Parameter abnormality detection method and semiconductor process apparatus

Abstract

Embodiments of the present application provide a parameter abnormality detection method and a semiconductor process equipment, the process of the semiconductor process equipment comprises a plurality of process steps performed in sequence, the parameter abnormality detection method comprises: collecting actual parameter values of a target parameter at each preset time point in a process step; discretizing the actual parameter values of the target parameter into a pre-acquired parameter relationship graph, and determining that the target parameter in the process step is abnormal in a case where the actual parameter values of the target parameter at the preset time points with a preset time precision change satisfy a preset abnormality condition with normal parameter values of the target parameter at the corresponding preset time points in the parameter relationship graph, the parameter relationship graph being a relationship graph of the normal parameter values of the target parameter in the process step changing with each preset time point; adjusting the process according to an adjustment scheme corresponding to the preset abnormality condition; and outputting prompt information corresponding to the preset abnormality condition. The parameter abnormality detection method of the present application can improve the parameter abnormality detection efficiency.

Description

Technical Field

[0001] This invention relates to the field of semiconductor technology, and in particular to a method for detecting abnormal parameters and semiconductor process equipment. Background Technology

[0002] The technological process, also known as the "processing flow" or "production flow," refers to the entire process of sequentially processing raw materials into finished products using specific production equipment or pipelines. A complete technological process typically includes several steps. In technological processes involving the manufacturing of process equipment, fluctuations in certain target parameters (such as temperature, flow rate, radio frequency, and pressure) directly affect the process results, and in severe cases, can lead to rework and impact equipment capacity.

[0003] Current parameter anomaly detection solutions are relatively simple and rigid in detecting abnormal parameters. As a result, anomalies are often only discovered in the production results after the process is completed. Moreover, the detection efficiency is not high and cannot meet the ever-increasing production requirements. Summary of the Invention

[0004] The technical problem to be solved by the embodiments of the present invention is that the detection of abnormal parameters is not timely and the efficiency is not high.

[0005] To address the aforementioned problems, this invention discloses a parameter anomaly detection method applied to semiconductor process equipment. The semiconductor process equipment comprises multiple sequentially performed process steps, and the method includes:

[0006] The actual parameter values ​​of the target parameter at each preset time point are collected during the process steps.

[0007] Discretize the actual parameter values ​​of the target parameter into a pre-acquired parameter relationship diagram. If the actual parameter value of the target parameter at a preset time point and the normal parameter value at the corresponding preset time point in the parameter relationship diagram meet the preset abnormal conditions, the target parameter in the process step is determined to be abnormal. The parameter relationship diagram is a relationship diagram of the normal parameter value of the target parameter in the process step with each preset time point.

[0008] Adjust the process according to the adjustment plan corresponding to the preset abnormal conditions;

[0009] Output the prompt message corresponding to the preset abnormal conditions.

[0010] Another embodiment of the present invention discloses a semiconductor process apparatus. The process of the semiconductor process apparatus includes multiple sequentially performed process steps. The semiconductor process apparatus includes:

[0011] The controller is used to collect the actual parameter values ​​of the target parameters at each preset time point in the process steps;

[0012] Discretize the actual parameter values ​​of the target parameter into a pre-acquired parameter relationship diagram. If the actual parameter value of the target parameter at a preset time point and the normal parameter value at the corresponding preset time point in the parameter relationship diagram meet the preset abnormal conditions, the target parameter in the process step is determined to be abnormal. The parameter relationship diagram is a relationship diagram of the normal parameter value of the target parameter in the process step with each preset time point.

[0013] Adjust the process according to the adjustment plan corresponding to the preset abnormal conditions;

[0014] Output the prompt message corresponding to the preset abnormal conditions.

[0015] According to the semiconductor process equipment provided by the present invention, by collecting the actual parameter values ​​of the target parameter at each preset time point in the process step, and discretizing the actual parameter values ​​of the target parameter into a pre-acquired parameter relationship diagram, if the actual parameter value of the target parameter at the preset time point, which varies with the preset time precision, and the normal parameter value at the corresponding preset time point in the parameter relationship diagram satisfy a preset abnormality condition, an abnormality in the target parameter in the process step is determined. Here, by comparing the actual parameter value of the target parameter in the process step with the normal parameter value of the target parameter in the process step, it is possible to accurately and quickly detect whether a parameter abnormality has occurred in the process. Thus, it is possible to quickly and efficiently trigger the corresponding adjustment scheme to adjust the process and output the prompt information corresponding to the preset abnormality condition. Therefore, by improving the detection speed of parameter abnormalities in the semiconductor process equipment, not only can the equipment capacity be increased, but also the equipment's ability to handle abnormalities can be improved, thereby enhancing safety. Attached Figure Description

[0016] Figure 1 A flowchart of a process provided in this embodiment is shown;

[0017] Figure 2 A flowchart of a parameter anomaly detection method provided in this embodiment is shown;

[0018] Figure 3 This embodiment shows a parameter relationship diagram and a schematic diagram of parameter values.

[0019] Figure 4 This embodiment illustrates a parameter anomaly diagram.

[0020] Figure 5 This embodiment illustrates another parameter anomaly.

[0021] Figure 6 This embodiment illustrates yet another parameter anomaly.

[0022] Figure 7 A schematic diagram of the spectrum curve of an abnormal parameter provided in this embodiment is shown;

[0023] Figure 8 This embodiment shows a schematic diagram of the spectrum curve of a normal parameter.

[0024] Figure 9 This embodiment shows a flowchart of a method for detecting parameter anomalies.

[0025] Figure 10 This embodiment shows a flowchart of another method for implementing parameter anomaly detection.

[0026] Figure 11 A schematic diagram of a semiconductor process equipment structure provided in this embodiment is shown. Detailed Implementation

[0027] The features and exemplary embodiments of various aspects of the present invention will now be described in detail. To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention and are not configured to limit the present invention. For those skilled in the art, the present invention can be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present invention by illustrating examples of the invention.

[0028] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0029] First, the technical terms involved in the embodiments of the present invention will be introduced.

[0030] A solar cell is a thin photovoltaic semiconductor wafer that generates electricity directly using sunlight. It is also known as a "solar chip" or "photovoltaic cell". As long as the illuminance meets certain conditions, it can instantly output voltage and generate current when there is a circuit.

[0031] The two most important directions in solar cell research are high efficiency and low cost. In the cell production process, since the number of wafers loaded in the graphite boat of the coating equipment is much smaller than the number of wafers loaded in the quartz boat of the diffusion and annealing equipment, the bottleneck of the workshop's capacity is often the coating equipment. Therefore, reducing the number of failed process boats and reworked wafers has become the research focus for improving the capacity of coating equipment.

[0032] The parameter anomaly detection method provided in this embodiment of the invention can be applied to at least the following application scenarios, which will be described below.

[0033] like Figure 1 As shown, this relates to the deposition step in the coating process of manufacturing crystalline silicon solar cells. The target parameters of temperature, flow rate, radio frequency (RF) pressure, and pressure in the deposition step directly affect the thickness of the silicon wafer. Changes in hardware and the environment can cause fluctuations in the values ​​of these four parameters. Examples include: loose RF power supply connections, decreased pump speed after prolonged use, fluctuations in plant flow rates, fluctuations in cavity temperature caused by RF discharge, or changes in the setting values ​​of process step skipping parameters.

[0034] There are two types of fluctuations in parameter values. One is a sudden jump, which can be filtered out if it occurs infrequently, but will affect the process results if it occurs frequently. The other is continuous fluctuation. If the fluctuation is within a certain range, it will have little impact on the process results. However, if the fluctuation range is large or the fluctuation lasts for a long time, it will affect the process results. In severe cases, it can cause silicon wafer rework and affect equipment capacity.

[0035] Currently, in semiconductor equipment manufacturing processes, each process step has only one baseline value for parameter fluctuation, which remains unchanged regardless of the process time. By comparing the actual parameter value with the baseline value, a timer is started when the fluctuation between the actual and baseline values ​​becomes significant. An alarm is triggered after a period of sustained significant fluctuation. If the fluctuation returns to normal within a certain time, the timer is cleared, and it restarts when the fluctuation becomes significant again. However, because only one baseline value is set for each process step, and the tolerable fluctuation range varies across multiple process steps due to different conditions at different times, existing technology cannot adequately address this situation. Furthermore, it can only detect continuous abnormal fluctuations, not fluctuations that affect the process results. Additionally, the judgment process cannot provide early warnings of hardware anomalies; judgment can only be made after an actual anomaly occurs.

[0036] Based on the above problems and application scenarios, the parameter anomaly detection method provided by the embodiments of the present invention will be described in detail below.

[0037] This invention provides a parameter anomaly detection method applied to semiconductor process equipment, the process of which includes multiple sequential process steps.

[0038] Figure 2 This is a flowchart of a parameter anomaly detection method provided in an embodiment of the present invention.

[0039] like Figure 2 As shown, the parameter anomaly detection method may include steps 210-240, as detailed below:

[0040] Step 210: Collect the actual parameter values ​​of the target parameter at each preset time point in the process steps.

[0041] Step 220: Discretize the actual parameter values ​​of the target parameter into a pre-acquired parameter relationship diagram. If the actual parameter value of the target parameter at a preset time point and the normal parameter value at the corresponding preset time point in the parameter relationship diagram meet the preset abnormal conditions, the target parameter in the process step is determined to be abnormal. The parameter relationship diagram is a relationship diagram of the normal parameter value of the target parameter in the process step with each preset time point.

[0042] Step 230: Adjust the process according to the adjustment plan corresponding to the preset abnormal conditions.

[0043] Step 240: Output the prompt message corresponding to the preset abnormal condition.

[0044] The parameter anomaly detection method provided by this invention collects the parameter values ​​of target parameters within a preset time period for each process step. When the relationship between the parameter value and a pre-acquired parameter relationship diagram meets preset anomaly conditions, an anomaly is determined in the process step. The parameter relationship diagram indicates the relationship between time and target parameter values ​​during the process. The collected actual parameter values ​​can be compared with the target parameter values ​​at the corresponding time points in the parameter relationship diagram, enabling accurate and rapid detection of parameter anomalies during the process. This allows for the rapid and efficient triggering of corresponding adjustment schemes to adjust the process and output of prompts corresponding to preset anomaly conditions. Therefore, by improving the detection speed of parameter anomalies, not only can equipment capacity be increased, but the equipment's ability to handle anomalies can also be improved, thereby enhancing safety.

[0045] The contents of steps 210-240 are described below:

[0046] Step 210: Collect the actual parameter values ​​of the target parameter at each preset time point in the process steps.

[0047] The target parameters mentioned above may include: temperature, flow rate, radio frequency, pressure, and butterfly valve angle.

[0048] Among them, such as Figure 3 As shown, the parameter relationship diagram illustrates the relationship between the target parameter values ​​of each process step and time in the process flow. The parameter relationship diagram can be determined based on the relationship between the set time and target parameter values ​​in each process step of a normal process flow. The first dimension (horizontal axis) of this parameter relationship diagram can be time, typically the process duration of a process step. The values ​​in the first dimension represent preset time points, and the interval between each preset time point is the preset time precision. The preset time point starts at 0 and ends at the step time of the process step, in seconds. For example, the preset time points are 0-23, and the preset time precision is 1 second. The second dimension (vertical axis) of the parameter relationship diagram can be the target parameter value, i.e., the curve showing the change of the target parameter value over time in the corresponding process step. The point corresponding to each preset time point on this curve is the normal parameter value of the target parameter at that preset time point. The precision and unit of this second dimension depend on the corresponding target parameter. Specifically, the parameter relationship diagram is determined based on the process formula and parameter type of the process step, as well as the required time and parameter precision requirements. Among them, the higher the accuracy and the higher the complexity, the higher the accuracy of calculating the fluctuation between the actual parameter value and the target parameter value, and the fewer false alarms and missed alarms.

[0049] When the process skips steps, that is, when the process proceeds from one process step to the next process step, the parameter relationship diagram of the corresponding process step is obtained again.

[0050] In one embodiment, step 220, where the target parameter changes with preset time precision at a preset time point and the corresponding normal parameter value at the preset time point in the parameter relationship graph meets preset abnormal conditions, determines that the target parameter in the process step is abnormal. This can specifically include the following steps:

[0051] Based on the parameter relationship diagram, determine the preset parameter fluctuation range corresponding to each preset time point for the normal parameter value;

[0052] In response to the detection that the actual parameter value at a preset time point exceeds the parameter fluctuation range at the preset time point, a count is performed to obtain the number of anomalies;

[0053] If the number of abnormalities exceeds the preset first threshold, the target parameter in the process step is determined to be abnormal, and the target parameter is in the first abnormal state.

[0054] If the number of abnormal events is less than a preset first threshold, determine the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point. If the deviation meets the preset abnormal conditions, then determine that the target parameter in the process step is abnormal.

[0055] Based on the parameter relationship diagram, determine the parameter fluctuation range corresponding to each preset time point for the normal parameter value. Specifically, this can be done as follows: Figure 3 As shown by the black curve in the middle. Each preset time point has an allowable deviation, namely... Figure 3 The vertical solid "I" shape in the image indicates that if the actual parameter value falls within the parameter fluctuation range, the target parameter at the preset time point is normal; if the actual parameter value exceeds the parameter fluctuation range, the target parameter at the preset time point is abnormal.

[0056] Typically, according to existing detection technologies, a sudden and abnormal increase in parameter values ​​during the process, followed by an immediate return to the normal range, is not considered an anomaly. Such anomalies are only discovered when the silicon wafer process results are tested at the end of the process.

[0057] The parameter anomaly detection method provided in this embodiment of the invention addresses this issue. For example... Figure 4 As shown, when parameter values ​​frequently exceed the parameter fluctuation range, it can also affect the process results. If this anomaly is detected during the process and handled in advance, it can save a lot of time and increase equipment capacity.

[0058] In response to this situation, if the actual parameter value corresponding to the preset time point is detected to exceed the parameter fluctuation range corresponding to the preset time point, a count is performed to obtain the number of abnormalities; if the number of abnormalities is greater than the preset first threshold, it is determined that the target parameter in the process step is abnormal and the target parameter is in the first abnormal state.

[0059] Specifically, an exception counter can be used. This counter is valid during the process steps and is reset to zero after a step is skipped. Each exception occurring during a process step increments the exception counter by 1, accumulating the number of exceptions. If the number of exceptions exceeds a preset first threshold, the target parameter in the process step is determined to be abnormal; if the number of exceptions is less than or equal to the first threshold, the target parameter in the process step is determined to be abnormal.

[0060] like Figure 4 As shown, a certain target parameter in this process step jumps a total of 7 times. Assuming the preset first threshold is 5 times, it means that the number of abnormal times is greater than the preset first threshold, and the target parameter in the process step is abnormal.

[0061] Generally, based on existing parameter detection technology, fluctuations in actual parameter values ​​at each point in time are considered normal if they are within a reasonable range or the number of anomalies is less than a preset first threshold. Anomalies in the target parameters during the process steps can only be detected when the silicon wafer process results are checked at the end of the process.

[0062] In response to this situation, the present application has designed a detection method. If the number of abnormalities is less than a preset first threshold, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point is determined. If the deviation meets the preset abnormality conditions, the target parameter in the process step is determined to be abnormal.

[0063] like Figure 5 As shown, before skipping a process step, the calculated cumulative deviation of that process step exceeded the tolerable parameter fluctuation range for that step. If this anomaly is detected before skipping to the next step and then handled, significant time can be saved and equipment capacity increased.

[0064] Specifically, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point can be determined. If the deviation meets the preset abnormal conditions, the target parameter in the process step is determined to be abnormal.

[0065] In one possible embodiment, if the target parameter is in a first abnormal state, the process is adjusted according to the adjustment scheme corresponding to the preset abnormal condition, including: terminating the process.

[0066] Output prompts corresponding to preset abnormal conditions, including: outputting a first prompt, which is used to prompt for checking the equipment wiring and / or checking the plant's gas source parameters.

[0067] If the number of anomalies exceeds a preset first threshold, the target parameter in the process step is determined to be abnormal. Such anomalies are mostly caused by hardware or environmental factors such as loose wiring or unstable pressure of various gas sources in the plant. Due to their unpredictability and quantifiability, process compensation cannot be performed based on fluctuation deviations. Once this type of anomaly handling mode is triggered, the process will be terminated directly regardless of the parameter type.

[0068] And input the first prompt message, which is used to prompt equipment personnel to check the equipment wiring or troubleshoot the gas source in the plant, etc.

[0069] It should be noted that, Figure 5 The example shown is a case where the number of abnormal occurrences of the actual parameter value exceeding the parameter fluctuation range corresponding to the preset time point is 0. This also falls under the case where the number of abnormal occurrences of the actual parameter value exceeding the parameter fluctuation range corresponding to the preset time point is less than a preset first threshold, where the preset first threshold is a positive integer.

[0070] In one embodiment, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point is determined. If the deviation meets a preset abnormality condition, the target parameter in the process step is determined to be abnormal, including:

[0071] If the actual parameter value at a preset time point is greater than the normal parameter value at the same preset time point, the cumulative larger value between the actual parameter value and the normal parameter value will be counted.

[0072] If the actual parameter value at a preset time point is less than the normal parameter value at the same preset time point, the cumulative smaller value between the actual parameter value and the normal parameter value will be counted.

[0073] If the difference between the cumulative excessive value and the preset excessive threshold is greater than the preset second threshold, then the target parameter in the process step is determined to be abnormal; and / or

[0074] If the difference between the cumulative underestimation value and the preset underestimation threshold is greater than the preset third threshold, then the target parameter in the process step is determined to be abnormal.

[0075] For example, the target parameter is temperature. For a given preset time point, the parameter value at that preset time point is 20 degrees Celsius. The target parameter value corresponding to that preset time point in the parameter relationship diagram is 15 degrees Celsius. Then, the deviation value at that preset time point is 5 degrees Celsius. And so on, accumulating the deviation values ​​corresponding to each time point within the preset time period to obtain the cumulative deviation value. If the difference between the cumulative deviation value (e.g., 50 degrees Celsius) and the preset deviation threshold (e.g., 30 degrees Celsius) (20 degrees Celsius) is greater than the preset second threshold (10 degrees Celsius), the target parameter in the process step is determined to be abnormal.

[0076] For example, the target parameter is pressure. For a given preset time point, the parameter value at that preset time point is 20 Pa. The target parameter value corresponding to that preset time point in the parameter relationship diagram is 30 Pa. Then, the deviation value at that preset time point is 10 Pa. And so on, accumulating the deviation values ​​corresponding to each preset time point within the preset time period to obtain the cumulative deviation value. If the difference between the cumulative deviation value (e.g., 50 Pa) and the preset deviation threshold (e.g., 30 Pa) (20 Pa) is greater than the preset third threshold (10 Pa), the target parameter in the process step is determined to be abnormal.

[0077] In one possible embodiment, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point is determined. If the deviation meets a preset abnormality condition, the target parameter in the process step is determined to be abnormal, including:

[0078] Determine the difference between the actual parameter value and the normal parameter value at each preset time point, and sum the differences to obtain the total deviation value.

[0079] If the total deviation value is greater than zero, and the difference between the total deviation value and the preset fourth threshold is greater than zero, then it is determined that the target parameter in the process step is abnormal and the target parameter is in the second abnormal state.

[0080] If the total deviation value is less than zero, and the difference between the absolute value of the total deviation value and the preset fifth threshold is greater than zero, then it is determined that the target parameter in the process step is abnormal and the target parameter is in the third abnormal state.

[0081] When the process has high requirements for parameter stability and the abnormality of large deviation and small deviation cannot cancel each other out, that is, when the difference between the total deviation value obtained by accumulating the deviation value corresponding to the preset time point of each collected parameter value and the preset fourth threshold is greater than zero, it is determined that the target parameter in the process step is abnormal and the target parameter is in the second abnormal state.

[0082] For example, if the cumulative deviation is 50 degrees Celsius and the cumulative deviation is -20 degrees Celsius, then the total deviation is 30 degrees Celsius. Since 30 degrees Celsius (total deviation value) is greater than zero, and the difference between 30 degrees Celsius (total deviation value) and 10 degrees Celsius (preset fourth threshold) is greater than zero (30 degrees Celsius - 10 degrees Celsius = 20 degrees Celsius), then it is determined that the target parameter in the process step is abnormal and the target parameter is in the second abnormal state.

[0083] For example, if the cumulative larger value is 50 degrees Celsius and the cumulative smaller value is -80 degrees Celsius, then the total deviation value is -30 degrees Celsius. Since -30 degrees Celsius (total deviation value) is less than zero, and the difference between the total deviation value and the preset fifth threshold is |-30 degrees Celsius - (-10 degrees Celsius)| = 20 degrees Celsius, and 20 degrees Celsius is greater than zero, then it is determined that the target parameter in the process step is abnormal and the target parameter is in the third abnormal state.

[0084] In one possible embodiment, the semiconductor process equipment is a deposition equipment, wherein if the target parameter is a temperature parameter and the process step is a deposition step, and the target parameter is in a second abnormal state, the process is adjusted according to an adjustment scheme corresponding to a preset abnormal condition, including:

[0085] Terminating the process;

[0086] If the target parameter is an RF parameter and the process step is a deposition step, and the target parameter is in the second abnormal state, then the process progress is adjusted according to the adjustment scheme corresponding to the preset abnormal conditions, including:

[0087] Perform process compensation processing on the process flow;

[0088] If the target parameter is a temperature parameter and the process step is a deposition step or a heating step, and the target parameter is in the third abnormal state, then the process should be adjusted according to the adjustment plan corresponding to the preset abnormal conditions, including:

[0089] Process compensation is performed on the process flow.

[0090] If the target parameter is a temperature parameter and the process step is a deposition step, the target parameter being in the second abnormal state will result in an excessively thick silicon wafer coating. In this case, the process should be terminated directly.

[0091] When the target parameter is an RF parameter and the process step is deposition, compensation is performed on the process progress. Since RF arcs only occur during the deposition step, the aforementioned target parameter being in a second abnormal state will result in a thinner silicon wafer coating. In this case, a compensation method can be adopted to compensate for a certain period of time in this process step before skipping steps to ensure the coating thickness. The compensation time can be determined based on the number of RF arcs that occur.

[0092] If the target parameter is a temperature parameter and the process step is a deposition step or a heating step, and the target parameter is in the third abnormal state, then process compensation processing will be performed on the process.

[0093] If the target parameter is in the third abnormal state during the heating or deposition step, it will result in a thinner silicon wafer coating. In this case, compensation measures can be taken to compensate for a certain amount of time in the process step before skipping the step to ensure the coating thickness. If it is another process step, the impact is smaller and can be ignored.

[0094] In one possible embodiment, if the actual parameter value at a preset time point and the normal parameter value at the corresponding preset time point in the parameter relationship graph, which shows the change of the target parameter with preset time precision, satisfy preset abnormal conditions, an abnormality in the target parameter in the process step is determined, including:

[0095] Compare the actual parameter values ​​at each preset time point with the corresponding normal parameter values;

[0096] If the actual parameter value corresponding to the preset time point is greater than the corresponding normal parameter value, and the larger value is greater than the preset sixth threshold, the number of larger values ​​is counted to obtain the total number of larger values.

[0097] If the actual parameter value corresponding to the preset time point is less than the corresponding normal parameter value, and the smaller value is less than the preset seventh threshold, the number of smaller values ​​is counted to obtain the total number of smaller values.

[0098] If the total number of times the parameter is too large and the total number of times the parameter is too small meet the preset number of times condition, perform Fourier transform on the actual parameter value to obtain the spectrum curve corresponding to the actual parameter value;

[0099] If the spectrum curve meets the preset discrete conditions, the abnormality of the target parameter in the process step is determined.

[0100] Generally, current parameter detection technology considers sawtooth-shaped periodic fluctuations—that is, parameter values ​​that alternate between abnormally large and abnormally small fluctuations—to be normal. Such anomalies typically do not affect the process results, but they do indicate a problem with the corresponding hardware control capabilities of the equipment.

[0101] Timely hardware maintenance and repair can prevent disruptions to the manufacturing process, saving significant time and increasing equipment productivity.

[0102] In this situation, the actual parameter values ​​at each preset time point can be compared with the corresponding normal parameter values; for example... Figure 6 As shown, if the actual parameter value corresponding to the preset time point is greater than the corresponding normal parameter value, and the value of the larger value is greater than the preset sixth threshold, the number of larger values ​​is counted to obtain the total number of larger values; if the actual parameter value corresponding to the preset time point is less than the corresponding normal parameter value, and the value of the smaller value is less than the preset seventh threshold, the number of smaller values ​​is counted to obtain the total number of smaller values.

[0103] The step of performing a Fourier transform on the actual parameter values ​​to obtain the spectrum curve corresponding to the actual parameter values, under the condition that the total number of over-optimizations and the total number of under-optimizations meet the preset number conditions, may specifically include:

[0104] If the total number of times the error is too large N1 > the expected threshold for being too large; and the total number of times the error is too small N2 > the expected threshold for being too small; and |N1-N2| < δ1 (δ1 is close to 0).

[0105] If both the total number of times the parameter is too large and the total number of times it is too small exceed the expected threshold, and the total number of times the parameter is too large and the total number of times it is too small are similar, then Fourier transform is performed on the parameter value to obtain the spectrum curve corresponding to the parameter value, so as to start judging whether the target parameter has periodicity.

[0106] Among them, Fourier transform is one of the most basic methods in time-domain and frequency-domain transformation analysis. When the spectrum curve corresponding to the parameter value obtained by Fourier transform meets the preset discrete conditions, the anomaly of the target parameter in the process step can be determined.

[0107] Among these steps, when the spectral curve meets preset discretization conditions, the anomalies of target parameters in the process steps are determined, including:

[0108] From the spectrum curve, determine the number of actual parameter values ​​whose frequency values ​​are less than the preset eighth threshold;

[0109] If the ratio of the number of actual parameter values ​​with a frequency value less than the preset eighth threshold to the number of preset time points is greater than the preset ninth threshold, then the target parameter in the process step is determined to be abnormal.

[0110] If the parameter values ​​change periodically, the spectrum curve will be discrete, such as... Figure 7 As shown, the parameter value only appears within a limited number of frequencies; this indicates that the target parameter only appears within a limited number of frequencies, which is an abnormal sawtooth-shaped periodic fluctuation. It can be determined that the target parameter in the process step is abnormal and an exception handling needs to be triggered.

[0111] If the actual parameter values ​​exhibit non-periodic changes, then the spectrum curve will be continuous, such as... Figure 8 As shown, the actual parameter values ​​appear randomly within various frequencies. This indicates that the target parameter appearing randomly within various frequencies is not an abnormal sawtooth-shaped periodic fluctuation, and it can be determined that the target parameter in the process step has not become abnormal.

[0112] in, Figure 7 , Figure 8 The horizontal axis represents frequency, and the vertical axis represents parameter values.

[0113] Compare each point on the spectrum curve N`(f) with the value 0. If N`(f)-0 < δ2 (δ2 is close to 0), then the point can be defaulted to 0. Determine the number of parameter values ​​N3 = N3+1 whose frequency value is less than the preset eighth threshold (δ2). If N`(f)-0 > δ2, then the value is defaulted to non-zero.

[0114] If the ratio of the number of samples (N3) to the number of samples (N4) is greater than the preset ninth threshold (90%), it indicates that the target parameter only appears within a limited number of frequencies, which is an abnormal sawtooth-shaped periodic fluctuation. It can be determined that the target parameter in the process step is abnormal and an abnormality handling needs to be triggered.

[0115] If the ratio of the number of samples (N3) to the number of samples (N4) is less than the preset ninth threshold (90%), it indicates that the target parameter appears randomly in various frequencies and does not belong to abnormal sawtooth periodic fluctuations. It can be determined that the process has not been abnormal.

[0116] In one possible embodiment, the target parameter is a temperature parameter, and the process is adjusted according to an adjustment scheme corresponding to preset abnormal conditions, including:

[0117] When the process is completed, an adjustment command is output to the temperature controller, which is used to adjust the temperature parameters of the temperature controller.

[0118] The target parameter is the angle parameter of the butterfly valve. The process is adjusted according to the adjustment plan corresponding to the preset abnormal conditions, including:

[0119] When the process is completed, a learning command is output to the butterfly valve. The learning command is used by the butterfly valve to learn and adjust the angle parameters.

[0120] On the one hand, when the process is completed, an adjustment command is output to the temperature controller, which is used to adjust the temperature parameters of the temperature controller.

[0121] The parameters that typically require handling sawtooth-shaped periodic abnormal fluctuations are temperature and butterfly valve angle. After prolonged operation, the equipment's stability deteriorates, and this type of anomaly indicates that the equipment needs maintenance. Therefore, the current process will not be processed, but the software will automatically activate the temperature and butterfly valve maintenance mode after the process ends. After maintenance, the process can continue.

[0122] For temperature as the target parameter, the temperature controller has a self-tuning function. After detecting that the spectrum curve meets the preset discrete conditions, it can automatically send an adjustment command to the temperature controller after the current process ends, so that the temperature controller can complete the adjustment of the temperature control PID controller (Proportion Integral Differential, PID). Among them, the PID controller, which controls according to the proportional (P), integral (I), and derivative (D) of the deviation, is the most widely used automatic controller in process control.

[0123] On the other hand, when the process is completed, a learning command is output to the butterfly valve. The learning command is used by the butterfly valve to learn and adjust the angle parameters.

[0124] A butterfly valve, also known as a flap valve, is a simple regulating valve used for on / off control of low-pressure pipeline media. A butterfly valve is characterized by its closing element (valve disc or butterfly plate) being a disk that rotates around a valve shaft to open and close. Valves can be used to control the flow of various types of fluids, including air, water, steam, various corrosive media, slurry, oil, liquid metals, and radioactive media. In pipelines, they primarily function as shut-off and throttling devices. The butterfly valve's opening and closing element is a disc-shaped butterfly plate that rotates around its own axis within the valve body, thereby achieving opening, closing, or regulation.

[0125] When the target parameter is an angle parameter, since the butterfly valve has a learning function, after detecting that the spectrum curve meets the preset discrete conditions, it can automatically send a learning command to the butterfly valve after the current process is completed, and complete the adjustment of PI (proportional (P) and integral (I) of the deviation).

[0126] During the process, the actual parameter values ​​collected can be compared with the normal parameter values ​​at the corresponding time in the parameter relationship diagram. This allows for more accurate detection of whether the target parameters in the process steps are abnormal. It also enables earlier detection of abnormalities and shorter processing time. Furthermore, it allows for more accurate detection of abnormal states corresponding to parameters and more accurate triggering of corresponding abnormality handling schemes. As a result, it can not only improve equipment capacity but also enhance the equipment's ability to handle abnormalities.

[0127] According to an embodiment of the present invention, by collecting the actual parameter values ​​of the target parameter at each preset time point in the process step, and discretizing the actual parameter values ​​of the target parameter into a pre-acquired parameter relationship diagram, if the actual parameter value of the target parameter at the preset time point, which varies with the preset time precision, meets the preset abnormal condition with the normal parameter value at the corresponding preset time point in the parameter relationship diagram, an abnormality in the target parameter in the process step is determined. Here, by comparing the actual parameter value of the target parameter in the process step with the normal parameter value of the target parameter in the process step, it is possible to accurately and quickly detect whether a parameter abnormality occurs in the process. This allows for the rapid and efficient triggering of corresponding adjustment schemes to adjust the process progress and output of prompt information corresponding to the preset abnormal condition. Thus, by improving the detection speed of parameter abnormalities, not only can equipment capacity be increased, but also the ability of the equipment to handle abnormalities can be improved, thereby enhancing safety.

[0128] In addition, based on the above Figure 2 The parameter anomaly detection method shown, and Figure 6 In response to the anomalies shown, this invention also provides a method for detecting parameter anomalies. Figure 9 A flowchart of a method for detecting parameter anomalies provided in an embodiment of the present invention is shown below:

[0129] 1. When the process starts or skips a step, first, clear the sum of the cumulative large values ​​S1, the sum of the cumulative small values ​​S2, and the sum of the total cumulative deviation values ​​S, that is, S1=0, S2=0, S3=0;

[0130] 2. Collect the actual values ​​of the target parameters according to the time precision of the two-dimensional matrix, and discretize the actual values ​​of the parameters into the two-dimensional matrix;

[0131] 3. Compare the actual value D(t) with the normal value B(t) at the corresponding time in the curve to obtain the deviation value N(t), where N(t) = D(t) - B(t);

[0132] 4. If N(t)>0, calculate the cumulative larger value S1, S1=S1+N(t);

[0133] 5. If N(t) < 0, calculate the cumulative smaller value S2, S2 = S2 + |N(t)|, where |N(t)| represents the absolute value of the deviation;

[0134] 6. If the process step has ended, proceed to step 7; otherwise, repeat steps 2, 3, 4, and 5.

[0135] 7. After the process step is completed, calculate the total cumulative deviation value S for that step, i.e., S = S1 - S2;

[0136] 8. Determine if the cumulative large value S1 is abnormal: Compare S1 with the expected threshold P1. If S1-P1 is greater than 0, it belongs to the (fluctuation anomaly 1) large value anomaly, which can trigger the anomaly handling module. If S1-P1 is less than or equal to 0, it is normal.

[0137] 9. Determine the cumulative smaller value S2 N Is there an anomaly? (S2) N Compared with the expected threshold P2, if S2 N If -P2 is greater than 0, it belongs to the (fluctuation anomaly 2) slightly smaller anomaly, which can trigger the anomaly handling module. If S2 N If -P2 is less than or equal to 0, then it is normal;

[0138] 10. Determine if the total cumulative deviation S in this step is abnormal. Compare S with the expected threshold P. If S... N >0 and S N If -P is greater than 0, it belongs to the (fluctuation anomaly 3) larger anomaly, which can trigger the anomaly handling module; if S N <0 and |S N If |-P is greater than 0, it belongs to the (fluctuation anomaly 4) small anomaly, which can trigger the anomaly handling module; otherwise, it is normal.

[0139] Here, in response to the detection of a mismatch between a parameter value and a target parameter value, the deviation between the two values ​​is determined; the time point corresponding to the parameter value coincides with the time point corresponding to the target parameter value; if the deviation value meets preset abnormality conditions, the target parameter in the process step is determined to be abnormal. During the process, this allows for more accurate detection of abnormalities; it also enables earlier detection and shorter processing times; and it allows for more accurate detection of parameter abnormality types, triggering corresponding abnormality handling schemes more precisely. Therefore, it not only improves equipment capacity but also enhances the equipment's ability to handle abnormalities.

[0140] In addition, based on the above Figure 2 The parameter anomaly detection method shown is for Figure 6 In response to the anomaly shown, this invention also provides another method for detecting parameter anomalies. Figure 10 A flowchart of another method for implementing parameter anomaly detection provided by an embodiment of the present invention is shown below:

[0141] 1. At the start of the process step, clear the cumulative overshoot counter N1, cumulative undershoot counter N2, periodic detection counter N3, and sampling counter N4, i.e., N1 = 0; N2 = 0; N3 = 0; N4 = 0; where:

[0142] Accumulated overage count counter N1: This counter increments by 1 for each overage anomaly.

[0143] Accumulated underestimation count counter N2: This counter increments by 1 for each underestimation anomaly.

[0144] Periodic detection counter N3: When detecting whether the actual value of the detection parameter has periodicity, the counter increments by 1 for each approximate 0 value that appears in the spectrum.

[0145] Sampling counter N4: The total number of samples in this process step. The counter increments by 1 for each sample taken.

[0146] 2. Collect the actual values ​​of the target parameters according to the time precision T of the two-dimensional matrix, and discretize the actual values ​​of the parameters into the two-dimensional matrix. Each time data is collected, the sampling counter N4 = N4 + 1.

[0147] 3. Compare the actual value of the target parameter D(t) with the normal value B(t) at the corresponding time in the curve to obtain the deviation value N(t), where N(t) = D(t) - B(t);

[0148] 4. Determine the fluctuation range of N(t). If N(t) is greater than 0 and greater than the expected threshold, the fluctuation is abnormally large, and the cumulative number of times it is abnormally large is counted as N1 = N1 + 1. If N(t) is less than 0 and less than the expected threshold, the fluctuation is abnormally small, and the cumulative number of times it is abnormally small is counted as N2 = N2 + 1.

[0149] 5. If the time for this process step has ended, proceed to step 6; otherwise, repeat steps 2, 3, and 4.

[0150] 6. After the process step time is completed, determine whether periodic detection is required. If N1 > expected large threshold and N2 > expected small threshold and |N1-N2| < δ1 (δ1 is close to 0), that is, the cumulative large and cumulative small count counters have exceeded the expected thresholds and are not much different, then start to detect whether the target parameter has periodicity.

[0151] 7. Perform a Fourier transform on the actual value curve D(t) of the target parameter to obtain the spectrum curve N'(f) of the actual value of the target parameter. If it is a periodic function, the spectrum curve N'(f) is discrete, such as... Figure 9 As shown, the parameter values ​​only appear within a finite number of frequencies; if it is a non-periodic function, then the N`(f) spectrum curve is continuous, as shown... Figure 10 As shown, the parameter values ​​appear randomly across various frequencies. Figure 9 , Figure 10 The horizontal axis represents frequency, and the vertical axis represents parameter values.

[0152] 8. Compare each point on the spectrum curve N`(f) with the value 0. If N`(f)-0 < δ2 (δ2 is close to 0), then the point can be defaulted to 0, and the period detection counter N3 = N3+1. If N`(f)-0 > δ2, then the value is defaulted to non-zero.

[0153] 9. If the ratio of the periodic detection counter N3 to the sampling counter N4 exceeds 90%, i.e., N3 / N4>90%, then the target parameter is considered to appear only within a limited number of frequencies, which is an abnormal sawtooth periodic fluctuation, and anomaly handling needs to be triggered.

[0154] 10. If the ratio of the periodic detection counter N3 to the sampling counter N4 does not exceed 90%, i.e., N3 / N4<90%, then the target parameter is considered to appear randomly in various frequencies and does not belong to abnormal sawtooth periodic fluctuations. In this case, it is ignored and no abnormality is triggered.

[0155] According to an embodiment of the present invention, by collecting the actual parameter values ​​of the target parameter at each preset time point in the process step, and discretizing the actual parameter values ​​of the target parameter into a pre-acquired parameter relationship diagram, if the actual parameter value of the target parameter at the preset time point, which varies with the preset time precision, meets the preset abnormal condition with the normal parameter value at the corresponding preset time point in the parameter relationship diagram, an abnormality in the target parameter in the process step is determined. Here, by comparing the actual parameter value of the target parameter in the process step with the normal parameter value of the target parameter in the process step, it is possible to accurately and quickly detect whether a parameter abnormality occurs in the process. This allows for the rapid and efficient triggering of corresponding adjustment schemes to adjust the process progress and output of prompt information corresponding to the preset abnormal condition. Thus, by improving the detection speed of parameter abnormalities, not only can equipment capacity be increased, but also the ability of the equipment to handle abnormalities can be improved, thereby enhancing safety.

[0156] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0157] Reference Figure 11 This diagram illustrates a structural block diagram of a semiconductor process apparatus 1110 according to an embodiment of the present invention. The semiconductor process apparatus 1110 includes...

[0158] The controller 1111 is used to collect the parameter values ​​of the target parameter within a preset time period in the process; when the relationship between the parameter value and the pre-acquired parameter relationship diagram meets the preset abnormal conditions, it determines that the target parameter in the process step is abnormal; the parameter relationship diagram is used to indicate the relationship between the time and the parameter value of the target parameter in each process step; the process is adjusted according to the adjustment plan corresponding to the preset abnormal conditions; and the prompt information corresponding to the preset abnormal conditions is output.

[0159] In an optional embodiment of the present invention, the controller 1111 is specifically used to determine the preset parameter fluctuation range corresponding to each preset time point of the normal parameter value according to the parameter relationship diagram.

[0160] In response to detecting that the actual parameter value corresponding to the preset time point exceeds the parameter fluctuation range corresponding to the preset time point, a count is performed to obtain the number of anomalies;

[0161] If the number of abnormal occurrences exceeds a preset first threshold, the target parameter in the process step is determined to be abnormal, and the target parameter is in a first abnormal state.

[0162] If the number of abnormalities is less than the preset first threshold, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point is determined. If the deviation meets the preset abnormality condition, the target parameter in the process step is determined to be abnormal.

[0163] In an optional embodiment of the present invention, if the target parameter is in the first abnormal state, the controller 1111 is specifically used to terminate the process; output a first prompt message, the first prompt message being used to prompt for checking the equipment wiring status, and / or checking the plant gas source parameters.

[0164] In an optional embodiment of the present invention, the controller 1111 is specifically used to count the cumulative larger value between the actual parameter value and the normal parameter value when the actual parameter value at the preset time point is greater than the normal parameter value corresponding to the preset time point.

[0165] If the actual parameter value at the preset time point is less than the normal parameter value corresponding to the preset time point, then the cumulative smaller value between the actual parameter value and the normal parameter value is counted.

[0166] If the difference between the cumulative excessive value and the preset excessive threshold is greater than the preset second threshold, then the target parameter in the process step is determined to be abnormal; and / or

[0167] If the difference between the cumulative underestimation value and the preset underestimation threshold is greater than the preset third threshold, then the target parameter in the process step is determined to be abnormal.

[0168] In an optional embodiment of the present invention, the controller 1111 is specifically configured to: determine the difference between the actual parameter value at each preset time point and the normal parameter value corresponding to each preset time point, and sum the differences to obtain a total deviation value;

[0169] If the total deviation value is greater than zero, and the difference between the total deviation value and the preset fourth threshold is greater than zero, then it is determined that the target parameter in the process step is abnormal and the target parameter is in a second abnormal state.

[0170] If the total deviation value is less than zero, and the difference between the absolute value of the total deviation value and the preset fifth threshold is greater than zero, then it is determined that the target parameter in the process step is abnormal and the target parameter is in the third abnormal state.

[0171] In an optional embodiment of the present invention, the semiconductor process equipment is a coating equipment, wherein if the target parameter is a temperature parameter and the process step is a deposition step, the controller 1111 is specifically used to: terminate the process.

[0172] If the target process parameter is a radio frequency parameter and the process step is a deposition step, and the target parameter is in a second abnormal state, the controller 1111 is specifically used to: perform process compensation processing on the process process.

[0173] If the target process parameter is a temperature parameter and the process step is a deposition step or a heating step, and the target parameter is in a third abnormal state, the controller 1111 is specifically used to: perform process compensation processing on the process.

[0174] In an optional embodiment of the present invention, the controller 1111 is specifically used to: compare the actual parameter value corresponding to each preset time point with the corresponding normal parameter value;

[0175] If the actual parameter value corresponding to the preset time point is greater than the corresponding normal parameter value, and the larger value is greater than the preset sixth threshold, the number of larger values ​​is counted to obtain the total number of larger values.

[0176] If the actual parameter value corresponding to the preset time point is less than the corresponding normal parameter value, and the smaller value is less than the preset seventh threshold, the number of smaller values ​​is counted to obtain the total number of smaller values.

[0177] If the total number of times the parameter is too large and the total number of times the parameter is too small meet the preset number of times condition, perform Fourier transform on the actual parameter value to obtain the spectrum curve corresponding to the actual parameter value;

[0178] If the spectral curve satisfies the preset discrete conditions, the target parameter in the process step is determined to be abnormal.

[0179] In an optional embodiment of the present invention, the controller 1111 is specifically configured to: determine from the spectrum curve the number of actual parameter values ​​whose frequency values ​​are less than a preset eighth threshold;

[0180] If the ratio of the number of actual parameter values ​​whose determined frequency value is less than the preset eighth threshold to the number of preset time points is greater than the preset ninth threshold, then the target parameter in the process step is determined to be abnormal.

[0181] In an optional embodiment of the present invention, the controller 1111 is specifically configured to: output an adjustment command to a temperature controller when the process is completed, the command being used by the temperature controller to adjust the temperature parameter; the target parameter is the angle parameter of the butterfly valve; and adjusting the process according to the adjustment scheme corresponding to the preset abnormal condition includes:

[0182] When the process is completed, a learning command is output to the butterfly valve, which is used to learn and adjust the angle parameter.

[0183] According to an embodiment of the present invention, by collecting the actual parameter values ​​of the target parameter at each preset time point in the process step, and discretizing the actual parameter values ​​of the target parameter into a pre-acquired parameter relationship diagram, if the actual parameter value of the target parameter at the preset time point, which varies with the preset time precision, meets the preset abnormal condition with the normal parameter value at the corresponding preset time point in the parameter relationship diagram, an abnormality in the target parameter in the process step is determined. Here, by comparing the actual parameter value of the target parameter in the process step with the normal parameter value of the target parameter in the process step, it is possible to accurately and quickly detect whether a parameter abnormality occurs in the process. This allows for the rapid and efficient triggering of corresponding adjustment schemes to adjust the process progress and output of prompt information corresponding to the preset abnormal condition. Thus, by improving the detection speed of parameter abnormalities, not only can equipment capacity be increased, but also the ability of the equipment to handle abnormalities can be improved, thereby enhancing safety.

[0184] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0185] This invention also provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, it implements the various processes of the above-described parameter anomaly detection method embodiment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0186] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described parameter anomaly detection method embodiment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0187] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0188] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0189] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0190] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0191] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0192] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0193] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0194] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0195] The present invention has provided a detailed description of a parameter anomaly detection method and a semiconductor process equipment. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A parameter anomaly detection method, applied to semiconductor process equipment, wherein the process of the semiconductor process equipment includes multiple sequentially performed process steps, characterized in that, The parameter anomaly detection method includes: Collect the actual parameter values ​​of the target parameters at each preset time point in the process steps; The actual parameter value of the target parameter is discretized into a pre-acquired parameter relationship diagram. If the actual parameter value of the target parameter at a preset time point, which varies with a preset time precision, and the normal parameter value at the corresponding preset time point in the parameter relationship diagram satisfy a preset abnormality condition, the target parameter in the process step is determined to be abnormal. The parameter relationship diagram is a relationship diagram of the normal parameter value of the target parameter in the process step with each preset time point. The abnormality of the target parameter is determined by comparing the magnitude between the actual parameter value corresponding to each preset time point and the corresponding normal parameter value. The process is adjusted according to the adjustment scheme corresponding to the preset abnormal conditions; Output the prompt information corresponding to the preset abnormal condition; Wherein, determining the abnormality of the target parameter in the process step when the actual parameter value at a preset time point, which varies with the preset time precision, and the normal parameter value at the preset time point corresponding to the parameter relationship diagram satisfy preset abnormality conditions include: If the actual parameter value corresponding to the preset time point is greater than the corresponding normal parameter value, and the larger value is greater than the preset sixth threshold, the number of larger values ​​is counted to obtain the total number of larger values. If the actual parameter value corresponding to the preset time point is less than the corresponding normal parameter value, and the smaller value is less than the preset seventh threshold, the number of smaller values ​​is counted to obtain the total number of smaller values. If the total number of times the parameter is too large and the total number of times the parameter is too small meet the preset number of times condition, perform Fourier transform on the actual parameter value to obtain the spectrum curve corresponding to the actual parameter value; If the spectrum curve meets the preset discrete conditions, the target parameter in the process step is determined to be abnormal.

2. The method according to claim 1, characterized in that, The step of determining the anomaly of the target parameter in the process step when the spectrum curve satisfies the preset discrete condition includes: From the spectrum curve, determine the number of actual parameter values ​​whose frequency values ​​are less than a preset eighth threshold; If the ratio of the number of actual parameter values ​​whose determined frequency value is less than the preset eighth threshold to the number of preset time points is greater than the preset ninth threshold, then the target parameter in the process step is determined to be abnormal.

3. The method according to claim 1, characterized in that, The target parameter is a temperature parameter, and adjusting the process according to the adjustment scheme corresponding to the preset abnormal conditions includes: Upon completion of the process, an adjustment command is output to the temperature controller, which is used to adjust the temperature parameter. The target parameter is the angle parameter of the butterfly valve. Adjusting the process according to the adjustment scheme corresponding to the preset abnormal conditions includes: Upon completion of the process, a learning command is output to the butterfly valve, which is used to learn and adjust the angle parameter.

4. A parameter anomaly detection method, applied to semiconductor process equipment, wherein the process of the semiconductor process equipment includes multiple sequentially performed process steps, characterized in that, The parameter anomaly detection method includes: Collect the actual parameter values ​​of the target parameters at each preset time point in the process steps; The actual parameter value of the target parameter is discretized into a pre-acquired parameter relationship diagram. If the actual parameter value of the target parameter at a preset time point, which varies with a preset time precision, and the normal parameter value at the corresponding preset time point in the parameter relationship diagram satisfy a preset abnormality condition, the target parameter in the process step is determined to be abnormal. The parameter relationship diagram is a relationship diagram of the normal parameter value of the target parameter in the process step with each preset time point. The abnormality of the target parameter is determined by comparing the magnitude between the actual parameter value corresponding to each preset time point and the corresponding normal parameter value. The process is adjusted according to the adjustment scheme corresponding to the preset abnormal conditions; Output the prompt information corresponding to the preset abnormal condition; Wherein, determining the abnormality of the target parameter in the process step when the actual parameter value at a preset time point, which varies with the preset time precision, and the normal parameter value at the preset time point corresponding to the parameter relationship diagram satisfy preset abnormality conditions include: Based on the parameter relationship diagram, the normal parameter value is determined within the preset parameter fluctuation range corresponding to each preset time point; In response to detecting that the actual parameter value corresponding to the preset time point exceeds the parameter fluctuation range corresponding to the preset time point, a count is performed to obtain the number of anomalies; If the number of abnormal occurrences exceeds a preset first threshold, the target parameter in the process step is determined to be abnormal, and the target parameter is in a first abnormal state. If the number of abnormalities is less than the preset first threshold, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point is determined. If the deviation meets the preset abnormality condition, the target parameter in the process step is determined to be abnormal. The step of determining the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point, and determining that the target parameter in the process step is abnormal if the deviation meets the preset abnormality condition, includes: If the actual parameter value at the preset time point is greater than the normal parameter value corresponding to the preset time point, then the cumulative larger value between the actual parameter value and the normal parameter value is counted. If the actual parameter value at the preset time point is less than the normal parameter value corresponding to the preset time point, then the cumulative smaller value between the actual parameter value and the normal parameter value is counted. If the difference between the cumulative excessive value and the preset excessive threshold is greater than the preset second threshold, then the target parameter in the process step is determined to be abnormal; and / or If the difference between the cumulative underestimation value and the preset underestimation threshold is greater than the preset third threshold, then the target parameter in the process step is determined to be abnormal.

5. The method according to claim 4, characterized in that, If the target parameter is in the first abnormal state, adjusting the process according to the adjustment scheme corresponding to the preset abnormal condition includes: terminating the process. The output of prompt information corresponding to the preset abnormal conditions includes: outputting first prompt information, which is used to prompt for checking the equipment wiring and / or checking the plant gas source parameters.

6. A parameter anomaly detection method, applied to semiconductor process equipment, wherein the process of the semiconductor process equipment includes multiple sequentially performed process steps, characterized in that, The parameter anomaly detection method includes: Collect the actual parameter values ​​of the target parameters at each preset time point in the process steps; The actual parameter value of the target parameter is discretized into a pre-acquired parameter relationship diagram. If the actual parameter value of the target parameter at a preset time point, which varies with a preset time precision, and the normal parameter value at the corresponding preset time point in the parameter relationship diagram satisfy a preset abnormality condition, the target parameter in the process step is determined to be abnormal. The parameter relationship diagram is a relationship diagram of the normal parameter value of the target parameter in the process step with each preset time point. The abnormality of the target parameter is determined by comparing the magnitude between the actual parameter value corresponding to each preset time point and the corresponding normal parameter value. The process is adjusted according to the adjustment scheme corresponding to the preset abnormal conditions; Output the prompt information corresponding to the preset abnormal condition; Wherein, determining the abnormality of the target parameter in the process step when the actual parameter value at a preset time point, which varies with the preset time precision, and the normal parameter value at the preset time point corresponding to the parameter relationship diagram satisfy preset abnormality conditions include: Based on the parameter relationship diagram, the normal parameter value is determined within the preset parameter fluctuation range corresponding to each preset time point; In response to detecting that the actual parameter value corresponding to the preset time point exceeds the parameter fluctuation range corresponding to the preset time point, a count is performed to obtain the number of anomalies; If the number of abnormal occurrences exceeds a preset first threshold, the target parameter in the process step is determined to be abnormal, and the target parameter is in a first abnormal state. If the number of abnormalities is less than the preset first threshold, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point is determined. If the deviation meets the preset abnormality condition, the target parameter in the process step is determined to be abnormal. Specifically, the deviation between the actual parameter value and the corresponding normal parameter value at each preset time point is determined. If the deviation meets the preset abnormality condition, the target parameter in the process step is determined to be abnormal, including: Determine the difference between the actual parameter value at each preset time point and the normal parameter value corresponding to each preset time point, and sum the differences to obtain the total deviation value; If the total deviation value is greater than zero, and the difference between the total deviation value and the preset fourth threshold is greater than zero, then it is determined that the target parameter in the process step is abnormal and the target parameter is in a second abnormal state. If the total deviation value is less than zero, and the difference between the absolute value of the total deviation value and the preset fifth threshold is greater than zero, then it is determined that the target parameter in the process step is abnormal and the target parameter is in the third abnormal state.

7. The method according to claim 6, characterized in that, The semiconductor process equipment is a deposition equipment. If the target parameter is a temperature parameter and the process step is a deposition step, and the target parameter is in a second abnormal state, then adjusting the process according to the adjustment scheme corresponding to the preset abnormal condition includes: The process is terminated; If the target parameter is an RF parameter and the process step is a deposition step, and the target parameter is in a second abnormal state, then adjusting the process according to the adjustment scheme corresponding to the preset abnormal condition includes: The process is compensated by performing process compensation. If the target parameter is a temperature parameter and the process step is a deposition step or a heating step, and the target parameter is in a third abnormal state, then adjusting the process according to the adjustment scheme corresponding to the preset abnormal condition includes: The process is compensated by process compensation.

8. A semiconductor process apparatus, characterized in that, The semiconductor process equipment includes multiple sequential process steps, and the semiconductor process equipment includes a controller for executing the parameter anomaly detection method according to any one of claims 1-7.

Images