Method for monitoring slow progressive deviation of product parameters

Abstract

The invention relates to a method for monitoring slow and gradual deviation of product parameters, which belongs to the technical field of silicon wafer processing and comprises the following operation steps of: 1, based on an SPC tool, adding a shift monitoring function by supplementing an online function; 2, controlling a common metering type SPC control chart to have an I graph amp; an MR graph and an Xbar graph amp; amp is carried out on an R graph or an Xbar graph; and combining three types of sigma diagrams. And 3, averaging the parameter values of the fixed monitoring number of the I graph or the Xbar graph, and taking the average value as the process level of batch processing to form a shift graph. And 4, performing movement range on the shift graph formed by the fixed monitoring number to form a shift R graph. And 5, taking 1.5 sigma of calculation data in the shiftR graph to set a control limit, establishing shift monitoring, and starting an alarm when the control limit is exceeded. The method has the advantages of convenient operation, high accuracy and good stability. The problems that in the silicon wafer processing process, the slow and progressive process drifting phenomenon is recognized from the long-term view angle, and reminding is given out are solved.

Description

Technical Field

[0001] The invention relates to the technical field of silicon wafer processing, and in particular to a method for monitoring the slow and progressive deviation of product parameters. Background Art

[0002] Statistical Process Control (SPC) is a process control tool that uses mathematical statistics. It analyzes and evaluates the production process, promptly discovers signs of systematic factors based on feedback information, and takes measures to eliminate their impact, so that the process is maintained in a controlled state affected only by random factors, in order to achieve the purpose of quality control.

[0003] Wafer manufacturing is a relatively complex process, involving complex chemical and physical reactions such as grinding, etching, and polishing. It is easily affected by the machine, fixture, chemical solution, time, temperature and other conditions, resulting in fluctuations in product parameters. Commonly used SPC tools can achieve online monitoring of continuous processing status, and are better at detecting sudden and significant changes, but are not sensitive enough to slow and gradual process drift. During the silicon wafer processing process, such gradual changes will cause the quality of the final product to not meet the requirements, and the SPC system cannot capture these changes in time. Summary of the invention

[0004] The present invention mainly solves the deficiencies in the prior art and provides a method for monitoring the slow and gradual drift of product parameters, which has the advantages of convenient operation, high accuracy and good stability. It solves the problem of identifying and issuing reminders for the slow and gradual process drift phenomenon in silicon wafer processing from a medium and long-term perspective.

[0005] The above technical problems of the present invention are mainly solved by the following technical solutions: A method for monitoring a slow and progressive deviation of a product parameter comprises the following steps: Step 1: Based on the SPC tool, the shift monitoring function is added by supplementing the online function.

[0006] Step 2: Combine the commonly used quantitative SPC control charts into three types: I chart & MR chart, Xbar chart & R chart, or Xbar chart & σ chart.

[0007] Step 3: Take the average value of the parameter values ​​of the fixed monitoring number of the I chart or Xbar chart as the process level of the processing batch to form a shift chart.

[0008] Step 4: Make a moving range for the shift graph formed by a fixed number of monitoring items to form a shift_R graph.

[0009] Step 5: Set the control limit of 1.5σ of the calculated data on the shift_R chart, establish shift monitoring, and enable the alarm for exceeding the control limit.

[0010] Preferably, the purpose of Shift monitoring is to identify the slow and gradual deviation of product parameters, especially the stratification phenomenon caused by the inherent differences of factors such as raw material batches, processing machines, and processing cycles. Therefore, when selecting the number of monitoring items Z, it should be considered to be representative of the identifiable inherent differences. The number of monitoring items Z is 5, 10, 20 or 25, and monitoring is carried out according to the processing batch, cycle or process characteristics.

[0011] Preferably, a processing batch is divided into 14 to 15 sub-groups for processing. When taking the monitoring number Z, the situation that can represent the raw material mother batch should be considered as much as possible, and Z≤14 should be satisfied, and Z=5 or 10 should be taken. The repair and maintenance cycle of a processing machine is generally 28 to 32 batches. When taking the monitoring number Z, the situation that can represent the machine cycle should be considered as much as possible, and Z≤28 should be satisfied, and Z=20 or Z=25 should be taken.

[0012] As a preferred option, the SPC software adds Shift Chart definition information, establishes data acquisition settings based on the main chart, forms ShiftX and Shift_R charts, and realizes the shift monitoring function: ShiftX takes the average value of a fixed number of main chart points to form a trend chart, and Shift_R takes the moving range value of ShiftX to form a trend chart.

[0013] As a preference, according to the sample capacity of the monitoring project, select the corresponding SPC control chart to establish management and control, and the specific control chart type selection rules are as follows: ① Each batch of test samples n=1: I&MR chart, I is a single value, MR is the absolute value of the difference between the current test sample result and the previous test sample result; ② Each batch of test samples n=2~8: Xbar&R chart, Xbar is the average value, R is the range of the current test sample result; ③ Each batch of test samples n≥9: Xbar&σ chart, Xbar is the average value, R is the standard deviation of the current test sample result; I chart and Xbar chart are the main charts, MR chart, R chart, and σ chart are auxiliary charts.

[0014] Preferably, according to the SPC control chart type, when the main chart is an I chart, the test result of each batch is taken as one monitoring point, and when the main chart is an Xbar chart, the average value of the test result of each batch is taken as one monitoring point, and the test results are continuously collected to form a trend chart of the parameter.

[0015] Preferably, the number of continuous monitoring points is selected as Z according to the processing batch, cycle or process characteristics, that is, the average value of every Z monitoring points is taken to obtain the shift monitoring value ShiftX, and the ShiftX shift trend chart of the parameter is continuously collected to form: ShiftX=ΣXi / Z (i=1,2,3,…,Z, Z∈N).

[0016] Preferably, when the ShiftX shift trend chart collects 2 or more values, the previous value is subtracted from the second value, and the absolute value is taken to obtain the moving range, which is continuously collected to form a moving range shift chart Shift_R: Shift_R=|ShiftXj+1-ShiftXj| (j∈N).

[0017] As a preferred method, according to the normal distribution trend, the probability that the monitoring value of the main chart meets ±3σ is 99.83%, and the processing state at this time can be regarded as a stable process. The Shift chart takes the average value of Z monitoring points to get shiftX, and the distribution probability of the collected shiftX value in the ±3σ interval also meets 99.83%; Shift_R takes the absolute value, so the control limit is set to +3σ.

[0018] The present invention can achieve the following effects: The present invention provides a method for monitoring the slow and gradual drift of product parameters, which has the advantages of convenient operation, high accuracy and good stability compared with the prior art. It solves the problem of identifying and issuing reminders for the slow and gradual process drift phenomenon in silicon wafer processing from a medium and long-term perspective. The purpose is to provide a simple and practical method for monitoring the stability of product parameters, identifying and issuing reminders for the slow and gradual process drift phenomenon. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a diagram of the interface of the monitoring product parameter system of the present invention.

[0020] Figure 2 It is the shift diagram of the present invention. DETAILED DESCRIPTION

[0021] The technical solution of the invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0022] Example: Figure 1 and Figure 2 As shown, a method for monitoring the slow and progressive deviation of product parameters includes the following steps: Step 1: Based on the SPC tool, the shift monitoring function is added by supplementing the online function. The SPC software adds the Shift Chart definition information, establishes the data acquisition settings based on the main chart, forms the ShiftX and Shift_R charts, and realizes the shift monitoring function: ShiftX takes the average value of a fixed number of main chart points to form a trend chart, and Shift_R takes the moving range value of ShiftX to form a trend chart.

[0023] Step 2: Combine the commonly used quantitative SPC control charts into three types: I chart & MR chart, Xbar chart & R chart, or Xbar chart & σ chart.

[0024] According to the sample capacity of the monitoring project, select the corresponding SPC control chart to establish management and control. The specific control chart type selection rules are as follows: ① The sample size of each batch of inspection n=1: I&MR chart, I is a single value, MR is the absolute value of the difference between the sample result of this inspection and the sample result of the previous inspection; ② The sample size of each batch of inspection n=2~8: Xbar&R chart, Xbar is the average value, R is the range of the sample result of this inspection; ③ The sample size of each batch of inspection n≥9: Xbar&σ chart, Xbar is the average value, R is the standard deviation of the sample result of this inspection; I chart and Xbar chart are the main charts, MR chart, R chart, and σ chart are auxiliary charts.

[0025] Step 3: Take the average value of the parameter values ​​of the fixed monitoring number of the I chart or Xbar chart as the process level of the processing batch to form a shift chart.

[0026] The purpose of shift monitoring is to identify the slow and gradual deviation of product parameters, especially the stratification phenomenon caused by the inherent differences of factors such as raw material batches, processing machines, and processing cycles. Therefore, when selecting the number of monitoring items Z, we should try to consider the stratification of the identifiable inherent differences; the number of monitoring items Z is 5, 10, 20 or 25, and the monitoring is carried out according to the processing batch, cycle or process characteristics.

[0027] A processing batch is divided into 14 to 15 sub-groups for processing. When taking the monitoring number Z, we should try our best to consider the situation that can represent the raw material mother batch, satisfy Z≤14, and take Z=5 or 10; the repair and maintenance cycle of a processing machine is generally 28 to 32 batches. When taking the monitoring number Z, we should try our best to consider the situation that can represent the machine cycle, satisfy Z≤28, and take Z=20 or Z=25.

[0028] According to the SPC control chart type, when the main chart is an I chart, the test result of each batch is taken as one monitoring point; when the main chart is an Xbar chart, the average value of the test result of each batch is taken as one monitoring point, and the test results are continuously collected to form a trend chart of the parameter.

[0029] According to the processing batch, cycle or process characteristics, the number of continuous monitoring points is selected as Z, that is, the average value of every Z monitoring points is taken to obtain the shift monitoring value ShiftX, and the ShiftX shift trend chart of the parameter is continuously collected to form: ShiftX=ΣXi / Z (i=1,2,3,…,Z, Z∈N).

[0030] Step 4: Make a moving range for the shift chart formed by a fixed number of monitorings to form a shift_R chart. When the ShiftX trend chart collects 2 or more values, subtract the previous value from the second value, take the absolute value to get the moving range, and continue to collect to form the moving range chart Shift_R: Shift_R=|ShiftXj+1-ShiftXj| (j∈N).

[0031] Step 5: The shift_R chart takes 1.5σ of the calculated data to set the control limit, establishes shift monitoring, and turns on the alarm for exceeding the control limit. According to the normal distribution trend, the probability that the monitoring value of the main chart meets ±3σ is 99.83%, and the processing state at this time can be regarded as a stable process. The shift chart takes the average value of Z monitoring points to get shiftX, and the distribution probability of the collected shiftX values ​​in the ±3σ interval also meets 99.83%; Shift_R takes the absolute value, so the control limit set is +3σ.

[0032] In summary, the method for monitoring the slow and gradual drift of product parameters has the advantages of convenient operation, high accuracy and good stability. Based on the SPC tool, on the basis of the original function of monitoring continuous short-term changes of the SPC tool, batch statistics for specific processing batches, cycles or process characteristics are added. The problem of identifying and issuing reminders for slow and gradual process drift phenomena from a medium and long-term perspective in silicon wafer processing is solved.

[0033] The above description is only a specific embodiment of the present invention, but the structural features of the present invention are not limited thereto. Any changes or modifications made by any technician in the field of the present invention are included in the patent scope of the present invention.

Claims

1. A method for monitoring the slow and progressive deviation of product parameters, characterized in that The steps are as follows: Step 1: Based on the SPC tool, the shift monitoring function is added by supplementing the online function; Step 2: Combine the commonly used quantitative SPC control charts into three types: I chart & MR chart, Xbar chart & R chart, or Xbar chart & σ chart; Step 3: Take the average value of the parameter values ​​of the fixed monitoring number of the I chart or Xbar chart as the process level of the processing batch to form a shift chart; Step 4: Make a moving range for the shift graph formed by a fixed number of monitoring items to form a shift_R graph; Step 5: Set the control limit of 1.5σ of the calculated data on the shift_R chart, establish shift monitoring, and enable the alarm for exceeding the control limit.

2. The method for monitoring the slow and progressive shift of product parameters according to claim 1, characterized in that: The purpose of shift monitoring is to identify the slow and gradual deviation of product parameters, especially the stratification phenomenon caused by the inherent differences of factors such as raw material batches, processing machines, and processing cycles. Therefore, when selecting the number of monitoring items Z, we should try to consider the stratification of the identifiable inherent differences; the number of monitoring items Z is 5, 10, 20 or 25, and the monitoring is carried out according to the processing batch, cycle or process characteristics.

3. The method for monitoring the slow and progressive shift of product parameters according to claim 2, characterized in that: A processing batch is divided into 14 to 15 sub-groups for processing. When taking the monitoring number Z, we should try our best to consider the situation that can represent the raw material mother batch, satisfy Z≤14, and take Z=5 or 10; the repair and maintenance cycle of a processing machine is generally 28 to 32 batches. When taking the monitoring number Z, we should try our best to consider the situation that can represent the machine cycle, satisfy Z≤28, and take Z=20 or Z=25.

4. The method for monitoring the slow and progressive shift of product parameters according to claim 1, characterized in that: The SPC software adds Shift Chart definition information, establishes data acquisition settings based on the main chart, forms ShiftX and Shift_R charts, and implements the shift monitoring function: ShiftX takes the average value of a fixed number of main chart points to form a trend chart, and Shift_R takes the moving range value of ShiftX to form a trend chart.

5. The method for monitoring the slow and progressive shift of product parameters according to claim 2, characterized in that: According to the sample capacity of the monitoring project, select the corresponding SPC control chart to establish management and control. The specific control chart type selection rules are as follows: ① The sample size of each batch of inspection n=1: I&MR chart, I is a single value, MR is the absolute value of the difference between the sample result of this inspection and the sample result of the previous inspection; ② The sample size of each batch of inspection n=2~8: Xbar&R chart, Xbar is the average value, R is the range of the sample result of this inspection; ③ The sample size of each batch of inspection n≥9: Xbar&σ chart, Xbar is the average value, R is the standard deviation of the sample result of this inspection; I chart and Xbar chart are the main charts, MR chart, R chart, and σ chart are auxiliary charts.

6. The method for monitoring the slow and progressive shift of product parameters according to claim 5, characterized in that: According to the SPC control chart type, when the main chart is an I chart, the test result of each batch is taken as one monitoring point; when the main chart is an Xbar chart, the average value of the test result of each batch is taken as one monitoring point, and the test results are continuously collected to form a trend chart of the parameter.

7. The method for monitoring the slow and progressive shift of product parameters according to claim 6, characterized in that: According to the processing batch, cycle or process characteristics, the number of continuous monitoring points is selected as Z, that is, the average value of every Z monitoring points is taken to obtain the shift monitoring value ShiftX, and the ShiftX shift trend chart of the parameter is continuously collected to form: ShiftX=ΣXi / Z (i=1,2,3,…,Z, Z∈N).

8. The method for monitoring the slow and progressive shift of product parameters according to claim 7, characterized in that: When the ShiftX shift trend chart collects 2 or more values, subtract the previous value from the second value, take the absolute value to get the moving range, and continue collecting to form the moving range shift chart Shift_R: Shift_R=|ShiftXj+1-ShiftXj| (j∈N).

9. The method for monitoring the slow and progressive shift of product parameters according to claim 8, characterized in that: According to the normal distribution trend, the probability that the monitoring value of the main chart meets ±3σ is 99.83%, and the processing state at this time can be regarded as a stable process; the Shift chart takes the average value of Z monitoring points to obtain shiftX, and the distribution probability of the collected shiftX values ​​in the ±3σ interval also meets 99.83%; Shift_R takes the absolute value, so the set control limit is +3σ.

Images