JP6281019B2 - Electron beam pattern inspection system - Google Patents

Description

  The present invention relates to an electron beam pattern inspection apparatus for semiconductor wafers, and more particularly to an apparatus for inspecting and reviewing hole patterns and groove patterns having a high aspect ratio.

  As described in Patent Document 1, as the first background art of the present invention, semiconductor device manufacturer companies have developed more detailed defects, In order to detect defects in high-aspect processes, we have introduced an electron beam wafer inspection system that has higher resolution and deeper focus than optical inspection equipment. A basic configuration of the electron beam type wafer inspection apparatus is shown in FIG.

  When the electron beam 262 is irradiated onto the sample wafer 261, secondary electrons 263 are generated from the wafer 261. The detector 264 detects some of the secondary electrons 263 as a bright part and a dark part, respectively, and acquires an image by scanning the electron beam on the wafer. Pattern defects are detected by image comparison of adjacent cells or the same circuit portion obtained from the die (usually the former is called cell comparison, the latter is called die comparison), and defect position coordinates, defect map, defect category, etc. Information is output.

  Further, as a second background art of the present invention, Patent Document 2 discloses a technique for detecting a minute open defect or a short defect in a fine pattern by irradiating the substrate on which the pattern is formed with an electron beam. The generated low-energy secondary electrons and high-energy reflected electrons are detected, and first and second SEM (Scanning Electron Microscope) images are generated from the respective electrons, and the second SEM image is generated. Pattern contour data is obtained by extracting a contour from the image, and applying the contour data to the first SEM image to determine an inspection region and performing binarization processing on the inspection region Techniques for detecting defects are disclosed.

  On the other hand, as a new inspection need, in the memory field, development of a technology (3D-NAND) for stacking memory cell arrays into a three-dimensional structure is accelerating, and deep holes and deep groove patterns with ultra-high aspect ratio (> 50) There is a need to inspect for defects in holes and grooves (such as pattern deformation and depth defects at the back of the grooves) with high accuracy.

JP 2002-250707 A JP 2011-174858 A

  In response to the above new needs, the techniques disclosed in Patent Documents 1 and 2 have the following problems.

  The first problem is that an electron beam used in the background art is an electron beam having an acceleration voltage of several kilovolts, and is a secondary electron image (referred to as a first electron beam image in Patent Document 2), a reflected electron image ( In any of Patent Documents 2 (denoted as a second electron beam image), it is a problem that it is difficult to reveal defects in deep holes and deep grooves having an aspect ratio of 20 or more, which is a new need.

  The second problem is mainly related to Patent Document 1. However, when inspection is performed by cell comparison or die comparison, the local fluctuation (edge roughness) of the edge position within the tolerance is erroneously detected as a defect. It is. This situation is shown in FIG. Assume that 280 and 282 are SEM images of hole patterns acquired by different dies, respectively. Although the edge roughness is represented by triangular protrusions (281, 283), since the images are images of holes different from each other, each edge roughness has a different position and size. When a difference image 286 is generated in the difference image generation step (285) using these images 280 and 282, and binarization processing (287) is performed using a threshold value for the difference image 286, an image 288 is obtained. Differences in roughness are also detected as defects. In order not to detect the difference in edge roughness as a defect, it is difficult to limit the inspection area (do not inspect the edge part) or reduce the defect detection sensitivity and to perform high-precision inspection. There was a problem.

  In addition, both Patent Documents 1 and 2 describe a configuration for detecting secondary electrons and reflected electrons (described as backscattered electrons in Patent Document 1) generated from a sample irradiated with a primary electron beam. . Information on the surface of the sample is included in a signal obtained by detecting secondary electrons, and surface defects can be detected from the secondary electron image. On the other hand, deep hole and deep groove information is generally included in the signals that detect reflected electrons, and reflected electron images detect defects in deep holes and deep grooves rather than secondary electron images. Suitable for However, neither of Patent Documents 1 and 2 describes the detection of a deep hole or deep groove defect from a reflected electron image.

  The present invention solves the above-described problems and can reveal defects in deep holes and deep grooves having an aspect ratio of 20 or more, and can perform high-precision inspection without being affected by edge roughness. An electron beam pattern inspection apparatus is provided.

  In order to solve the above problems, in the present invention, an electron beam pattern inspection apparatus includes an electron beam irradiation unit that irradiates a sample on which a pattern is formed with a converged electron beam, and an electron that is converged by the electron beam irradiation unit. A backscattered electron detector for detecting backscattered electrons having a relatively high energy generated from the sample irradiated with the beam, and a secondary having a relatively low energy generated from the sample irradiated with the electron beam focused by the electron beam irradiator. A secondary electron detector that detects electrons, a reflected electron image generator that generates a reflected electron image from a signal obtained by detecting reflected electrons at the reflected electron detector, and a secondary electron that is detected by the secondary electron detector The secondary electron image generation unit that generates a secondary electron image from the signal obtained in this way, the reflected electron image generated by the reflected electron image generation unit, and the secondary electron image generated by the secondary electron image generation unit are processed. To check for sample defects. And a calculation unit that includes an inspection region extraction unit that extracts an inspection region from the secondary electron image, and a region corresponding to the inspection region extracted by the inspection region extraction unit using the reflected electron image. And a defect detection unit configured to inspect the set area and detect a defect.

  In order to solve the above problems, in the present invention, in a pattern inspection method using an electron beam pattern inspection apparatus, a focused electron beam is irradiated to a sample on which a pattern is formed, scanned, and converged. A reflected electron generated from a sample scanned by scanning with an electron beam is detected to generate a reflected electron image, and a relatively high energy generated from a scanned sample irradiated by a focused electron beam. A secondary electron image is generated by detecting a secondary electron with a low level, an inspection region is extracted from the generated secondary electron image, and a corresponding reflected inspection region is extracted from the secondary electron image using the generated reflected electron image The area was inspected to detect defects.

  According to the present invention, a BSE image obtained by electron beam irradiation at a high acceleration voltage is used for inspection. According to this configuration, defects in deep holes and deep grooves can be made obvious, so defects in deep holes and deep grooves that can not be detected by electron beam irradiation with an acceleration voltage of several kilovolts can be detected. Become. Furthermore, since SE images acquired at the same location with the same edge roughness are used as a reference, there is a problem that the edge roughness, which is a problem in the conventional cell comparison method and die comparison method, is erroneously detected as a defect. No longer.

  That is, according to the present invention, defects in deep holes and deep grooves having an aspect ratio of 20 or more can be made visible, and high-precision inspection can be performed without being affected by edge roughness. .

1 is a block diagram showing a schematic configuration of an electron beam pattern inspection apparatus according to a first embodiment of the present invention. It is a block diagram which shows the structure of the comparison calculating part of the electron beam type pattern inspection apparatus which concerns on 1st Example of this invention. It is a principle diagram of a conventional electron beam type wafer inspection apparatus. It is a flowchart explaining the problem of the conventional defect determination system. It is a figure explaining the difference between this invention and background art. It is a figure explaining the difference between this invention and background art. It is sectional drawing of the sample explaining the principle of this invention. It is sectional drawing of the sample explaining the principle of this invention. It is sectional drawing of the sample explaining the principle of this invention. It is sectional drawing of the sample explaining the principle of this invention. 6B is a graph showing the relationship between the electron beam irradiation position from the center of the hole pattern and the secondary electron detection signal intensity when the pattern of FIGS. 6A and 6B is scanned with an electron beam. 6B is a graph showing the relationship between the electron beam irradiation position from the center of the hole pattern and the reflected electron detection signal intensity when the pattern of FIGS. 6A and 6B is scanned with an electron beam. It is a graph which shows the relationship between an acceleration voltage and a reflected-electron detection signal intensity | strength by making the depth of a hole into a parameter. It is sectional drawing of the pattern which took the data of FIG. 7A. It is a graph which shows distribution of the reflection electron which generate | occur | produces when an electron beam is irradiated to the pattern of FIG. 7B. It is a flowchart of the process in the comparison calculating part which concerns on 1st Example of this invention. It is a flowchart of the process in the comparison calculating part at the time of applying the 1st Example of this invention to the inspection of another hole pattern defect. It is a flowchart of the process in the comparison calculating part at the time of applying a 1st Example of this invention to the test | inspection of a groove pattern defect. It is a block diagram which shows the structure of the comparison calculating part of the electron beam type pattern inspection apparatus which concerns on 2nd Example of this invention. It is a flowchart of the process in the comparison calculating part which concerns on 2nd Example of this invention. It is a block diagram which shows the structure of the comparison calculating part of the electron beam type pattern inspection apparatus which concerns on 3rd Example of this invention. It is a flowchart of the process in the comparison calculating part which concerns on 3rd Example of this invention. It is a block diagram which shows the structure of the comparison calculating part of the electron beam type pattern inspection apparatus which concerns on 4th Example of this invention. It is a flowchart of the process in the comparison calculating part which concerns on 4th Example of this invention. It is a graph showing the brightness frequency distribution of the BSE image which concerns on 4th Example of this invention. It is a front view of the screen explaining the user interface according to the fifth embodiment of the present invention. It is a block diagram which shows the schematic structure of the electron beam type pattern inspection apparatus which concerns on the 6th Example of this invention. It is explanatory drawing of the normal pattern which concerns on 7th Example of this invention. It is explanatory drawing of the bending defect which concerns on 7th Example of this invention. It is a flowchart of the process in the comparison calculating part which concerns on 7th Example of this invention. It is explanatory drawing of the pattern feature-value based on 7th Example of this invention. It is explanatory drawing of another pattern feature-value based on 7th Example of this invention. It is a flowchart of the process in the comparison calculating part which concerns on 8th Example of this invention. It is explanatory drawing of the edge detection method which concerns on 8th Example of this invention. It is explanatory drawing of the image composition method which concerns on 9th Example of this invention. It is a test result display screen based on 9th Example of this invention. It is a flowchart of the process in the comparison calculating part which concerns on 10th Example of this invention.

The present invention provides an electron beam pattern inspection apparatus that irradiates a sample having a deep hole or deep groove having an aspect ratio of 20 or more with a focused electron beam, and generates reflected electrons with relatively high energy generated from the sample. Secondary electrons with relatively low energy are simultaneously detected by a separate detector to generate a backscattered electron image and a secondary electron image. By utilizing the characteristic that defects appear only in the backscattered electron image, A defect is detected by comparing secondary electron images.
Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1A is a diagram showing a basic configuration of an electron beam pattern inspection apparatus 100 to which the present invention is applied. The electron beam pattern inspection apparatus 100 includes an electron gun 101 that emits a primary electron beam 102, a condenser lens 103 that converges the primary electron beam 102, a polarizing lens 104 that deflects the primary electron beam 102, and an objective that converges the primary electron beam 102. A lens 105, a stage 108 on which a sample 200 is mounted and movable in a plane, an annular scintillator 106 for detecting reflected electrons 110 generated from the sample 200 irradiated with the primary electron beam 102, and an optical signal output from the annular scintillator 106 An optical fiber 111 for transmitting the light, a photomultiplier tube 112 for inputting an optical signal sent from the optical fiber 111, a signal output from the photomultiplier tube 112, and a back scattered electron (BSE). BSE image generator 113 for generating an image, primary electron beam 1 The E × B deflector 107 that changes the trajectory of the secondary electrons 114 generated from the sample 200 irradiated with 02, the photomultiplier tube 115 that detects the secondary electrons whose trajectory has been changed by the E × B deflector 107, and photoelectrons. An SE image generation unit 116 that processes a signal output from the multiplier 115 to generate an image by secondary electrons (SE), a storage unit 021 that stores data, a comparison operation unit 022, and an input / output unit 024 And a control unit 023 for controlling the whole.

  An electron gun 101, a condenser lens 103, a polarizing lens 104, an objective lens 105, a stage 108, an annular scintillator 106, an optical fiber 111, a photomultiplier tube 112, and a photomultiplier tube 115 to be fired are vacuumed inside. It is installed in an evacuable column.

  In such a configuration, the primary electron beam 102 generated by the electron gun 101 at a high acceleration voltage (for example, 15 kilovolts or more) is focused by the condenser lens 103 and further focused on the surface of the sample 200 by the objective lens 105. The sample is scanned two-dimensionally by the deflector 104. The sample targeted by the present invention is a deep hole / groove pattern in which the hole / groove depth is about several microns with respect to the hole / groove diameter of about 50 nm.

  The backscattered electrons (BSE) 110 emitted from the sample 200 irradiated with the primary electron beam 102 are detected by the annular scintillator 106 and converted into an optical signal, and are guided to the high electron multiplier 112 by the optical fiber 111. A digital image is formed in the BSE image generator 113 from the signal output from the electron multiplier 112. As the annular scintillator 106 that detects the reflected electrons, an annular YAG scintillator, an annular semiconductor detector, or a Robinson detector can be applied. Moreover, the structure which arrange | positions a detector in multiple directions instead of cyclic | annular form may be sufficient.

  The secondary electrons (SE) 114 emitted from the sample 200 irradiated with the primary electron beam 102 are guided to the high electron multiplier 115 by changing the trajectory by the E × B deflector 107, and are supplied to the high electron multiplier 115. A digital image is formed by the SE image generator 116 from the signal output from 115. A feature of this configuration is that a BSE image and an SE image at the same location on the sample 200 are simultaneously captured. By moving the stage 108, an image is taken at an arbitrary position of the sample. The captured image is stored in the storage unit 021.

  The control unit 022 adjusts the voltage applied to the periphery of the electron gun 101, the focus adjustment of the condenser lens 103 and the objective lens 105, the scanning of the primary electron beam 02 on the surface of the sample 200 by the deflection electrode 104, the movement of the stage 108, and the image generation unit 113. , 116 is controlled. The comparison calculation unit 022 performs defect detection processing using the captured images generated by the image generation units 113 and 116. Input of sample information, input of imaging conditions, output of inspection results, and the like are performed by the input / output unit 024.

  FIG. 1B shows the configuration of the comparison calculation unit 022. The comparison calculation unit 022 includes a region extraction unit 0221 that extracts an inspection region from the image obtained by the SE image generation unit 116, and a defect detection unit that detects a defect by processing the image of the inspection region set by the inspection region setting unit 0221. 0223, a feature amount calculation unit 0224 that calculates the feature amount of the defect detected by the defect detection unit 0223, and outputs information on the feature amount of the defect calculated by the feature amount calculation unit 0224 to the storage unit 021 and the input / output unit 024 To do.

  4A is a schematic diagram of an SE image of a hole pattern having an internal defect, and FIG. 4B is a schematic diagram of a BSE image of a hole pattern having an internal defect (detailed description of the drawing will be described later). In the BSE image of FIG. 4B, the defect portion (253) becomes apparent, but since the defect does not become apparent in the SE image of FIG. 4A, the comparison operation unit 022 uses the SE image as a reference of the normal shape, and the BSE image. It is possible to detect a defect by comparing.

Details of the present embodiment will be described below with reference to FIGS.
FIG. 5A shows hole observation by an SE image, and FIG. 5B is a diagram schematically showing the state of hole observation by a BSE image. FIG. 5A shows a cross-sectional view of the sample 200 in which an upper layer film 201 and a lower layer film 202 are formed, and a hole 203 having a relatively high aspect ratio is formed in the upper layer film 201. The reason why the defects in the hole are not revealed in the SE image is that, as shown in FIG. 5A, most of the secondary electrons 114 generated in the hole 203 strike the hole sidewall 204 and disappear.

  On the other hand, when observing a hole with the BSE image shown in FIG. 5B, when the primary electron beam 102 is highly accelerated (high energy), the energy of the reflected electrons 110 is high, so that part of the electron beam 102 passes through the hole sidewall 204. Since the reflected electrons 110 transmitted through the hole sidewall 204 have a low angle (= the zenith angle is large), the annular scintillator 106 that detects the reflected electrons at a low angle functions favorably, thereby revealing defects inside the hole 203. Is possible.

  FIG. 6A to FIG. 6D are the results of confirming by electron beam simulation (Monte Carlo simulation) that defects appear in the BSE image but defects do not appear in the SE image. In the cross-sectional shape of the sample 600 corresponding to the sample 200 shown in FIG. 6A, an upper layer film 601 and a lower layer film 602 are formed, and a hole 603 is formed in the upper layer film 601. The hole 603 is a hole having a top diameter of 70 nm as the diameter of the hole inlet 6031, a bottom diameter 6032 as a diameter of the hole bottom 6032 of 70 nm, and a hole depth 6033 of 3.2 μm (hereinafter referred to as a t70b70 hole).

  On the other hand, in the cross-sectional shape of the sample 600 shown in FIG. 6B, an upper layer film 601 and a lower layer film 602 are formed, and a hole 604 is formed in the upper layer film 601. The holes 604 are two types of holes (hereinafter referred to as t70b30 holes) in which the top diameter 6041, which is the diameter of the hole inlet 6041, is 70 nm, the bottom diameter 6042 is 30 nm, and the hole depth 6043 is 3.2 um.

  6C shows a signal waveform of an image obtained when the sample 600 having the cross-sectional shape as shown in FIG. 6A is imaged by the SEM, and FIG. 6D shows the sample 600 having the cross-sectional shape as shown in FIG. 6B. The signal waveform of the image obtained when it images with SEM is shown. The acceleration voltage of the primary electron beam 102 was 30 kV, and FIG. 6C shows the SE signal waveform detected by the high electron multiplier 115 in the configuration of FIG. 1 and was obtained by detecting electrons having an energy of 50 eV or less. It is a waveform.

  On the other hand, FIG. 6D shows a low-angle BSE signal waveform detected by the annular scintillator 106 in the configuration of FIG. 1 and is obtained by detecting an electron having an energy of 5000 eV or more and an emitted electron zenith angle of 15 to 65 degrees. It is a waveform.

  The horizontal axis of the graphs in FIGS. 6C and 6D is the distance (x) from the center of the hole 603 or 604. x = 35 nm corresponds to the edge of the hole bottom 6032 of the t70b70 hole (hole 603), and x = 15 nm corresponds to the edge end of the hole bottom 6042 of the t70b30 hole (hole 604). The vertical axis of the graphs in FIGS. 6C and 6D represents the detected signal intensity (Yield) of the signal detected by the high electron multiplier 115 or the annular scintillator 106.

  Comparing FIG. 6C and FIG. 6D, in the low-angle BSE signal waveform of FIG. 6D, there is a clear difference in the waveform of the t70b70 hole (hole 603) and the t70b30 hole (hole 604), but the SE signal waveform of FIG. The difference in waveform is slight. This result can be obtained from the SE signal waveform, although the SE signal waveform cannot capture the difference in shape between the hole bottoms 6032 and 6042, that is, the BSE image obtained from the BSE signal waveform reveals a defect at the back of the hole. In the SE image, it means that it does not appear.

  7A to 7C are simulation results showing the necessity of the configuration of the electron optical system of the present embodiment shown in FIG. 1A. FIG. 7A shows two types of holes 701 having a cross-sectional shape formed in the sample 710 as shown in FIG. 7B, with a hole diameter of 50 nm and a depth of 3.2 microns, and the same hole diameter and a depth of 5.1 microns. In this case, the relationship between the acceleration voltage of the primary electron beam 102 and the detected signal intensity (Yield) when the reflected electrons 110 generated from the hole bottom 711 are detected by the annular scintillator 106 is obtained by simulation.

  As shown in the figure, in order to detect a defect inside the hole 701, an acceleration voltage corresponding to the depth of the hole 701 is required. It can be seen that in order to inspect a sample having a deep hole with an aspect ratio of 20 or more, which is the main target of the present invention, an acceleration voltage of the primary electron beam 102 of at least 15 kilovolts is required.

  FIG. 7C shows the result of calculating the zenith angle distribution of electrons emitted from the sample by simulation. Here, the zenith angle is defined as an angle from the incident direction of the primary electron beam 102 as shown in FIG. 7B. In the figure, the horizontal axis represents the zenith angle of the emitted electrons, and the vertical axis represents the detection signal intensity. The detection signal intensity (Yield) is large when the zenith angle is around 30-50 degrees, and in order to detect this without missing, an annular scintillator capable of detecting reflected electrons at low angles in all directions is suitable. Is shown.

  FIG. 8 shows specific calculation contents in the comparison calculation unit (022 in FIG. 1). The input is the SE image and BSE image shown in FIGS. 4A and 4B. In FIG. 4A, 221 is a hole, 222 is the outside of the hole, and 224 is a hole edge. In the SE image, the edge portion is detected brightly by the edge effect. In the figure, edge roughness is represented by triangular protrusions.

  FIG. 4B is a BSE image acquired simultaneously. Reference numeral 251 denotes a hole, 252 denotes an outside of the hole, 254 denotes a hole edge, and 253 denotes an in-hole defect. As described with reference to FIGS. 5A, 5B, and 6A to 6D, the in-hole defect becomes apparent in the BSE image but not in the SE image. Since the SE image and the BSE image obtained by the electron beam pattern inspection apparatus 100 having the configuration shown in FIG. 1 are obtained at the same location on the sample, both have the same edge roughness. is there.

Returning to FIG. 8, the calculation contents (processing flow) in the comparison calculation unit (022 in FIG. 1) will be described.
First, the region extraction unit 0221 performs region extraction using the SE image 220 created by the SE image generation unit 116 (S301), and creates inspection region data 225. In the SE image, the brightness of the hole is sufficiently lower than the outside of the hole or the edge, so that a dark area is extracted by binarization and used as an inspection area. By using the fact that the brightness of the edge portion is high, a contour line may be extracted and a region surrounded by the contour line may be set as an inspection region. The inspection area data 225 is data for designating the inspection target area. In FIG. 8, the inspection target area is shown in white and the other areas are shown in two colors.

  Next, in the inspection area setting unit 0222, the BSE image 250 generated by the BSE image generation unit 113 and the inspection area data 225 generated by the area extraction unit 0221 are ANDed (S302), and the inspection area ( 226) is set. In FIG. 8, the inspection area 226 is shown in gray and the outside of the inspection area is shown in black.

  Next, the defect detection unit 0223 applies a preset threshold value (showing a setting method in the fifth embodiment) to the inspection region 226 set by the inspection region setting unit 0222 (S303). 227) is detected. Then, the feature amount calculation unit 0224 calculates feature amounts such as the position of the defect portion detected by the defect detection unit 0223, the luminance of the defect portion (luminance before binarization), and the area (S304).

  Finally, information on the calculated feature amount of the defect is stored in the storage unit 021 and is output to the input / output unit 024 (S305).

  The embodiment described above is the basic configuration of the present invention. According to the present embodiment, it becomes possible to detect a defect in a deep hole, which could not be detected by a conventional electron beam inspection apparatus having an acceleration voltage of several kilovolts. For this purpose, acceleration voltage of 15 kV or more and low-angle BSE detection function effectively. Furthermore, since the SE image acquired at the same location having the same edge roughness is used as a reference, the problem of erroneously detecting edge roughness as a defect as shown in FIG. Inspection can be realized.

The present embodiment is also effective for inspecting deep hole shape defects by the processing flow shown in FIG.
FIG. 9 shows the flow of processing for inspecting the sample in which the eccentricity and the reduction of the hole diameter occur in the back of the deep hole by the comparison calculation unit 022. The formation defect at the back of the hole does not appear in the SE image 220, but the BSE image 260 reveals that the position displacement of the hole bottom and the size of the hole bottom diameter are small. In the flow of FIG. 9, the processes from S301 to S305 are the same as the steps described in FIG.

  That is, in the processing flow shown in FIG. 9, the inspection area data 901 is obtained as a result of the area extraction process of S301 in the area extraction unit 0221, and the inspection is performed as a result of logical product in the processing of S302 in the inspection area setting unit 0222. A region image 902 is obtained. Further, in S303, the threshold value processing is performed on the image 902 using the threshold value stored in the storage unit 021 in advance by the defect extraction unit 0223, thereby obtaining a binarized image 903 in which the defect is manifested. Finally, in S304, the feature amount calculation unit 0224 calculates the feature amount of the defect. Here, if the BSE image is inspected using the SE image as a reference, it is possible to quantitatively detect the eccentricity and the change in the hole diameter as the defect feature amount extracted by the feature amount calculation unit 0224 in S304. is there.

  Furthermore, the present embodiment is also effective for inspecting deep groove formation defects as shown in FIG. FIG. 10 shows a case where the groove is deformed at the corner of the deep groove pattern (for example, a partial depth of the groove is insufficient). The formation defect at the back of the groove does not appear in the SE image 270, but appears in the BSE image 280 as a pattern deformation 1010 at the groove bottom. In the flow of FIG. 10, the processes from S301 to S305 are the same as the steps described in FIG.

  That is, in the processing flow shown in FIG. 10, inspection region data 1001 is obtained as a result of the region extraction processing in S301 in the region extraction unit 0221, and inspection is performed as a result of logical product in the processing in S302 in the inspection region setting unit 0222. A region image 1002 is obtained. In S303, the threshold value processing is performed on the image 1002 using the threshold value stored in advance in the storage unit 021 by the defect extraction unit 0223, thereby obtaining a binarized image 1003 in which the defect 1011 is made obvious. Finally, in S304, the feature amount calculation unit 0224 calculates the feature amount of the defect. Here, if the BSE image is inspected using the SE image as a reference, it is possible to quantitatively detect pattern formation defects.

  A second embodiment according to the present invention is shown in FIGS. 11A and 11B. In the present embodiment, the comparison operation unit (022 in FIG. 1) described in the first embodiment is replaced with a comparison operation unit 022-1 as shown in FIG. 11A. Since the configuration other than the comparison calculation unit 022-1 is the same as that described with reference to FIG. 1A in the first embodiment, the description thereof is omitted.

  The comparison calculation unit 022-1 according to the present embodiment illustrated in FIG. 11A extracts an inspection region from the SE image obtained by the SE image generation unit 116, and the SE obtained by the SE image generation unit 116. A difference image generation unit 0225 that generates a difference image between the image and the image generated by the BSE image generation unit 113, an inspection region extracted by the region extraction unit 0221, and a difference image generated by the difference image generation unit 0225 are used to detect defects. A defect detection unit 0226 that performs detection and a feature amount calculation unit 0224 that calculates the feature amount of the defect detected by the defect detection unit 0226, and inputs / outputs information regarding the feature amount of the defect calculated by the feature amount calculation unit 0224 to and from the storage unit 021. Output to the unit 024.

  Next, processing performed in the comparison operation unit 022-1 will be described. The input is the SE image and BSE image shown in FIGS. 4A and 4B, as in the first embodiment. The region extraction unit 0221 performs region extraction (S301) from the SE image to create the inspection region data 225 as in the first embodiment.

  Next, in the present embodiment, the difference image generation unit 0225 generates a difference image 401 between the SE image 220 and the BSE image 250 (S305), and the defect detection unit 0226 calculates the difference image 401 and the inspection area data 225. A logical product is taken (S306), and a threshold value set in advance is applied to the image 402 obtained as a result, and binarization processing is performed to detect a defect 403 (S307). Then, the feature amount calculation unit 0224 calculates the feature amount such as the position of the defect portion, the luminance of the defect portion (luminance before binarization), the area, etc. (S308), and stores information on the feature amount of the defect in the storage unit 021. And output to the input / output unit 024 (S309).

  Since the BSE image and the SE image are acquired simultaneously, changes in the luminance of the image due to changes in the amount of irradiation light are synchronized. Therefore, the luminance of the difference image does not change even if the irradiation light quantity changes. In this embodiment, since a threshold value is applied to the difference image, there is an advantage that the inspection sensitivity is not affected by the change in the amount of irradiation light.

  When the inspection apparatus is operated for a long time, the amount of irradiation light may fluctuate. However, in this embodiment, since the inspection sensitivity is not affected by this, a more stable inspection can be realized.

  The present embodiment is also effective for the inspection of defective formation of deep holes as shown in FIG. 9 and the inspection of defective formation of deep grooves as shown in FIG.

  A third embodiment according to the present invention is shown in FIGS. 12A and 12B. In this embodiment, the comparison operation section (022 in FIG. 1) described in the first embodiment is replaced with a comparison operation section 022-2 as shown in FIG. 12A. Since the configuration other than the comparison calculation unit 022-2 is the same as that described with reference to FIG. 1A in the first embodiment, the description thereof is omitted.

  The comparison operation unit 022-2 according to the present embodiment illustrated in FIG. 12A uses the SE image generated by the SE image generation unit 116 for the image 411 of the in-hole region that is the inside of the hole and the out-of-hole region that is outside the hole. An image dividing unit 0227 that divides the image 414 and the image 411 in the hole area divided by the image dividing unit 0227 and the BSE image generated by the BSE image generating unit 113 are calculated and the first inspection region is set. The second inspection region is set by taking the logical product of the image 414 of the non-hole region divided by the first inspection region setting unit 0228 and the image dividing unit 0229 and the BSE image generated by the BSE image generation unit 113. The second inspection area setting unit 0229, the first inspection area set by the first inspection area setting unit 0228 is threshold-processed to detect a defect in the hole, 2 inspection area setting A feature value of the defect in the hole detected by the out-of-hole defect detecting unit 02111 and the in-hole defect detecting unit 02210 that detects the defect outside the hole by performing threshold processing on the second inspection area set in 0229 is calculated. In-hole defect feature quantity calculation unit 02212, and an out-of-hole defect feature quantity calculation unit 02213 that calculates the feature quantity of the defect outside the hole detected by the out-of-hole defect detection unit 02111. Information regarding the feature amount of the defect calculated by the feature amount calculation unit 02213 is output to the storage unit 021 and the input / output unit 024.

  Next, processing performed by the comparison operation unit 022-2 will be described. As in the first embodiment, the input is an SE image 220 and a BSE image 410 obtained by simultaneously imaging the same portion with the optical system shown in FIG. 1A. On the BSE image, 421 is a defect in a hole, and 422 is an internal defect in a region outside the hole (defect not exposed on the surface). Neither is manifested in the SE image.

  First, region division is performed using the image dividing unit 0227SE image 220 (S310), and in-hole region data 411 and out-hole region data 414 are created. The region division by the image dividing unit 0227 may be performed using the relationship between the luminance in the SE image 220, the inside of the hole <the outside of the hole <the edge portion, and the ternary processing may be performed, or the high luminance of the edge portion may be used. Thus, the contour line may be extracted, and the region surrounded by the contour line may be defined as the in-hole region, and the region outside the contour line may be defined as the out-hole region.

  Next, the first inspection area 412 is set for the BSE image by taking the logical product of the hole area data 411 and the BSE image 410 by the hole defect detection unit 02210 (S311). In FIG. 12B, the outside of the inspection area is shown in black. The first inspection area 412 is binarized by applying a preset threshold value a to detect a defect 413 (S312).

  Then, feature amounts such as the position of the defect portion, the luminance before the binarization of the defect portion, and the area are selected (S313) and output. Further, the second inspection region 415 is set for the BSE image by taking the logical product of the out-of-hole region data 414 and the BSE image 410 (S314). In FIG. 12B, the outside of the inspection area is shown in black. The second inspection area 415 is binarized by applying a preset threshold value b to detect the defect 410 (S315). Then, feature amounts such as the position of the defective portion, the luminance of the defective portion (luminance before binarization), and the area are calculated (S316) and output.

According to this embodiment, it is possible to detect defects inside and outside the hole with sensitivity suitable for each.
Note that this embodiment is also effective for the inspection of deep hole formation defects as shown in FIG. 9 and the inspection of deep groove formation defects as shown in FIG.

  A fourth embodiment according to the present invention is shown in FIGS. 13A to 13C. In the present embodiment, the comparison operation unit (022 in FIG. 1) described in the first embodiment is replaced with a comparison operation unit 022-3 as illustrated in FIG. 13A. Since the configuration other than the comparison calculation unit 022-3 is the same as that described with reference to FIG. 1A in the first embodiment, the description thereof is omitted.

  Similar to the first embodiment, the comparison operation unit 022-3 according to the present embodiment illustrated in FIG. 13A extracts a region extraction unit 0221 and a region extraction unit 0221 that extract the inspection region from the SE image generated by the SE image generation unit 116. The inspection area setting unit 0222 for setting the inspection area from the inspection area extracted in step B and the BSE image generated by the BSE image generation unit 113, and the average value of the brightness of the BSE image in the inspection area set by the inspection area setting unit 0222 A luminance average value / standard deviation calculation unit 02214 for obtaining a standard deviation, a defect detection unit 02215 for detecting defects using information on the average value and standard deviation of the luminance of the BSE image obtained by the luminance average value / standard deviation calculation unit 02214, A feature amount calculation unit 0224 that calculates the feature amount of the defect detected by the defect detection unit 02215 is provided.

  Next, processing performed by the comparison operation unit 022-3 will be described. The input is the SE image and BSE image shown in FIGS. 4A and 4B, as in the first embodiment. The region extraction unit 0221 performs region extraction on the SE image (S301), creates inspection region data 225, and the inspection region setting unit 0222 obtains a logical product with the BSE image 250 and sets the inspection region 251 in the BSE image. Up to the setting (S302) is the same as that in the first embodiment.

In the present embodiment, the luminance average value / standard deviation value calculation unit 02214 calculates the luminance histogram (FIG. 13C) in the BSE image inspection region unit 251 set in S302, and calculates the average luminance value (avg). The standard deviation (σ) is calculated. Then, a threshold value is calculated by (Equation 1) using the calculated average value (avg) of brightness and standard deviation (σ) (S321). Next, the defect detection unit 02215 binarizes the inspection area 251 of the BSE image using the threshold value calculated by Expression 1 to detect the defect 411 (S322). Threshold = avg + r × σ (Equation 1)
Here, r is a coefficient by which a preset dispersion value is multiplied.

  For example, if r = 3 is set, a region where the difference from the average luminance is three times the standard deviation or more is detected as a defect. That is, in the present embodiment, a singular region of brightness within the inspection region is detected as a defect. If the brightness of the entire area in the inspection area changes uniformly, it does not become a defect.

  The feature amount calculation unit 0224 calculates feature amounts such as the position of the detected defect portion, the luminance of the defect portion (luminance before binarization), and the area (S323), and information about the feature amount of the defect is obtained. The data is output to the storage unit 021 and the input / output unit 024 (S324).

  According to the present embodiment, the threshold value is automatically adjusted following the overall luminance change, which is advantageous for detecting a defect that appears as a local luminance change in the inspection region. .

  Note that this embodiment is also effective for the inspection of deep hole formation defects as shown in FIG. 9 and the inspection of deep groove formation defects as shown in FIG.

  FIG. 14 shows a fifth embodiment according to the present invention. This embodiment is a user interface 500 for setting inspection conditions. The defect determination mode is selected at 501. When “difference image” is selected, the SE image is used as a reference, and the difference image between the BSE image and the SE image is set as an inspection target (see Example 2). When “BSE image” is selected, the mode is a mode in which the SE image is used as a reference and the BSE image is used as an inspection target (see Examples 1, 3, and 4). In FIG. 14, the difference image mode is selected.

  The screen includes a BSE image (502), an SE image (505), a difference image between the BSE image and the SE image (503), an inspection area (506) created from the SE image, and an overlay of the difference image and the inspection area (504). , The defect detection result (507) is displayed. The user can adjust the threshold value for creating the inspection region from the SE image with the slider 506 while checking the result of the action. It is also possible to adjust the threshold value for defect detection while confirming the result of the action with the slider 509.

  FIG. 15 shows a schematic configuration of an electron beam pattern inspection apparatus 800 in the present embodiment. The same parts as those in the electron beam pattern inspection apparatus 100 described in FIG.

  The electron beam pattern inspection apparatus 100 in Example 1 of FIG. 1A has a configuration using the annular scintillator 106 and the photomultiplier tube 115 as detectors, but the electron in this example shown in FIG. In addition to the above two detectors, the linear pattern inspection apparatus 800 includes an E × B deflector 117 that bends the trajectory of reflected electrons generated from the sample 200 in the optical axis direction (vertical direction) of the primary electron beam 102, and E The point which provided the upper detector 118 which detects the reflected electron 120 by which the orbit was bent by * B deflector 117, and the upper BSE image generation part 119 which processes the detection signal of this upper detector 118, and produces | generates a BSE image It is in.

  The method of inspecting the sample 200 using the electron beam pattern inspection apparatus 800 in the present embodiment is basically the same as the method described in the first embodiment, but S304 of the processing flow described in FIG. The defect feature amount calculation process is different.

  There may be a relationship between the luminance information of the upper BSE image generated by the upper BSE image generation unit 119 and the material of the region irradiated with the primary electron beam 102 of the sample 200. Therefore, in this embodiment, various materials of a known type that may exist on the sample 200 are irradiated with the primary electron beam 102 and the reflected electrons are detected by the upper detector 118 to generate an upper BSE image. The brightness information of the BSE image generated by the unit 119 is stored in advance in the database. In the case of actual inspection, reflected electrons from the sample are detected by the upper detector 118, luminance information is extracted from the BSE image generated by the upper BSE image generation unit 119, and the extracted luminance information is collated with a database. The material (material) of the defect can be specified.

  That is, in this embodiment, in the defect feature amount calculation step in S304 of the processing flow in Embodiment 1 described in FIG. 8, the position of the defect portion, the luminance of the defect portion (luminance before binarization), and the area In addition to the feature quantities such as, it is also possible to obtain information on the material (material) of the defect.

  According to the present embodiment, since the defect material information is also obtained as a feature of the defect, it is possible to specify the defect generation process and the cause of occurrence in a shorter time than in the past.

  In the seventh embodiment according to the present invention, bending (bending) inside the hole is detected. FIGS. 16A and 16D are schematic views of a normal hole 801 and a hole 802 in which bending is generated in the sample 1610, respectively. Sectional views (FIGS. 16A, 16D), SE images (FIGS. 16B, 16E), and BSE images (FIGS. 16C, 16F) of the hole portion from the top.

  Since bending is a bending of a hole at a deep part, there is no change in the SE image FIGS. 16B and 16E (see 802 in FIG. 16B and 812 in FIG. 16E), but pattern deformation and displacement occur in the BSE image. (See 803 in FIG. 16C and 813 in FIG. 16F). In this embodiment, bending is detected using this characteristic. In the explanatory diagrams of the first to sixth embodiments, the edge roughness is shown as a triangular protrusion, but is omitted in FIGS. 16A to 16F for the sake of simplicity.

  In this embodiment, the calculation contents in the comparison calculation section (022 in FIG. 1) described in the first embodiment are replaced with the calculation contents shown in FIG. Since the configuration other than the comparison operation unit is the same as that described with reference to FIG. 1A in the first embodiment, the description thereof is omitted.

  The inputs are the SE image 812 in FIG. 16D shown in FIG. 16E and the BSE image 813 in FIG. 16D shown in FIG. 16F. In the processing flow, as shown in FIG. 17B, first, pattern areas of the SE image 812 in FIG. 16E and the BSE image 813 in FIG. 16F are extracted (S321, S331), and the SE pattern 8121 and the BSE pattern 8131 are extracted. In the figure, the pattern is white and the background is black. The region extraction method is the same as the pattern region extraction in the first embodiment.

  Next, feature quantities of the extracted SE pattern 8121 and BSE pattern 8131 are calculated (S322, S332). An important feature amount in bending detection is the pattern center (8122, 8132). Based on these feature amounts, a bending feature amount as shown in FIG. 17C is calculated (S333).

  In FIG. 17C, dx and dy are deviations of the pattern center 8132 of the BSE image from the pattern center 8122 of the SE image. This is an index that represents the direction and size of bending. In addition to this, as detailed information, as shown in FIG. 17C, distances (L1 to L8) in each direction from the pattern center 8122 of the SE image to the BSE pattern edge 8100 may be calculated. When there is no bending, L1 to L8 have substantially the same value, but when bending occurs, the values of L1 to L8 vary depending on the direction of bending.

  Further, as shown in FIG. 17D, the SE pattern 8200 and the BSE pattern 8100 are compared (S342), a difference image 8140 is generated, and the distance between edges (t1) in each direction as shown in FIG. 17C in S343. ˜t8) may be calculated.

Since the pattern edge 8200 of the SE image corresponds to the top edge 8111 of the hole 811 and the pattern edge 8100 of the BSE image corresponds to the bottom edge 8112 of the hole 811, t1 to t8 correspond to the taper width in each direction. . As in L1 to L8 of FIG. 17C, in FIG. 17D, when there is no bending, t1 to t8 are substantially equal values. Arise.
According to the present embodiment, quantitative inspection of bending can be performed.

  The eighth embodiment according to the present invention is another bending detection method using the characteristic that the bending becomes apparent only in the BSE image, as in the seventh embodiment. Since the configuration other than the comparison operation unit is the same as that described with reference to FIG. 1A in the first embodiment, the description thereof is omitted.

  The processing flow in the comparison operation unit is shown in FIG. 18A. The inputs are the SE image 812 of the hole 1602 in FIG. 16D shown in FIG. 16E and the BSE image 813 shown in FIG. 16F.

  In this embodiment, instead of extracting a pattern area, edge detection is performed using an image profile (S351, S361). Various methods can be applied to the edge detection. As a specific example, the case where the “threshold method” is applied is shown in FIGS. 18B to 18E.

  The profile 1801 in FIG. 18C is the edge detection of the SE image 812 shown in FIG. 18B, and the profile 1802 in FIG. 18E is the edge detection of the BSE image 813 shown in FIG. 18D. In the threshold method, Max 1811, 1812, Min 1821, and 1822 of the waveform are detected, and internal values of these appropriate ratios are calculated as threshold values 1831 and 1832, and threshold values 1831 and 1832 and profiles 1801 and 1802 are calculated. This is a method using the intersections 1841 and 1842 as the edge points.

  As shown in FIGS. 18B and 18D, profiles 1811 and 1812 in each direction (16 directions indicated by straight lines in FIGS. 18B and 18D) are generated from the images 812 and 813, and edges are formed in the profiles 1811 and 1812. By performing detection, an edge point sequence conforming to the pattern shape such as 815 and 816 in FIG. 18A can be obtained. Similar to the seventh embodiment, the feature amount such as the pattern center is calculated from the edge point sequence (S322, S332), and the bending feature amount is calculated (S333). Although this embodiment is not suitable for detecting irregularly shaped defects as shown in FIG. 4B, in addition to bending, a type in which the edge position is shifted as shown in FIG. It is effective against defects of

  This embodiment relates to a composite image generation method suitable for a user to review inspection results.

  For example, even if the BSE image (813 in FIG. 16F) of the hole pattern in which bending occurs is displayed on the screen, it is difficult for the user to visually confirm the deformation of the pattern and the positional deviation (eccentricity). If displayed side by side with the SE image (812 in FIG. 16E), it can be visually confirmed if the degree of deformation or displacement is large, but if the degree is slight, it is still difficult to visually confirm. The object of the present embodiment is to generate an observation image so that the positional relationship between the pattern of the SE image and the BSE image can be understood at a glance.

  FIG. 19A shows a composite image generation flow. The input is an SE image 812 and a BSE image 813. First, by using the fact that the brightness of the portion corresponding to the inside of the hole is sufficiently lower than that of the outside of the hole or the edge portion in the SE image, the region corresponding to the inside of the hole is extracted by binarization or the like. Extract (S321). In the image 8120 from which holes are extracted, the in-hole region is shown in white, and the others are shown in black.

  Next, using the hole area information of the image 8120, image synthesis is performed in which a BSE image is inserted into the in-hole area and an SE image is inserted into the other area (S350), thereby generating a synthesized image 817. When compositing images, the brightness of each image is adjusted (offset adjustment, contrast adjustment) in consideration of visibility, and the blend ratio of both images near the boundary is set so that the seam of the images becomes more natural. You may make it change in steps. Alternatively, edge enhancement processing may be performed.

  According to the composite image 817 generated according to the present embodiment, pattern deformation and positional deviation (eccentricity) can be easily visually confirmed, so that inspection results can be reviewed efficiently. Needless to say, this embodiment is also effective for reviewing defect detection results other than bending as shown in FIGS.

  FIG. 19B shows an example of the inspection result review screen 1900. On the screen, an SE image (812), a BSE image (813), and a composite image (817) are displayed, and various feature amounts are displayed as a list as inspection results (900, 901, 902). If there are a plurality of patterns on the image, the average values (tabs Avg.) 920 to 922 are also displayed in addition to the feature amounts (tabs 1 to 6) 910 to 912 of the individual patterns.

“Eccentricity (x)” 9121 of the list display 902 is dx in FIG. 17C, “Eccentricity (y)” 9122 is dy in FIG. 17C, “Eccentric distance” 9123 is the square root of dx ² + dy ² , and “Eccentric angle” 9124 is tan ⁻¹ (dy / dx), “edge width by direction” 9125 is t1 to t8 in FIG. 17D. According to the present embodiment, since the quantitative inspection result is displayed together with the composite image, the inspection result can be easily reviewed.

  In the present embodiment, the processing contents in the comparison operation unit (022 in FIG. 15) described in the sixth embodiment are replaced with the processing contents shown in FIG. In the seventh embodiment, the bending inspection is performed using the SE image 812 and the BSE image 813. In the tenth embodiment, in addition to this, the high-angle BSE image 814 and the SE image 812 generated by the upper BSE generation unit in FIG. Conduct a bending inspection using.

  The inspection using the SE image 812 and the BSE image 813 in FIG. 20 is the same as the inspection described in FIG. 17B, and the first bending feature amount is calculated using the pattern feature amounts of the SE image 812 and the BSE image 813. (S333). In the present embodiment, the pattern feature amount of the high-angle BSE image 814 and the SE image 812 is further calculated (S342), and the second bending feature amount is calculated using the calculated pattern feature amount (S343). The calculated first bending feature value and the second bending feature value calculated in S343 are integrated to obtain a final output (S344).

  For the integration, the average value of the first bending feature value and the second bending feature value may be obtained, or defective only when both the first bending feature value and the second bending feature value exceed the reference value. It may be used as such.

  Further, in FIG. 20, two processes of FIG. 17B are provided, but instead of this, two processes of FIG. 18A may be provided.

  According to the present embodiment, since two types of defect determination are performed, it is possible to realize a more reliable inspection.

  021: Storage unit 022, 022-1, 022-2, 022-3 ... Comparison operation unit 023 ... Control unit 024 ... Input / output unit 101 ... Electron gun 103 ... Condenser lens DESCRIPTION OF SYMBOLS 104 ... Deflector 105 ... Objective lens 106 ... Low angle BSE detection annular scintillator 108 ... Stage 111 ... Optical fiber 112 ... Photomultiplier tube 113 ... BSE image generation part 115 ... Photomultiplier tube 200 ... Sample 0221 ... Area extraction unit 0222 ... Inspection region setting unit 0223, 0226 ... Defect detection unit 0224 ... Feature quantity calculation unit 0225 ... Difference Image generation unit 0227 ... image division unit 0228 ... first inspection region setting unit 0229 ... second inspection region setting unit 02210... In-hole defect detection unit 02211... Out-of-hole defect detection unit 02212... In-hole defect feature quantity calculation unit 02213.

Claims (14)

An electron beam irradiation unit that irradiates the sample on which the pattern is formed with the converged electron beam; and
A reflected electron detector for detecting reflected electrons having a relatively high energy generated from the sample irradiated with the electron beam converged by the electron beam irradiation unit;
A secondary electron detector that detects secondary electrons with relatively low energy generated from the sample irradiated with the electron beam focused by the electron beam irradiator;
A backscattered electron image generation unit that generates a backscattered electron image from a signal obtained by detecting backscattered electrons with the backscattered electron detection unit;
A secondary electron image generation unit that generates a secondary electron image from a signal obtained by detecting secondary electrons in the secondary electron detection unit;
An electron beam pattern inspection comprising: an operation unit that detects a defect of the sample by processing the reflected electron image generated by the reflected electron image generation unit and the secondary electron image generated by the secondary electron image generation unit A device,
The computing unit is
An inspection region extraction unit for extracting an inspection region from the secondary electron image;
An electron beam comprising: a defect detection unit that sets an area corresponding to the inspection area extracted by the inspection area extraction unit using the reflected electron image and inspects the set area to detect a defect Expression pattern inspection device.   The electron beam pattern inspection apparatus according to claim 1, wherein the calculation unit further includes a feature amount calculation unit that calculates a feature amount of a defect detected by the defect detection unit. Inspection device.   The electron beam pattern inspection apparatus according to claim 1, wherein the calculation unit further includes a difference image generation unit that creates a difference image between the reflected electron image and the secondary electron image, and the defect detection unit. The electron beam type is characterized in that an area corresponding to the inspection area extracted by the inspection area extraction unit is set for the difference image generated by the difference image generation unit, and the defect is detected by inspecting the set area. Pattern inspection device.   The electron beam pattern inspection apparatus according to claim 1, wherein the inspection area extraction unit of the arithmetic unit sets an inspection area inside the pattern and an inspection area outside the pattern from the secondary electron image, and detects the defect. The unit detects defects by inspecting an inspection area inside the pattern and an inspection area outside the pattern set using the secondary electron image with respect to the reflected electron image, respectively. Expression pattern inspection device.   The electron beam pattern inspection apparatus according to claim 1, wherein the defect detection unit of the arithmetic unit sets a threshold value for defect detection using information on brightness of the reflected electron image, and sets the set value. An electron beam pattern inspection apparatus, wherein a defect is detected from the reflected electron image using a threshold value.   2. The electron beam pattern inspection apparatus according to claim 1, wherein the reflected electron detection unit is generated in an oblique direction with respect to the electron beam among the reflected electrons generated from the sample irradiated with the electron beam. An electron beam pattern inspection apparatus comprising: a first backscattered electron detector for detecting backscattered electrons; and a second backscattered electron detector for detecting backscattered electrons generated in a direction along the electron beam. . A system for inspecting a deep hole or deep groove pattern formed on a substrate,
An electron beam irradiation unit for irradiating the sample with the converged electron beam;
A backscattered electron detector, a backscattered electron detector, and a backscattered electron detector that simultaneously acquire a relatively high energy backscattered electron generated from the sample irradiated with the electron beam and a relatively low energy secondary electron;
A backscattered electron image generation unit that generates a backscattered electron image from a signal obtained by detecting backscattered electrons with the backscattered electron detection unit;
A secondary electron image generation unit that generates a secondary electron image from a signal obtained by detecting secondary electrons in the secondary electron detection unit;
An arithmetic unit for detecting a difference by comparing the reflected electron image and the secondary electron image is provided.   8. The electron beam pattern inspection apparatus according to claim 7, wherein the calculation unit calculates a shift amount of a pattern edge position of the reflected electron image with respect to a pattern edge position of the secondary electron image. .   8. The electron beam pattern inspection apparatus according to claim 7, wherein the calculation unit calculates a deviation amount of a center position of the reflected electron image pattern from a center position of the pattern of the secondary electron image. And   8. The electron beam pattern inspection apparatus according to claim 7, wherein the calculation unit further includes an image composition unit that generates a composite image of the secondary electron image and the reflected electron image. Pattern inspection device. A system for inspecting a deep hole or deep groove pattern formed on a substrate,
An electron beam irradiation unit for irradiating the sample with the converged electron beam;
The low zenith angle component of the relatively high energy reflected electrons generated from the sample irradiated with the electron beam, the high zenith angle component of the relatively high energy reflected electrons, and the relatively low energy secondary electrons. Obtained by detecting reflected electrons with the first reflected electron detector, the second reflected electron detector, the secondary electron detector and the first reflected electron detector, respectively. A first reflected electron image generation unit that generates a first reflected electron image from the signal;
A second backscattered electron image generation section that generates a second backscattered electron image from a signal obtained by detecting backscattered electrons with the second backscattered electron detection section;
A secondary electron image generation unit that generates a secondary electron image from a signal obtained by detecting secondary electrons in the secondary electron detection unit;
An arithmetic unit for detecting a difference by comparing the first reflected electron image, the second reflected electron image, and the secondary electron image is provided.   The electron beam pattern inspection apparatus according to claim 11, wherein the calculation unit includes a shift amount of a pattern edge position of the first reflected electron image with respect to a pattern edge position of the secondary electron image, and the second A shift amount of the pattern edge position of the second reflected electron image with respect to the pattern edge position of the secondary electron image is calculated.   The electron beam pattern inspection apparatus according to claim 11, wherein the calculation unit includes a shift amount of a center position of the pattern of the first reflected electron image with respect to a gravity center position of the pattern of the secondary electron image, and A deviation amount of the center position of the pattern of the second reflected electron image from the center position of the pattern of the secondary electron image is calculated.   12. The electron beam pattern inspection apparatus according to claim 11, wherein the calculation unit generates a composite image of the secondary electron image, the first reflected electron image, and the second reflected electron image. The electron beam type pattern inspection apparatus further comprising:

Images