives of utilizing these real-time data and designs a judgment method that can make timely and accurate judgments on the internal environment of the furnace chamber through these real-time data from different physical quantities without relying on direct component analysis.

[0027] In some embodiments of the present application, the target furnace chambers mainly targeted by the judgment method described in the present application are process furnaces / chambers in the semiconductor device manufacturing process; specifically including diffusion furnaces, oxidation furnaces, annealing furnaces, chemical vapor deposition furnaces (CVD furnaces, including LPCVD, PECVD, ALD, etc.), etching reaction chambers, etc.

[0028] In some embodiments of the present application, in step 2, characteristic parameters are acquired in real time from a programmable logic controller (PLC) or sensor network installed in the target furnace chamber. These characteristic parameters include the target furnace chamber volume, target furnace chamber gas type, exhaust gas characteristics, and data acquired by a mass flow controller. The characteristic parameters are selected based on the application scenario of the target furnace chamber.

[0029] In some embodiments of the present application, the definition of safety includes the types of safe gases and their corresponding concentration thresholds, and the types of unsafe gases and their corresponding concentration thresholds. Different furnace chambers have different definitions of safety. For example, taking an LPCVD furnace (such as silicon nitride deposition) as an example, when making a safety judgment on its purge, it is necessary to ensure that the residual concentration of toxic and flammable gases such as DCS and NH3 in the furnace tube drops below the safe level before subsequent operations can be carried out; but for a plasma etching chamber (such as polysilicon etching), when making a standby safety judgment on it, it is necessary to ensure that the wall is in an inert atmosphere, and at the same time monitor whether there is any process gas leaking into the chamber from the pipeline valve. It can be seen from this that it is necessary to set a targeted definition of safety based on the application scenario of the target furnace chamber, which also shows that the judgment method described in this application is highly flexible and adaptable.

[0030] In some embodiments of the present application, the algorithm model includes basic parameters, mixing efficiency, replacement model selection, judgment rules, PLC linkage address, etc. The selection of the algorithm model is also adjusted according to different application scenarios of the furnace chamber.

[0031] In some embodiments of the present application, in step 3, the following operations are performed by the gas state inference engine: Step 1: Determine the initial gas concentration based on the target furnace chamber inlet gas accumulation calculator; Step 2: Call at least one of the algorithm models to calculate the change in gas concentration; wherein, the model is selected based on the specific application scenario of the furnace chamber, for example, the gas types involved in the furnace chamber based on its corresponding safety definition and their corresponding concentration thresholds.

[0032] Step 3: The target chamber’s pressure and temperature impact corrector is used to correct the gas flow and volume parameters, ensuring that the gas state inference engine’s analysis of the process data in the target chamber accurately reflects the current state of the target chamber. Step 4: Analyze abnormal gas fluctuations through flow balance and preliminary leakage diagnosis logic.

[0033] In some embodiments of the present application, in step 4, the safety state determiner retrieves a pre-defined safety definition for the target furnace chamber and compares it with the target furnace chamber gas state data output in step 3. Different furnace chambers have different definitions of safety, so the safety state determiner's criteria must be adjusted based on the target furnace chamber. In this application, adjustments to the safety state determiner only require adjustments to the relevant safety definitions, eliminating the need for large-scale adjustments to the production line.

[0034] In some embodiments of the present application, in step 5, if the judgment result is a safe state, the next process is carried out; if the judgment result is an unsafe state, the control system displays that the target furnace chamber is in an unsafe state, and at the same time, determines whether to trigger an alarm signal based on the process characteristics of the target furnace chamber. Different furnace chambers have different requirements for safe states, and the subsequent processing methods for safe or unsafe states are also different. For example, the source chamber of an ion implanter needs to switch different ion source gases between different batch processes, such as switching from a phosphorus source gas (such as PH3) to an arsenic source gas (such as AsH3) or a boron source gas (such as BF3). These gases are usually highly toxic and / or flammable and explosive. During the switching process, it is necessary to ensure that the previous gas in the source chamber is completely replaced and removed, and before maintenance or opening the chamber, it must be ensured that the concentration of all hazardous gases drops below the safe level. Therefore, for the source chamber of the ion implanter, before the gas switching begins, it is necessary to estimate the type and concentration of the residual process gas in the source chamber based on the operating parameters of the previous process. After purging with N2, the concentration of the process gas of the previous process in the source chamber drops below the safety threshold to indicate that the purging is completed, and the subsequent normal production process operations can continue. At this time, the state can be considered a safe state; but If the pressure does not drop below the safety threshold, it indicates that the purge is still in progress. At this time, the state in the source chamber should be considered unsafe and the source chamber should continue to be purged. However, this will not immediately trigger an alarm signal. The control system will start timing after receiving the first "unsafe state" stability result. If the "safe state" can be reached within the preset maximum purge time, the alarm signal will not be triggered, and the control system will prompt that the normal production process can continue. If the "safe state" is not reached within the preset maximum purge time, an alarm signal will be triggered, an alarm will be displayed on the control interface, and the operator will be prompted to take timely action. For plasma etching chambers (such as polysilicon etching), after the etching process is completed and it is purged, it is necessary to ensure that the inert atmosphere in the chamber is maintained before the normal production process can proceed. However, the judgment result of the safety state will change at this time. Once a potential leak is detected in the furnace chamber and it is judged to be "unsafe", an alarm will be immediately triggered and displayed on the interface. It can be seen that when the judgment result is an unsafe state, while displaying that the target furnace chamber is in an unsafe state through the control system, it is necessary to determine whether to trigger an alarm signal based on the process characteristics of the target furnace chamber.

[0035] 2. A monitoring system for the gas environment safety of a general-purpose semiconductor manufacturing furnace The monitoring system is used to execute the judgment method; the monitoring system includes a database, a real-time data acquisition module, a gas state inference module, a safety state determination module, and a result output and response module; The database is used to store and manage the specific parameters, safety definitions, and required algorithm models for the target chambers. This allows the gas state inference module to select the corresponding algorithm model from the database, and the safety state determination module to retrieve the safety definitions from the database. The database stores a series of data required for different processes and scenarios for each target chamber, ensuring that the monitoring system can obtain the required data from the database when applied to different target chambers. Operators can also adjust and modify the different target chambers stored in the database.

[0036] The real-time data acquisition module acquires real-time data from the target chamber's PLC or sensor network via an interface and sends the data to the gas state inference module. The real-time data acquired here corresponds to the process parameters of the target chamber. For example, in the source chamber of an ion implanter, different ion source gases need to be switched between different batch processes, such as switching from a phosphorus source gas (such as PH3) to an arsenic source gas (such as AsH3) or a boron source gas (such as BF3). During the switching process, the concentrations of these gases need to be monitored to ensure that the previous gas in the chamber is completely replaced and cleared. Therefore, the data collected at this time includes process parameters such as N2 flow rate, chamber pressure, chamber temperature, and exhaust system exhaust flow.

[0037] The gas state inference module receives real-time data from the real-time data acquisition module, analyzes the real-time data based on an algorithm model obtained from the database, and sends the analysis results to the safety state determination module. The gas state inference module is targeted when obtaining the algorithm model from the database. The operator inputs the scene that needs to be monitored in the target furnace chamber. For example, for the standby safety and leakage monitoring of the plasma etching chamber, the gas state inference module can obtain the plasma etching chamber process gas leakage monitoring model from the database. This model includes a leakage determination model and a leakage rate estimation model. The leakage determination model can be established in advance, collecting the leakage volume of the chamber in the past and the cavity pressure change rate caused by the leakage volume. The two are used to construct a leakage determination model for real-time monitoring of the cavity pressure change rate of the target furnace chamber, and then determine the possible leakage volume.

[0038] The Safety Status Determination Module receives the analysis results from the Gas Status Inference Module and compares them with the safety definitions in the database. It determines whether the target furnace chamber is in a safe or unsafe state and sends the result to the Result Output and Response Module. The Safety Status Determination Module retrieves the corresponding data from the database based on the target furnace chamber's application scenario.

[0039] The result output and response module receives the judgment result sent by the safety status judgment module and transmits the judgment result to the upper control system; at the same time, when the judgment result is unsafe, the control system displays that the target furnace chamber is in an unsafe state, and determines whether to trigger an alarm signal based on the process characteristics of the target furnace chamber. If the target furnace chamber needs to trigger an alarm immediately for an unsafe state, the result output and response module will immediately send an alarm signal when receiving the judgment result of unsafe. If the unsafe state can last for a period of time for the target furnace chamber to be transformed into a safe state, the result output and response module will not trigger an alarm signal immediately, but will start timing from the first time it receives the judgment result of the unsafe state. If the judgment result of the safe state can be received within the preset maximum time, the alarm will not be triggered, and the control system will display that subsequent normal production process operations can be carried out; if the judgment result received after exceeding the preset maximum time is still an unsafe state, an alarm will be triggered, and the control system will display the alarm information on the interface.

[0040] 3. Examples Example 1: Gas safety switching and monitoring for ion implanter source chamber Scenario Description: The source chamber of an ion implanter needs to switch between different ion source gases between different batch processes, such as switching from a phosphorus source gas (such as PH3) to an arsenic source gas (such as AsH3) or a boron source gas (such as BF3). These gases are often highly toxic and / or flammable and explosive. During the switching process, it is necessary to ensure that the previous gas in the source chamber is completely replaced and purged. Before maintenance or opening the chamber, the concentration of all hazardous gases must be reduced to below a safe level.

[0041] Step 1: Parameter configuration Furnace chamber characteristics: ion source chamber volume (V_Source), for example 10 liters.

[0042] Gas list and properties: N2 / Ar: Safety gas / purge gas.

[0043] PH3: Unsafe / process gas (highly toxic, flammable), safety threshold C_PH3_safe = 0.3 ppm.

[0044] AsH3: Unsafe / process gas (highly toxic, flammable), safety threshold C_AsH3_safe = 0.05 ppm.

[0045] BF3: Unsafe / process gas (corrosive, toxic), safety threshold C_BF3_safe = 1 ppm.

[0046] MFC information: identification and range of MFC_N2 (through N2), MFC_PH3 (through PH3), MFC_AsH3 (through AsH3), MFC_BF3 (through BF3).

[0047] Exhaust characteristics: effective pumping speed or pressure-flow characteristic curve of the source cavity exhaust system.

[0048] Algorithm Model: Select "Multi-Region Gas Displacement Model" and configure mixing parameters for different regions within the source chamber (such as near the electrode and near the chamber wall).

[0049] Step 2: Real-time data collection During the gas switching and purge phases, the system collects real-time data: F_N2_in (N2 flow rate of MFC_N2, for example 5 slm).

[0050] P_Source (source chamber pressure, e.g. 1-10 Pa).

[0051] T_Source (source cavity temperature, which may be measured at multiple points, e.g. the electrode area may be hundreds of °C, while the cavity wall area may be lower).

[0052] F_exhaust_out (exhaust flow rate of the source chamber exhaust system).

[0053] Step 3: Gas state inference Initial concentration assessment: Before gas switching begins, the initial concentration distribution of the process gas (such as PH3) remaining in the source chamber is estimated based on the operating parameters of the previous process (such as PH3 flow rate, operating time, and source chamber temperature distribution).

[0054] Multi-zone displacement model: Considering the possible temperature gradients and complex structures (such as electrodes and shielding plates) in the ion source cavity, a multi-zone model is used to estimate the gas displacement efficiency in different zones: For each region i (i = 1, 2, ..., n) in the source cavity: C_gas_i(t) = C_gas_i_initial * exp(-(k_i * F_N2_in_std / V_i_std) * t) Where C_gas_i(t) is the concentration of a hazardous gas in region i at time t, C_gas_i_initial is the initial concentration of the region, k_i is the mixing efficiency factor of the region, and V_i_std is the volume of the region under standard conditions.

[0055] Adsorption / Desorption Consideration: For gases that may adsorb on the chamber wall or component surface (such as PH3, AsH3), a temperature-dependent desorption term is added to the model: R_desorption_i = A_i * exp(-E_a / (R*T_i)) * θ_i Where, R_desorption_i is the desorption rate of region i, A_i is the frequency factor, E_a is the desorption activation energy, T_i is the temperature of this region, and θ_i is the surface coverage.

[0056] Step 4: Safety Judgment IF (C_gas_i(t) < C_gas_safe for all regions i for all hazardous gases to be monitored) THEN Judge as "Safe State (Gas switching / purging completed)".

[0057] ELSE Judge as "Unsafe State (Gas switching / purging in progress)", and indicate which regions and gases have not yet reached the safe level.

[0058] Step 5: Result Output and Response The interface displays the estimated concentrations of hazardous gases, purging progress, and overall safety status in each region.

[0059] When the "Safe State" is reached, the next operation is allowed (such as introducing a new process gas to start a new process, or allowing maintenance personnel to safely open the chamber).

[0060] If the "Safe State" is not reached after the preset maximum purging time, an alarm is triggered, and it is recommended to increase the purging time or check for possible system abnormalities (such as MFC failure, reduced efficiency of the pumping system, abnormal adsorption / desorption phenomena, etc.).

[0061] For gas switching operations, the system can write to the specified address of the PLC to ensure that the device only allows the introduction of a new hazardous gas after the previous hazardous gas has been fully replaced, preventing the mixing of incompatible gases.

[0062] Example 2: Standby Safety and Leakage Monitoring for Plasma Etching Chambers (such as polysilicon etching) Scenario Description: After completing an etching process using corrosive and toxic gases such as Cl2 and HBr in a plasma etching chamber, it is usually purged with N2 or Ar and enters the standby state. In the standby state, it is necessary to ensure that the chamber maintains an inert atmosphere and monitor whether there is a slight leakage of process gas from the pipeline valve into the chamber, or air leakage from the seal.

[0063] Step 1: Parameter Configuration Furnace chamber characteristics: Etching chamber volume (V_Etch), for example 50 liters.

[0064] Gas list and properties: N2: Safety gas.

[0065] Ar: safety gas.

[0066] Cl2: Unsafe / process gas, safety threshold C_Cl2_safe = 1 ppm.

[0067] HBr: Unsafe / process gas, safety threshold C_HBr_safe = 3 ppm.

[0068] Air (O2 as indicator): Unsafe / contaminated gas (in inert standby), O2 safety threshold C_O2_safe = 1000 ppm (example).

[0069] MFC information: MFC_N2, MFC_Ar, MFC_Cl2, MFC_HBr, etc.

[0070] Exhaust gas characteristics: Pumping speed of turbomolecular pump for N2.

[0071] Algorithm model: configured as "trace gas accumulation model" and "flow balance mismatch monitoring model".

[0072] Step 2: Real-time data collection In standby mode (for example, the chamber door is closed and a small amount of N2 is introduced to maintain positive pressure or low vacuum): F_N2_in (N2 flow rate of MFC_N2, for example 0.1 slm).

[0073] P_Etch (cavity pressure, e.g. 10 Pa or 100 kPa + slightly positive pressure).

[0074] T_Etch (chamber temperature, e.g. room temperature).

[0075] F_exhaust_out (the flow rate out of the exhaust system, which may be close to zero or equal to a trace amount of N2 flow).

[0076] Step 3: Gas state inference (1) Process gas leakage monitoring: Assume the Cl2 and HBr MFC valves are closed. If, while a steady stream of N2 is flowing through the P_Etch, an unexplained, slow but sustained pressure rise occurs, exceeding normal fluctuations caused by temperature changes. Furthermore, the rate of pressure rise is similar to that caused by historically known minor leaks (e.g., 1 sccm of Cl2) (a signature library or model must be established in advance), then a minor Cl2 or HBr leak is likely present.

[0077] Alternatively, if the system is equipped with an extremely low-flow MFC (capable of detecting minute reverse flows or pressure changes), this can also be used for judgment.

[0078] Estimate the leak rate and calculate the cumulative concentration based on V_Etch. If C_Cl2_cumulative (cumulative, increasing) > C_Cl2_safe or C_HBr_cumulative > C_HBr_safe, an alarm is issued.

[0079] (2) Air leakage monitoring (vacuum standby): If P_Etch cannot maintain the set low pressure when the vacuum pump is continuously operating and the N2 / Ar flow rate is zero or very small, but instead slowly recovers, and the recovery rate exceeds the normal cavity wall gas outflow rate, it is inferred that there may be an air leak.

[0080] The air leakage rate is estimated by the pressure recovery rate, and then the O2 concentration is estimated. If C_O2_cumulative>C_O2_safe, an alarm is triggered.

[0081] (3) Flow balance monitoring (micro-positive pressure standby): Compare F_N2_in with F_exhaust_out (if measured) and the stability of P_Etch. If F_N2_in>F_exhaust_out and P_Etch continues to rise abnormally, it may indicate that the exhaust is not smooth or there is extra gas entering.

[0082] Step 4: Safety Assessment IF (no process gas leaks detected AND no significant air leaks detected AND P_Etch stabilizes in a safe inert atmosphere range) THEN Determined to be in "safe standby state".

[0083] ELSE Determine it as an "unsafe state (potential leakage)" and indicate the possible cause.

[0084] Step 5: Result Output and Response The safety status of the chamber is continuously displayed; once a potential leak is detected and judged to be "unsafe", an alarm is immediately triggered, the interface displays the alarm, or writes to the specified address of the PLC.

[0085] Example 3: Process safety monitoring for diffusion furnaces (e.g., wet oxygen oxidation) Scenario Description: When performing a wet oxygen oxidation process in a diffusion furnace, O2 and H2 must be introduced (either through combustion to generate water vapor or by directly introducing water vapor), with N2 used as both a carrier gas and a balance gas. The ratio of H2 to O2 must be maintained within a safe range to prevent the formation of an explosive mixture, and the integrity of the furnace tubes must be monitored.

[0086] Step 1: Parameter configuration Furnace chamber characteristics: diffusion furnace tube volume (V_Diffusion).

[0087] Gas list and properties: N2: safety gas / carrier gas.

[0088] O2: Process gas, not unsafe in itself, but there are risks when mixed with H2.

[0089] H2: Process gas / unsafe gas (flammable and explosive), its ratio with O2 and accidental leakage into non-combustion areas must be monitored.

[0090] Safety thresholds: for example, the upper limit of H2 concentration in the mixture (non-precombustion zone) and the lower limit of O2 concentration in the H2 pipeline (anti-backfire).

[0091] MFC information: MFC_N2, MFC_O2, MFC_H2.

[0092] Tail gas characteristics: Diffusion furnaces are usually at normal pressure or slightly positive pressure, and the tail gas is discharged directly or treated.

[0093] Algorithm model: Configured as "gas ratio monitoring model" and "pressure anomaly monitoring model".

[0094] Step 2: Real-time data collection In the wet oxygen oxidation process: F_N2_in, F_O2_in, F_H2_in.

[0095] P_Diffusion (pressure inside the furnace tube, usually close to atmospheric pressure).

[0096] T_Diffusion (furnace temperature, e.g. 900-1100°C).

[0097] Step 3: Gas state inference H2 / O2 Ratio Monitoring: Calculates the ratio of incoming H2 and O2 flows in real time to ensure it remains within process settings and safe operating windows. If the ratio exceeds safe limits, an alarm is issued, potentially triggering MFC adjustment or shutdown.

[0098] Total flow and pressure monitoring: Monitors the relationship between the total inlet flow (F_N2_in + F_O2_in + F_H2_in) and the furnace tube pressure P_Diffusion. If P_Diffusion increases abnormally while the flow rate is stable, it may indicate a blockage in the exhaust port; if it decreases abnormally, it may indicate a ruptured furnace tube or seal leak.

[0099] Judgment of leakage to the surrounding environment (indirect): If an external H2 sensor is equipped (not mandatory in the present invention, but can be used as an enhancement), when the pressure in the furnace is abnormal and the external sensor alarms, it can be highly suspected that the furnace tube is broken.

[0100] Step 4: Safety Assessment IF (H2 / O2 ratio is within the safe range AND P_Diffusion is stable and normal) THEN It is judged as "process operation is safe".

[0101] ELSE It is judged as "unsafe process operation (proportional imbalance or abnormal pressure)".

[0102] Step 5: Result Output and Response Displays key parameters such as H2 / O2 ratio, furnace pressure, and safety status.

[0103] If it is judged that "process operation is unsafe", an alarm is triggered and displayed on the interface, or the data is written to the specified address of the PLC.

[0104] Comparative Example Existing technologies usually use three methods to monitor the status of the target furnace chamber: (1) Real-time detection of the target substance in the target furnace chamber by installing high-precision detection instruments, such as mass spectrometers, gas chromatographs, etc. However, in actual production, these devices are not only expensive in themselves, but also have very high maintenance costs. They require professional analysts to operate, which makes the entire monitoring system more complicated and has higher operating costs; (2) Assisted judgment by mechanical instruments and meters. In actual production, this monitoring method has a lag, and operators are required to regularly inspect and observe the parameter changes of the instruments and meters, which will lead to untimely inspections and thus failure to discover accident points in time; (3) Determine the status of the target furnace chamber based on the operator's experience. For example, for the judgment of the status of the target furnace chamber after purging, due to the gap in work experience among operators, the judgment results are often inaccurate, and in order to ensure safety, operators usually perform excessive purging, which will cause a lot of resource waste and increase production costs.

[0105] The aforementioned examples, using different types of furnace chambers and different scenarios, demonstrate that the judgment method framework provided in this application is highly flexible and adaptable compared to the comparative examples. By tailoring the parameters in the configuration database (furnace chamber characteristics, gas properties, safety thresholds, algorithm model selection and parameters, etc.), the same core judgment logic can be effectively applied to gas environment safety monitoring in a variety of semiconductor manufacturing process chambers, achieving the method's versatility.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit the technical solutions. Those skilled in the art should understand that modifications or equivalent replacements of the technical solutions of the present application that do not depart from the purpose and scope of the technical solutions of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for determining the safety of a general furnace gas environment in semiconductor manufacturing, characterized in that: The specific steps are as follows: Step 1: Based on the process design characteristic parameters of the target furnace chamber, the definition of safety in the safety state determiner, and the algorithm model in the gas state inference engine; Step 2: Acquire the characteristic parameters of the target furnace chamber in real time; Step 3: The process data obtained in step 2 is analyzed by the gas state inference engine, and the analysis results are sent to the safety state determiner; Step 4: The safety state determiner compares the result output from step 3 with the definition of safety to determine whether the target furnace chamber is in a safe state or an unsafe state; Step 5: Send the judgment result obtained in step 4 to the control system, and perform the next operation based on the judgment result.

2. The judgment method according to claim 1, characterized in that: The target furnace chamber is a process furnace chamber in a semiconductor device manufacturing process.

3. The judgment method according to claim 2, characterized in that: The process furnace chamber includes a diffusion furnace, an oxidation furnace, an annealing furnace, a chemical vapor deposition furnace and an etching reaction chamber.

4. The judgment method according to claim 1, characterized in that: In step 2, characteristic parameters are obtained in real time from the PLC or sensor network of the target furnace chamber; wherein the characteristic parameters include the target furnace chamber volume, the type of gas in the target furnace chamber, the exhaust gas characteristics and the data obtained by the mass flow controller.

5. The judgment method according to claim 1, characterized in that: The definition of safety includes the types of safe gases and their corresponding concentration thresholds, and the types of unsafe gases and their corresponding concentration thresholds.

6. The judgment method according to claim 1, characterized in that: The algorithm model includes mixing efficiency, replacement model selection and judgment rules.

7. The judgment method according to claim 1, characterized in that: In step 3, the gas state inference engine performs the following operations: determining an initial gas concentration based on a target furnace chamber inlet gas accumulation calculator; Call at least one model from the configurable gas displacement and concentration estimation model library to calculate the change in gas concentration; Correct the gas flow and volume parameters through the target furnace chamber pressure and temperature influence corrector; Analyze abnormal gas fluctuations through flow balance and preliminary leakage diagnosis logic.

8. The judgment method according to claim 1, characterized in that: In step 5, if the judgment result is a safe state, the next process is carried out; if the judgment result is an unsafe state, the control system displays that the target furnace chamber is in an unsafe state, and at the same time, determines whether to trigger an alarm signal based on the process characteristics of the target furnace chamber.

9. A monitoring system for the gas environment safety of a general-purpose semiconductor manufacturing furnace, characterized in that: Execute any judgment method described in claims 1 to 8; the monitoring system includes a database, a real-time data acquisition module, a gas state inference module, a safety state determination module, and a result output and response module; The database is used to store and manage specific parameters, safety definitions, and algorithm models of the target furnace chamber, so that the gas state inference module can select the corresponding algorithm model from the database and the safety state determination module can obtain the safety definitions in the database; The real-time data acquisition module obtains real-time data from the PLC or sensor network of the target furnace chamber through the interface and sends the data to the gas state inference module; The gas state inference module receives the real-time data sent by the real-time data acquisition module, analyzes the real-time data according to the algorithm model obtained from the database, and sends the analysis results to the safety state determination module; The safety state determination module receives the analysis results sent by the gas state inference module, compares them with the safety definition in the database, determines whether the state in the target furnace chamber is safe or unsafe, and sends the judgment result to the result output and response module; The result output and response module receives the judgment result sent by the safety status judgment module and transmits the judgment result to the upper control system; at the same time, when the judgment result is unsafe, it determines whether to trigger an alarm signal based on the process characteristics of the target furnace chamber.

Images