JP5468535B2 - Anomaly detection apparatus, method, and program for causing computer to execute the method - Google Patents

Description

  Embodiments described herein relate generally to an abnormality detection apparatus, a method, and a program that causes a computer to execute the method.

  In the manufacturing process of a semiconductor device, in order to detect an abnormality of the semiconductor manufacturing device, device parameters representing various operating states of the semiconductor manufacturing device are monitored by a sensor. When an apparatus parameter obtained from the sensor shows an abnormal value, an alarm is generated on the assumption that an abnormality has occurred in the semiconductor manufacturing apparatus.

JP 2010-87459 A

  However, in the prior art, it is determined that the device parameter is abnormal when the device parameter is out of the predetermined range. However, for example, even when the device parameter suddenly shows an abnormal value, the manufacturing yield of the semiconductor device. May have little effect.

  An object of one embodiment of the present invention is to provide an abnormality detection apparatus and method capable of accurately detecting a specific abnormal behavior of an apparatus that adversely affects the manufacture of a semiconductor device.

According to one embodiment of the present invention, there is provided an abnormality detection device comprising device parameter acquisition means, data processing means, trend acquisition means, fitting error acquisition means, and abnormality determination means. The device parameter acquisition unit acquires a device parameter representing an operation state of the device. The data processing means performs a fitting process on the time series data of the device parameters. The trend acquisition means performs second-order differentiation on the data subjected to the fitting process, and acquires a second-order differential value at the time of the device parameter acquisition as trend change information. The fitting error acquisition means acquires a fitting error between the acquired device parameter and the data subjected to the fitting process as change information of the device parameter. Then, the abnormality determining means uses a combination of the trend change information acquired by the data behavior acquiring means and the apparatus parameter change degree information, and the operation state of the apparatus has an adverse effect on the manufacture of the semiconductor device. compared to failure determination reference information for defining a range in which it is determined that a particular abnormality of the apparatus determines whether the operation state of the apparatus is a specific abnormality adversely affect the device for manufacturing a semiconductor device.

FIG. 1 is a diagram illustrating an example of a trend of apparatus parameters. FIG. 2 is a diagram illustrating the relationship between the abnormality of the apparatus parameter and the state of the apparatus state. FIG. 3 is a block diagram schematically illustrating an example of the configuration of the abnormality detection device according to the embodiment. FIG. 4 is a flowchart illustrating an example of the procedure of the abnormality detection method according to the embodiment. FIG. 5 is a diagram schematically illustrating an example of an abnormality detection process using apparatus parameters. FIG. 6 is a diagram illustrating an example of the abnormality determination criterion information. FIG. 7 is a diagram schematically illustrating an example of an abnormality detection process using apparatus parameters. FIG. 8 is a diagram schematically illustrating an example of an abnormality detection process using apparatus parameters.

  Exemplary embodiments of an abnormality detection apparatus and method and a program for causing a computer to execute the method will be described below in detail with reference to the accompanying drawings. In addition, this invention is not limited by this embodiment. Further, in the following, after briefly explaining the general problem about the abnormality detection of the apparatus, the embodiment will be described.

  For example, when monitoring the device status of semiconductor manufacturing equipment that constitutes a semiconductor manufacturing system, the behavior of the device parameter indicating the device status is a baseline indicating the change over time of the value that the device parameter can take when the device status is normal. On the other hand, there are a baseline rising / falling type change that gradually shifts and a sudden type change that suddenly changes with respect to the baseline. FIG. 1 is a diagram showing an example of a trend (trend) of device parameters over time, FIG. 1 (a) is a diagram showing an example of a baseline rising / falling type, and FIG. It is a figure which shows an example of a sudden type. FIG. 2 is a diagram illustrating the relationship between the abnormality of the apparatus parameter and the state of the apparatus. Here, examples of apparatus parameters include focus followability and exposure amount in an exposure apparatus, and He gas leakage on the back surface of a cooling wafer stage in a CVD (Chemical Vapor Deposition) apparatus.

  In the case of the baseline rising / falling type, as shown in FIG. 1 (a), the device parameters distributed near the average value y1 until time t1 and the base line are formed after time t1. The value gradually increases and deviates from the baseline. As shown in FIG. 2, when the focus followability and exposure amount in the exposure apparatus and the leak amount of He gas in the CVD apparatus show such fluctuations as apparatus parameters, all of them are serious problems in the manufacture of the semiconductor device. For example, it has been found that a defect occurs in a wide area of a semiconductor device and the manufacturing yield deteriorates.

  On the other hand, in the case of the sudden type, as shown in FIG. 1 (b), the apparatus parameters distributed near the average value y2 at most times and forming the baseline are suddenly changed at, for example, the time t2. It shows a behavior that takes a value far from the baseline and then takes a value around the average value y2 again). As shown in FIG. 2, when the focus followability of the exposure apparatus shows such a variation as an apparatus parameter, it is mild in terms of manufacturing yield, but the semiconductor device (chip) has a dust defect (low area). It is known that a defect occurs. In addition, when the exposure amount of the exposure apparatus shows such fluctuations as an apparatus parameter, it is due to a setting change accompanying maintenance, not due to a trouble in the exposure apparatus, and even if processing is continued as it is, the manufacturing yield decreases. It is known that such problems do not occur. Furthermore, when the amount of He gas leak in the CVD apparatus shows such fluctuations as an apparatus parameter, it is a temporary increase in leakage due to stage dust, and is not caused by a trouble in the CVD apparatus. However, it has been found that problems such as a decrease in manufacturing yield do not occur.

  In a general anomaly detection device, when detecting a baseline rising / falling type anomaly, it is determined that an anomaly has occurred in the device when the device parameter changes by 10%, and a sudden anomaly is detected. In the case of detection, when an apparatus parameter that is out of a predetermined range from the baseline is detected, it is determined that an abnormality has occurred in the apparatus. However, in the case of the baseline rising / falling type, if a device parameter changes by 10% after time t1 in FIG. 1A, a considerable time has passed since a serious abnormality occurred, Many defective products may occur in the semiconductor device manufactured during that time. In the case of the sudden type, even when an apparatus parameter that is out of a predetermined range from the baseline is detected, there is a case where the manufacturing apparatus is not abnormal and the semiconductor device is not greatly affected. Furthermore, it may be difficult to distinguish between a baseline rising / falling type and a sudden type.

  Therefore, in the following embodiment, whether or not to show a baseline rising / falling type behavior or a sudden type behavior is quickly distinguished from the device parameter data acquired in time series, and the type is determined. A description will be given of an abnormality detection apparatus and method capable of grasping an abnormal state occurring in the apparatus and notifying the administrator or user of the apparatus and a program for causing the computer to execute the method.

  FIG. 3 is a block diagram schematically illustrating an example of the configuration of the abnormality detection device according to the embodiment. The abnormality detection apparatus 10 is connected to a sensor that collects the state of a semiconductor manufacturing apparatus such as a CVD apparatus, an exposure apparatus, and an etching apparatus constituting a semiconductor manufacturing system, for example, and acquires a value obtained from the sensor as an apparatus parameter. Detects whether semiconductor manufacturing equipment is exhibiting abnormal behavior when the trend of data in which semiconductor manufacturing equipment parameters are arranged in time series deviates from the baseline or when semiconductor manufacturing equipment parameters suddenly deviate from the baseline To do. Here, when a trend changes or suddenly an abnormal value of a device parameter occurs when the semiconductor manufacturing device to be detected is operating in a normal (normal) state. Next, the case where it is determined whether or not an abnormality has occurred in the semiconductor manufacturing apparatus will be described as an example.

  The abnormality detection apparatus 10 includes an apparatus parameter acquisition unit 11, an apparatus parameter storage unit 12, a data processing unit 13, a trend acquisition unit 14, a fitting error acquisition unit 15, an abnormality determination reference information storage unit 16, and an abnormality determination. A unit 17, an alarm output unit 18, and a control unit 19 that controls these processing units are provided.

  The device parameter acquisition unit 11 is connected to a sensor that detects a device state provided in the device to be monitored, acquires an output from the sensor as a device parameter, and stores the device parameter in the device parameter storage unit 12. The device parameter storage unit 12 stores device parameters together with time information. The apparatus parameter is a parameter indicating a predetermined apparatus state in which an abnormality of an apparatus to be monitored (hereinafter referred to as an object apparatus) can be detected. For example, focus followability in an exposure apparatus, exposure amount, and cooling of a CVD apparatus The amount of helium gas leak on the back surface of the wafer stage can be exemplified.

  The data processing unit 13 performs a fitting process for approximating the time series data of the device parameters stored in the device parameter storage unit 12, so-called raw data, with a primary expression or a secondary expression, and obtains a fitting result. When performing the fitting process, for example, the fitting may be performed for each predetermined time interval.

  The trend acquisition unit 14 uses the fitting result generated by the data processing unit 13 to process the trend change information that indicates the change in the trend of the device parameters. Here, a second-order differential value is taken for the fitting result. By taking the second-order differential value, when a peak appears in the second-order differential value, it is possible to detect that the trend of the target device has changed. Here, the trend acquisition unit 14 outputs the acquired second-order differential value to the abnormality determination unit 17.

  The fitting error acquisition unit 15 indicates the degree of change from the baseline of the device parameter acquired by the device parameter acquisition unit 11, specifically, the degree of deviation from the fitting result generated by the device parameter data processing unit 13. Get the fitting error. As the fitting error, for example, a residual between the actually acquired apparatus parameter and the fitting result, a square value of the residual, a chi-square value, or the like can be used. Here, the fitting error acquisition unit 15 acquires the residual, and outputs the acquired fitting error to the abnormality determination unit 17. The data processing unit 13, trend acquisition unit 14, and fitting error acquisition unit 15 serve as data behavior acquisition means.

  The abnormality determination criterion information storage unit 16 stores abnormality determination criterion information that defines a second-order differential value and a fitting error range in which the target device is determined to be abnormal. In the abnormality determination reference information, a second-order differential value and a range (combination) of fitting errors are defined for each type of abnormality.

  The abnormality determination unit 17 compares the combination of the second-order differential value acquired from the trend acquisition unit 14 and the fitting error acquired from the fitting error acquisition unit 15 with the abnormality determination reference information in the abnormality determination reference information storage unit 16. When the second-order differential value and the fitting error exist in the range determined to be abnormal, it is determined that an abnormality has occurred. Further, when an abnormality occurs, the type of abnormality is acquired from the abnormality determination reference information, and is output to the alarm output unit 18 together with the occurrence of the abnormality.

  When the alarm output unit 18 obtains a signal including the occurrence of the abnormality of the target device and the type of abnormality from the abnormality determination unit 17, for example, an abnormality has occurred in a work computer or an alarm device arranged close to the target device. An alarm indicating that is output. At this time, by outputting the type of abnormality, it is possible to notify the operator what kind of abnormality has occurred. The alarm output unit 18 may not be provided as a configuration of the abnormality detection device 10.

  Next, an abnormality detection method in such an abnormality detection apparatus 10 will be described. FIG. 4 is a flowchart illustrating an example of the procedure of the abnormality detection method according to the embodiment.

  First, the device parameter acquisition unit 11 of the abnormality detection device 10 acquires the device parameter of the target device to be processed from the sensor (step S11) and stores it in the device parameter storage unit 12. The device parameters are stored in the device parameter storage unit 12 together with time information.

  Next, the data processing unit 13 performs a fitting process on the time series data of the device parameters stored in the device parameter storage unit 12 (step S12). As the fitting process, for example, primary approximation or secondary approximation is used.

  Thereafter, the trend acquisition unit 14 acquires a first-order differential value for the fitting result obtained by the fitting process (step S13), further differentiates the result of the first-order differentiation, and obtains a second-order differential value for the fitting result. Obtain (step S14). Then, the trend acquisition unit 14 outputs the second-order differential value at the time when this processing is performed to the abnormality determination unit 17.

  Further, the fitting error acquisition unit 15 calculates a fitting error for the fitting result obtained by the fitting process (step S15), and outputs the result to the abnormality determination unit 17. Here, as the fitting error, for example, a residual that is a difference between the apparatus parameter acquired in step S11 and the fitting result calculated in step S12 is calculated.

  Next, the abnormality determination unit 17 determines whether the second-order differential value acquired from the trend acquisition unit 14 and the fitting error acquired from the fitting error acquisition unit 15 are included in the apparatus abnormality range defined in the abnormality determination reference information. Determination is made (step S16).

  Thereafter, when the determination result shows that the acquired second-order differential value and the fitting error are not included in the range considered abnormal (No in step S16), the abnormality determination unit 17 has an abnormality in the target device. It determines with having not generate | occur | produced (step S17), and sets it as the state which is operating the object apparatus as it is.

  On the other hand, as a result of the determination, when the acquired second-order differential value and fitting error are included in the range that is considered abnormal (Yes in step S16), the abnormality determination unit 17 generates an abnormality in the target device. (Step S18), the content of the abnormality specified by the combination of the second-order differential value and the fitting error is acquired from the abnormality determination reference information (step S19), and the content is output to the alarm output unit 18. . And the alarm output part 18 outputs the warning which shows that abnormality of the acquired abnormality content has generate | occur | produced (step S20). Thus, the process ends.

  In the abnormality determination reference information, the severity of the abnormality of the target device is further defined for the abnormal range, and the output method in the alarm output unit 18 is changed according to the severity of the abnormality. Good.

  Next, specific examples of the anomaly detection processing will be described by way of example in which the device parameters are normal, the device parameters exhibit baseline rising / falling behavior, and the device parameters exhibit sudden behavior. explain.

<When device parameters are normal>
FIG. 5 is a diagram schematically illustrating an example of an abnormality detection process when the apparatus parameter is normal. Here, an example of the abnormality detection process at time t11 will be described. The abnormality detection process is continuously performed at predetermined intervals until an abnormality is detected.

  First, as shown in FIG. 5A, it is assumed that the device parameter is acquired within a normal range, and the value obtained at time t11 is also within the normal range. When the fitting process is performed by the data processing unit 13 at time t11, as shown in FIG. 5B, an approximate curve having a predetermined value and approximately parallel to the horizontal axis is obtained.

  Next, the trend acquisition unit 14 acquires the first-order differential value and the second-order differential value at time t11 using the approximate curve. FIG. 5C shows a curve obtained by first-order differentiation of the approximate curve with time, and FIG. 5D shows a curve obtained by second-order differentiation of the approximate curve with time. Since the fitting result is a straight line parallel to the horizontal axis, both the first-order differential value and the second-order differential value at time t11 are zero.

  Thereafter, the fitting error acquisition unit 15 calculates a fitting error between the device parameter value and the approximate curve at time t11. FIG. 5E shows a fitting error. As shown in this figure, within the range in which the apparatus parameter is normal, the approximate curve is substantially equal to the normal value, and the fitting error is within a predetermined range.

  The abnormality determination unit 17 determines whether or not an abnormality has occurred in the apparatus by comparing the combination of the second-order differential value in FIG. 5D and the fitting error in FIG. 5E with the abnormality determination reference information. To do. 6A and 6B are diagrams showing an example of abnormality determination reference information, where FIG. 6A shows a region showing fluctuations in the baseline rising / falling type and sudden type, and FIG. 6B shows an abnormality related to the focus followability of the exposure apparatus. An example of the criterion information is shown, and (c) shows an example of abnormality criterion information relating to the exposure amount of the exposure apparatus and the He gas leak amount of the CVD apparatus. In this example, the abnormality determination reference information is obtained by taking the second order differential value on the horizontal axis and the fitting error on the vertical axis. For example, when a combination of the second-order differential value and the fitting error is included in the region R1 where the second-order differential value is a1 or less and the fitting error is b1 or more, it indicates that a sudden variation has occurred. When the combination of the second-order differential value and the fitting error is included in the region R2 where the second-order differential value is a2 or more and the fitting error is b2 or less, a baseline rising / falling type fluctuation occurs. It is shown that.

  For example, as shown in FIG. 6B, when it is determined whether or not an abnormality of the target apparatus has occurred with respect to the focus followability of the exposure apparatus, it indicates that a sudden fluctuation has occurred. The region R1 is associated with the occurrence of a mild dust defect, and the region R2 indicating the occurrence of the baseline rising / falling type variation is associated with the occurrence of a severe area defect. It is done. And it is matched with the area | regions other than area | region R1, R2 that a target apparatus is normal.

  In addition, as shown in FIG. 6C, when it is determined whether an abnormality of the target apparatus has occurred with respect to the exposure amount of the exposure apparatus and the leak amount of the He gas of the CVD apparatus, The occurrence of a severe area defect is associated with the region R2 indicating that a falling type fluctuation has occurred. And it is matched with the area | regions other than area | region R2 that a target apparatus is normal.

  In the example of FIG. 5, the point P11 that is a combination (0, b11) of the second order differential value and the fitting error is neither a sudden fluctuation nor a baseline rising / falling fluctuation from FIG. Exists in the area indicating the device status. Further, as shown in FIGS. 7B and 7C, when the focus followability of the exposure apparatus, the exposure amount, or the leak amount of the He gas of the CVD apparatus is monitored, a normal apparatus state is indicated. Is determined by the abnormality determination unit 17.

<When device parameters indicate baseline rising / falling fluctuation>
FIG. 7 is a diagram illustrating an example of an anomaly detection process when the apparatus parameter exhibits a baseline rising / falling behavior. Here, an example of the abnormality detection process at time t13 will be described. As shown in FIG. 7 (a), it is assumed that the apparatus parameters behave normally as shown in FIG. 5 until time t12, and the apparatus is determined to be normal. Further, it is assumed that until the time t12, the baseline is configured within the range in which the apparatus parameters are normal, but after the time t12, the baseline is gradually removed and the time t13 is reached. Here, the behavior after time t13 is also shown for convenience.

  As shown in FIG. 7B, when the fitting process is performed by the data processing unit 13, the straight line has a predetermined value until time t12 and is substantially parallel to the horizontal axis, and from time t12 to t13 An approximate curve that is an upward curve is obtained. Thus, when the baseline rising / falling behavior is shown, the time t12 becomes an inflection point, but the apparatus parameter (approximate curve) changes continuously. The approximate curve at time t13 is indicated by a dotted line.

  Next, the trend acquisition unit 14 acquires the first-order differential value and the second-order differential value at time t13 using the approximate curve. FIG. 7C is a curve obtained by first-order differentiation of the approximate curve with time, and FIG. 7D is a curve obtained by second-order differentiation of the approximate curve with time. As shown in FIG. 7C, until the time t12, the first-order differential value is 0, but after the time t12, it has a positive value indicating the slope of the corresponding portion of the approximate curve, and then a predetermined value. I take the. Further, as shown in FIG. 7D, the second-order differential value has a shape showing a positive peak near time t13.

  Thereafter, the fitting error acquisition unit 15 calculates a fitting error between the device parameter value and the approximate curve at time t13. FIG. 7E shows a fitting error. As described above, in the case of the baseline rising / falling type, since the trend of the apparatus parameter continuously changes at the inflection point, the approximate curve also changes accordingly. For this reason, as shown in FIG. 7 (e), some fitting error occurs in the vicinity of the inflection point at time t12, but a considerably large fitting error does not occur.

  The abnormality determination unit 17 compares the combination of the second-order differential value in FIG. 7D and the fitting error in FIG. 7E with the abnormality determination reference information to determine whether or not an abnormality has occurred in the apparatus. To do. In the example of FIG. 7, the point P13, which is a combination of the second-order differential value and the fitting error (a13, b13), is included in the region R2 of FIG. 7A, and shows the baseline rising / falling type behavior. I understand that. Then, when the abnormality determination unit 17 performs abnormality detection on the focus followability of the exposure apparatus, as shown in FIG. 7B, the exposure amount of the exposure apparatus or the He gas of the CVD apparatus When the leak amount is monitored, it is determined that the area is seriously defective as shown in FIG.

<When device parameters show sudden behavior>
FIG. 8 is a diagram illustrating an example of an abnormality detection process in the case where the apparatus parameter exhibits a sudden behavior. Here, an example of the abnormality detection process at time t15 will be described. As shown in FIG. 8A, it is assumed that until the time t14, the apparatus parameter exhibits a normal behavior as shown in FIG. 5, and the apparatus is determined to be normal. Also, until time t14, the apparatus parameter was configured within the normal range, but at time t14, the apparatus parameter suddenly deviated greatly from the baseline was observed, and then gradually toward the baseline. It is assumed that the device parameters change and time t15 is reached. Here, the behavior after time t15 is also shown for convenience.

  As shown in FIG. 8B, when the fitting process is performed by the data processing unit 13, the straight line has a predetermined value until time t14 and is substantially parallel to the horizontal axis. Since the parameters are greatly deviated, an approximate curve that shifts in the direction of the apparatus parameters is obtained. In this way, when sudden behavior is shown, it suddenly changes greatly from the baseline at time t14, but the fitting process is actually performed because it includes the device parameters accumulated up to time t14. The value of the device parameter is far from the approximate curve. The approximate curve after time t13 is indicated by a dotted line.

  Next, the trend acquisition unit 14 acquires the first-order differential value and the second-order differential value at time t15 using the approximate curve. FIG. 8C is a curve obtained by first-order differentiation of the approximate curve with time, and FIG. 8D is a curve obtained by second-order differentiation of the approximate curve with time. As shown in FIG. 8C, the first-order differential value is 0 until time t14, but has a positive value indicating the slope of the corresponding portion of the approximate curve after time t14. Although it has a positive value even at time t15, after that, it reaches a peak as shown by the dotted line, the value decreases, and at a certain point, it becomes a straight line substantially parallel to the horizontal axis. Further, as shown in FIG. 8 (d), the second-order differential value shows a substantially positive peak at time t15, and then becomes zero after showing a negative peak at a point of decreasing thereafter.

  Thereafter, the fitting error acquisition unit 15 calculates a fitting error between the device parameter value and the approximate curve at time t15. FIG. 8E shows the fitting error. In the case of the sudden type, since the trend of the device parameter changes discontinuously, the approximate curve cannot represent the actual device parameter status with high accuracy (the approximate curve may follow the actual device parameter). become unable). Therefore, as shown in FIG. 8 (e), the fitting error becomes large in the vicinity of time t14, which is a point that changes discontinuously.

  The abnormality determination unit 17 determines whether or not an abnormality has occurred in the apparatus by comparing the combination of the second-order differential value in FIG. 8D and the fitting error in FIG. 7E with the abnormality determination reference information. To do. In the example of FIG. 8, it can be seen that the point P15, which is a combination of the second-order differential value and the fitting error (a15, b15), is included in the region R1 of FIG. 7A and exhibits a sudden behavior. Then, the abnormality determination unit 17 determines that there is a mild dusty defect as shown in FIG. 7B when performing abnormality detection on the focus followability of the exposure apparatus, and exposure is also performed. When the exposure amount of the apparatus or the leak amount of the He gas of the CVD apparatus is monitored, it is determined as normal as shown in FIG.

  In the abnormality determination reference information, the severity of the abnormality of the target device is further defined for the abnormal range, and the output method in the alarm output unit 18 is changed according to the severity of the abnormality. Good.

  Also, the abnormality detection method shown in the above-described embodiment, specifically, a program for causing a computer to execute the processing steps shown in steps S11 to S19 in FIG. A program for causing a computer to execute this abnormality detection method is a file in an installable format or an executable format, and is a CD-ROM (Compact Disc Read Only Memory), a floppy (registered trademark) disc, a DVD (Digital Versatile Disc, or And recorded on a computer-readable recording medium such as Digital Video Disc). In addition, a program for causing a computer to execute the abnormality detection method shown in the above-described embodiment is stored on a computer connected to a network such as the Internet, and is provided by being downloaded via the network. Also good.

  By using a program that causes a computer to execute an abnormality detection method in this way, the above-described abnormality detection apparatus 10 has a CPU (Central Processing Unit) computing means, a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. Storage means, external storage means such as HDD (Hard Disk Drive) and CD-ROM drive devices, display means such as display devices, input means such as keyboards and mice, and network boards as required And an information processing apparatus such as a personal computer equipped with network interface means. In this case, the above-described method is performed by expanding a program for causing the computer to execute the abnormality detection method installed in the external storage unit on a storage unit such as a RAM and executing the program on the calculation unit.

  In this embodiment, after fitting the time series data of the device parameters obtained from the target device, the behavior of the change from the baseline is captured by the second-order differential value, and the difference between the actual device parameters and the fitting result is obtained. The amount of change from the baseline was captured from a certain fitting error. As a result, the target device being monitored shows a baseline rising / falling behavior in which the device parameters are gradually deviating from the baseline, or a suddenly different value with respect to the baseline. It can be distinguished at the initial stage of the change whether or not it shows a sudden behavior. As a result, it is possible to determine whether the target device exhibits a serious abnormality, a minor abnormality, or a behavior with no problem. There is an effect that the manufacturing yield of the semiconductor device can be improved without leaving the operating state.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

  DESCRIPTION OF SYMBOLS 10 ... Abnormality detection apparatus, 11 ... Apparatus parameter acquisition part, 12 ... Apparatus parameter storage part, 13 ... Data processing part, 14 ... Trend acquisition part, 15 ... Fitting error acquisition part, 16 ... Abnormality criterion information storage part, 17 ... Abnormality determination unit, 18 ... alarm output unit, 19 ... control unit.

Claims (7)

Device parameter acquisition means for acquiring device parameters representing the operating state of the device;
Data processing means for fitting the time series data of the device parameters;
Trend acquisition means for performing second order differentiation on the fitting-processed data, and obtaining second-order differential values at the time of device parameter acquisition as trend change information;
Fitting error acquisition means for acquiring a fitting error between the acquired device parameter and the fitting-processed data as the degree-of-change information of the device parameter;
A combination of the trend change information acquired by the trend acquisition means and the device parameter change degree information acquired by the fitting error acquisition means, the operation state of the device adversely affects the manufacture of the semiconductor device. An abnormality that determines whether the operation state of the device is a specific abnormality of the device that adversely affects the manufacture of a semiconductor device , as compared with abnormality determination reference information that defines a range that is determined to be a specific abnormality of the device A determination means;
An abnormality detection device comprising:   The abnormality detection apparatus according to claim 1, wherein the abnormality determination reference information is associated with the abnormality type corresponding to the abnormality range.   The abnormality detection apparatus according to claim 1, wherein the apparatus parameter is a focus followability of an exposure apparatus, an exposure amount, or a gas leak amount of a CVD apparatus. An apparatus parameter acquisition step for acquiring an apparatus parameter representing an operation state of the apparatus;
A data processing step of fitting the time series data of the device parameters;
A trend acquisition step of performing second-order differentiation on the fitting-processed data and acquiring second-order differential values at the time of device parameter acquisition as trend change information;
A fitting error acquisition step of acquiring a fitting error between the acquired device parameter and the fitting-processed data as device parameter change degree information;
An abnormality determination criterion that defines a range in which the operation state of the device is determined as a specific abnormality of the device that adversely affects the manufacture of a semiconductor device , by combining the trend change information and the device parameter change degree information An abnormality determination step of determining whether the operation state of the device is a specific abnormality of the device that adversely affects the manufacture of the semiconductor device , as compared with the information;
An abnormality detection method comprising:   The abnormality detection method according to claim 4, wherein the abnormality determination reference information is associated with the abnormality type according to the abnormality range.   6. The abnormality detection method according to claim 4, wherein the apparatus parameter is a focus followability of an exposure apparatus, an exposure amount, or a gas leak amount of a CVD apparatus.   A program for causing a computer to execute the abnormality detection method according to any one of claims 4 to 6.

Images