JP5583962B2 - Charged particle beam equipment - Google Patents

Description

  The present invention relates to a charged particle beam apparatus, for example, a charged particle beam apparatus used for evaluation of a semiconductor manufacturing process.

  A charged particle beam apparatus is frequently used for dimensional inspection of fine patterns on semiconductor wafers and inspection of fine defects. The semiconductor wafer size is 300 [mm] at present, but a large diameter is being studied to 450 [mm] in the near future. Also, the pattern dimension is mainly 35 [nm] or less, and the miniaturization of semiconductor processes is progressing. Accordingly, a charged particle beam apparatus that performs these dimension inspections and pattern defect inspections is desired to have higher accuracy measurement and higher resolution.

  Among such charged particle beam devices, for charged particle beam devices that perform particularly high-precision inspection and measurement, the device performance is maintained at a high level, and a plurality of devices in a semiconductor production line It is important to make the performance gap as close to 0 as possible. The use of a charged particle beam device with guaranteed performance in this way improves the yield by immediately detecting a defect due to fluctuations in the semiconductor manufacturing process and feeding it back or feed-forward to the semiconductor manufacturing device.

  However, as the performance of the charged particle beam apparatus is improved, the mechanism and processing algorithm for realizing it become more complex, and the maintenance and adjustment man-hours necessary to maintain the performance are gradually increasing. Therefore, it is extremely difficult to evaluate the absolute achievement level of the charged particle beam device adjustment and to confirm its sustainability.

  As a typical solution to the evaluation of the absolute achievement level of device adjustment, for example, Patent Document 1 is available. As shown here, by preparing a standard sample and observing it, the absolute performance state of the charged particle beam apparatus is grasped and the performance is adjusted by adjustment. Many other proposals have been made regarding what samples are prepared and what is used as an index.

JP 2007-122995 A JP-A-4-269613

  However, the above-mentioned Patent Document 1 does not have a specific measure as to when absolute adjustment using a standard sample should be performed (execution time) as a method for confirming sustainability after adjustment of a charged particle beam apparatus.

  In the prior art, in many cases, it is performed periodically, or the performance is deteriorated so that it cannot be used, or triggered by a serious error of the charged particle beam apparatus. At this time, the apparatus must be stopped and removed from the semiconductor manufacturing process, which requires a great deal of time and money. As a result, the same cost is generated for all charged particle beam apparatuses that have not actually deteriorated in performance or have a short actual operation time, and are wasteful. In addition, the charged particle beam apparatus that has caused the performance deterioration to withstand use may cause a defective lot to be passed to a subsequent process, which greatly deteriorates the yield.

  Furthermore, it is very difficult to determine whether a serious error is caused by a device that is an internal factor or a semiconductor manufacturing process that is an external factor. Therefore, the user takes a lot of time to analyze a serious error, and cannot immediately find a defect due to a variation in the semiconductor manufacturing process, and the yield is deteriorated.

  The present invention has been made in view of the above problems, and by maintaining the performance of a highly-adjusted charged particle beam apparatus without specially preparing a special sample and special work time, a semiconductor It is an object of the present invention to provide a charged particle beam apparatus that can immediately find defects caused by variations in manufacturing processes.

  In order to solve the above-described problems, the present invention predicts the timing of absolute adjustment of a charged particle beam apparatus by continuously monitoring the performance of the apparatus and continuously evaluating the deterioration tendency. In addition, it properly discriminates between defects caused by fluctuations in the semiconductor manufacturing process and device performance deterioration, and warns the user to prevent performance deterioration of the charged particle beam device.

That is, the charged particle beam apparatus according to the present invention uses the apparatus control function to move the electron optical system to the normal focus position of the inspection target on the sample set as the processing procedure parameter, and acquires the image information of the inspection target. the a charged particle beam apparatus for performing a pattern search to be inspected, the coordinates before Symbol image processing of the inspection object moved by using the device control function on the basis of the image information of the inspection object by using an image processing function a difference value calculating unit for obtaining a difference value between specified the inspection target coordinates by using the function, by the difference value calculation unit every time the acquisition and pattern searching of the image information of said object is performed in the procedure parameter The difference values calculated from different samples are classified and managed in correspondence with the coordinates of the inspection object moved using the apparatus control function, and are divided into the same coordinates of the inspection object. A difference value analysis unit for reading the performance state of the charged particle beam apparatus from variations in the difference value, based on the difference value analysis of the results, and a warning generating unit for performing a warning to the user in a predetermined case, Is provided.

  According to the charged particle beam apparatus of the present invention, by maintaining the performance of the highly-adjusted charged particle beam apparatus without preparing a special sample and a special work time, the semiconductor manufacturing process can be changed. It is possible to provide a charged particle beam apparatus that can immediately find a defect caused by the defect.

  As a result, the cost required for maintenance and inspection can be reduced, the usefulness of the apparatus can be maintained at a high level, and the yield and the overall efficiency of the facility can be increased.

Schematic configuration of onboard charged particle beam equipment Drawing depicting the difference between the wafer coordinates moved using the device control function and the wafer coordinates specified using the pattern detection function Drawing depicting the difference between the wafer coordinates moved using the device control function and the wafer coordinates specified using the pattern detection function (parameter failure example) A drawing depicting the difference between the wafer coordinates moved using the system control function and the wafer coordinates specified using the pattern detection function Drawing depicting the difference between the wafer coordinate moved using the device control function and the wafer coordinate specified using the pattern detection function (device failure example) Drawing depicting the difference between the wafer coordinates moved using the equipment control function and the wafer coordinates specified using the pattern detection function (process variation example) At each inspection point on the wafer, the control amount of the lens determined from the potential height by the device control function and the control amount of the lens determined from the physical height, and the sum of these control amounts and the auto focus function Drawing the difference value with the controlled amount of the lens At each inspection point on the wafer, the control amount of the lens determined from the potential height by the device control function and the control amount of the lens determined from the physical height, and the sum of these control amounts and the auto focus function Drawing the difference value from the controlled amount of the lens (example with large difference) At each inspection point on the wafer, the control amount of the lens determined from the potential height by the device control function and the control amount of the lens determined from the physical height, and the sum of these control amounts and the auto focus function Drawing the difference value from the controlled amount of the lens (example where there is a feature in the difference value) At each inspection point on the wafer, the control amount of the lens determined from the potential height by the device control function and the control amount of the lens determined from the physical height, and the sum of these control amounts and the auto focus function Drawing of the difference value with the controlled amount of the lens (the difference value has a characteristic part but is not applicable) Adjustment time prediction chart based on error appearance ratio Integrated system configuration diagram on inspection line Flow chart showing a method for analyzing the cause of variation in the difference value and warning the user of the cause Flow chart showing a method to warn the user of deterioration of the performance state of the charged particle beam device Internal configuration diagram of the control computer 112 Internal configuration diagram of the host computer 110

The present invention relates to a charged particle beam apparatus used for evaluation of a semiconductor manufacturing process.
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be noted that this embodiment is merely an example for realizing the present invention, and does not limit the technical scope of the present invention. In each drawing, the same reference numerals are assigned to common components.

<Configuration of charged particle beam device>
FIG. 1 is a diagram showing a schematic configuration of a semiconductor charged particle beam apparatus according to an embodiment of the present invention.
The charged particle beam device includes an electron optical system 101, an optical microscope 102 used for observation at a low magnification, a charge measuring device 104 for measuring a carry-in charging potential of a wafer 108, a charge eliminating device 106 for removing the charge measuring device 104, The sample chamber 107 is provided with a sample height measuring device 105 using a Z sensor or the like for accurately measuring the distance in the height direction between the electron optical system 101 and the wafer 108. A stage 109 for transferring a wafer 108 to be a sample to a target observation position is installed inside the sample chamber 107 maintained in a high vacuum, and a stage using a linear scale or the like for measuring the position of the stage. A position measuring device 103 is installed with respect to the operation axis of the stage.

  The control computer 112 is connected to the apparatus via the control circuit 113, and is obtained from the stage position information obtained from the stage position measuring device 103, the sample height information obtained from the sample height measuring device 105, and the charge measuring device 104. It is designed to control the electron optical system based on the potential information of the sample. For example, based on the stage position information, the difference from the position coordinate of the wafer 108 to be originally observed is calculated, and the amount is applied to the coil of the electron optical system 101, thereby changing the irradiation position of the electron beam and inspecting. Observe the target coordinates. Further, based on the sample height information, the amount of current to be applied to the electromagnetic lens is calculated and applied so as to form a positive focus on the sample surface, thereby performing beam irradiation at the positive focus. Further, based on the potential information of the sample, the time for performing neutralization by the neutralization device 106 is determined, and the necessary and sufficient neutralization process is performed, or the electrostatic electrode of the electron optical system 101 is used without performing neutralization. Beam irradiation at the focal point. As described above, the control computer 112 controls the electron optical system so that the correct position can be observed with the correct focus by using various measuring devices. In general, this function is called an apparatus control function.

  Further, control of an electron optical system that is difficult to perform physical measurement, such as optical axis correction of an electron beam, is adjusted in advance according to a designed correction algorithm, and the adjustment data parameter 115 is stored in a storage device held by the host computer 110. 111 is stored. The control computer 112 reads this out in a timely manner, matches the current control preconditions of the electron optical system 101, and performs optical axis correction control of the electron beam according to the designed correction algorithm. The current control preconditions for the electron optical system 101 are adopted so that an appropriate electron beam energy, resolution, and depth of focus can be obtained according to the semiconductor manufacturing process to be observed and measured. Since the control precondition depends on the semiconductor manufacturing process, it is generally managed in a processing procedure parameter 114 which is a file or a group of files called a recipe.

  FIG. 15 is an internal configuration diagram of the control computer 112. The control computer 112 includes an electron optical system control unit 1501, a feature location determination unit 1502, a height measurement unit 1503, an error occurrence count unit 1504, a linear regression model creation unit 1505, a warning generation unit 1506, a difference value calculation unit 1507, a difference value. An analysis unit 1508.

  The user selects a recipe according to the purpose and object to be observed, so that the control preconditions described in the processing procedure parameter 114 are applied and interpreted by the control computer 112 to control the electron optical system 101.

  On the other hand, the charged particle beam device converts an electronic signal obtained by the detector into image data 116.

  The host computer 110 has a function of reading information necessary for controlling the electron optical system 101 from the image data 116 by performing various processes on the converted image data 116. In general, this function is called an image processing function.

  FIG. 16 is an internal configuration diagram of the host computer 110. The host computer 110 includes an image correction unit 1601 and a lens control amount calculation unit 1602.

<Functions of charged particle beam equipment>
In the present embodiment, the charged particle beam apparatus used in the semiconductor manufacturing process calculates the difference value of the control amount to the electron optical system, which is calculated using each of the apparatus control function and the image processing function that are mounted as standard. By tracking, the performance state of the charged particle beam apparatus is objectively determined.

  In many cases, the device control function is a function for controlling the electron optical system so that observation can be performed with a correct focus by adjusting various parameters based on the control conditions of the electron optical system corresponding to the semiconductor manufacturing process. Further, the measurement result by the hardware sensor may be input, and an appropriate control amount may be calculated by the computer accordingly to control the electron optical system. In many cases, many calculation parameters are used for calculating the control amount.

  The image processing function, on the other hand, takes image information created by the electron optical system of the charged particle beam device controlled by the device control function as input, and interprets the image information by applying various image processing algorithms to it. This function corrects image information so as to be in focus. In addition, the lens control amount calibrated based on the image information is calculated and fed back to the apparatus control function. The apparatus control function receives feedback from the image processing function and finely adjusts the electron optical system so that the focus is correct.

  Generally, an image is acquired in a state where the electron optical system is roughly placed at a position where the focus is achieved by the device control function, and then the image correction function is performed by the image processing function, or the image processing function is controlled by the device. Fine-tune the electron optical system by feeding back to the function. As described above, conventionally, the relationship between the device control function and the image processing function is the relationship between the sparse control and the fine control, or the relationship between the sparse adjustment and the fine adjustment, and it is considered that there is a difference between the two. .

  However, as the demand for higher throughput of the apparatus gradually increases, the accuracy of the apparatus control function gradually improves, and the image correction amount by the image processing function, or feedback to the apparatus control function for fine adjustment of the electron optical system Control amount has decreased. As a result, it has become possible to determine minute abnormalities and tendencies of the electron optical system that could not be determined due to the excessively large correction amount and feedback control amount.

  In other words, if the device control function is as accurate as the image processing function, it can be said that the image processing function can objectively determine the image acquired by the electron optical system controlled by the device control function.

  Therefore, by continuously collecting and managing the difference value of the control amount to the electron optical system by both functions and statistically processing it in association with the semiconductor manufacturing process, an external cause failure, for example, the semiconductor manufacturing process Specimen failure due to sample fluctuations and internal factor performance degradation, such as degradation of the electron optics system over time, sample position and height, sensor degradation for potential measurement, etc. Degradation and the like can be separated. Then, it is possible to predict the execution time of absolute adjustment necessary for recovery of the performance state and inform the user of it.

  Thereby, the user can improve the yield by immediately detecting the defect of the sample due to the fluctuation of the semiconductor manufacturing process and feeding back or feedforward it to the semiconductor manufacturing apparatus.

<Performance inspection based on wafer coordinate difference values>
One of image processing functions frequently performed in a charged particle beam apparatus is a pattern detection function. This function is a function for searching where a pre-given pattern image is in an observation image and specifying the position thereof, and is used for detecting the position of a semiconductor pattern or the like to be inspected. .

  On the other hand, when operating the apparatus control function, the electron optical system control unit 1501 places the wafer at a predetermined position on the wafer coordinates set in the processing procedure parameter 114 based on the signal from the stage position measuring device 103. The stage is controlled as described above, and the electron beam is shifted so as to be interlocked with the stage. That is, the wafer coordinates to be inspected should be accurately realized on the charged particle beam apparatus so as to be directly under the electron beam as described in the processing procedure parameter 114.

  Assuming that, the wafer coordinates to which the electron optical system has moved based on the processing procedure parameter 114 are continuously statistically investigated to determine how far the wafer coordinates are identified from the wafer coordinates specified using the pattern detection function. This makes it possible to read information on the semiconductor manufacturing process and the performance state of the charged particle beam apparatus.

  Here, the position of the inspection object (coordinates on the wafer) moved to the wafer coordinates set in the processing procedure parameter 114 using the apparatus control function calculated by the difference value calculation unit 1507 and specified using the pattern detection function By investigating the difference value from the inspection target position (coordinates on the wafer), it is possible to discriminate between a sample defect due to a variation in the semiconductor manufacturing process and a deterioration in the performance state of the charged particle beam apparatus. When the difference value analysis unit 1508 analyzes the difference value and detects the deterioration of the performance state of the charged particle beam apparatus, the warning generation unit 1506 gives a warning to the user. A user who has received a warning can maintain the performance state of the charged particle beam apparatus in a highly adjusted state by performing maintenance of the charged particle beam apparatus in accordance with the warning. As a result, the user can immediately find a defect in the sample due to fluctuations in the semiconductor manufacturing process without bothering the deterioration of the performance state of the charged particle beam apparatus.

  2 to 6 are diagrams in which difference values between the wafer coordinates moved using the apparatus control function and the wafer coordinates specified using the pattern detection function are drawn.

  The coordinate difference value is classified and managed for each processing procedure parameter 114 (recipe) created depending on the semiconductor manufacturing process. Further, the difference value is managed for each wafer coordinate determined using the pattern detection function.

  For example, as shown in FIG. 2, the recipe “Photo01001XX” alignment point No. 1. Classification and management using information such as wafer coordinates (X, Y) = (21312230, 15623) [nm] as a key. The difference value is drawn on the X and Y planes as a vector from the origin (0, 0) to the black point.

  Hereinafter, with respect to the variation patterns of the difference values in FIGS. 2 to 6, how the deterioration of the performance state of the charged particle beam apparatus is detected in each pattern and a warning is given to the user will be described. Hereinafter, the difference value variation is calculated by the difference value calculation unit 1507, the difference value analysis unit 1508 is analyzed, and the warning generation unit 1506 issues a warning to the user.

  FIG. 13 is a flowchart showing a method of analyzing a factor of variation in the difference value and warning the user of the factor.

  The difference value calculation unit 1507 calculates the difference value, and determines whether or not the variation in the difference value is within a certain range (S1301). If the variation of the difference value is within a certain range, the difference value analysis unit 1508 follows the center of gravity of the variation in the operation time direction of the charged particle beam apparatus. Then, the center of gravity is analyzed in the time direction, and it is determined whether the center of gravity is near the origin in the short term (S1302). If the center of gravity is in the vicinity of the origin, it is determined whether the center of gravity is in the vicinity of the origin over the long term (S1303). Again, if the center of gravity is near the origin, the wafer coordinates of the alignment point set in the processing procedure parameter 114 are appropriate for the sample of the semiconductor manufacturing process, and both the stage accuracy and the beam shift accuracy are well adjusted. Therefore, it is determined that the performance state of the charged particle beam apparatus is normal (S1304). In this case, the warning generation unit 1506 may notify the user that it is normal. FIG. 2 is an example in which the variation of the difference value is within a certain range near the origin (0, 0) for both the X and Y components. Note that the variation in the difference value is distributed near the origin (0, 0) due to the stage stop error.

  If it is determined in S1301 that the variation in the difference value is within a certain range, but it is determined in S1302 that the center of gravity is not near the origin in the short term, the alignment point set in the processing procedure parameter 114 Is determined to be not appropriate for the sample of the semiconductor manufacturing process (S1305). This means that although the semiconductor manufacturing process and the performance state of the charged particle beam apparatus are sound, they have no likelihood for determining the position to be inspected. Therefore, the warning generation unit 1506 can work to prevent a stop due to an error in the inspection process by automatically issuing a warning requesting the user to correct the processing procedure parameter 114 (S1310). . FIG. 3 is an example in which the variation itself is within a certain range as compared with FIG. 2, but the center of gravity of the variation in the difference value is far from the origin (0, 0).

  If it is determined in S1301 that the variation in the difference value is within a certain range and the center of gravity is in the vicinity of the origin in S1302, but the center of gravity is not in the vicinity of the origin in S1303 Then, it is determined that the sign accompanying the deterioration of the device with time is appearing (S1306). In such a case, the control of the stage 109 and its position measuring device 103, or the control algorithm for shifting to the beam irradiation position according to the measurement value of the stage position and the parameter are incompatible. If there is a tendency for the center of gravity movement of the difference value variation to be performed in a relatively short period of time, it is caused by the adjustment parameter of the electron optical system 101 for calculating the beam shift amount and is performed in a considerably long period. If there is a tendency, there is a backlash of the mounting of the stage position measuring device 103. As described above, if the time factor is also taken into consideration regarding the movement of the center of gravity of the difference value difference, the cause location can be estimated in more detail. Furthermore, it is possible to roughly predict when the pattern detection limit 501 will be exceeded from the relationship between the transition speed of the center of gravity movement and short-term variations. For this reason, the warning generation unit 1506 automatically issues a warning indicating how much time is left to execute absolute adjustment to the user so that the adjustment work is planned at an appropriate time. Work can be done (S1310). FIG. 4 shows an example in which the center of gravity continuously moves in the long term although the short-term variation does not vary greatly.

  If it is determined in S1301 that the variation in the difference value is not within a certain range, it is determined whether there are a plurality of variation sets (S1307). When the variation set is independent and the center of gravity of the variation does not show any tendency, the performance state of the charged particle beam apparatus that is already an internal factor (for example, the stage 109, its position measuring device 103, or beam shift accuracy) ) Is determined to have already failed (S1308). In other words, the device control function is not functioning normally, the image processing function has corrected the image and has barely recovered, and the inspection has only been performed. Therefore, the warning generation unit 1506 can automatically issue a warning prompting the user to adjust the device as soon as possible, thereby making it possible to reduce the device downtime as much as possible (S1310). FIG. 5 shows an example in which the variation of the difference value is extremely large and the variation set is single.

  If it is determined in S1301 that the variation in the difference value is not within a certain range, and it is determined in S1307 that there are a plurality of variation sets, the charged particle beam apparatus has undergone a large environmental change or the semiconductor manufacturing process. It is determined that some variation has occurred (S1309). For this reason, it is possible to encourage the user to minimize the fatal damage by automatically issuing a warning that prompts the user to judge whether the environmental change or process change is appropriate. (S1310). In detecting whether there are multiple variation sets, various known algorithms are used to classify the statistic. For example, the Mahalanobis general distance is used to detect that the lever ratio has increased rapidly. And can be classified. This applies to cases where the calculated difference value continues to be drawn as a newly classified group after detecting that the lever ratio has suddenly increased, and the tendency can be seen simultaneously at all pattern detection locations. . FIG. 6 shows an example in which a set of classifications completely different from the set of variations of the difference values has appeared before a certain timing.

  As described above, by using the difference value between the result of the apparatus control function using the sensor for specifying the position on the X and Y planes with respect to the sample and the result of the image processing function for detecting the semiconductor pattern, It is possible to clearly identify the cause of the defect and the defect in the performance state of the charged particle beam apparatus main body caused by fluctuations in the semiconductor manufacturing process.

<Difference value of lens control amount calculated from each function>
As described above (FIG. 5), by collecting the difference values of the wafer coordinates in the time direction, it is simply that either the sample height measuring instrument 105 or the charge measuring instrument 104 or both have started to deteriorate in performance. It can be recognized and is very effective.

  However, it is difficult to determine whether there is a problem with either the sample height measuring instrument 105 or the charge measuring instrument 104 or the problem with the sample wafer 108 itself due to semiconductor manufacturing process variations. Both the physical height by the sample height measuring instrument 105 and the potential height by the charging measuring instrument 104 affect the focal length, so that it is not possible to focus by simply collecting the difference value of the wafer coordinates. This is because it is not possible to clearly distinguish between the physical height and the potential height.

  Therefore, the difference between the control amount for the lens determined from the physical height and the potential height and the control amount for the lens determined by the image processing function from the normal focus image is added in the time direction to the sample. Information on the performance state of the semiconductor manufacturing process and the charged particle beam apparatus is read by evaluating on the surface of the wafer 108, detecting characteristic points, and collecting them.

  There are roughly two factors that affect the focal length. One is the physical height from the sample to the electron optical system, and is used as the focal length. This is measured by the sample height measuring device 105. The other is the potential height held by the sample wafer 108. Since the potential charged on the sample wafer 108 causes a potential distance with respect to the electron optical system 101, the focal distance is affected. This is measured by the charge measuring device 104.

  First, when operating the apparatus control function, the height measurement unit 1503 calculates the sample height and potential of each inspection object in the sample chamber 107 in accordance with the wafer coordinates set in the processing procedure parameter 114 and the determined procedure. Each measurement is performed, and the electron optical system control unit 1501 controls the electron optical system so that observation can be performed with a correct focus. The control amount (current amount) actually applied to the electromagnetic lens is the sum of the current amounts calculated based on the sample height and potential.

  Next, when operating the image processing function, there is an autofocus function as a function that enables observation with the correct focus. The autofocus function samples the image while operating the electromagnetic or electrostatic lens for focus adjustment in order to perform automatic focusing, and determines the control amount of the lens that becomes the normal focus according to the focus evaluation value by differential method etc. It is a function to decide.

  Here, the difference between the control amount of the lens determined from the physical height and potential height by the device control function and the control amount of the lens determined from the normal focus image by the autofocus function is obtained, and the sample wafer is obtained. If a characteristic part is detected on the surface 108, information on the semiconductor manufacturing process and the performance state of the charged particle beam apparatus can be read.

  FIGS. 7 to 10 show the lens control amount 701 (hereinafter, the control amount 701 based on the potential height) and the physical height determined from the potential height by the apparatus control function at each inspection point on the wafer. The control amount 702 of the lens determined from the following (hereinafter referred to as the control amount 702 based on the physical height), and the difference value 703 (hereinafter referred to as the difference value 703) between the sum of these control amounts and the control amount of the lens determined by the autofocus function. ). In FIGS. 7 to 10, only one-dimensional (X direction) data is drawn for simplification. However, since the measurement is actually performed at each inspection point on the plane of the sample wafer 108, the data is drawn two-dimensionally. Is desirable. As a technique for determining a lens control amount with such an autofocus function, there is, for example, Patent Document 2.

  As shown in FIG. 7, when the control amount 701 based on the potential height and the control amount 702 based on the physical height have no characteristics for each inspection object, what the difference value 703 indicates. I do not know clearly.

  Further, in FIG. 8, the difference value 703 is uniformly large, but this graph alone causes the control amount 701 based on the potential height to be too large, or the control amount 702 based on the physical height. It is not possible to distinguish whether is too large. However, if uniformity depending on the semiconductor manufacturing process is observed in the absolute value of the physical height and the potential height, it is collected at each inspection object on the wafer as in the first embodiment. By processing statistically, it becomes possible to read information regarding the semiconductor manufacturing process and the performance state of the charged particle beam apparatus. In this case, the difference value 703 is effective as a trigger for notifying an abnormal value.

  FIG. 9 shows a behavior in which the difference value 703 is characteristic at the characteristic location 901. If the sample height measuring device 105 and the charge measuring device 104 are ideally adjusted performance states, the difference value 703 falls within a certain range of variation with the average value 0 as the center. Here, a characteristic portion 901 that deviates from this is analyzed.

FIG. 14 is a flowchart showing a method for warning the user of deterioration of the performance state of the charged particle beam apparatus.
First, the feature location determination unit 1502 determines the feature location 901 on the difference value 703 (S1401). Although several statistical processes can be applied to the determination method, a difference value 703 exceeding a preset threshold with respect to the average value is simply set as the feature location 901 (S1402). Alternatively, a method of determining an outlier by using a technique such as the Smirnov-Grubbs test more accurately may be employed. If the difference value 703 does not exceed the threshold value, the feature location determination unit 1502 determines the next feature location (S1404).

  When the feature location 901 is determined on the difference 703, the height measurement unit 1503 determines whether the cause is the physical height or the potential height. The control amount 701 and the control amount 702 based on the physical height are measured (S1403). In FIG. 9, the control amount 701 based on the potential height shows a behavior opposite to the difference 703 in the characteristic portion 901. That is, in order to eliminate the influence from the control amount 701 due to the potential height, it is considered that it appeared as a characteristic result in the characteristic location 901 on the difference 703. As described above, when the control amount 701 based on the potential height or the control amount 702 based on the physical height deviates from the approximate line, the cause is either the physical height or the potential height. Therefore, the error occurrence counting unit 1504 counts the number of potential measurement errors as the number of occurrences (S1405, S1406).

  However, in FIG. 9, a portion 902 where the lens control amount 701 depending on the potential height is similarly deviated from the approximate line is seen, but here the difference value 703 does not indicate a characteristic value. Therefore, since it cannot be determined whether the physical height or the potential height is the cause, it is not counted.

  In FIG. 10, even if the characteristic location 1001 is determined on the difference value 703, the cause cannot be clearly determined. Therefore, in this case, it is not counted. In FIG. 10, the noise performance of the measuring device that occurs frequently is considered as the cause.

  In this manner, the characteristic portion 901 is extracted from the difference value 703, and obvious height measurement errors and potential measurement errors are continuously counted. From the counted errors, the linear regression model creation unit 1505 creates a growth line 1102 as a linear regression model using the number of appearances or the appearance ratio. The warning generation unit 1506 provides an appropriate threshold value 1103 for this, predicts when to reach the threshold value 1103, and warns the user of the absolute adjustment necessity time 1105 corresponding to the threshold value (S1407, S1408). Thereby, the performance state of the highly-adjusted charged particle beam apparatus can be maintained.

<Application to integrated system>
The method of maintaining the performance state of the semiconductor manufacturing process and charged particle beam equipment is not performed on a single device basis, but is integrated for multiple devices in the inspection line for each semiconductor manufacturing process, so that there is a gap between the devices. Alternatively, it is possible to efficiently manage fluctuations in the semiconductor manufacturing process.

  FIG. 12 shows an integrated system configuration diagram in the inspection line. Every time a recipe is executed based on the processing procedure parameter 114 associated with the semiconductor manufacturing process from all the devices connected in the inspection line, the device control function and the image processing function obtained in the normal processing are obtained. The difference data 1205 is automatically transmitted to the data server computer 1201 via the network connection 1203.

  The transmitted difference data 1205 is classified and stored using a process name as a key in a storage area such as a database system of the data server computer 1201, and data processing is performed using the difference data obtained by the above method.

  The processed data is displayed on a screen 1202 on the monitor as a performance state index of the charged particle beam apparatus. Further, a warning is displayed on the screen 1202 as necessary, or is remotely transferred to a necessary administrator or service company, thereby realizing continuous monitoring.

  By introducing such a system management method, we monitor the fluctuations in the semiconductor manufacturing process and continuously monitor the performance status of each charged particle beam device and the expected maintenance / adjustment timing. Display warnings.

  By introducing this management system, the user can perform appropriate maintenance according to the actual state of the equipment without preparing special samples or special processes, and integrating the entire inspection line. By performing the scheduling, it is possible to maintain high equipment efficiency of the entire inspection line.

<Summary>
The charged particle beam apparatus according to the present invention obtains a difference value between the coordinates on the semiconductor wafer moved using the apparatus control function and the coordinates on the semiconductor wafer specified using the image processing function. The performance state of the charged particle beam apparatus is read from the difference value difference, and a warning is given to the user in a predetermined case. By this processing, the user can continue to maintain the performance state of the charged particle beam apparatus in a highly adjusted state, and can immediately find a defective sample due to fluctuations in the semiconductor manufacturing process.

  Further, when the variation in the difference value exceeds the limit area where the image correction function can be performed by the image processing function, a warning prompting the user to adjust the apparatus immediately is automatically issued. By this process, the user can suppress the apparatus non-operation time.

  In addition, by tracking the trend and speed of the centroid movement of the difference value variation in the time direction, it is predicted when the centroid of the difference value variation will exceed the limit area where the image correction function can perform image correction processing. Then, a warning indicating how much time is left to execute absolute adjustment of the charged particle beam apparatus is automatically issued to the user. By this process, the user can plan the adjustment work of the charged particle beam apparatus at an appropriate time.

  In addition, by tracking the tendency and speed of the centroid movement of the difference value variation in the time direction to indicate a set different from the difference value variation set at a predetermined timing, A warning is automatically issued to prompt the determination of whether the change or the semiconductor manufacturing process is appropriate. By this processing, the user can immediately find a defect in the sample due to a change in the semiconductor manufacturing process, and can minimize the fatal damage to the sample.

  Further, the charged particle beam device according to the present invention includes a lens control amount determined from a physical height and a potential height using an apparatus control function, and a lens determined from a positive focus image using an image processing function. The control value of the lens determined from the physical height and potential height at the feature location is determined. When the lens control amount determined from the physical height or potential height is out of the predetermined range, it is counted as the number of error occurrences, and the cumulative number of occurrences or the appearance ratio of error occurrences is used. Predict when the absolute adjustment of the charged particle beam device is necessary, and warn the user. By this processing, the user can know the necessary time for absolute adjustment of the charged particle beam apparatus at an appropriate time, and can maintain the performance state of the highly adjusted charged particle beam apparatus.

  In the charged particle beam apparatus system according to the present invention, the difference value between the apparatus control function and the image processing function is automatically transmitted to the data server computer via the network connection, and the transmitted difference value is stored in the data server computer. It is stored in the database system it has, predetermined processing is executed, and a warning is given to the user in a predetermined case. By introducing this system, the user can perform appropriate maintenance according to the actual state of the charged particle beam device without preparing a special sample or special process. By performing integration and scheduling, it is possible to maintain high equipment efficiency of the entire inspection line.

DESCRIPTION OF SYMBOLS 101 ... Electro-optical system 102 ... Optical microscope 103 ... Stage position measuring device 104 ... Charge measuring device 105 ... Sample height measuring device 106 ... Static elimination apparatus 107 ... Sample chamber 108 ... Sample (wafer)
109 ... Stage 110 ... Host computer 111 ... Storage device 112 ... Control computer 113 ... Control circuit 114 ... Processing procedure parameters (recipe)
115 ... adjustment data parameter 116 ... image data 501 ... pattern detection limit 601 ... Mahalanobis generalized distance 701 ... lens control amount 702 depending on potential height ... physical height Lens control amount 703 by ... Difference value 901 from lens control amount corresponding to the normal focus determined by the image processing function ... Characteristic location 1101 ... Appearance ratio drawing 1102 ... Growth line 1103 ... Threshold value 1104 ... Current 1105 ... Necessary timing 1201 ... Data server computer 1202 ... Screen display example 1203 ... Network connection 1204 ... Charged particle beam device 1205 ... Differential data

Claims (6)

The apparatus control function is used to move the electron optical system to the normal focus position of the inspection target on the sample set as the processing procedure parameter, to acquire the image information of the inspection target , and the image processing function is used to acquire the inspection target A charged particle beam apparatus that performs a pattern search of the inspection object based on image information ,
A difference value calculating unit for obtaining a difference value between the inspection target coordinates identified using coordinates before Symbol image processing function inspection object has moved by using the device control function,
The difference values calculated by the difference value calculation unit each time the image information to be inspected and pattern search are performed using the processing procedure parameters are moved using the apparatus control function. A differential value analysis unit that performs classification management corresponding to each coordinate of the inspection target, and reads the performance state of the charged particle beam device from the variation of the differential value classified into the same coordinate of the inspection target ;
Based on the result of the difference value analysis unit, a warning generation unit that warns the user in a predetermined case, and
A charged particle beam apparatus comprising: The warning generation unit promptly provides a device to the user when the variation of the difference value classified into the same coordinates of the inspection object exceeds a limit region where the image correction function can be performed by the image processing function. The charged particle beam apparatus according to claim 1, wherein a warning for prompting adjustment is automatically issued. The warning generation unit tracks the centroid movement tendency and speed of the variation of the difference value classified in the same coordinates of the inspection target in the time direction, so that the centroid of the variation of the difference value is the image processing function. Automatically predicts when it will exceed the limit area where image correction processing can be performed, and automatically issues a warning to the user indicating how much time is left to execute absolute adjustment of the charged particle beam device. The charged particle beam apparatus according to claim 1. The warning generation unit tracks the tendency and speed of the center of gravity movement of the variation of the difference value classified into the same coordinate of the inspection target in the time direction, so that the variation of the difference value can be measured at a predetermined timing. 2. A warning is automatically issued to urge a user to judge whether or not an environmental change or a semiconductor manufacturing process is appropriate when a set different from the set is indicated. Charged particle beam equipment. A plurality of charged particle beam devices, a data server computer, a database system provided in the data server computer, and a network connection for connecting the plurality of charged particle beam devices and the data server computer,
At least one of the plurality of charged particle beam devices is the charged particle beam device according to any one of claims 1 to 4 ,
The difference value between the apparatus control function and the image processing function calculated by the charged particle beam apparatus according to any one of claims 1 to 4 is automatically transmitted to the data server computer via the network connection. Transmitted
The transmitted difference value is stored in the database system. The program for functioning a computer as a charged particle beam apparatus of any one of Claims 1-4 .

Images