JP5538955B2 - Method and system for predicting apparatus abnormality in semiconductor manufacturing - Google Patents

Description

  The present invention relates to a semiconductor device manufacturing technique, and more particularly to a method and a system for detecting a sign of apparatus abnormality during processing by a manufacturing apparatus.

  In the manufacture of semiconductor devices, with the miniaturization of circuit patterns transferred to semiconductor wafers, the allowable range of manufacturing conditions that must be satisfied by the semiconductor manufacturing apparatus used in each process is becoming smaller year by year. In addition, if the manufacturing conditions deviate from the allowable fluctuation range, an operation failure occurs in the processed semiconductor wafer circuit. Therefore, it is necessary to monitor the manufacturing status of the semiconductor manufacturing apparatus and monitor whether it deviates from the allowable fluctuation range or whether there is a sign of deviating.

  In a semiconductor device manufacturing method, a semiconductor device having a multilayer structure is formed on a silicon wafer by repeatedly performing many manufacturing processes such as a film forming process, an exposure process, and an etching process according to a process flow. Therefore, many types and many manufacturing apparatuses are used in the manufacture of semiconductor devices.

  Therefore, as a technique for comprehensively monitoring the manufacturing status of a large number of manufacturing devices, data that monitors the manufacturing status of each manufacturing device (hereinafter referred to as the device log) can be statistically analyzed from the fluctuation status of past device logs, or There is a technique for detecting an abnormality in a device log being processed or a sign of abnormality by setting an upper and lower threshold value by a signal processing method.

  In addition, there is a case where a device log necessary for detecting a device abnormality or a sign of abnormality is not monitored. In such a case, there is a technique for attaching a detachable sensor, acquiring a device log to be detected, and detecting a sign.

  For example, there is Patent Document 1 as the above-described prior art of sign detection. In Patent Document 1, a management value of a device log, a management upper limit value, and a management lower limit value are calculated from a past device log by statistical processing, and abnormality detection of the device log is performed based on a management width determined by them. It is an example.

  Further, Patent Document 2 regularly monitors the current value of a crossing breaker, determines the state of the breaker from the transition of the current value, and detects abnormality using an abnormality determination criterion set for each state of the breaker. It is a well-known example to implement.

JP-A-8-202775 JP 2008-290549 A

  In the manufacture of semiconductor devices, when data is acquired using a detachable sensor, the device sampling (hereinafter referred to as periodic sampling data) is acquired according to a preset cycle regardless of the manufacturing process stage. Therefore, data at different processing stages and data at the time when the product is not started are mixed.

  For this reason, if a uniform standard is set for periodic sampling data as shown in Patent Document 1 and predictive detection is performed, the processing stage is not distinguished, so that there is no false alarm (a level change that does not become abnormal is considered abnormal). Or overlooked (a change in the level that becomes abnormal cannot be regarded as abnormal).

  Further, in the semiconductor manufacturing apparatus, a plurality of types of apparatus logs are controlled / managed, and if any one of them is different, a different operation state is obtained. Therefore, the detection target described in Patent Document 2 is detected for the periodic sampling data. When the technology using only the device log is applied, the operation state cannot be estimated, and there is a possibility that false information or oversight may occur.

  Therefore, an object of the present invention is to make it possible to distinguish between operating states that were difficult to distinguish in the prior art, and in semiconductor manufacturing that can only detect data for a period in a predetermined operating state. The object is to provide a sign detection technique for device abnormality.

  The above and other objects and novel features of the present invention will be apparent from the description of this specification and the accompanying drawings.

  Of the inventions disclosed in the present application, the outline of typical ones will be briefly described as follows.

In other words, the outline of typical ones is
“Event log condition, event establishment condition information consisting of an operation type that specifies an operation in which the device log name, threshold value, device log value and threshold value intersect or touch each other”, Corresponding event section correspondence information "and" detection target section type information that sets whether the data sandwiched between the start point and the end point is the detection target section or the data after the start point is the detection target section " Input means for input;
Using “at least one of the apparatus log information, maintenance history information, and start history information” and “the event establishment condition information”, “data to be executed before the start / end of a predetermined operation state” The occurrence time of at least one of the events of “change in transition”, “record of start history information”, and “record of maintenance history information” is specified, and the occurrence time of the event corresponds to the event section From the information and the detection target section type information, a process for automatically determining the start / end of a predetermined driving state, a detection target section extracting means for specifying the detection target section that was in the predetermined driving state,
Using the device log and threshold value in the detection target section, a sign determination unit that generates a sign determination result that determines the presence or absence of a change that is a sign;
Output means for outputting the sign determination result.

Specifically, the detection target section extraction unit and the sign determination unit include
"Device log information consisting of the device log name of the target device", "Maintenance history information consisting of the maintenance work content, maintenance work start time and maintenance work end time of the target device", and "Starting recipe and start of the target device" Device log information, maintenance history information, and start history information collection step for acquiring `` start history information consisting of time and start end time '',
Using event definition information set in advance consisting of “the device log name”, “threshold”, and “operation type that specifies an operation in which the device log values of the device log name intersect or touch each other”, An event that is set as event occurrence time information by combining at least one of “the time when the device log information crosses the threshold value”, “the maintenance history information”, and “the construction history information” Configuration steps;
From “the event occurrence time information”, “the event occurrence order information in which the preset event occurrence order, the start point of the processing section and the end point of the processing section are associated”, and “the event occurrence time information” A detection target section calculating step for generating detection target section information;
A detection target section data acquisition step for acquiring detection target section data from the detection target section information and the device log information;
Statistical processing step for statistically processing the detection target section data and generating detection target statistical processing processed data;
A sign determination step of determining whether the detection target statistical processing processed data includes a sign of an apparatus failure.

  Among the inventions disclosed in the present application, effects obtained by typical ones will be briefly described as follows.

  In other words, the effect obtained by a representative one is to use a plurality of device logs to determine the start / end of a predetermined operation state from the order of events that must be performed before the start / end of the predetermined operation state. It becomes possible to distinguish the operating state that was difficult to distinguish in the prior art. Thereby, since only the data of the period in a predetermined driving | running state can be made into a detection target, it becomes possible to implement | achieve the improvement of detection accuracy.

It is the figure which showed an example of the apparatus abnormality sign detection method in the semiconductor manufacture by one embodiment of this invention. It is the figure which showed an example of the manufacturing method of the semiconductor device by one embodiment of this invention. In one embodiment of this invention, it is the figure which showed an example of event formation condition information. In one embodiment of the present invention, it is a figure showing an example of event section correspondence information. In one embodiment of this invention, it is the figure which showed an example of the detection object area according to detection object area type. In one embodiment of the present invention, it is a figure showing an example of a sign detection processing flow. In one embodiment of the present invention, it is a figure showing an example of an event setting processing flow. In one embodiment of this invention, it is the figure which showed an example of the search flow of a detection object area. It is the figure which showed an example of the apparatus abnormality sign detection system in semiconductor manufacture by one embodiment of this invention. In one embodiment of the present invention, it is a figure showing an example of a screen for performing detection condition setting.

  DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of a semiconductor device manufacturing method and manufacturing system including a method and system for predicting apparatus abnormality in semiconductor manufacturing according to the present invention will be described with reference to the drawings. Note that components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiment, and the repetitive description thereof will be omitted.

<Semiconductor device manufacturing method>
An example of a method for manufacturing a semiconductor device according to an embodiment of the present invention will be described with reference to FIG. FIG. 2 is a diagram showing an example of a semiconductor device manufacturing method according to the present embodiment.

  The present embodiment is applied to a method for manufacturing a semiconductor device including a process for determining whether or not a wafer processing process of a manufacturing apparatus used in a wafer manufacturing process of a semiconductor device is abnormal or a sign of abnormality is as follows. It has characteristics.

  In the method for manufacturing a semiconductor device, as shown in FIG. 2, a film forming process for forming an insulating film, an oxide film, a metal film, etc. on a semiconductor wafer, or a resist (photosensitive agent) is applied on the semiconductor wafer and light is applied to the resist. Many manufacturing processes such as an exposure process for shaping a resist into a predetermined pattern by applying a chemical process, an etching process for shaping a film shaped in a film formation process into a predetermined pattern along the resist, A semiconductor device having a multilayer structure is formed on a silicon wafer by repeatedly performing the process according to the flow (steps 1 to 3 for layer 1, steps 4 to 6 for layer 2,..., Steps 3N-2 to 3N for layer N). .

  In this semiconductor device manufacturing method, as the circuit pattern transferred to the semiconductor wafer is miniaturized, the allowable range of fluctuations in the manufacturing conditions of the semiconductor manufacturing apparatus used in each process is narrowing year by year. If the manufacturing conditions deviate from the allowable range due to incorrect input of semiconductor manufacturing process procedures or manufacturing conditions, deterioration or failure of parts of the manufacturing equipment, or fluctuations caused by material consumption, the processed semiconductor wafer is A malfunction of the target operation occurs. Usually, in the manufacturing process of a semiconductor device, the product quality of the semiconductor device is managed by performing a sampling inspection of a semiconductor wafer processed in each manufacturing process, a periodic inspection of a manufacturing apparatus, or monitoring an apparatus log.

  A defect in the manufacturing apparatus in which the manufacturing condition deviates from the allowable fluctuation range can be detected by the sampling inspection, the QC inspection by the periodic inspection of the manufacturing apparatus, or the monitoring of the apparatus log.

  However, in QC inspections by sampling inspections and periodic inspections of manufacturing equipment, detection may occur after a product failure occurs or after manufacturing conditions deviate from the allowable fluctuation range. Normally, in semiconductor manufacturing, each process is performed in a unit called a lot in which a plurality of semiconductor wafers are collected. Therefore, the detection method described above may produce a large number of defective products.

  Therefore, in the present embodiment, a sign detection method using periodic sampling data acquired by attaching a detachable sensor and its system in order to suppress the occurrence of product defects due to abnormal operation of a manufacturing apparatus in a semiconductor device manufacturing line The present invention provides a method for manufacturing a semiconductor device using the above. This will be specifically described below.

<Method for detecting signs of equipment abnormalities in semiconductor manufacturing>
1 and 3 to 8, an example of an apparatus abnormality sign detection method in semiconductor manufacturing according to an embodiment of the present invention will be described. FIG. 1 is a diagram showing an example of a sign detection method according to the present embodiment. FIG. 3 is a diagram showing an example of event establishment condition information. FIG. 4 is a diagram showing an example of event section correspondence information. FIG. 5 is a diagram illustrating an example of a detection target section for each detection target section type. FIG. 6 is a diagram illustrating an example of the sign detection processing flow. FIG. 7 is a diagram showing an example of an event setting process flow. FIG. 8 is a diagram illustrating an example of a search flow for the detection target section.

  In FIG. 1, the present embodiment is an input means used when inputting information necessary for executing detection in advance, a detection target section extracting means for selecting device data in a predetermined operating state, and extracted data. Is composed of a sign determination means for determining whether or not a fluctuation that is a sign is included, and an output means for confirming the sign determination result.

  In order to implement the present embodiment, information set in advance is event establishment condition information, event section correspondence information, and detection target section type information. Further, a detection target section flag is included in information that is updated each time the detection process is performed and is fed back when the next detection process is performed.

  Information set in advance and the detection target section flag will be described.

  First, the detection target section flag will be described. The detection target section flag is information indicating whether or not the last data (latest data) of the input data at the previous detection is data in the detection target section. The detection target section flag is updated and fed back every time the detection process is performed.

  Next, three types of information set in advance will be described.

  The event establishment condition information is information about event conditions that always occur before and after a desired operation state, and FIG. 3 shows setting information of event establishment conditions. The event establishment condition information includes five pieces of information.

  The first is an event number to be defined (hereinafter referred to as an event number). The second is an event type. The event type can be set from the device log, arrival completion, and maintenance. The third is the log name. When the device log is set by event type, when the device log name of the relevant device is set, "Complete" is set by the event type, when maintenance is set by the event type from the recipe name or all recipes, maintenance work Each can be set from the type information. The fourth is operation. If the device log is set by event type, start from the top, bottom, touch from top, bottom-to-top, touch from bottom, event type, complete or maintenance Each of the two can be set from the end. The fifth is a threshold value. When the device log is set by event type, an arbitrary real number can be input.

  The event section correspondence information is information for determining a start point or an end point of a period in a predetermined driving state, and FIG. 4 shows setting information of event section correspondence information. The event section correspondence information includes two pieces of information. The first is end point information of the section. It can be set from the start point and the end point. The second is Event No. It is. No. in the event establishment condition information described above. Can be set from

  The detection target section type information is set from two types, a variation type and a constant type, according to the operation method of the device log to be detected.

  The variation type is mainly used when detecting a device log having a different value at the time of processing and a value at the time of idling. Normally, such device logs detect data being processed and idle separately. Therefore, the start point and end point of the detection target section are generated each time the process is performed (for example, when processing is set as the detection target section, the process start is the start point of the detection target section and the process end is the end point of the detection target section). Therefore, it is possible to set the data sandwiched between the start point and the end point as a detection target.

  On the other hand, the fixed type is used to detect a device log having the same value at the time of processing and the value at the time of idling. Normally, such device logs have no difference in value except during maintenance, so detection is performed without distinction between processing time and idle time. Therefore, the start point of the detection target section is the maintenance end time, and the end point of the detection target section is the maintenance start time. Therefore, if both the start point and the end point are searched and only the sandwiched period is set as the detection target, the detection target section is not determined until the maintenance is started, and the detection cannot be performed. Therefore, in the fixed type, even after the event set as the end point occurs, the data after the start point needs to be detected.

  The difference of the section selection by these two detection object section type information is demonstrated using FIG. FIG. 5 is a table summarizing sections to be selected when section determination is performed with a variation type and a fixed type according to the state of input data. For input data on the vertical axis, No. The detection target section type information is shown on the horizontal axis.

  There are five possible positional relationships between the position of the start / end point of the detection target section and the range of the input data (the start time of the detection target section: a, the end time: b, and the start time of the input data range: A, end time: B).

Case 1: Input data includes both start and end points (A ≦ a ≦ b ≦ B)
Case 2: The input data includes only the starting point (A ≦ a ≦ B ≦ b)
Case 3: Input data is sandwiched between the start point and end point (a ≦ A ≦ B ≦ b)
Case 4: The input data includes only the end point (a ≦ A ≦ b ≦ B)
Case 5: No common part between input data and detection target section (a ≦ b ≦ A ≦ B)
Of these, in case 5, there is no common part between the detection target section and the input data range, and therefore the description is omitted.

  As described above, in the variation type, when both the start point and the end point are confirmed, data sandwiched between the start point and the end point is set as a detection target. Therefore, if both the start point and the end point are not included in the input data, it is not regarded as a detection target. Therefore, in the case of the variation type, only case 1 is determined to have a target section, and cases 2 to 4 are determined to have no target section. In the fixed type, the data after the start point is detected even before the event set as the end point occurs. Therefore, not only the case 1 but also the cases 2 and 3 are considered to be a desired operation state after the start point, and are determined to be detection target sections. In case 4, since the input data is in the range after the start point, the period until the end point condition is satisfied is considered as a desired operation state and is set as a detection target section.

  FIG. 6 is a flowchart of the detection target section extraction unit and the sign determination unit in FIG. Steps 601 to 605 are steps executed by the detection target section extraction unit, and steps 606 to 609 are steps executed by the sign determination unit.

  First, in step 601, information on a target manufacturing apparatus, a target chamber, and a target manufacturing recipe (an operation program name of a manufacturing apparatus used in the corresponding product / process) is selected. The manufacturing apparatus information includes the name of the manufacturing apparatus or the code name. Similarly, the chamber information includes a chamber name or a code name. Similarly, the manufacturing recipe information includes a manufacturing recipe name or code name. Depending on the type of manufacturing apparatus, there are one having a plurality of chambers in the same manufacturing apparatus, and one having only a single chamber in the manufacturing apparatus. For single chamber manufacturing equipment, there is no need for chamber information.

  In step 602, detection target section information is acquired. Information to be acquired includes event establishment condition information, event section correspondence information, setting information (user input items) about detection target section type information, and detection target section flags (feedback items).

  In step 603, maintenance history information and start history information are acquired. The maintenance history information is information on the work start and end date / time of the maintenance type. The construction history information is information on the start and end date / time of the recipe for each recipe.

  In step 604, device log information of the corresponding device is acquired. The device log information includes a device log name, a device log, and a device log acquisition date / time.

  In step 605, the detection target section is calculated based on the acquired information. Step 605 is divided into two steps, step 6051 and step 6052.

  Step 6051 performs event setting processing based on the behavior of the apparatus log, the start history, and the maintenance history according to the flow shown in FIG.

  In step 701, it is confirmed whether maintenance (maintenance history information) is set as the event type of the event establishment condition information. If it is set (use), the process proceeds to step 702. If it is not set (not used), the process proceeds to step 703.

  In step 702, an event is set based on the maintenance history information. If the maintenance work type set in the log name is targeted and the operation is started, the time when the corresponding work is started is set as an event. When the operation is finished, the time when the corresponding work is finished is set as an event.

  Next, in step 703, it is confirmed whether or not completion (starting history information) is set as the event type of the event establishment condition information. If it is set (use), the process proceeds to step 704. If it is not set (not used), the process proceeds to step 705.

  In step 704, an event is set based on the start history information. If the recipe is set to the log name and the operation is started, the time when the recipe starts is set as an event. When the operation is finished, the time when the start of the recipe is finished is set as an event. If all recipes are set in the log name, all recipes are targeted.

  Next, in step 705, it is confirmed whether or not a device log (device log information) is set as the event type of the event establishment condition information. If it is set (use), the process proceeds to step 706. If it is not set (not used), the process is terminated.

  In step 706, an event is set based on the behavior of the device log information. If the device log set in the log name is the target, and the content set in the operation item is "over from top to bottom", the value set for the threshold for the corresponding device log is from top to bottom. Set the time that passed the time as an event. Similarly, when it is “Contact from above”, the time when the corresponding device log touches the threshold from above is exceeded. When it is “Over from bottom to top”, the corresponding device log falls below the threshold. When the time exceeding the top is “contact from below”, the time when the corresponding device log contacts the threshold from the bottom is set as an event. In step 706, if necessary, the device log may be subjected to preprocessing by statistical processing such as moving average processing, moving median processing, or signal processing such as frequency decomposition using Fourier transform.

  Next, step 6052 in FIG. 6 searches for the start and end of the detection target section by searching the processing sequence according to the flow shown in FIG.

  In step 801, a search target event and search target start / end point information are selected in order from the event section correspondence information acquired in step 602 in FIG.

  Next, in step 802, the time is sequentially advanced in order to search for an event from the past time of the device log acquisition date.

  In step 803, it is confirmed whether an event has occurred at the target time. If no event has occurred (none), the process returns to step 802 to advance the time. If an event has occurred (Yes), the process proceeds to step 804, and it is determined whether the end point being searched is the start point or the end point. If it is the start point, the process proceeds to step 805, and if it is the end point, the process proceeds to step 807.

  First, a case where the end point being searched is the start point will be described.

  In step 805, it is determined whether or not all the events set as the start point have been found by the found event. If all have been found (established), the process proceeds to step 806, where the start point is set, and the process returns to step 801. If all the start point conditions have not been found (not established), the process returns to step 801.

  Next, a case where the end point being searched is an end point will be described.

  In step 807, it is determined whether all the events set as end points have been found by the found event. If all have been found (established), the process proceeds to step 808, the end point is set, and the process proceeds to step 809. If all the end point conditions have not been found (not established), the process proceeds to step 809.

  In step 809, it is determined whether the input data has been searched to the end. If input data that has not yet been searched remains (not established), the process returns to step 801. If all input data has been searched (established), the process proceeds to step 810.

  In step 810, the setting information about the detection target section type information acquired in step 602 in FIG. 6 is confirmed. If it is a variation type, the process proceeds to step 812. If it is a fixed type, the process proceeds to step 811, the latest data acquisition date / time of the input data is set as the end point, and the process proceeds to step 812.

  In step 812, it is determined whether the latest data of the input data is within the detection target section or outside the detection target section, the detection target section flag is updated, and the process ends.

  Next, the process proceeds to step 606 in FIG. In step 606, the detection target apparatus log information in the detection target section calculated in step 605 is acquired.

  Next, in step 607, representative value processing is performed by calculating a statistic such as an average or median value of the acquired apparatus log information. Note that this step does not require processing.

  Next, in step 608, a detection threshold value is calculated using the apparatus log information that has undergone the representative value processing (or the apparatus log information acquired in step 606).

  Next, in step 609, the detection threshold value calculated in step 608 is set for the apparatus log information (or the apparatus log information acquired in step 606) subjected to the representative value processing in step 607. A detection process for determining the presence / absence of a sign is executed, and the process ends.

  The above flow is repeatedly executed according to a cycle T set independently of the periodic sampling data acquisition cycle t. As a result, it is only necessary to perform detection processing at a frequency that can be processed using data sampled at a frequency at which signs can be detected, and even a server with low processing capacity can be detected compared to a method that performs detection processing at a data sampling cycle. Processing can be performed. Thereby, it is possible to maintain the detection accuracy while suppressing the investment amount.

  When the above flow is executed and it is determined that a sign has occurred, the engineer is notified by e-mail. Information to be notified includes, but is not limited to, the target detection target device name, the target detection result, the target device log information, the threshold value used for the target determination, and the start and end times of the detection target section. . The engineer can also check the sign detection result from the terminal in the room at any time.

  Each process of the above sign detection processing flow is realized by software using programs such as a main unit 905, a detection processing unit 906, and a detection target section determination unit 907 that configure a sign detection system of FIG.

<Early detection system for equipment abnormalities in semiconductor manufacturing>
With reference to FIGS. 9 and 10, an example of an apparatus abnormality sign detection system in semiconductor manufacturing according to an embodiment of the present invention will be described. FIG. 9 is a diagram showing an example of the sign detection system according to the present embodiment. FIG. 10 is a diagram showing an example of a screen for setting detection conditions.

  The sign detection system according to the present embodiment includes a sensor unit (data logger) that constitutes a plurality of detachable sensors connected to one or more types of manufacturing apparatuses in the manufacturing apparatus group, apparatus data acquired in the past, and manufacturing. A database section that stores device operation information, manufacturing device maintenance history information, event definition information for each device, event occurrence order information, detection target determination information, etc., and whether or not the device data contains a sign of failure The detection unit (main unit, detection processing unit, detection target section determination unit) to determine, and the interface unit (main unit) that can notify the engineer of the sign determination result or input information necessary for detection by the engineer ) Etc.

  In FIG. 9, the manufacturing apparatus group 901 includes at least one manufacturing apparatus (1 to N) 902. Further, depending on the type of manufacturing apparatus group 901, there are those having a plurality of chambers in the same manufacturing apparatus 902 and those having only a single chamber in the manufacturing apparatus 902. Then, the semiconductor wafer is processed in each chamber of the manufacturing apparatus 902.

  Data loggers (1 to N) 903 are connected to each manufacturing apparatus 902, respectively. The data logger 903 periodically samples the apparatus log of the manufacturing apparatus 902. All the data loggers 903 are connected to a PLC (programmable logic controller) 904. The PLC 904 periodically acquires a device log sampled by the data logger 903. The PLC 904 stores the acquired device log and the device log acquisition date and time in the memory. At this time, if necessary, statistical processing such as an average value, a maximum value, and a minimum value is performed. The PLC 904 is connected to the main unit 905 via a network.

  The main unit 905 acquires the device log stored in the memory of the PLC 904 and the device log acquisition date and time for each preset period. The main unit 905 is connected to the database unit 908 via a network. The main unit 905 associates a device name and a device log name with the device log acquired from the PLC 904 and the device log acquisition date and time, and stores them in the database 909 in the database unit 908. In addition, the main unit 905 performs the sign detection process for each manufacturing apparatus 902 according to a preset period. The main unit 905 detects the detection condition information of the corresponding device from the database 911 in the database unit 908, the start history information of the corresponding device from the database 910, the maintenance history information of the corresponding device from the database 912, and the database 909. The device log information of the corresponding device is acquired. Here, the detection condition information includes a data extraction condition, a detection threshold value calculation condition, and a detection condition. Of the acquired information, the main unit 905 transmits data extraction conditions, start history, maintenance history, and device log history to the detection target section determination unit 907.

  When the detection target section determination unit 907 acquires information from the main unit 905, the detection target section determination unit 907 calculates a detection target section using the acquired data. The detection target section determination unit 907 transmits the calculated detection target section to the main unit 905. When the main unit 905 acquires the detection target section, the main unit 905 transmits the apparatus log corresponding to the detection target section, maintenance history information, detection threshold calculation conditions, and detection conditions to the detection processing unit 906 in the apparatus log history.

  When the detection processing unit 906 acquires information from the main unit 905, the detection processing unit 906 calculates a detection threshold according to the detection threshold calculation condition, performs detection of the target data according to the detection condition, and whether the corresponding device is normal or abnormal. Detect if there was. The detection processing unit 906 transmits the detection result to the main unit 905. The main unit 905 stores the detection result in the database 913. The main unit 905 is connected to the terminal unit 914 via a network.

  The terminal unit 914 includes at least one terminal (1 to N) 915. The engineer can check the detection result acquired by the main unit 905 through the terminal 915. The main unit 905 can also transmit the detection result to the terminal 915 by mail. The engineer can also input detection conditions to the terminal 915. The input detection condition is acquired by the main unit 905 and stored in the database 911 of the database unit 908.

  At this time, an example of a screen on which the engineer inputs detection conditions (conditions for the start and end of the detection target section) is shown in FIG. On this screen, for each device, it is possible to specify the definition and order of events for determining the start point of the detection target section, and the definition and order of events for determining the end point. In the apparatus log column, not only the apparatus log name but also a maintenance work type and a start recipe can be specified.

<Effect of Embodiment>
According to the present embodiment, in the sign detection system having the data logger 903, the main unit 905, the detection processing unit 906, the detection target section determination unit 907, the database unit 908, etc., the apparatus log information, maintenance history information, and start history information Collection step (601 to 604), event setting step (6051), detection target section calculation step (6052), detection target section data acquisition step (606), statistical processing step (607), predictive judgment step (608, 609), etc. By executing the predictive detection method having the above, the detection target section of the apparatus log information periodically acquired from the data logger connected to the manufacturing apparatus used in the semiconductor wafer manufacturing process includes a sign leading to a failure. Can be determined.

  In this way, it is possible to automatically select data in a desired operating state using data acquired periodically and determine whether or not a sign of a change in the device state leading to product failure has occurred. .

  As a result, since the start / end of the predetermined operation state is determined from the sequence of events that are always performed before the start / end of the predetermined operation state using a plurality of device logs, it is difficult to distinguish them with the prior art. It becomes possible to distinguish the operation state which existed. Thereby, since only the data of the period in a predetermined driving | running state can be made into a detection target, it becomes possible to implement | achieve the improvement of detection accuracy.

  As mentioned above, the invention made by the present inventor has been specifically described based on the embodiment. However, the present invention is not limited to the embodiment, and various modifications can be made without departing from the scope of the invention. Needless to say.

  The semiconductor device manufacturing technology of the present invention can be used in particular for a method and system for detecting a sign of apparatus abnormality during processing by a manufacturing apparatus.

601 ... Target manufacturing device information selection step, 602 ... Detection target section information acquisition step, 603 ... Maintenance history information / starting history information acquisition step, 604 ... Device log information acquisition step, 605 ... Detection target section calculation Step 6051 ... Event setting process step 6052 ... Detection target section search step 606 ... Device log information acquisition step in detection target section 607 ... Representative value processing step 608 ... Detection threshold calculation step 609 ... Detection processing execution step,
701 ... Maintenance history information determination step, 702 ... Event setting step based on maintenance history information, 703 ... Start history information determination step, 704 ... Event setting step based on start history information, 705 ... Device log information determination step, 706 ... Device log information Event setting step by behavior,
801 ... Search target event, start / end point information selection step, 802 ... Time advance step, 803 ... Event presence / absence determination step, 804 ... End point determination step during search, 805 ... Start point condition determination step, 806 ... Start point setting step, 807 ... end point condition determination step, 808 ... end point setting step, 809 ... end condition determination step, 810 ... detection target section type determination step, 811 ... end point setting step, 812 ... update step of detection target section flag,
901: Manufacturing device group, 902 ... Manufacturing device, 903 ... Data logger, 904 ... PLC, 905 ... Main unit, 906 ... Detection processing unit, 907 ... Detection target section determination unit, 908 ... Database unit, 909 ... Device log information database 910 ... Construction history information database, 911 ... Detection condition information database, 912 ... Maintenance history information database, 913 ... Detection result database, 914 ... Terminal unit, 915 ... Terminal.

Claims (4)

A method for determining whether a detection target section of apparatus log information periodically acquired from a sensor connected to a manufacturing apparatus used in a semiconductor wafer manufacturing process includes a sign leading to a failure,
Device log information consisting of the device log name of the target device, maintenance history information consisting of the maintenance work content, maintenance work start time, and maintenance work end time of the target device, the start recipe, start start time, and start end time of the target device Device log information, maintenance history information, and start history information collection step for acquiring start history information consisting of:
The device log name, a threshold value for determining an event necessary for detection, and an operation type that specifies an operation in which the device log value of the device log name intersects or touches the threshold value are set in advance. The event occurrence time information is obtained by combining at least one of the time when the device log information intersects the threshold value, the maintenance history information, and the start history information using the event definition information. An event setting step to set;
Detection target section information is generated from the event generation time information, event generation order information that associates a predetermined event generation order with the start point of the processing section and the end point of the processing section, and the event generation time information. Detecting target section calculating step,
A detection target section data acquisition step for acquiring detection target section data from the detection target section information and the device log information;
Statistical processing step for statistically processing the detection target section data and generating detection target statistical processing processed data;
And a sign determination step for determining whether or not the detection target statistical processing processed data includes a sign of an apparatus failure. A method for detecting a sign of an apparatus abnormality in semiconductor manufacturing. In the method for predicting apparatus abnormality in semiconductor manufacturing according to claim 1,
In the detection target section calculating step, the event generation time information, event generation order information in which a preset event generation order, the start point of the processing section and the end point of the processing section are associated with each other, and the previous detection process An apparatus abnormality sign detection method in semiconductor manufacturing, wherein detection object section information is generated from detection target determination information of data at the latest time among apparatus log information at the time and the event occurrence time information. In the method for predicting apparatus abnormality in semiconductor manufacturing according to claim 1,
In the apparatus log information / maintenance history information / starting history information collecting step, the acquired apparatus log information includes at least one of information related to past driving and information related to an internal state of a processing chamber of the manufacturing apparatus. An apparatus abnormality sign detection method in semiconductor manufacturing, comprising: A system for determining whether a detection target section of apparatus log information periodically acquired from a sensor connected to a manufacturing apparatus used in a semiconductor wafer manufacturing process includes a sign leading to a failure,
A sensor unit comprising a plurality of removable sensors connected to one or more types of manufacturing devices;
A database unit that stores device data acquired in the past, manufacturing device operation information, manufacturing device maintenance history information, and event definition information, event occurrence order information, and detection target determination information according to claim 1 for each device. When,
A detector that determines whether the device data includes a sign leading to failure;
The engineer is notified of the sign determination result, the threshold value for the engineer to determine the event necessary for detection, the event occurrence order information for determining the start point of the section, and the section An interface unit capable of inputting event generation order information for determining an end point and information on whether only data between the start point and the end point is to be detected or whether all data after the start point is to be detected An apparatus abnormality sign detection system in semiconductor manufacturing, characterized by comprising:

Images