JP2013088415A - Semiconductor pattern measuring method and semiconductor pattern measuring device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problems that it is difficult for a user to determine the application propriety just by looking at an arbitrary pattern shape when selecting a measuring method and a method suitable for a pattern shape of a measuring object is not sometimes selected when automatically selecting the measuring method, in a method of measuring a pattern dimension on a semiconductor wafer.SOLUTION: In the method of measuring a dimension of a semiconductor pattern image captured by an electron beam, a measurement object part on a pattern is specified by a user, shape information is extracted from a pattern measuring part, an applicable measuring method is selected from a prepared measuring method library on the basis of the extracted shape information, the dimension of the pattern is measured by the selected method, and the measured dimension is outputted.

Description

  The present invention relates to a method and an apparatus for selecting a measurement technique suitable for an arbitrary pattern shape in the field of measuring a pattern dimension on a semiconductor wafer.

  In dimension measurement of a semiconductor pattern, pattern edges on an image are detected based on a high-resolution image captured by a scanning electron microscope, and dimension measurement is performed from the edge coordinates. Since the shape of the conventional measurement target pattern is basically limited to lines and holes, it has been possible to obtain high measurement reproducibility by using a measurement method that has been established.

JP 2006-302952 A

  With the miniaturization of semiconductor pattern circuits, complex shape patterns different from conventional lines and holes tend to increase. Since there are many cases in which the measurement accuracy of these patterns is lowered by the conventional measurement method, many measurement methods corresponding to the pattern shape have been developed. These measurement methods can often be applied to pattern shapes that are similar to the pattern that was developed, but it is not possible to determine the applicability of the measurement method simply by looking at the shape. We have examined the effectiveness and set up an appropriate measurement method through trial and error. However, in recent years, the diversification of pattern shapes has further progressed, and there is an increasing need to select a measurement method suitable for new shape patterns more efficiently.

  Patent Document 1 discloses a method of performing measurement processing by selecting a measurement recipe based on incidental information of image data captured by a scanning electron microscope. The incidental information is acquired image data, lot information of the image data, measurement conditions, and the like. In the measurement method, even if images are acquired by CDSEMs from different manufacturers, the pattern measurement method is the same. There is no description about a method for quantitatively capturing the pattern shape on the image. Therefore, the measurement recipe to be selected is not necessarily a method suitable for the pattern shape to be measured.

  An object of the present invention is to provide a method for selecting a measurement technique suitable for an arbitrary pattern shape in the field of measuring a pattern dimension on a semiconductor wafer.

  In particular, the present invention relates to a method and apparatus for supporting efficient selection of a measurement method according to the shape of a new pattern, and in particular, solves the above problems by the following means.

  That is, in the present invention, in order to solve the above problem, in a method for measuring the size of a semiconductor pattern image captured by an electron beam, a step of specifying a measurement target part on the pattern by a user, and a shape from the pattern measurement part A step of extracting information, a step of selecting an applicable measurement method from a measurement method library prepared in advance based on the extracted shape information, a step of measuring a dimension of the pattern by the selected method, Provided is a method for measuring a semiconductor pattern dimension, comprising a step of outputting a measured dimension.

  According to the present invention, efficiency is improved in selecting a measurement method for measuring a pattern dimension on a semiconductor wafer.

It is a figure which shows the example of the whole structure of a semiconductor pattern measuring device. It is a figure which shows an example of this measurement process step. It is a figure which shows an example of the captured image of a measurement object pattern. It is a figure which shows the measurement target pattern image after measurement box installation. It is a figure which shows an example of user input GUI. It is a figure which shows the mode of the edge detection on a measurement object pattern. It is a figure which shows an example of the signal waveform which calculates a pattern edge position. It is a figure which shows a mode that the parallelism of the left-right edge point sequence of a measurement object pattern is evaluated. It is a figure which shows a mode that the linearity of the right-and-left edge point sequence of a measurement object pattern is evaluated. It is a figure which shows a mode that the curve closeness of the left-right edge point sequence of a measurement object pattern is evaluated. It is a figure which shows an example of the total result of the acquired pattern information. It is a figure which shows an example of the selection process of an applicable measurement method. It is a figure which shows an example of the applicability determination flowchart of a measurement method. It is a figure which shows an example of the applicability determination flowchart of a measurement method. It is a figure which shows an example of pattern dimension measurement. It is a figure which shows an example of the output screen of pattern measurement. It is a figure which shows an example of the captured image of a measurement object pattern. It is a figure which shows the measurement target pattern image after measurement box installation. It is a figure which shows an example of the selection process of an applicable measurement method. It is a figure which shows a mode when the waveform projection method which is one of the measurement methods is applied. It is a figure which shows an example of the signal waveform which calculates a pattern edge position. It is a figure which shows an example of the output screen of pattern measurement.

  FIG. 1 is an overall configuration diagram of a semiconductor pattern measuring apparatus according to the present invention. The length measuring SEM 10 according to this embodiment includes a sample stage on which a measurement wafer is placed, an irradiation optical system that controls the emitted electron beam 100, and a detection system that detects secondary electrons emitted from the sample. The sample stage is composed of a load lock chamber 101 for transferring a measurement wafer and a stage table 102 for fixing the transferred wafer. The irradiation optical system includes a condenser lens 104, an objective movable diaphragm 105, an alignment coil 106, a stigma coil 107, a deflection coil 108, and an objective lens 109 on the path of the electron gun 103 and the electron beam 100. The emitted electron beam 100 passes through the condenser lens 104, the irradiation position deviation and astigmatism of the electron beam 100 are corrected by the alignment coil 106 and the stigma coil 107, the irradiation position on the sample is controlled by the deflection coil 108, and the objective lens The light is condensed by 109 and irradiated on the sample wafer on 102. The stage controller 110 controls the irradiation position of the electron beam 100 onto the wafer on the stage table 102. The detection system includes E × B 111 and a detector 112. In E × B 111, an electric field and a magnetic field are applied to the secondary electrons generated from within the sample by irradiation of the electron beam 100, thereby guiding the secondary electrons in the direction of the detector 112. The bent secondary electrons are detected by the detector 112. The detected secondary electrons are converted into a digital signal by the A / D converter 113 and stored as an image in the memory 1141 in the image processing unit 114. The image processing unit 114 extracts measurement technique data from the measurement technique library stored in the data storage device 115 using the image stored in the memory 1141 and information input from the control terminal 116, and performs pattern dimension measurement and the like. Do. In the measurement method library, one or more measurement method data is stored in a library in advance. The measurement method library records a measurement program for executing measurement, and corresponds to the measurement method. Here, the measurement program records a measurement procedure for executing measurement by the measurement method, and is synonymous with the measurement method. Here, the measurement method data is configured by information on the measurement method application requirement indicating the measurement method and the pattern conditions to which the measurement method can be applied. An example of the measurement method application information is shown in a measurement method application requirement storage table 1201 in FIG. Details of the image processing unit 114 will be described later. Reference numeral 116 denotes a control terminal of the entire apparatus having an input device such as a mouse and a keyboard and a display device such as a monitor, and controls the entire length measuring SEM 10 to adjust the irradiation condition of the electron beam 100 and the irradiation position on the wafer. Etc., and the control of the processing method of the image processing unit 114 and the output of the processed image can be performed. Note that the processing in the image processing unit 114 is controlled according to a program recorded in the recording medium 117 having a preset image processing program.

  FIG. 2 shows an example of a measurement processing flow according to this embodiment. This processing is performed using the image processing unit 114, the data storage device 115, and the control terminal 116. The image processing unit 114 includes a memory 1141 and a CPU 1142 and is controlled by the control terminal 116. First, in step 201, the CPU 1142 reads one or more images into the memory 1141. Next, in step 202, pattern information is created by extracting the user input information input from the control terminal 116 and the pattern shape of the image read by the CPU 1142, and stores it in the memory 1141. Here, the pattern information includes user input information and pattern shape information, and is shown as pattern information 1101 in FIG. 12 as an example. In step 203, the measurement method application requirement information stored in the data storage device 115 is read into the memory 1141, the CPU 1142 collates the pattern information with the measurement method application requirement information, and the determination result on whether the measurement method is applicable is stored in the memory. It is stored in 1141. An example of the determination result is shown in the applicability determination result storage table 1202 in FIG. In step 204, the measurement method determined to be applicable is read from the library 115 into the memory 1141, and the CPU 1142 measures the dimensions. Finally, in step 205, the obtained result is transmitted to the control terminal 116, and the measurement result is output. Hereinafter, the specific contents of each step will be described in detail.

  First, an example of a captured image read in step 201 is shown in 301. A line pattern in which left and right edges meander periodically as shown in FIG. 3 is imaged on the captured image 301, and this pattern is taken as a measurement target pattern 302. Here, the line width of the locally narrowed portion of the measurement target pattern 302 is set as the measurement target portion 303.

  Next, the user input step 202-a for inputting user input information will be described. The user input information is information based on a user instruction and includes information that is difficult or impossible to obtain by image processing. For example, the position or measurement value of the measurement target unit 303 in the image is an average value or For example, the minimum value. Here, the measurement position and measurement range on the captured image 301, the measurement item, the measurement target, and the repeatability are described as examples of user input information. However, the user input information is not limited to this, and these items are always input. It's not necessary. FIG. 4 is a captured image 301 displayed on the display device of the control terminal 116 and shows a state in which the measurement box 401 is designated. The measurement box 401 is a measurement region that includes the measurement target unit 303, and the position and range of the measurement box 401 indicate the position in the captured image 301 shown on the display device of the control terminal 116 using an input device such as a mouse. This is specified. Here, although the shape of the measurement box 401 is a rectangle, according to the shape of the measurement target pattern 302 such as a circular shape or a semicircular shape, for example, by indicating with the icon indicated by 402 in FIG. A shape can also be selected. Furthermore, here, as shown in FIG. 4, it is assumed that the measurement box 401 is installed for each left and right edge, and for convenience, the one installed on the left side is referred to as 401L and the one installed on the right side is referred to as 401R. It may be installed as a box 401. In addition, the measurement box 401 can be installed by being rotated. For example, when the measurement box 401 is installed by being rotated 90 degrees, the measurement boxes 401L and 401R are arranged vertically. This is used when the measurement target portion is the distance between the upper and lower edges.

  FIG. 5 is an example of the GUI 501 for inputting user input information to be displayed on the display device of the control terminal 116. In the measurement item 502, not only the dimension of the pattern but also roughness such as LER (Line Edge Roughness) can be designated. The measurement target 503 is an item for designating the type of shape of the measurement target unit 303. Here, there are a wide variety of input pattern shapes, but the measurement target portion 303 can be broadly divided into either a pattern width, a gap, or a diameter regardless of the shape. The method is different. Therefore, a line is designated for the pattern width, a space is designated for the gap, and a hole is designated for the diameter. Although it is specified on the GUI 501 here, the measurement target 503 may be automatically specified by designating the shape of the measurement box 401 with the measurement box shape specification icon 402 or the like. Further, the repeatability 504 designates whether or not the pattern within the same image has repeatability. Here, since the measurement target part 303 is the line width of a locally narrowed part on the line, the measurement item 502 specifies the minimum width, the measurement target 503 specifies the line, and this locally narrow part is Since it repeatedly appears on the image 301, the repeatability 504 indicates a state where “Yes” is selected.

Next, the shape information acquisition step 202-b will be described. The shape information is information obtained by quantitatively evaluating the shape of the pattern measurement target portion that can be extracted from the captured image 301. For example, the shape information includes whether the measurement target portion is a straight line. Extraction may be performed automatically. Here, a method using pattern edges will be described as an example of the extraction method. As shown in FIG. 6, pattern shape information in the measurement boxes 401L and R is extracted. Here, description will be made using the measurement box 401L. For convenience, the horizontal direction of the captured image 301 is assumed to be the i direction and the vertical direction is assumed to be the j direction. First, a signal waveform 701 in the i direction shown in FIG. 7 is acquired at a certain coordinate j ₀ in the j direction in the measurement box 401L. The acquisition range is from the left end to the right end of the measurement box 401L. In FIG. 7, since the pixel value increases near the edge on the pattern, the signal waveform 701 has a shape in which a signal peak occurs near the edge. Next, an edge signal value 703 is calculated based on an edge threshold value 702 specified in advance. The edge threshold value 702 is a parameter that can be arbitrarily specified by the user, and represents a signal intensity ratio when the maximum signal value 704 on the signal waveform 701 is 100% and the minimum signal value 705 is 0%. The edge threshold value 702 may be specified at any timing, but here, it is assumed that it is input together with the user input information in the user input step 202-a. The calculation formula for obtaining the edge signal value 703 is shown below.
(Equation 1)

  Finally, coordinates that coincide with the edge signal value 703 are detected as the edge position 706 from the signal waveform 701. This processing is performed at each coordinate in the j direction in the measurement box 401L, and an edge point sequence 601L in the measurement box 401 is calculated. In this example, the acquisition direction of the signal waveform 701 is the i direction depending on the pattern shape, but the acquisition direction may be changed by the user according to the shape of the measurement target unit 303. This processing is performed individually in the left and right measurement boxes 401L and R, and edge point sequences 601L and R are detected. Here, it is assumed that the number of detected edge points in both measurement boxes 401L and R is the same, and the number is N points.

Pattern shape information is extracted using the obtained edge point sequences 601L and R. As an example of the shape information extraction method, the evaluation of the parallelism, linearity, and curve approximation of the edge point sequences 601L and R will be described with reference to FIGS. First, a method for determining parallelism will be described with reference to FIG. Parallelism is a result of determining whether or not the left edge point sequence 601L and the right edge point sequence 601R are parallel. First, edge numbers are assigned respectively from the upper end to the lower end of the left edge point sequence 601L and the right edge point sequence 601R, and the distances between edges (801-1 to 801-N) between the same edge numbers are calculated. Next, the variance of the edge distances 801-1 to 801-N is obtained. If the left and right edge point sequences 601L and R are close to parallel, the dispersion of the distance between the edges becomes small. Therefore, the dispersion of the distance between the edges becomes an index value indicating the parallelism of the left and right edge point sequences 601L and R. Therefore, a parallel threshold value for determining the presence or absence of parallelism is set in advance, and when the variance of the distance between edges is equal to or less than the parallel threshold value, it is determined that there is parallelism, and when it exceeds, it is determined that there is no parallelism. The calculation formula is shown below.
if (Edge-to-edge distance dispersion ≤ Parallel threshold) Parallelism
else No parallelism Next, a method for determining linearity will be described with reference to FIG. The linearity is an evaluation result of the possibility of linear approximation of each of the left and right edge point sequences 601L and R. Here, the left edge point sequence 601L will be described as an example, but the same processing is performed for the right edge point sequence 601R. First, an approximate straight line 901 is fitted to the left edge point sequence 601L. As a fitting method, a generally known method such as a least square method that minimizes the distance between the edge point sequence 601L and the approximate straight line 901 is used. Next, an average distance on the approximate straight line 901 and the edge point sequence 601L is obtained as a linear residual. The straight line residual is an index value indicating the fitting accuracy of the approximate straight line 901 with respect to the edge point sequence 601L, and the distance from the upper end edge point to the lower end edge point of the approximate straight line 901 edge point sequence 601L is sequentially set to l ₁ , l ₂ , ..., _LN , the linear residual is calculated by the following equation.
(Equation 2)

A linear threshold value for determining the presence / absence of linearity is set in advance. When the linear residual is equal to or smaller than the linear threshold value, linearity is determined, and when it is higher, it is determined that there is no linearity. The calculation formula is shown below.
if (Linear residual ≤ Linear threshold) Linearity
else No linearity Finally, a method for determining curve approximation will be described with reference to FIG. Curve approximation is the result of evaluating whether each of the left and right edge point sequences 601L and R can be approximated to a known curve. This evaluation will also be described by taking the left edge point sequence 601L as an example. There are various types of known curves. Here, a parabola which is one of the most common curves will be described. First, the approximate curve 1001 is fitted to the left edge point sequence 601L. Next, an average distance between the approximate curve 1001 and the edge point sequence 601L is obtained as a curve residual. The curve residual is an index value indicating the fitting accuracy with respect to the edge point sequence 601L of the approximate curve 1001, and the distance from the upper end edge point to the lower end edge point of the approximate curve 1001 and the edge point sequence 601L is sequentially set to r ₁ , r _2. ,..., R _N , the curve residual is calculated by the following equation.
(Equation 3)

A curve threshold value for determining the presence or absence of curve approximation is set in advance. If the curve residual is equal to or less than the curve threshold value, curve approximation is determined, and if it exceeds, it is determined that there is no curve approximation. The calculation formula is shown below.
if (curve residual <curve threshold) with curve approximation
else There is no curve approximation. Here, a parabola is used as an example of an approximation curve. However, similar processing is possible for any curve such as an elliptic curve or a curve represented by a higher-order equation. Curve approximation may be evaluated.

  The above are examples of evaluation methods for parallelism, linearity, and curve approximation, but the present invention is not limited to these methods, and pattern shape information may naturally be extracted using another evaluation method. By using these results for measurement method selection, applicable measurement methods can be selected more quantitatively.

  Next, step 203 which is a measurement method selection process will be described. In this step, the pattern information acquired in step 202 and the measurement method application requirement information in the measurement method library stored in the data storage device 115 are collated, and the measurement method data satisfying the measurement method application requirement information. Select the measurement method. First, the user input information and the pattern shape information acquired by the user input information acquisition 202-a and the pattern shape information extraction 202-b are tabulated as pattern information and stored in the memory 1141. FIG. 11 shows a pattern information storage table 1101 that stores pattern information. In this embodiment, the information to be selected from a plurality of options such as the measurement item and measurement target is “◯” for the selected information, blank for the information not selected, repeatability, parallelism, linearity, curve approximation Information for identifying the presence or absence of pattern characteristics such as sex is expressed as “present” or “absent”. In the example of FIG. 11, “minimum width” is selected as the measurement item of the user input information, “line” is selected as the measurement target, “existence” is selected as the repeatability, and parallelism and linearity are “none” as the pattern shape information. It is assumed that the curve approximation is determined to be “present”.

  When the pattern information storage table 1101 is saved, the measurement technique application requirement storage table 1201 in which information on the measurement technique application requirements is stored is extracted from the data storage device 115. The columns of the table 1201 are the same items as the pattern information storage table 1101, and the contents of the table are configured by applicability information of each item for each measurement technique data. The applicability information indicates “applicable” for the applicable condition, “impossible” for the inapplicable condition, and “required” for the indispensable condition. In other words, the measurement item in the pattern information storage table 1101, the item selected for the measurement target, and the item for which “Yes” is selected from the repeatability, parallelism, linearity, and curve approximation property are “possible” in the table 1201. "Or" necessary ", it is applicable to the item, and if it is" impossible ", it is not applicable. Conversely, items that are not selected in the measurement items in the table 1101 or items to be measured, and items that are selected as “None” among repeatability, parallelism, linearity, and curve approximation are “OK” in the table 1201. Or, if it is “impossible”, it is applicable for the item, and if it is “necessary”, it is not applicable. Therefore, items that are “impossible” or “necessary” in the table 1201 are actually determination items for applicability.

  The method (1) described in the measurement method application requirement storage table 1201 shown in FIG. 12 will be specifically described as an example. In method (1), the average, minimum, and maximum values of the measurement items are all “measurable” because they can be measured, while LER (Line Edge Roughness), which is an index value indicating the roughness of the pattern, cannot be measured. Because it is, it is “impossible”. Similarly, the measurement target can be applied to all lines, spaces, and holes, so it can be measured regardless of the parallelism, linearity, and curve approximation. Yes. In addition, the repeatability is “necessary” because measurement is impossible without it. Similarly, after the method (2), information on applicability to the measurement items, measurement objects, information on the shape, etc. of the method is stored. The items in the measurement technique application requirement storage table 1201 and the pattern information storage table 1101 are collated to determine whether the measurement technique can be applied. The methods (1) and (3) will be described as an example with reference to the flowcharts of FIGS. In method (1), since all the patterns to be measured, that is, lines, spaces, holes, and the parallelism, linearity, and curve approximation of the pattern shape information are “possible”, the determination item for applicability is measurement. Item and repeatability. First, in step 1301, since the measurement item in the pattern information storage table 1101 is “minimum width”, the process proceeds to step 1302, and in step 1302, the repeatability is “present”. ) Is “applicable”. On the other hand, in the method (3), since the repeatability and the curve approximation are “possible”, the determination items for applicability are the measurement item, the measurement object, the presence / absence of parallelism, and the presence / absence of linearity. Here, since “minimum value” is selected as the measurement item in the pattern information storage table 1101, “applicable” is determined in step 1401, and the method (3) cannot be applied without passing through steps 1402 to 1404. Become. Similarly, the applicability of each measurement method is determined, and a table 1202 storing the applicability determination result is generated. When the applicability determination result storage table 1202 is generated, one or more measurement techniques applicable in the table 1202 are read from the measurement technique library stored in the data storage device 115 to the memory 1141. The above is the content of the measurement method selection processing step 203.

  Next, the dimension measurement step 204 will be described. In this step, the applicability determination result storage table 1202 in which the adaptability determination result is stored is read, and the dimension of the measurement target unit 303 is measured by each measurement method determined to be applicable. Here, an example of processing contents when measured by the parabolic approximation method, which is one of them, will be described with reference to FIG. In this embodiment, the left parabola 1501L and the right parabola 1501R are approximately fitted to the left and right edge point sequences 601L and R of the measurement target unit 303, respectively. Next, the right end coordinates on the left parabola 1501L and the left end coordinates on the right parabola 1501R are measured as the left and right measurement points 1502L and R, and the distance between both measurement points is measured as the pattern dimension 1503. Also, the left and right fitting accuracy of the left and right parabolas 1501L and R are calculated and transmitted to the control terminal 116. Here, the fitting accuracy is a curve residual 1002 between the left and right parabolas 1501L and R and the edge point sequence 601L and R. In addition, when measuring time by a parabolic approximation method or when measuring a plurality of measurement target parts, reproducibility and the like are also calculated using a plurality of measurement values, and transmitted to the control terminal 116 as measurement results. Here, the reproducibility refers to variation in dimension values when dimension measurement is performed a plurality of times, and can be calculated, for example, by calculating the variance of a plurality of measurement values. In this way, it is possible to check the reliability of measurement by calculating and recording the variance of a plurality of measurements.

  Finally, in the measurement result output 205, the measurement result obtained by the dimension measurement 204 is stored in the data recording device in the control terminal 116, read from the recording device, and displayed on the monitor of the control terminal 116. The display destination is not limited to the page or the monitor. Here, an example of the output screen 1601 is shown in FIG. 16 as being displayed on the monitor. On the output screen 1601, in addition to the image name 1602 of the captured image 301, measurement results for each measurement method such as measurement values, measurement reproducibility, and measurement time as shown in the measurement results 1603 are shown. Information other than the measurement result 1603 may be displayed. A measurement position 1604 on the measurement target portion 303 is displayed. Thereby, the user can see the output screen 1601, compare the measurement results of the respective measurement methods, and select an appropriate measurement method.

  A series of processing procedures for the captured image 301 described above is stored in the recording medium 117, and the recording medium 117 is read according to an instruction from the control terminal 116 and processing is performed.

  By using the present invention, the user can easily select a method suitable for the user's purpose from various measurement methods. Further, this method can be applied not only to a conventional general shape pattern such as a line or a hole but also to a pattern having a complicated shape such as a measurement target pattern 302. By outputting the measurement value and measurement accuracy of each applicable measurement method, the user can compare the measurement method quantitatively and select the method.

  In the present embodiment, an example in which the present process is applied to a shape pattern different from the first embodiment will be described. FIG. 17 shows a measurement target pattern 1702 on the captured image 1701 read into the memory 1141 in step 201. The measurement target pattern 1702 has a shape in which four vertices of a square are recessed 90 degrees inward, and is repeatedly formed on the captured image 1701. In addition, the measurement target unit 1703 of this pattern is a distance between two measurement target patterns 1702 adjacent in the vertical direction.

  Next, the user input step 202-a will be described. FIG. 18 shows a state in which the measurement box 1801 is designated from the captured image 1701 displayed on the display device of the control terminal 116 and the measurement box icon 402. Here, it is assumed that “rectangular” is selected from the measurement box icon 402, and the selected measurement box 1801 is rotated 90 degrees and is individually installed on the upper and lower edges of the measurement target unit 1703, respectively. The measurement box is 1801U, and the measurement box on the lower edge is 1801D. Furthermore, since the measurement target unit 1703 is the distance between two adjacent patterns 1702, the user input GUI 501 uses “average value” for the measurement item 502, “space” for the measurement target 503, and “ It is assumed that “Yes” is selected.

  The shape information extraction method in the next shape information acquisition step 202-b is the same as that in the first embodiment. As the extraction result, it is determined that the parallelism and linearity are “present” and the curve approximation is “absent”, and the result is stored in the pattern information storage table 1901 together with the user input information input in 202-a. It is stored in 1141.

  Furthermore, step 203 which is a measurement method selection process will be described. In this step, as in the first embodiment, the pattern information in the pattern information storage table acquired in step 202 is collated with the measurement method application requirement information in the measurement method application requirement storage table 1201, and the measurement method application requirement information is obtained. Select the measurement method of the measurement method data to be satisfied. Similar to the first embodiment, the methods (1) and (3) of the measurement method application requirement storage table 1201 will be described as an example with reference to FIGS. In method (1), since the measurement item is “average value” in step 1301, the process proceeds to step 1302, and in step 1302, repeatability is “Yes”. Possible ". In method (3), since the measurement item is “average value” in step 1401, the process proceeds to step 1402. In step 1402, the measurement object is “space”, and thus the process proceeds to step 1403. In 1403, since the parallelism is “Yes”, the process proceeds to Step 1404. In 1401, the linearity is also “Yes”, so it is determined as “Applicable”. Similarly, applicability of each measurement method is determined, and an applicability determination result storage table 1902 is generated. Measurement methods that are “applicable” in this table 1902 are read out from the measurement method library stored in the data storage device 115 to the memory 114.

  The dimension measurement step 204 will be described. In this step, the measurement method read to the memory 114 is applied. Here, a case where the waveform projection method is applied as one of them will be described. In this method, for example, signals in the measurement box 1801U and 1801D regions are projected in specific directions, edge positions are obtained from the projected waveform 2001, and the distance between the edges is measured. Here, it is assumed that the specific projection direction is determined when the measurement box 1801 is installed from the measurement box icon 402. That is, as described in the first embodiment, since the measurement box is rotatable, the rotation angle is made to coincide with the projection direction. For example, when the rotation angle is 0 degree, projection is performed in the j direction (from the top to the bottom on the image), and when the rotation angle is 90 degrees, projection is performed in the i direction (from the right to the left on the image). Here, since the measurement box 1801 is rotated 90 degrees and installed, the projection direction is the i direction. Further, as a method of detecting an edge from the projection waveform 2001, for example, as shown in FIG. 7 of the first embodiment, an edge threshold 2101 is set in advance, and the maximum signal value and the minimum signal value of the projection waveform 2001 are calculated (equation 1 ) Is used to obtain the edge signal value 2102, and the coordinates corresponding to 2102 are regarded as the edge position 2103. Here, the projected waveform 2001 is divided into two, and for convenience, the upper waveform in the drawing of the projected waveform 2001 is called 2001U, and the lower waveform is called 2001D. The edge detection described above is performed for each of the projection waveforms 2001U and 2001D, and the distance between the edges is set to a dimension value 2003.

  Finally, the measurement result output 205 outputs the measurement result obtained by the dimension measurement 204 to the monitor of the control terminal 116 at the same time as it is stored in the data recording device in the control terminal 116. The display destination may be other than the monitor as in the first embodiment. An example of the output screen 2201 is shown in FIG. On the output screen 2201, as in the first embodiment, the measurement result 2203, the measurement position 2204 on the measurement target unit 1703, and the like are displayed in addition to the image name 2202 of the captured image 1701. As a result, the user can compare measurement results of measurement methods applicable to the measurement target pattern 1702 based on the output screen 2201, and each user can select an effective measurement method for an arbitrary shape pattern. Become.

  As described above, this embodiment can be applied not only to an arbitrary shape pattern different from that of the first embodiment but also to a conventional general shape pattern such as a line or a hole. It is possible to easily select a technique according to the user's purpose from the technique. In addition, by outputting the measurement values and measurement accuracy of each applicable measurement method, the user can compare the measurement methods quantitatively and select the method.

DESCRIPTION OF SYMBOLS 10 ... SEM apparatus main body, 100 ... Electron beam, 101 ... Load lock chamber, 102 ... Stage stand, 103 ... Electron gun, 104 ... Condenser lens, 105 ... Objective movable diaphragm, 106 ... Alignment coil, 107 ... Stigma coil, 108 ... Deflection coil, 109 ... objective lens, 110 ... stage controller, 111 ... ExB, 112 ... detector, 113 ... A / D converter, 114 ... image processing unit, 115 ... data storage device, 116 ... control terminal, 117 ... Recording medium 201 ... Image reading step 202 ... Pattern shape acquisition step 203 ... Measurement method selection processing step 204 ... Dimension measurement step 205 ... Measurement result output step 301 ... Captured image 302 ... Measurement target pattern 303 ... Measurement target part 401 ... Measurement box 402 ... Measurement box designation Icon, 501 ... User input GUI, 502 ... Measurement item, 503 ... Measurement target, 504 ... Repeatability, 601 ... Edge point sequence, 701 ... Signal waveform, 702 ... Edge threshold, 703 ... Edge signal value, 704 ... Maximum Signal value, 705 ... Minimum signal value, 706 ... Edge position, 801-1 to 801-N ... Distance between edges, 901 ... Approximate line, 902 ... Linear residual, 1001 ... Approximate curve, 1002 ... Curve residual, 1101 ... Pattern information storage table, 1201 ... Measurement method application requirement storage table, 1202 ... Applicability determination result storage table, 1501 ... Parabola, 1502 ... Measurement points, 1503 ... Pattern dimensions, 1601 ... Output screen, 1602 ... Image name, 1603 ... Measurement Result, 1604 ... Measurement position, 1701 ... Captured image, 1702 ... Measurement target pattern, 1703 ... Measurement target part, 1 801 ... Measurement box, 1901 ... Pattern information storage table, 1902 ... Applicability determination result storage table, 2001 ... Projected waveform, 2101 ... Edge threshold value, 2102 ... Edge signal value, 2103 ... Edge position, 2201 ... Output screen, 2202 ... Image name 2203 ... Measurement result 2204 ... Measurement position

Claims (20)

A method for measuring an image of a semiconductor pattern imaged by an electron beam,
We have a measurement method library that stores multiple measurement methods in advance.
Designating a measurement target portion from an image of a captured semiconductor pattern;
Extracting shape information of the specified measurement target part;
Selecting the measurement technique for measuring the measurement target portion using the shape information from the measurement technique library;
Measuring the dimensions of the measurement target portion using the selected measurement method;
Outputting the measured dimensions;
A semiconductor pattern measurement method comprising: The semiconductor pattern measurement method according to claim 1,
The measurement method library stores the measurement method and measurement method application requirements as measurement method data. The semiconductor pattern measurement method according to claim 1,
Inputting user input information of the measurement target portion specified in the step of specifying the measurement target portion;
A semiconductor pattern measurement method comprising: The semiconductor pattern measurement method according to claim 3,
The step of selecting the measurement method includes
A semiconductor pattern measurement method, wherein the measurement method for measuring the measurement target portion is selected from the measurement method library using the input user input information and the extracted shape information. The semiconductor pattern measurement method according to claim 2,
Inputting user input information of the measurement target part specified in the step of specifying the measurement target part;
Using the input user input information and the extracted shape information to create pattern information of the measurement target part;
A semiconductor pattern measurement method comprising: The semiconductor pattern measurement method according to claim 5,
The step of selecting the measurement method includes
A semiconductor pattern measurement method, comprising: comparing the measurement method data with the pattern information and selecting the measurement method for measuring the measurement target portion from the measurement method library. The semiconductor pattern measurement method according to claim 3,
The user input information includes at least one of measurement position information, measurement range information, measurement item information, measurement target information, and extracted shape information. Semiconductor pattern measurement method. The semiconductor pattern measurement method according to claim 7,
The measurement item information includes at least one of an average value, a maximum value, and a minimum value. The semiconductor pattern measurement method according to claim 7,
The measurement target information includes at least one of a line, a space, and a hole. The semiconductor pattern measurement method according to claim 7,
The shape information is composed of at least one of pattern parallelism, linearity, and curve approximation. The semiconductor pattern measurement method according to claim 1,
In the step of measuring the dimensions,
A method for measuring a semiconductor pattern, comprising: measuring a dimension of a measurement target portion using the selected measurement technique; and obtaining measurement reproducibility from the measurement of the dimension. The semiconductor pattern measurement method according to claim 11,
In the step of outputting the dimensions,
A semiconductor pattern measuring method, wherein the measured dimensions and the obtained measurement reproducibility are output. The semiconductor pattern measurement method according to claim 1,
Edges to be detected based on the edge point sequence obtained by detecting the edge of the specified measurement target portion using the shape information extracted in the extracting step, obtaining an edge point sequence of the detected edge A semiconductor pattern measurement method comprising a point sequence extraction step. A semiconductor pattern measuring device for measuring a semiconductor pattern formed on a wafer,
A measurement method storage means for storing a measurement method library having a plurality of pattern measurement methods prepared in advance;
Captured image acquisition means for acquiring an image of a semiconductor pattern by an electron beam;
Means for designating a measurement target portion from an image of a semiconductor pattern;
Means for extracting shape information of the specified measurement target part;
Means for selecting a measurement technique for measuring the measurement target portion from a measurement technique library stored in the measurement technique storage means using the extracted shape information;
Means for measuring the dimensions of the measurement object using the selected measurement method;
Means for outputting the measured dimensions;
A semiconductor pattern measuring apparatus comprising: The semiconductor pattern measurement device according to claim 14,
The said memory | storage means memorize | stores the said measurement method and measurement method application requirements as measurement method data, The semiconductor pattern measuring device characterized by the above-mentioned. The semiconductor pattern measurement device according to claim 14,
Means for inputting user input information of the measurement target portion specified in the means for specifying the measurement target portion;
A semiconductor pattern measuring apparatus comprising: The semiconductor pattern measurement apparatus according to claim 16,
In the means for selecting a measurement method for measuring the measurement target part,
A semiconductor pattern measurement apparatus, wherein the measurement technique for measuring the measurement target portion is selected from the measurement technique library using the input user input information and the extracted shape information. The semiconductor pattern measurement apparatus according to claim 15,
Means for inputting user input information of the measurement target portion specified in the means for specifying the measurement target portion;
The means for extracting the shape information is:
A semiconductor pattern measurement apparatus, wherein pattern information of the measurement target part is created using the input user input information and the extracted shape information. The semiconductor pattern measurement apparatus according to claim 18,
In a means for selecting a measurement technique for measuring the measurement target part,
A semiconductor pattern measurement apparatus that compares the measurement technique data with the pattern information and selects the measurement technique for measuring the measurement target portion from the measurement technique library. The semiconductor pattern measuring device according to claim 14,
Using the shape information extracted in the means for extracting the shape information of the specified measurement target portion, the edge of the specified measurement target portion is detected, and an edge point sequence of the detected edge is obtained, A semiconductor pattern measuring device comprising an edge point sequence extracting means for extracting based on the edge point sequence.

Images