A method and system for endpoint determination for cleaning of a semiconductor process chamber

Abstract

The application discloses a kind of for semiconductor process chamber's cleaning endpoint determination method and system, and cleaning endpoint determination method includes the following steps, the optical parameter of characteristic gas generated in cleaning process is obtained, data processing is carried out to optical parameter, and the normalized absorption A of characteristic gas generated in cleaning process is obtained;Electrical parameter V is obtained;Data processing is carried out to optical parameter and electrical parameter V, and the optical endpoint index, optical weight, electrical endpoint index and electrical weight are obtained;Decision index S is obtained;The first threshold of normalized absorption A and the second threshold of decision index S are set;When decision index S is successively lower than the second threshold, and, determine that cleaning is completed.Cleaning endpoint determination system includes semiconductor chamber, tail gas discharge pipeline and detection device;Detection device is arranged in tail gas discharge pipeline to detect the concentration of characteristic gas in real time.The application can improve the anti-interference ability in detection process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of semiconductor detection, and in particular to a cleaning endpoint judgment method and system for a semiconductor process chamber. BACKGROUND

[0002] In semiconductor manufacturing, after a certain number of film growths are completed in a process chamber of a film deposition device such as a chemical vapor deposition (CVD) or an atomic layer deposition (ALD), reaction byproduct films will accumulate on the inner wall, the gas distribution plate, and the quartz window. When the thickness of the film reaches the micron level, micro cracks or peeling may easily occur due to thermal stress differences, forming particle sources that directly fall on the surface of a subsequent wafer, causing irreparable defects and a sharp drop in yield. At the same time, the attached film changes the thermal radiation and gas flow field of the chamber, resulting in drifts in film thickness, refractive index, and stress between batches, and poor consistency in device electrical properties. In addition, when different processes are switched, residual materials (such as metals and dopants) from the previous process may be incorporated into the next film layer, forming impurity energy levels or leakage channels. Therefore, efficient and thorough dry cleaning must be performed on the chamber during the process gap to restore the initial surface state.

[0003] In the prior art, the cleaning endpoint detection of a semiconductor film growth chamber generally uses a single optical signal such as optical emission spectroscopy (OES) or infrared absorption spectroscopy (IR). Once the optical path is contaminated by particles, the mirror surface is fogged, or the window is coated, baseline drift will cause the fixed threshold to be misjumped, resulting in premature cleaning stoppage of residual film or excessive etching damage to the chamber wall. At the same time, the fixed threshold cannot adapt to fluctuations in process pressure, temperature, and power, causing differences in cleaning time between different batches of the same device and a decrease in device utilization.

[0004] Therefore, it is necessary to provide a new cleaning endpoint judgment method and system for a semiconductor process chamber to solve the above problems in the prior art. SUMMARY

[0005] The technical problem to be solved by the present application is how to provide a cleaning endpoint judgment method and system for a semiconductor process chamber with strong anti-interference capability.

[0006] To solve the above technical problems, according to an embodiment of the present application, a cleaning endpoint judgment method for a semiconductor process chamber is provided, comprising the following steps,

[0007] obtaining an optical parameter of a characteristic gas generated in a cleaning process, and performing data processing on the optical parameter to obtain a normalized absorption rate A;

[0008] obtaining an electrical parameter V of the characteristic gas generated in the cleaning process;

[0009] performing data processing on the optical parameter and the electrical parameter V to obtain an optical endpoint index Optical weight Electrical endpoint indicators and electrical weight ;

[0010] For the optical endpoint index The optical weights The electrical endpoint indicators and the electrical weight Data processing is performed to obtain the decision indicator S;

[0011] Set a first threshold for the normalized absorbance A and the second threshold of decision indicator S ;

[0012] When the decision indicator S is continuously lower than the second threshold ,and When the cleaning is complete, M is a constant with a value ranging from 0.5 to 3.

[0013] According to an embodiment of this application, the step of acquiring the optical parameters of the characteristic gas generated during the cleaning process, and processing the optical parameters to obtain the normalized absorptivity A, includes:

[0014]

[0015] in, The absorption intensity of the characteristic gas generated during the cleaning process; The reference light intensity is used by the optical system that generates the characteristic gas during the cleaning process.

[0016] According to an embodiment of this application, the optical parameters and the electrical parameter V are processed to obtain optical endpoint indicators. Optical weight Electrical endpoint indicators and electrical weight ,include,

[0017] The optical parameters are processed to obtain the optical signal quality factor Q;

[0018] The optical signal quality factor Q is processed to obtain the optical weights. ;

[0019] According to the optical weight The electrical weights are obtained .

[0020] According to an embodiment of this application, the step of processing the optical parameters to obtain the optical signal quality factor Q includes:

[0021] The reference light intensity of the optical system that generates the characteristic gas during the cleaning process is obtained. ;

[0022] Calculate reference light intensity The standard deviation S,

[0023]

[0024] in, Let be the reference light intensity at the i-th sampling time. N is the number of sampling points within a 30-second window. When the sampling frequency is 10Hz, then N=300. Reference light intensity within a 30-second window The arithmetic mean;

[0025] Data processing is performed on the standard deviation S to obtain the optical signal quality factor Q.

[0026]

[0027] Where k is a constant, taking values ​​from 0.5 to 2.

[0028] According to an embodiment of this application, the optical signal quality factor Q is processed to obtain optical weights. ,include,

[0029] ;

[0030] According to the optical weight The electrical weights are obtained ,include,

[0031] =1- .

[0032] According to an embodiment of this application, the optical parameters and the electrical parameter V are processed to obtain optical endpoint indicators. Optical weight Electrical endpoint indicators and electrical weight It also includes,

[0033] The optical parameters are processed to obtain optical endpoint indicators. ;

[0034]

[0035] The electrical parameter V is processed to obtain the electrical endpoint index. ;

[0036] .

[0037] According to embodiments of this application, the optical endpoint index is... The optical weights The electrical endpoint indicators and the electrical weight Data processing yields decision indicators S, including:

[0038]

[0039] in, Optical endpoint indicators during the cleaning cycle The maximum value; Electrical endpoint indicators during the cleaning cycle The maximum value.

[0040] According to an embodiment of this application, the first threshold for setting the normalized absorbance A is... and the second threshold of decision indicator S ,include,

[0041] The first threshold for,

[0042]

[0043] in, for The value of the normalized assimilation rate A at time t.

[0044] According to an embodiment of this application, the first threshold for setting the normalized absorbance A is... and the second threshold of decision indicator S It also includes,

[0045] Second threshold for,

[0046]

[0047] in, It represents the real-time maximum value of the absolute value of the first derivative of the normalized absorbance A within the cleaning cycle.

[0048] A cleaning endpoint determination system for cleaning semiconductor process chambers, used to implement the above-mentioned cleaning endpoint determination method, the cleaning endpoint determination system including a semiconductor chamber, an exhaust gas emission pipeline and a detection device;

[0049] The exhaust gas pipeline is connected to the semiconductor chamber to discharge the products and by-products inside the semiconductor chamber;

[0050] The detection device is installed in the exhaust gas pipeline to detect the concentration of characteristic gases in real time.

[0051] By adopting the above technical solution, optical and electrical parameters are collected, and optical endpoint indicators are calculated based on the optical and electrical parameters. Optical weight Electrical endpoint indicators and electrical weight According to optical endpoint indicators Optical weight Electrical endpoint indicators and electrical weight The decision index S is calculated, and the cleaning endpoint is determined based on the decision index S and the normalized absorbance A. In other words, by combining optical and electrical methods, the cleaning process is determined under multiple index conditions, reducing the probability of misjudgment. This involves real-time acquisition and fusion of optical and plasma electrical parameters to construct a decision model that resists signal interference and improves the reliability of endpoint determination. Attached Figure Description

[0052] Fig. 1 This is a step diagram of a cleaning endpoint determination method according to an embodiment of the present invention;

[0053] Fig. 2 This is a schematic diagram of signal comparison under interference scenario according to an embodiment of the present invention, where the horizontal axis represents the reaction time in seconds and the vertical axis represents the concentration of the characteristic gas in ppm; the black dashed line represents the concentration of the characteristic gas affected by interference, and the black solid line represents the concentration of the characteristic gas not affected by interference.

[0054] Fig. 3 This is a graph showing the rate of change of concentration of a characteristic gas according to an embodiment of the present invention, where the horizontal axis represents the reaction time in seconds and the vertical axis represents the rate of change of concentration of the characteristic gas in ppm / s. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Unless otherwise defined, the technical or scientific terms used herein should have the ordinary meaning understood by those skilled in the art. The terms "comprising" and similar expressions used herein mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but does not exclude other elements or objects.

[0056] The following is in conjunction with the appendix Figs. 1-3 The specific embodiments of the present invention will be further described in detail below.

[0057] This invention provides a method for determining the cleaning endpoint in semiconductor process chamber cleaning. Specifically, existing methods for detecting the cleaning endpoint in semiconductor process chamber cleaning often rely on a single signal source, such as optical emission spectroscopy (OES) or infrared absorption spectroscopy (IR), which has significant limitations, namely poor anti-interference capability, insufficient reliability, and weak adaptability. For example, optical signals are easily affected by mirror contamination or particulate scattering interference, and fixed threshold judgment logic cannot adapt to process fluctuations. Moreover, its fixed threshold judgment cannot handle changes in different cleaning stages, increasing the risk of over-cleaning or under-cleaning. Therefore, this invention provides a method for determining the cleaning endpoint in semiconductor process chamber cleaning. Specifically, the cleaning endpoint determination method includes the following steps:

[0058] S1. Obtain the optical parameters of the characteristic gas generated during the cleaning process, process the optical parameters, and obtain the normalized absorptivity A.

[0059] S2. Obtain the electrical parameters V of the characteristic gas generated during the cleaning process;

[0060] S3. Perform data processing on the optical and electrical parameters V to obtain the optical endpoint index. Optical weight Electrical endpoint indicators and electrical weight ;

[0061] S4. Optical endpoint indicators Optical weight Electrical endpoint indicators and electrical weight Data processing is performed to obtain the decision indicator S;

[0062] S5. Set the first threshold for the normalized absorbance rate A. and the second threshold of decision indicator S ;

[0063] S6. When the decision indicator S is continuously lower than the second threshold ,and When the cleaning is complete, M is a constant with a value ranging from 0.5 to 3.

[0064] In some embodiments, the cleaning gas selected varies depending on the deposited film within the process chamber; for example, when the deposited film is... or At that time, the cleaning gas can When the deposited film is or At that time, the cleaning gas can Specific details are not limited here; the focus is on the deposited film. The clean gas is For example, the products generated by the reaction are as follows:

[0065]

[0066] Its main characteristic gas is That is, in the products generated, The concentration is the highest and most easily detected, therefore it is used... As a characteristic gas.

[0067] In some embodiments, it is necessary to monitor the normalized absorptivity A of the optical parameter and the electrical parameter V of the characteristic gas. Both the normalized absorptivity A and the electrical parameter V can indicate whether the cleaning process within the semiconductor process chamber has ended, i.e., whether the cleaning process within the semiconductor process chamber is complete. Whether the reaction is complete. More specifically, the normalized absorbance A reflects the real-time concentration of the characteristic gas; that is, the normalized absorbance A can be used to determine the concentration of the characteristic gas within the semiconductor process chamber. Whether the reaction is complete; and the electrical parameter V can reflect the energy state of the plasma, that is, to determine the density of fluorine atoms participating in the reaction. When the density of fluorine atoms decreases, it indicates that the reaction is likely to stop. Therefore, by combining the normalized absorbance A and the electrical parameter V, it is possible to determine whether the cleaning process in the semiconductor process chamber is complete.

[0068] In some embodiments, after obtaining the normalized absorptivity A and the electrical parameter V, it is necessary to calculate the optical endpoint index based on the normalized absorptivity A and the electrical parameter V. Optical weight Electrical endpoint indicators and electrical weight Specifically, although the normalized absorptivity A and the electrical parameter V can be used for judgment, the roughness, dimensions, and noise of the curves formed by A and V differ, leading to errors and inaccurate results. Therefore, it is necessary to transform them into comparable indicators with the same dimensions and weights, namely, optical endpoint indicators. Optical weight Electrical endpoint indicators and electrical weight Among them, optical endpoint indicators Capable of determining the concentration change rate of characteristic gases within semiconductor process chambers, and electrical endpoint indicators. It can determine the rate of plasma energy drop within the semiconductor process chamber, i.e., the rate of change of the plasma spatial potential. And optical weights... and electrical weight It can prevent misjudgment.

[0069] In some embodiments, after obtaining optical endpoint indicators Optical weight Electrical endpoint indicators and electrical weight After that, it is also necessary to check the optical endpoint indicators. Optical weight Electrical endpoint indicators and electrical weight Data processing is performed to obtain the decision index S; specifically, based on the optical endpoint index... It can determine the changes in the concentration of characteristic gases during the reaction process, based on optical weighting. Capable of quantifying optical endpoint indicators The reliability of the electrical endpoint indicators The rate of change of plasma spatial potential during the reaction can be determined based on the electrical weight. Electrical endpoint indicators The reliability of the indicators is crucial; however, when determining the cleaning endpoint, there are only two states: cleaning complete and cleaning incomplete. Therefore, it is necessary to combine the above indicators into a single indicator that can determine whether the cleaning is complete, i.e., an optical endpoint indicator. Optical weight Electrical endpoint indicators and electrical weight Data processing is performed to obtain the decision indicator S.

[0070] In some embodiments, after obtaining the decision index S, it is also necessary to set a first threshold for the normalized uptake rate A. and the second threshold of decision indicator S When the decision indicator S is continuously below the second threshold ,and When the cleaning is complete, the first threshold is considered to be... As a limiting indicator for normalized uptake rate A This ensures that over 80% of the reactants are consumed during the cleaning process; second threshold. As a limiting indicator for decision indicator S, the two work together to ensure that over-cleaning does not occur when determining the cleaning endpoint.

[0071] In some specific embodiments, the optical parameters of the characteristic gas generated during the cleaning process are acquired, and the optical parameters are processed to obtain the normalized absorptivity A, including...

[0072]

[0073] in, The absorption intensity of the characteristic gas generated during the cleaning process; This is the reference light intensity for the optical system that detects the characteristic gas generated during the cleaning process. Specifically, it is based on the absorption intensity of the characteristic gas generated during the cleaning process. This allows us to determine which components participated in the reaction during the cleaning process. The amount of gas is used to determine the cleaning progress; the reference light intensity of the optical system that generates the characteristic gas during the cleaning process is also used to determine the cleaning progress. Will not be Absorption occurs, but it can be affected by window contamination or particle scattering during the reaction process; therefore, the reference light intensity will be used. As a real-time benchmark, it is used in conjunction with the absorption intensity of the characteristic gas. The normalized absorbance A is calculated so that it is not affected by window contamination or particle scattering, thus improving accuracy.

[0074] In some embodiments, optical and electrical parameters V are processed to obtain optical endpoint indices. Optical weight Electrical endpoint indicators and electrical weight ,include,

[0075] The optical parameters are processed to obtain the optical signal quality factor Q;

[0076] The optical signal quality factor Q is processed to obtain the optical weights. ;

[0077] According to optical weight Obtain electrical weights .

[0078] In some specific embodiments, optical parameters are processed to obtain the optical signal quality factor Q, including,

[0079] The reference light intensity of the optical system that generates the characteristic gas during the cleaning process is obtained. ;

[0080] Calculate reference light intensity The standard deviation S,

[0081]

[0082] in, Let be the reference light intensity at the i-th sampling time. N is the number of sampling points within a 30-second window. When the sampling frequency is 10Hz, then N=300. Reference light intensity within a 30-second window The arithmetic mean; specifically, in the calculation process, the sampling frequency is set to 1-10Hz, preferably 10Hz, and the detection window is 30s, that is, the reference light intensity within the most recent 30s. The sampled data sequence; for example, a sampling frequency of 10 Hz with 300 sampling points within a 30s window, the standard deviation of the 300 sampling points is calculated to determine the optical signal quality factor Q, thus judging the quality of the optical signal; specifically, the reference light intensity... The standard deviation S can be used as a quantitative indicator of the degree of interference, i.e., the reference light intensity. The smaller the standard deviation S, the higher the reference light intensity. The more stable, the better. The larger the standard deviation S value, the greater the reference light intensity. The more drastic the fluctuations.

[0083] Data processing is performed on the standard deviation S to obtain the optical signal quality factor Q.

[0084]

[0085] Where k is a constant, ranging from 0.5 to 2. Specifically, the optical signal quality factor Q is calculated exponentially, giving it nonlinear sensitivity and auto-saturation characteristics. Nonlinear sensitivity means that small fluctuations have little impact on the optical signal quality factor Q, while large fluctuations cause a sharp drop in Q, thus increasing the stability of Q and filtering out small fluctuations that do not affect the endpoint judgment result. Auto-saturation means that regardless of the range of the standard deviation S, the optical signal quality factor Q remains within the range of 0 to 1. More specifically, in this embodiment, k is set to 1; when k is greater than 1, for example, k is 2, it is suitable for processes with higher purity; when k is less than 1, for example, k is 0.5, it is suitable for reaction environments with more impurities.

[0086] In some specific embodiments, the threshold of the optical signal quality factor Q is 0-1. When it approaches 1, it represents good optical signal quality and reliable signal; when it approaches 0, it represents poor optical signal quality and unreliable signal.

[0087] In some embodiments, the optical signal quality factor Q is processed to obtain optical weights. ,include,

[0088] ;

[0089] The optical signal quality factor Q with a threshold of 0-1 is mapped to optical weights. This allows the threshold of the optical weight to be set to 0-1, thereby determining the reliability of the optical signal quality factor Q.

[0090] According to optical weight Obtain electrical weights ,include,

[0091] =1-

[0092] Among them, optical weight and electrical weight The sum is 1, therefore the electrical weight =1- .

[0093] In some embodiments, optical and electrical parameters V are processed to obtain optical endpoint indices. Optical weight Electrical endpoint indicators and electrical weight It also includes,

[0094] Optical parameters are processed to obtain optical endpoint indicators. ;

[0095]

[0096] The electrical parameter V is processed to obtain the electrical endpoint index. ;

[0097] .

[0098] Among them, when optical endpoint index >0 represents As the concentration increases, the reaction becomes more vigorous; when the optical endpoint index... <0 means The concentration decreases, indicating the reaction is nearing completion. Electrical endpoint indicators. >0 indicates that the endpoint is near, meaning the reaction is about to end; electrical endpoint index <0 indicates that the reaction is still in progress.

[0099] In some specific embodiments, the electrical parameter V is the plasma space potential, which is obtained by a Langmuir probe at a detection frequency of 0.1s.

[0100] In some embodiments, optical endpoint indicators Optical weight Electrical endpoint indicators and electrical weight Data processing yields decision indicators S, including:

[0101]

[0102] in, Optical endpoint indicators during the cleaning cycle The maximum value; Electrical endpoint indicators during the cleaning cycle The maximum value. Specifically, This can be understood as taking the moment of maximum vibration within the cleaning cycle as the maximum point, i.e. ;at this time Optical endpoint indicators during the cleaning cycle Compared to optical endpoint indicators within the cleaning cycle The proportion of the maximum value. That is, the endpoint index of light learning. Reliability and electrical endpoint indicators The reliability is limited to the range of (-1)-1. Therefore, by combining multiple parameters, the decision index S can be calculated, and then a judgment can be made based on the decision index to determine whether the cleaning process has reached the cleaning endpoint.

[0103] In some embodiments, a first threshold for the normalized absorbance A is set. and the second threshold of decision indicator S ,include,

[0104] First threshold for,

[0105]

[0106] in, for The value of the normalized assimilation rate A corresponding to time t. Specifically, the first threshold. The actual value is the real-time maximum value of the normalized absorbance A within this cleaning cycle, meaning the value within the semiconductor process chamber. At maximum concentration; due to The value of M ranges from 0.5 to 3. Taking M = 0.2 as an example, during the judgment process... The cleaning endpoint is determined only when the concentration decreases by 80% from the maximum concentration value; that is... .

[0107] In some embodiments, a first threshold for the normalized absorbance A is set. and the second threshold of decision indicator S It also includes,

[0108] Second threshold for,

[0109]

[0110] in, This represents the real-time maximum value of the absolute value of the first derivative of the normalized absorbance A within the cleaning cycle. Specifically, the second threshold... The decision criterion S is set to -10% of the real-time maximum absolute value of the first derivative of the normalized absorbance ADE within this cleaning cycle. In other words, during the judgment process, the cleaning reaction is considered complete only when the decision criterion S is less than -10% of the real-time maximum absolute value of the first derivative of the normalized absorbance ADE. This means that the decision criterion S must simultaneously be continuously below the second threshold. ,and When the cleaning is complete, it is determined that the cleaning is finished.

[0111] An embodiment of the present invention also discloses a cleaning endpoint determination system for semiconductor process chamber cleaning, which is used to implement the above-described cleaning endpoint determination method; the cleaning endpoint determination system includes a semiconductor chamber, an exhaust gas emission pipeline, and a detection device;

[0112] The exhaust gas pipeline is connected to the semiconductor chamber to discharge the products and by-products inside the semiconductor chamber;

[0113] The detection device is installed in the exhaust gas pipeline to detect the concentration of characteristic gases in real time.

[0114] In some embodiments, a semiconductor process chamber is used to place wafers, where the wafers undergo reactions. An exhaust gas duct is located at the bottom of the semiconductor process chamber and communicates with the interior of the chamber. The exhaust gas duct discharges processing products and byproducts generated during wafer processing, thereby keeping the chamber in a state that does not affect the production process. Specifically, the exhaust gas duct can be installed by bonding, snap-fitting, or integral molding, etc., without limitation, as long as the exhaust gas duct can be fixed to the semiconductor process chamber and communicate with the interior of the semiconductor process chamber to discharge processing products and byproducts.

[0115] In some specific embodiments, the detection device is located inside the exhaust gas emission pipe. The method of installation can be adhesive, snap-fit, or bolted, etc., without limitation, as long as it can detect the characteristic gases of the exhaust gas emission pipe.

[0116] The implementation principle of the cleaning endpoint determination method and system for semiconductor process chambers disclosed in this application is as follows: optical parameters and electrical parameters are collected, and optical endpoint indicators are calculated based on the optical and electrical parameters. Optical weight Electrical endpoint indicators and electrical weight According to optical endpoint indicators Optical weight Electrical endpoint indicators and electrical weight The decision index S is calculated, and the cleaning endpoint is determined based on the decision index S and the normalized absorbance A. In other words, by combining optical and electrical methods, the cleaning process is determined under multiple index conditions, reducing the probability of misjudgment. This involves real-time acquisition and fusion of optical and plasma electrical parameters to construct a decision model that resists signal interference and improves the reliability of endpoint determination.

[0117] While embodiments of the present invention have been described in detail above, it will be apparent to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations fall within the scope and spirit of the invention as set forth in the claims. Furthermore, the invention described herein may have other embodiments and can be implemented or carried out in various ways.

Claims

1. A method for determining the cleaning endpoint in a semiconductor process chamber, characterized in that, Includes the following steps, The optical parameters of the characteristic gas generated during the cleaning process are obtained, and the optical parameters are processed to obtain the normalized absorptivity A. Obtain the electrical parameters V of the characteristic gas generated during the cleaning process; Data processing is performed on the optical parameters and the electrical parameter V to obtain the optical endpoint index. Optical weight Electrical endpoint indicators and electrical weight ; For the optical endpoint index The optical weights The electrical endpoint indicators and the electrical weight Data processing is performed to obtain the decision indicator S; Set a first threshold for the normalized absorbance A and the second threshold of decision indicator S ; When the decision indicator S is continuously lower than the second threshold ,and When the cleaning is complete, M is a constant with a value ranging from 0.5 to 3.

2. The method for determining the cleaning endpoint according to claim 1, characterized in that, The process of acquiring optical parameters of the characteristic gas generated during the cleaning process, and processing these optical parameters to obtain the normalized absorptivity A, includes: in, The absorption intensity of the characteristic gas generated during the cleaning process; The reference light intensity is used by the optical system that generates the characteristic gas during the cleaning process.

3. The method for determining the cleaning endpoint according to claim 1, characterized in that, The optical parameters and electrical parameter V are processed to obtain the optical endpoint index. Optical weight Electrical endpoint indicators and electrical weight ,include, The optical parameters are processed to obtain the optical signal quality factor Q; The optical signal quality factor Q is processed to obtain the optical weights. ; According to the optical weight The electrical weights are obtained .

4. The method for determining the cleaning endpoint according to claim 3, characterized in that, The process of processing the optical parameters to obtain the optical signal quality factor Q includes, The reference light intensity of the optical system that generates the characteristic gas during the cleaning process is obtained. ; Calculate reference light intensity The standard deviation S, in, Let be the reference light intensity at the i-th sampling time. N is the number of sampling points within a 30-second window. When the sampling frequency is 10Hz, then N=300. Reference light intensity within a 30-second window The arithmetic mean; Data processing is performed on the standard deviation S to obtain the optical signal quality factor Q. Where k is a constant, taking values ​​from 0.5 to 2.

5. The method for determining the cleaning endpoint according to claim 4, characterized in that, The optical signal quality factor Q is processed to obtain the optical weights. ,include, ; According to the optical weight The electrical weights are obtained ,include, =1- 。 6. The method for determining the cleaning endpoint according to claim 3, characterized in that, The optical parameters and electrical parameter V are processed to obtain the optical endpoint index. Optical weight Electrical endpoint indicators and electrical weight It also includes, The optical parameters are processed to obtain optical endpoint indicators. ; The electrical parameter V is processed to obtain the electrical endpoint index. ; 。 7. The method for determining the cleaning endpoint according to claim 1, characterized in that, The optical endpoint index The optical weights The electrical endpoint indicators and the electrical weight Data processing yields decision indicators S, including: in, Optical endpoint indicators during the cleaning cycle The maximum value; Electrical endpoint indicators during the cleaning cycle The maximum value.

8. The method for determining the cleaning endpoint according to claim 1, characterized in that, The first threshold value for the normalized absorbance A is set. and the second threshold of decision indicator S ,include, The first threshold for, in, for The value of the normalized assimilation rate A at time t.

9. The method for determining the cleaning endpoint according to claim 1, characterized in that, The first threshold value for the normalized absorbance A is set. and the second threshold of decision indicator S It also includes, Second threshold for, in, It represents the real-time maximum value of the absolute value of the first derivative of the normalized absorbance A within the cleaning cycle.

10. A cleaning endpoint determination system for semiconductor process chamber cleaning, characterized in that, For implementing the cleaning endpoint determination method according to any one of claims 1-9, the cleaning endpoint determination system includes a semiconductor chamber, an exhaust gas pipeline, and a detection device; The exhaust gas pipeline is connected to the semiconductor chamber to discharge the products and by-products inside the semiconductor chamber; The detection device is installed in the exhaust gas pipeline to detect the concentration of characteristic gases in real time.

Images