Charged particle beam apparatus

Abstract

The present invention provides a charged particle beam apparatus capable of stabilizing automatic MS. The charged particle beam apparatus is a charged particle beam apparatus that automatically produces a specimen piece from a specimen, and includes: a charged particle beam irradiation optical system that irradiates a charged particle beam; a specimen stage that mounts and moves the specimen; a specimen piece transfer unit that holds and transports the specimen piece separated and extracted from the specimen; a holder fixing stage that holds a specimen piece holder for transferring the specimen piece; and a computer that controls the position based on the result of a second determination of the position made according to the result of a first determination of the position related to the object, and information including an image obtained by irradiating the charged particle beam.

Description

Technical Field

[0001] The present invention relates to a charged particle beam apparatus. Background Art

[0002] Conventionally, there has been known an apparatus that extracts a specimen piece produced by irradiating a specimen with a charged particle beam composed of electrons or ions, and processes the specimen piece into a shape suitable for various processes such as observation, analysis, and measurement using a transmission electron microscope (TEM: Transmission Electron Microscope) or the like (Patent Document 1). In the apparatus described in Patent Document 1, in the case of observation using a transmission electron microscope, so-called micro-sampling (MS: Micro-sampling) is performed: after taking out a fine thin film specimen piece from a specimen as an object to be observed, the thin film specimen piece is fixed to a specimen holder to produce a TEM specimen.

[0003] In the production of a thin specimen for TEM observation, there has been known a charged particle beam apparatus that detects an object such as the tip of a microprobe, the pickup position of a thin specimen, and the end of a support on a grid by template matching (Patent Document 2). In the charged particle beam apparatus described in Patent Document 2, position control related to the object is performed based on a template produced from an image of the object obtained by irradiating the object with a charged particle beam and position information obtained from the image of the object. Thus, in the charged particle beam apparatus described in Patent Document 2, MS (automatic MS) can be automatically executed.

[0004] Prior Art Documents

[0005] Patent Documents

[0006] Patent Document 1: Japanese Unexamined Patent Application Publication No. 2019-102138

[0007] Patent Document 2: Japanese Unexamined Patent Application Publication No. 2016-157671 Summary of the Invention

[0008] Problems to be Solved by the Invention

[0009] In the charged particle beam apparatus described in Patent Document 2, when the contrast or focus is different between the image of the object obtained by irradiating the object with a charged particle beam and the template image, or when the surface shape of the object (including the attachment of foreign matter) is different between the image of the object and the template image, template matching sometimes fails. In the case of template matching failure, in the charged particle beam apparatus described in Patent Document 2, automatic MS stops.

[0010] Thus, the stability of conventional automatic MS has been insufficient, and it is desired to stabilize automatic MS and increase the production volume.

[0011] The present invention has been completed in view of the above aspects, and provides a charged particle beam apparatus capable of stabilizing an automatic MS.

[0012] Means for solving the problem

[0013] In order to solve the above problems and achieve the above object, the present invention adopts the following solutions.

[0014] (1) One aspect of the present invention is a charged particle beam apparatus that automatically produces a specimen piece from a specimen, wherein the charged particle beam apparatus includes: a charged particle beam irradiation optical system that irradiates a charged particle beam; a specimen stage that mounts the specimen and moves; a specimen piece transfer unit that holds and conveys the specimen piece separated and extracted from the specimen; a holder fixing stage that holds a specimen piece holder for transferring the specimen piece; and a computer that controls the position based on the result of a second determination for the position based on the result of a first determination for a position related to an object, and information including an image obtained by irradiating the charged particle beam.

[0015] In the composite charged particle beam apparatus according to the above (1), the position of the object can be detected based on the result of the second determination based on the result of the first determination, and thus, the automatic MS can be stabilized.

[0016] Here, in the case where the detection of the position of the object fails and the automatic MS stops, every time the automatic MS stops, the user has to take measures, resulting in a decrease in production volume. In the composite charged particle beam apparatus according to the above (1), even when the first determination fails, the position of the object can be detected based on the result of the second determination, and thus, the success rate of position detection can be improved and recovery can be achieved when template matching fails.

[0017] (2) Based on the charged particle beam apparatus according to the above (1), the first determination is a determination based on template matching using a template for the object, and the second determination is a determination based on a machine learning model learned from second information including a second image of a second object.

[0018] In the composite charged particle beam apparatus according to the above (2), the position of the object can be detected based on the result of the determination based on the machine learning model based on the result of the determination based on template matching, and thus, based on template matching and the machine learning model, the automatic MS can be stabilized. In particular, in the composite charged particle beam apparatus according to the above (2), even when template matching fails, the position of the object can be detected based on the machine learning model.

[0019] (3) Based on the charged particle beam device described in (1) or (2) above, the computer selects the type for at least one of the first determination and the second determination according to the result of the third determination for selecting the type of determination.

[0020] In the composite charged particle beam device according to the scheme described in (3) above, it is possible to select the type of determination (appropriate image processing algorithm) for determining the position of the object, and therefore, the detection accuracy of the position of the object can be improved.

[0021] (4) Based on the charged particle beam device described in any one of (1) to (3) above, the computer controls the position according to the result of the fourth determination selected based on at least one of the result of the first determination and the result of the second determination, and information including an image obtained by irradiating with the charged particle beam.

[0022] In the composite charged particle beam device according to the scheme described in (4) above, it is possible to detect the position of the object according to the result of the fourth determination selected based on at least one of the result of the first determination and the result of the second determination, and therefore, compared with the case of detecting the position of the object based on the result of the second determination, the automatic MS can be stabilized.

[0023] Effects of the Invention

[0024] According to the present invention, automatic micro-sampling can be stabilized. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 FIG. is an example showing the structure of a charged particle beam device and an image processing computer according to the first embodiment of the present invention.

[0026] Figure 2 FIG. is an example showing the structure of a charged particle beam device according to the first embodiment of the present invention.

[0027] Figure 3 FIG. is a top view of a specimen wafer according to the first embodiment of the present invention.

[0028] Figure 4 FIG. is a top view of a specimen wafer holder according to the first embodiment of the present invention.

[0029] Figure 5 FIG. is a side view of a specimen wafer holder according to the first embodiment of the present invention.

[0030] Figure 6 FIG. is an example showing the structure of an image processing computer according to the first embodiment of the present invention.

[0031] Figure 7 This is a diagram showing an example of the initial setting process of the first embodiment of the present invention.

[0032] Figure 8 This is a top view of the columnar part of the first embodiment of the present invention.

[0033] Figure 9 This is a side view of the columnar part of the first embodiment of the present invention.

[0034] Figure 10 This is a diagram showing an example of the learning image of the columnar part of the first embodiment of the present invention.

[0035] Figure 11 This is a diagram showing an example of the columnar part in which the pillar of the first embodiment of the present invention does not form a stepped structure.

[0036] Figure 12 This is a diagram showing an example of the learning image of the columnar part in which the pillar of the first embodiment of the present invention does not form a stepped structure.

[0037] Figure 13 This is a diagram showing an example of the specimen piece picking process of the first embodiment of the present invention.

[0038] Figure 14 This is a diagram showing an example of the movement process of the needle in the first embodiment of the present invention.

[0039] Figure 15 This is a diagram showing an example of the needle tip position determination process of the first embodiment of the present invention.

[0040] Figure 16 This is a diagram showing an example of the SEM image data including the tip of the needle in the first embodiment of the present invention.

[0041] Figure 17 This is a diagram showing an example of the SIM image data including the tip of the needle in the first embodiment of the present invention.

[0042] Figure 18 This is a diagram showing an example of the tip of the needle in the first embodiment of the present invention.

[0043] Figure 19 This is a diagram showing an example of the learning image of the needle in the first embodiment of the present invention.

[0044] Figure 20 This is a diagram showing an example of the specimen piece attached to the tip of the needle in the first embodiment of the present invention.

[0045] Figure 21 This is a diagram showing an example of the learning image for abnormal situations in the first embodiment of the present invention.

[0046] Figure 22 This is a diagram showing an example of the removal of foreign matter in the first embodiment of the present invention.

[0047] Figure 23 This is a diagram showing an example of the pick-up position determination process in the first embodiment of the present invention.

[0048] Figure 24 This is a diagram showing an example of SIM image data including a specimen piece in the first embodiment of the present invention.

[0049] Figure 25 This is a diagram showing an example of the learning image of the specimen piece in the first embodiment of the present invention.

[0050] Figure 26 This is a diagram showing an example of the learning image in the first embodiment of the present invention.

[0051] Figure 27 This is a diagram showing an example of the additional image in the first embodiment of the present invention.

[0052] Figure 28 This is a diagram showing an example of the image in which feature points are determined in the first embodiment of the present invention.

[0053] Figure 29 This is a diagram showing an example of the image in which feature points are determined in the first embodiment of the present invention.

[0054] Figure 30 This is a diagram showing the cutting position of the support portion of the specimen and the specimen piece in the SIM image data in the first embodiment of the present invention.

[0055] Figure 31 This is a diagram showing an example of the specimen piece mounting process in the first embodiment of the present invention.

[0056] Figure 32 This is a diagram showing an example of the structure of the image processing computer in the second embodiment of the present invention.

[0057] Figure 33 This is a diagram showing an example of the bare part in the second embodiment of the present invention.

[0058] Figure 34 This is a diagram showing an example of the pattern image in the second embodiment of the present invention.

[0059] Figure 35 This is a diagram showing an example of the pseudo image in the second embodiment of the present invention.

[0060] Figure 36 This is a diagram showing an example of the detection process of the pick-up position in the second embodiment of the present invention.

[0061] Explanation of reference numerals

[0062] 10, 10a... charged particle beam device, S... specimen, Q... specimen piece, 14... convergent ion beam irradiation optical system (charged particle beam irradiation optical system), 15... electron beam irradiation optical system 15 (charged particle beam irradiation optical system), 12... specimen stage, 18... needle (specimen piece transfer unit), 19... needle drive mechanism (specimen piece transfer unit), P... specimen piece holder, 12a... holder fixing table, 22... control computer (computer) Detailed implementation manners

[0063] (First Embodiment)

[0064] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Figure 1 FIG. is an example showing the structure of the charged particle beam device 10 and the image processing computer 30 of the present embodiment. The control computer 22 disposed in the charged particle beam device 10 can acquire image data obtained by irradiation with a charged particle beam. The control computer 22 performs data transmission and reception with the image processing computer 30. The image processing computer 30 first determines an object included in the image data received from the control computer 22 based on template matching using a template T. When the determination based on template matching fails, the image processing computer 30 determines the object based on the machine learning model M. The control computer 22 controls the position related to the object based on the determination result of the image processing computer 30.

[0065] The control computer 22 is an example of a computer that makes a second determination (determination based on the machine learning model M) for the position of the object according to the result of the first determination (template matching) for the position related to the object, and controls the position related to the second object based on the result of the second determination and information including an image obtained by irradiation with a charged particle beam.

[0066] In addition, the image processing computer 30 may be disposed in the charged particle beam device 10.

[0067] Here, refer to Figure 2 to describe the structure of the charged particle beam device 10. (Charged particle beam device)

[0068] Figure 2FIG. 0 is a diagram showing an example of the structure of a charged particle beam apparatus 10 according to an embodiment. The charged particle beam apparatus 10 includes a specimen chamber 11, a specimen stage 12, a stage drive mechanism 13, a focused ion beam irradiation optical system 14, an electron beam irradiation optical system 15, a detector 16, a gas supply unit 17, a needle 18, a needle drive mechanism 19, a drain current detector 20, a display device 21, a control computer 22, and an input device 23.

[0069] The specimen chamber 11 maintains the inside in a vacuum state. The specimen stage 12 fixes a specimen S and a specimen wafer holder P inside the specimen chamber 11. Here, the specimen stage 12 includes a holder fixing stage 12a that holds the specimen wafer holder P. The holder fixing stage 12a may be configured to be able to mount a plurality of specimen wafer holders P.

[0070] The stage drive mechanism 13 drives the specimen stage 12. Here, the stage drive mechanism 13 is housed inside the specimen chamber 11 while being connected to the specimen stage 12, and displaces the specimen stage 12 relative to a predetermined axis according to a control signal output from the control computer 22. The stage drive mechanism 13 includes a movement mechanism 13a that moves the specimen stage 12 in parallel at least along an X-axis and a Y-axis that are parallel to and orthogonal to each other in a horizontal plane, and a Z-axis that is orthogonal to the X-axis and the Y-axis. The stage drive mechanism 13 includes a tilt mechanism 13b that tilts the specimen stage 12 about the X-axis or the Y-axis, and a rotation mechanism 13c that rotates the specimen stage 12 about the Z-axis.

[0071] The focused ion beam irradiation optical system 14 irradiates a focused ion beam (FIB) onto an irradiation target within a predetermined irradiation region (i.e., a scanning range) inside the specimen chamber 11. Here, the focused ion beam irradiation optical system 14 irradiates a focused ion beam onto an irradiation target such as the specimen S, the specimen wafer Q, and the needle 18 existing within the irradiation region placed on the specimen stage 12, from above in a vertical direction downward.

[0072] The focused ion beam irradiation optical system 14 includes an ion source 14a that generates ions, and an ion optical system 14b that converges and deflects the ions extracted from the ion source 14a. The ion source 14a and the ion optical system 14b are controlled according to a control signal output from the control computer 22, and the irradiation position and irradiation conditions of the focused ion beam are controlled by the control computer 22.

[0073] The electron beam irradiation optical system 15 irradiates an electron beam (EB) onto an irradiation target within a predetermined irradiation region inside the specimen chamber 11. Here, the electron beam irradiation optical system 15 can irradiate an electron beam onto an irradiation target such as the specimen S, the specimen wafer Q, and the needle 18 fixed to the specimen stage 12, from above in an inclined direction that is inclined by a predetermined angle (e.g., 60°) with respect to the vertical direction downward.

[0074] The electron beam irradiation optical system 15 includes an electron source 15a that generates electrons, and an electron optical system 15b that converges and deflects the electrons emitted from the electron source 15a. The electron source 15a and the electron optical system 15b are controlled according to a control signal output from the control computer 22, and the irradiation position and irradiation conditions of the electron beam are controlled by the control computer 22.

[0075] Alternatively, the configurations of the electron beam irradiation optical system 15 and the focused ion beam irradiation optical system 14 may be swapped, with the electron beam irradiation optical system 15 arranged in the vertical direction and the focused ion beam irradiation optical system 14 arranged in an inclined direction inclined at a predetermined angle with respect to the vertical direction.

[0076] The detector 16 detects secondary charged particles (secondary electrons, secondary ions) R generated from the irradiation target by the irradiation of the focused ion beam or the electron beam. The gas supply unit 17 supplies a gas G to the surface of the irradiation target. The needle 18 extracts a minute specimen piece Q from the specimen S fixed to the specimen stage 12, holds the specimen piece Q, and transfers it to the specimen piece holder P. The needle drive mechanism 19 drives the needle 18 to transport the specimen piece Q. Hereinafter, the needle 18 and the needle drive mechanism 19 may also be collectively referred to as the specimen piece transfer unit.

[0077] The absorption current detector 20 detects the inflow current (also referred to as the absorption current) of the charged particle beam flowing into the needle 18, and outputs the detected result as an inflow current signal to the control computer 22.

[0078] The control computer 22 controls at least the stage drive mechanism 13, the focused ion beam irradiation optical system 14, the electron beam irradiation optical system 15, the gas supply unit 17, and the needle drive mechanism 19. The control computer 22 is arranged outside the specimen chamber 11, and is connected to a display device 21 and input devices 23 such as a mouse and a keyboard that output signals corresponding to the input operations of the operator. The control computer 22 uniformly controls the operation of the charged particle beam apparatus 10 according to signals output from the input devices 23 or signals generated by a preset automatic operation control process.

[0079] Here, as described above, the control computer 22 performs control of the position related to the object based on the determination result of the image processing computer 30. The control computer 22 includes a communication interface for communicating with the image processing computer 30.

[0080] In addition, the control computer 22 forms an image of the inflow current signal output from the absorption current detector 20 as absorption current image data. Here, the control computer 22 converts the detection amount of the secondary charged particles R detected by the detector 16 while scanning the irradiation position of the charged particle beam into a luminance signal corresponding to the irradiation position, and generates absorption current image data representing the shape of the irradiation object based on the two-dimensional position distribution of the detection amount of the secondary charged particles R. In the absorption current image mode, the control computer 22 generates absorption current image data representing the shape of the needle 18 based on the two-dimensional position distribution (absorption current image) of the absorption current detected while scanning the irradiation position of the charged particle beam and detecting the absorption current flowing in the needle 18. The control computer 22 causes the generated image data to be displayed on the display device 21.

[0081] The display device 21 displays image data and the like based on the secondary charged particles R detected by the detector 16.

[0082] The charged particle beam apparatus 10 can perform imaging of an irradiation object, various processes based on sputtering (such as milling, trimming, etc.), formation of a deposited film, etc. by irradiating while scanning a focused ion beam on the surface of the irradiation object.

[0083] Figure 3 is a top view of a specimen piece Q before being detached from the specimen S, which is formed by irradiating a focused ion beam onto the surface (hatched portion) of the specimen S in the charged particle beam apparatus 10 according to an embodiment of the present invention. Reference numeral F indicates a processing frame based on the focused ion beam, that is, the scanning range of the focused ion beam, and the inside thereof (white portion) indicates a processed region H that is milled by sputtering processing by irradiating the focused ion beam. Reference mark Ref is a reference point indicating the position where the specimen piece Q (retained without being milled) is formed. A deposited film is used to know the approximate position of the specimen piece Q, and a micro hole is used for precise alignment. In the specimen S. The specimen piece Q is subjected to etching processing so that a support portion Qa connected to the specimen S is retained, and the peripheral portions on the side portion and the bottom portion are milled and removed, and is cantilever-supported on the specimen S through the support portion Qa.

[0084] Next, refer to Figure 4 and Figure 5 to describe the specimen piece holder P.

[0085] Figure 4 is a top view of the specimen piece holder P, Figure 5It is a side view. The specimen piece holder P includes a substantially semi-circular plate-shaped base portion 42 having a cutout portion 41, and a specimen stage 43 fixed to the cutout portion 41. As an example, the base portion 42 is formed of a circular plate-shaped metal. The specimen stage 43 has a comb shape and includes a plurality of columnar portions (hereinafter also referred to as struts) 44 that are separated and protrude for transferring the specimen piece Q.

[0086] (Computer for image processing)

[0087] Next, refer to Figure 6 and describe the computer 30 for image processing. Figure 6 It is a diagram showing an example of the structure of the computer 30 for image processing according to the present embodiment. The computer 30 for image processing includes a control unit 300 and a storage unit 305.

[0088] The control unit 300 includes a learning data acquisition unit 301, a learning unit 302, a determination image acquisition unit 303, and a determination unit 304.

[0089] The learning data acquisition unit 301 acquires learning data. The learning data is information for machine learning. The learning data is a set of a learning image and information indicating the position of an object in the learning image. As an example, the objects in the learning image include a specimen piece, a needle, columnar portions provided on the specimen piece holder, etc. Here, the types of objects in the learning image are the same as the types of objects in the determination image. For example, when the types of objects in the learning image are a specimen piece, a needle, or columnar portions, the types of objects in the determination image are a specimen piece, a needle, or columnar portions, respectively.

[0090] Here, in the present embodiment, as the learning image, a SIM image and an SEM image obtained in advance by irradiating an object with a charged particle beam are used. The charged particle beam is irradiated onto the object from a specified direction. In the charged particle beam apparatus 10, the direction of the lens barrel of the charged particle beam irradiation system is fixed, and thus, the direction of irradiating the charged particle beam onto the object is determined in advance.

[0091] As an example, the information indicating the position of the object in the learning image is the coordinates indicating the position of the object in the learning image. The coordinates indicating the position in the learning image are, for example, two-dimensional orthogonal coordinates, polar coordinates, etc.

[0092] The learning images include both the SIM image and the SEM image of the object. The learning images are the SIM image when observing the object in an inclined direction that is inclined by a predetermined angle from the vertical direction with respect to the specimen stage 12 and the SEM image when observing the object in the vertical direction of the specimen stage 12. That is, the learning images include the image when observing the object in the first direction based on the specimen stage 12 and the image when observing the object in the second direction. The second direction is a direction different from the first direction based on the specimen stage 12.

[0093] The learning unit 302 performs machine learning based on the learning data obtained by the learning data acquisition unit 301. The learning unit 302 stores the learned result as a machine learning model M in the storage unit 305. As an example, the learning unit 302 performs machine learning for each type of object in the learning images included in the learning data. Therefore, the machine learning model M is generated for each type of object in the learning images included in the learning data. In addition, the learning unit 302 may not perform machine learning for each type of object. That is, it may perform common machine learning regardless of the type of object. For example, according to the setting input to the control computer 22, it is set in the image processing computer 30 whether the learning unit 302 performs machine learning for each type of object.

[0094] In addition, the machine learning model M includes multiple models. The multiple models included in the machine learning model M are distinguished not only by the set of learning data used to generate the models but also by the machine learning algorithm.

[0095] In addition, in the following description, the object photographed or depicted in the image is sometimes referred to as the object of the image.

[0096] Here, the machine learning performed by the learning unit 302 is, for example, deep learning using a convolutional neural network (CNN: Convolutional Neural Network) or the like. In this case, in the machine learning model M, a multi-layer neural network in which the weights between nodes are changed according to the correspondence between the learning image and the position of the object in the learning image is included. The multi-layer neural network includes an input layer having nodes corresponding to each pixel of the image and an output layer having nodes corresponding to each position in the image. When the brightness values of each pixel of the SIM image and the SEM image are input to the input layer, a set of values representing the positions in the image is output from the output layer.

[0097] The determination image acquisition unit 303 acquires a determination image. The determination image refers to a SIM image and a SEM image output from the control computer 22. The determination image includes an image of the above object. Among the objects in the determination image, there are objects related to the irradiation of the charged particle beam, such as the specimen wafer Q and the used needle 18.

[0098] The determination image is both a SIM image when observing the object from an inclined direction that is inclined by a specified angle from the vertical direction with respect to the specimen stage 12 and a SEM image when observing the object from the vertical direction of the specimen stage 12. That is, the determination image includes an image when observing the object from a first direction and an image when observing the object from a second direction. Here, the first direction is a direction based on the specimen stage 12, and the second direction is a direction different from the first direction based on the specimen stage 12.

[0099] The determination unit 304 determines the position of the object included in the determination image acquired by the determination image acquisition unit 303 based on template matching. Here, in the template matching, the determination unit 304 uses a template T for the object. The template T is pre-made based on an image of the object obtained by irradiating with a charged particle beam. As an example, the template T is stored in the storage unit 305.

[0100] When the template matching fails, the determination unit 304 determines the position of the object included in the determination image acquired by the determination image acquisition unit 303 based on the machine learning model M that has been learned by the learning unit 302.

[0101] Here, the position of the object included in the determination image, for example, includes the pick-up position of the specimen wafer in the SIM image and the SEM image, the position of the tip of the needle in the SIM image and the SEM image, and the position of the columnar part 44 in the SIM image and the SEM image. As an example of the position of the object included in the determination image, the determination unit 304 determines the coordinates of the object in the image.

[0102] In addition, in the present embodiment, as an example, when the object is the specimen wafer Q, the determination unit 304 determines the pick-up position of the specimen wafer Q based on template matching, and when the template matching fails, determines the pick-up position based on the machine learning model M. On the other hand, when the object is the columnar part 44 or the needle 18, the determination unit 304 determines the position of the columnar part 44 or the position of the tip of the needle 18 based on the machine learning model M. Even when the object is other than the specimen wafer Q, such as the columnar part 44 or the needle 18, the determination unit 304 can make a determination based on template matching in the same way as in the case of the specimen wafer Q, and when the determination fails, determine the position of the object based on the machine learning model M.

[0103] For example, it is preset by the user which algorithm to use in the determination of the position of the object.

[0104] The determination based on template matching is an example of the first determination, and the determination based on the machine learning model M is an example of the second determination.

[0105] In addition, the image processing computer 30 can also obtain the template T and the learned machine learning model from an external database, for example. In this case, the control unit 300 may not include the learning data acquisition unit 301 and the learning unit 302.

[0106] Hereinafter, regarding the operation of automatic micro-sampling (MS: Micro-sampling) performed by the control computer 22, that is, the operation of automatically transferring the specimen piece Q formed by processing the specimen S with a charged particle beam (convergent ion beam) to the specimen piece holder P, it is roughly divided into an initial setting process, a specimen piece picking process, and a specimen piece mounting process and will be described in sequence. (Initial setting process)

[0107] Figure 7 FIG. is a diagram showing an example of the initial setting process of the present embodiment. Step S10: The control computer 22 sets the mode and processing conditions. The setting of the mode refers to the setting of the presence or absence of the posture control mode described later according to the operator's input at the start of the automatic timing. The setting of the processing conditions is the setting of the processing position, size, the number of specimen pieces Q, etc.

[0108] Step S20: The control computer 22 registers the position of the columnar portion 44. Here, the control computer 22 sends the SIM image and the SEM image including the columnar portion 44 as the object to the image processing computer 30.

[0109] In the present embodiment, the absorption current image data including the object is a set of the SIM image of the object and the SEM image of the object. That is, the SIM image and the SEM image including the object are a set of the SIM image when observing the object from an inclined direction inclined by a predetermined angle with respect to the vertical direction of the specimen stage 12 and the SEM image when observing the object from the vertical direction of the specimen stage 12.

[0110] The determination image acquisition unit 303 acquires the SIM image and the SEM image as determination images from the image processing computer 30. The determination unit 304 determines the position of the columnar portion 44 included in the determination images acquired by the determination image acquisition unit 303 based on the machine learning model M. The determination unit 304 outputs the position information indicating the determined position of the columnar portion 44 to the control computer 22.

[0111] Here, the determination unit 304 determines the two-dimensional coordinates of the object in the specimen stage 12 based on the SIM image when observing the object in the tilting direction that is tilted by a predetermined angle from the vertical direction with respect to the specimen stage 12. On the other hand, the determination unit 304 determines the two-dimensional coordinates of the object in the plane perpendicular to the tilting direction based on the SEM image when observing the object in the tilting direction that is tilted by a predetermined angle from the vertical direction with respect to the specimen stage 12. The determination unit 304 determines the position of the object as the value of the three-dimensional coordinates based on the determined two-dimensional coordinates in the specimen stage 12 and the two-dimensional coordinates in the plane perpendicular to the tilting direction.

[0112] In addition, the determination unit 304 uses the direction information, which is the information on the direction in which the electron beam irradiation optical system 15 and the focused ion beam irradiation optical system 14 are arranged in the charged particle beam apparatus 10 and the angle between the two, for the calculation of the value of the three-dimensional coordinates. The determination unit 304 reads the direction information stored in the storage unit 305 in advance, or obtains the direction information from the control computer 22.

[0113] Here, in step S20, the object refers to the columnar part 44. In the following processes, except for the case where the object is the specimen piece Q, the process of the determination unit 304 determining the position of the object is the same.

[0114] Here, refer to Figures 8 to 12 to explain the columnar part 44 and the learning image of the columnar part 44 for generating the machine learning model M.

[0115] Figure 8 and Figure 9 are diagrams showing an example of the columnar part 44 of the present embodiment. Figure 8 and Figure 9 The columnar part A0 shown is an example of the design structure of the columnar part 44. Here, Figure 8 is a top view of the columnar part A0, Figure 9 is a side view of the columnar part A0. The columnar part A0 has a structure in which a pillar A01 with a stepped structure is bonded to a base A02.

[0116] Figure 10 are diagrams showing an example of the learning image of the columnar part 44 of the present embodiment. The learning images X11, X12, and X13 are used for learning the position of the columnar part 44. In the learning images X11, X12, and X13, the information indicating the position of the columnar part is shown in the form of a circle.

[0117] Among the learning images X11, X12, and X13, the shapes of the columns A11, A21, and 31 are different respectively. On the other hand, among the learning images X11, X12, and X13, the shapes of the bases A12, A22, and A32 are the same.

[0118] In addition, as an example, the learning images X11, X12, and X13 are learning images for determining the position of the columnar part 44 included in the SIM image and the SEM image when observing the columnar part 44 from the horizontal direction of the specimen stage 12. In Figure 2 the convergent ion beam irradiation optical system 14 and the electron beam irradiation optical system 15 do not face the specimen stage 12 from the horizontal direction of the specimen stage 12, but either the convergent ion beam irradiation optical system 14 or the electron beam irradiation optical system 15 may also face the specimen stage 12 from the horizontal direction, and the learning images X11, X12, and X13 are learning images for determining the position of the columnar part 44 in this case.

[0119] Figure 11 It is a diagram showing an example of the columnar part 44 in which the column of the present embodiment does not have a stepped structure. Figure 11 The shown columnar part A4 is a side view of an example of the design structure of the columnar part 44 in which the column does not have a stepped structure.

[0120] Figure 12 It is a diagram showing an example of the learning image of the columnar part 44 in which the column of the present embodiment does not have a stepped structure. As an example, the learning images X21, X22, and X23 are learning images for determining the position of the columnar part 44 included in the SEM image when observing the columnar part 44 from the vertical direction of the specimen stage 12.

[0121] Among the learning images X21, X22, and X23, the shapes of the columns A51, A61, and 71 are different respectively. On the other hand, among the learning images X21, X22, and X23, the shapes of the bases A52, A62, and A72 are the same.

[0122] In the conventional template matching, when the shapes of the columns are different, it is sometimes impossible to determine the position of the columnar part. On the other hand, since the machine learning model M is generated based on machine learning using the learning image including the base of the columnar part 44, in the machine learning model M, for example, the shape of the base is learned as a feature amount. Therefore, in the charged particle beam device 10, even when the shapes of the columns are different, the accuracy of determining the columnar part is improved.

[0123] The object object of the learning image preferably includes a portion having the same shape among the object objects of a plurality of learning images.

[0124] Return Figure 7 Continue with the description of the initial setting process.

[0125] The control computer 22 registers the position of the columnar portion 44 based on the position information indicating the position of the columnar portion 44 determined by the image processing computer 30.

[0126] In addition, in the learning image of the columnar portion 44, it preferably includes images of the columnar portions at both ends of the specimen stage 43 located in the columnar portion 44. The image processing computer 30 detects the columnar portions at both ends of the specimen stage 43 in the columnar portion 44 and the columnar portions other than the two ends separately based on the machine learning model M generated using the learning data including this learning image. The control computer 22 can also calculate the inclination of the specimen holder P based on the positions of the detected columnar portions at both ends. The control computer 22 can also correct the coordinate value of the position of the object based on the calculated inclination.

[0127] Step S30: The control computer 22 controls the focused ion beam irradiation optical system 14 to process the specimen S.

[0128] (Specimen piece picking process)

[0129] Figure 13 is a diagram showing an example of the specimen piece picking process of the present embodiment. Here, picking means separating and extracting the specimen piece Q from the specimen S by machining or a needle based on a focused ion beam.

[0130] Step S40: The control computer 22 adjusts the position of the specimen. Here, the control computer 22 moves the specimen stage 12 through the stage drive mechanism 13 so that the specimen piece Q as the object enters the field of view of the charged particle beam. Here, the control computer 22 uses the relative positional relationship between the reference mark Ref and the specimen piece Q. After the specimen stage 12 moves, the control computer 22 performs alignment of the specimen piece Q.

[0131] Step S50: The control computer 22 executes the movement of the needle 18.

[0132] Here, referring to Figure 14 , the process for the movement of the needle 18 executed by the control computer 22 will be described. Figure 14 is a diagram showing an example of the movement process of the needle 18 of the present embodiment. Figure 14 Steps S510 to S540 of Figure 13 correspond to step S50 of

[0133] Step S510: The control computer 22 executes a needle movement (coarse adjustment) that moves the needle 18 through the needle driving mechanism 19. Step S520: The control computer 22 detects the tip of the needle 18. Here, the control computer 22 transmits absorption current image data including the needle 18 as an object to the image processing computer 30.

[0134] The determination image acquisition unit 303 acquires a SIM image and a SEM image as determination images from the image processing computer 30. The determination unit 304 determines the position of the needle 18 included in the determination image acquired by the determination image acquisition unit 303 as the position of the object based on the machine learning model M. The determination unit 304 outputs position information indicating the determined position of the needle 18 to the control computer 22.

[0135] Next, the control computer 22 executes a needle movement (fine adjustment) that moves the needle 18 through the needle driving mechanism 19 based on the position information indicating the position of the needle 18 determined by the image processing computer 30.

[0136] Here, with reference to Figures 16 to 19 , the needle 18 and the learning image of the needle 18 for generating the machine learning model M will be described.

[0137] Figure 16 is a diagram showing an example of SEM image data including the tip of the needle 18 in the present embodiment. Figure 17 is a diagram showing an example of SIM image data including the tip of the needle 18 in the present embodiment.

[0138] Figure 18 is a diagram showing an example of the tip of the needle 18 in the present embodiment. In Figure 18 , as an example of the needle 18, the needle B1 is shown when observed from an inclined direction inclined by a predetermined angle in the vertical direction with respect to the specimen stage 12.

[0139] Figure 19 is a diagram showing an example of the learning image of the needle 18 in the present embodiment. The learning images Y31, Y32, and Y33 are used for learning the position of the tip of the needle 18. In the learning images Y31, Y32, and Y33, information indicating the position of the tip of the needle 18 is shown in the form of a circle. In the learning images Y31, Y32, and Y33, the thickness of the tip of the needle is different respectively. On the other hand, in the learning images Y31, Y32, and Y33, the shape of the tip of the needle is the same.

[0140] The thickness of the tip of the actual needle 18 changes due to cleaning. In conventional template matching, when the thickness of the tip of the needle is different, it is sometimes impossible to determine the position of the tip of the needle. On the other hand, since the machine learning model M is generated based on machine learning using a learning image including the tip of the needle 18, in the machine learning model M, for example, the shape of the tip of the needle is learned as a feature quantity. Therefore, in the charged particle beam device 10, even when the thickness of the tip of the needle is different, the accuracy of determining the tip of the needle is improved.

[0141] Here, with reference to Figure 15 , the detailed process of the image processing computer 30 determining the position of the tip of the needle 18 will be described. Figure 15 FIG. is a diagram showing an example of the needle tip position determination process of the present embodiment. Figure 15 The needle tip position determination process shown is executed in Figure 14 step S520.

[0142] Step S5210: The determination unit 304 determines the position of the tip of the needle 18 included in the determination image acquired by the determination image acquisition unit 303 as the position of the object based on the machine learning model M.

[0143] Step S5220: The determination unit 304 determines whether it is possible to determine the position of the tip of the needle 18. When the determination unit 304 determines that it is possible to determine the position of the tip of the needle 18 (step S5220; YES), the position information indicating the determined position of the tip of the needle 18 is output to the control computer 22, and the needle tip position determination process ends. On the other hand, when the determination unit 304 determines that it is impossible to determine the position of the tip of the needle 18 (step S5220; NO), the process of step S5230 is executed.

[0144] A case where it is impossible to determine the position of the tip of the needle 18 is, for example, a case where a part of the specimen piece Q cut off adheres to the tip of the needle 18 and the position of the tip of the needle 18 cannot be accurately determined. Figure 20 FIG. is a diagram showing an example of the specimen piece Q2 attached to the tip of the needle B2 in the present embodiment.

[0145] Hereinafter, a case where it is impossible to determine the position of the tip of the needle 18 may be referred to as an abnormal situation.

[0146] Step S5230: In the current pick-up position determination process, the determination unit 304 determines whether the foreign object detection has been completed. When the determination unit 304 determines that the foreign object detection has been completed (step S5230; YES), the process of step S5240 is executed. On the other hand, when the determination unit 304 determines that the foreign object detection has not been completed (step S5230; NO), the process of step S5250 is executed.

[0147] Step S5240: The determination unit 304 causes the control computer 22 to stop the automatic MS. Here, the determination unit 304 outputs a stop signal for stopping the automatic MS to the control computer 22. After that, the determination unit 304 ends the needle tip position determination process.

[0148] Step S5250: The determination unit 304 determines a foreign object included in the determination image acquired by the determination image acquisition unit 303 based on the machine learning model M. Here, the foreign object means a part of the specimen piece Q attached to the tip of the needle 18.

[0149] Here, refer to Figure 21 , and a learning image for determining an abnormal situation by machine learning will be described. Figure 21 is a diagram showing an example of a learning image for abnormal situations in the present embodiment. In the learning images Y41, Y42, Y43, Y44, Y45, and Y46, a part of the specimen piece (specimen piece Q41, specimen piece Q42, specimen piece Q43, specimen piece Q44, specimen piece Q45, specimen piece Q46) is attached to the tip of the needle (needle B41, needle B42, needle B43, needle B44, needle B45, needle B46).

[0150] Step S5260: The determination unit 304 determines whether the foreign object can be determined. When the determination unit 304 determines that the foreign object can be determined (Step S5260; Yes), it executes the process of Step S5270. On the other hand, when the determination unit 304 determines that the foreign object cannot be determined (Step S5260; No), it executes the process of Step S5240.

[0151] Step S5270: The determination unit 304 causes the control computer 22 to remove the foreign object. Here, the determination unit 304 outputs a control signal for executing the removal of the foreign object to the control computer 22. After that, the determination unit 304 executes the process of Step S5210 again. That is, the determination unit 304 determines the position of the tip of the needle 18 from which the foreign object has been removed.

[0152] The removal of the foreign object means removing a part of the specimen piece Q attached to the tip of the needle 18 by cleaning the needle 18. Figure 22 is a diagram showing an example of the removal of the foreign object in the present embodiment. In Figure 22 the removal of the foreign object shown, a processing frame FR6 for cleaning the needle 18 is provided to remove the foreign object Q6 of the needle B6.

[0153] Return Figure 14Next, the description of the movement process of the needle 18 will be continued. Step S530: The control computer 22 detects the pickup position of the specimen wafer Q. Here, the control computer 22 transmits the SIM image and the SEM image including the specimen wafer Q as the object to the image processing computer 30.

[0154] Here, referring to Figure 23 , the process of determining the pickup position by the image processing computer 30 will be described.

[0155] Figure 23 is a diagram showing an example of the pickup position determination process of the present embodiment. Figure 23 Each process from step S5310 to step S5370 shown corresponds to Figure 14 the process of step S530.

[0156] Step S5310: The determination unit 304 determines the pickup position of the specimen wafer Q included in the determination image obtained by the determination image acquisition unit 303 based on template matching. Here, in the template matching, the determination unit 304 uses the template T stored in the storage unit 305.

[0157] Step S5320: The determination unit 304 determines whether it is possible to determine the pickup position of the specimen wafer Q based on template matching. The determination unit 304 determines that it is possible to determine the pickup position when the score of the template matching is equal to or higher than a specified value.

[0158] When the determination unit 304 determines that it is possible to determine the pickup position (step S5320; Yes), the position information indicating the determined pickup position is output to the control computer 22, and the pickup position determination process ends. On the other hand, when the determination unit 304 determines that it is not possible to determine the pickup position (step S5320; No), the process of step S5330 is executed.

[0159] Step S5330: The determination unit 304 selects the machine learning model M-j for determining the pickup position. Here, the determination unit 304 selects one machine learning model M-j for determining the pickup position from the machine learning models M-i (i = 1, 2, ···, N: N is the number of models) included in the machine learning model M. In the present embodiment, as an example, the determination unit 304 selects, based on a specified order, the machine learning model M-i (i = 1, 2, ···, N: N is the number of models) included in the machine learning model M that has not been selected in the current pickup position determination process. The specified order is, for example, the ascending order of the index i of the machine learning model M-i.

[0160] Step S5340: The determination unit 304 determines the pick-up position based on the selected machine learning model M-j. This determination process is the same as the process in which the determination unit 304 determines the position of the object in the above steps S20 and the like.

[0161] Step S5350: The determination unit 304 determines whether it is possible to determine the pick-up position of the specimen piece Q based on the selected machine learning model M-j.

[0162] When the determination unit 304 determines that the pick-up position can be determined (Step S5350; Yes), it outputs the position information indicating the determined pick-up position to the control computer 22 and ends the pick-up position determination process. On the other hand, when the determination unit 304 determines that the pick-up position cannot be determined (Step S5350; No), it executes the process of Step S5360.

[0163] Step S5360: The determination unit 304 determines whether all the machine learning models M-i (i = 1, 2, ···, N: N is the number of models) included in the machine learning model M have been used. When the determination unit 304 determines that all the machine learning models have been used (Step S5360; Yes), it executes the process of Step S5370. On the other hand, when the determination unit 304 determines that not all the machine learning models have been used (Step S5360; No), it executes the process of Step S5330 again.

[0164] Step S5370: The determination unit 304 stops the automatic MS of the control computer 22. Here, the determination unit 304 outputs a stop signal for stopping the automatic MS to the control computer 22. After that, the determination unit 304 ends the pick-up position determination process.

[0165] Here, refer to Figure 24 and Figure 25 to describe the specimen piece Q and the learning image of the specimen piece Q for generating the machine learning model M.

[0166] Figure 24 is a diagram showing an example of SIM image data of the specimen piece Q including the present embodiment. In Figure 24 as an example of the specimen piece Q, the specimen piece Q71 is shown together with a circle indicating the pick-up position.

[0167] Figure 25This is a diagram showing an example of a learning image of the sample piece Q of the present embodiment. The learning images Z11, Z12, and Z13 are used for learning the position of the front end of the sample piece Q. In the learning images Z11, Z12, and Z13, information indicating the pick-up position of the sample piece Q is shown in the form of a circle. In the learning images Z11, Z12, and Z13, the size and the shape of the surface of the sample piece are different respectively. On the other hand, in the learning images Z11, Z12, and Z13, the shape at the pick-up position of the sample piece is the same.

[0168] The shape of the surface of the actual sample piece is different for each individual. In the conventional template matching, when the shape of the surface of the sample piece is different, it is sometimes impossible to determine the pick-up position of the sample piece. On the other hand, since the machine learning model M is generated based on machine learning using learning images including the pick-up position of the sample piece Q, in the machine learning model M, for example, the shape of the pick-up position of the sample piece Q is learned as a feature amount. Therefore, in the charged particle beam device 10, even when the shape of the surface of the sample piece is different, the accuracy of determining the pick-up position of the sample piece Q is improved.

[0169] In addition, in Figure 23 the step S5330 shown, it is also possible to change the order in the process of selecting one machine learning model M-j for determining the object from the machine learning models M-i (i = 1, 2, ···, N: N is the number of models) included in the machine learning model M. For example, when initially executing the process of the image processing computer 30 determining the object (in Figure 23 an example, the pick-up position), it can be set to the above-mentioned prescribed order, and in the subsequent processes, the order is changed according to whether the object of the previous time can be determined.

[0170] For example, when the object can be determined based on the machine learning model M-k in the process of determining the object of the previous time, the determination unit 304 can also set the order of the machine learning model M-k to the first order among the machine learning models M-i (i = 1, 2, ···, N: N is the number of models) included in the machine learning model M. Or the determination unit 304 can also advance the order of the machine learning model M-k by a prescribed number of positions (for example, one position). In addition, when the object cannot be determined based on the machine learning model M-m in the process of determining the object of the previous time, the determination unit 304 can also set the order of the machine learning model M-m to the last order. Or the determination unit 304 can also make the order of the machine learning model M-m retreat by a prescribed number of positions (for example, one position).

[0171] In addition, in Figure 23In the processing of the object to be determined shown, when it is impossible to determine the object after using all the machine learning models M-i (i = 1, 2, ···, N: N is the number of models) included in the machine learning model M, the learning unit 302 can also include the determination image of the object that cannot be determined in the learning image and re-execute machine learning to update the machine learning model M. In this case, for example, the learning unit 302 adds the determination image of the object that cannot be determined to the learning data, re-executes machine learning, and updates the machine learning model M. The number of determination images of the object that cannot be determined added to the learning data can be multiple.

[0172] Updating the machine learning model M means adding the model obtained as a result of re-executing learning to the machine learning model M. Alternatively, updating the machine learning model M can also mean replacing any one of the multiple models included in the machine learning model M with the model obtained as a result of re-executing learning.

[0173] The timing for the learning unit 302 to update the machine learning model M is, for example, at regular intervals of days. For example, the learning unit 302 updates the machine learning model M every 7 days. In addition, the learning unit 302 can also update the machine learning model M when the image processing computer 30 receives an operation to update the machine learning model M from the user of the charged particle beam apparatus 10.

[0174] The learning unit 302 can also calculate the determination accuracy based on the updated machine learning model M after updating the machine learning model M. In this case, for example, a test image set is pre-stored in the storage unit 305. The test image set is a plurality of images including images of objects of the same type as the object (the pick-up position in the Figure 23 example) included in the determination image. Since the test image set is pre-stored in the storage unit 305, it can also be changed by the user of the charged particle beam apparatus 10.

[0175] For example, the learning unit 302 uses the machine learning model M before the update to cause the determination unit 304 to determine the object included in the image included in the test image set, and calculates the determination accuracy based on the determination result. Next, the learning unit 302 uses the updated machine learning model M to cause the determination unit 304 to determine the object included in the image included in the test image set, and calculates the determination accuracy based on the determination result. The learning unit 302 calculates, for example, the ratio of the images for which the determination of the object is successful to the images included in the test image set as the determination accuracy. When the updated machine learning model M improves the determination accuracy compared to the machine learning model M before the update, the learning unit 302 replaces the machine learning model M stored in the storage unit 305 with the updated machine learning model M. On the other hand, when the updated machine learning model M does not improve the determination accuracy compared to the machine learning model M before the update, the learning unit 302 discards the updated machine learning model M.

[0176] In addition, the machine learning model M can also be generated by the user. In this case, for example, the user operates the image processing computer 30 to generate the machine learning model M. The user prepares learning images in advance. The learning data acquisition unit 301 acquires the learning images prepared in advance by the user. The learning images prepared in advance are generated, for example, by photographing SIM images and SEM images using a charged particle beam device. Here, regarding the learning images prepared in advance, it is preferable to generate them while changing the parameters for the images within the same range as in the case where the charged particle beam device 10 actually generates SIM images and SEM images as determination images. The parameters for the images include contrast, brightness, magnification, focus, and beam conditions, etc.

[0177] When the user prepares learning images in advance, it is not preferable that the proportion of the images of which the types are determined increases among the multiple images included in the learning images. When the user prepares learning images in advance, it is preferable that the learning images include these multiple images in such a way that the number of images of multiple types is equal to each other. Here, the types of images are distinguished by the above-mentioned parameters for the images, for example.

[0178] In addition, the learning images may also include the pseudo-images described later.

[0179] In addition, in the case where the machine learning model M is generated by the user, the user determines the appropriateness of the learning images for machine learning. In this case, when the user determines the appropriateness of the learning images, the user can utilize XAI (Explainable AI). Explain the process of determination by the machine learning model in XAI. The learning unit 302, based on XAI, determines the region used as the feature point indicating the position of the object in the image during the process of the machine learning model M determining the position of the object in the image including the object. The learning unit 302 uses, for example, a method such as Layerwise relevancepropagation (LRP) as XAI. The user visually confirms the region determined by the learning unit 302 as the feature point and determines the appropriateness of the learning images.

[0180] Here, with reference to Figures 26 to 29 , the determination of the appropriateness of the learning images based on XAI will be described. Figure 26 FIG. shows an example of the learning image Y5 of the present embodiment. In the learning image Y5, the images Y51 to Y54 are included. The images Y51 to Y54 can be any images such as SEM images, SIM images, or pseudo-images described later. In the images Y51 to Y54, the needles B41 to B44 are respectively included. In the machine learning model M5 learned based on the learning image Y5, in the images Y51 to Y54, machine learning is performed by imposing the condition that the respective regions R41 to R44 are regions representing the tips of the needles. In Figure 26 , as an example, the shapes of the respective regions R41 to R44 are ellipses.

[0181] Figure 27 FIG. shows an example of the additional image I1 of the present embodiment. The additional image I1 is a learning image to be added to the learning image Y5 and is an object for determining the appropriateness of whether it should be set as the learning image Y5. As an example, the additional image I1 includes an image of a needle. The additional image I1 can be any image such as an SEM image, a SIM image, or a pseudo-image described later.

[0182] Figure 28 and Figure 29 FIG. shows an example of the image in which the feature points are determined in the present embodiment. In Figure 28 , the image O1 is shown, and this image O1 shows, for example, the region R1 used as the feature point by the machine learning model M1 when determining the tip position of the needle included in the additional image I1 shown in Figure 27 . In Figure 29Image O2 is shown, which shows the regions R21 and R22 used as feature points by the machine learning model M2 in the case where, for example, the position of the tip of the needle included in the additional image I1 shown is determined based on the machine learning model M2 which is a machine learning model M. The machine learning model M1 and the machine learning model M2 are each generated by performing machine learning based on the learning image Y5. Figure 27 Based on the image O1, the machine learning model M1 uses the region R1 as a feature point for determination. The region R1 corresponds to the position of the tip of the needle. As described above, in the learning image Y5 shown, the region indicating the tip of the needle is shown, and therefore, there is no need to add the additional image I1 to the machine learning model M1. In this case, the user determines that it is inappropriate to add the additional image I1 to the learning image Y5.

[0183] Figure 26

[0184]

[0185] Based on the image O2, the machine learning model M2 uses the regions R21 and R22 as feature points for determination. The region R21 corresponds to the position of the tip of the needle. On the other hand, the region R22 corresponds to a position other than the tip of the needle. If the additional image I1 is used for the learning of the machine learning model M2 under the condition that the region R22 corresponds to a position other than the tip of the needle, it is expected to suppress the determination of the position other than the tip of the needle shown in the region R22 as the tip of the needle. In this case, the user determines that it is appropriate to add the additional image I1 to the learning image Y5.

[0185] In addition, in the processing of the determination object shown, an example in the following case is described: in the case where the object cannot be determined based on template matching in step S5310, the image processing computer 30 selects one machine learning model M-j for determining the object from the machine learning models M-i (i = 1, 2, ···, N: N is the number of models) included in the machine learning model M, that is, an example in the case of retrying, but not limited thereto. The image processing computer 30 may also perform the determination based on template matching and the determination based on the machine learning model M in parallel, and select the appropriately determined result as the determination result of the object. Figure 23

[0186] Figure 14 Return , and continue the description of the movement process of the needle 18. Step S540: The control computer 22 moves the needle 18 to the detected pickup position.

[0187] The control computer 22 has completed the movement process of the needle 18 as described above.

[0188] Figure 13 Return , and continue the description of the specimen piece pickup process.

[0189] ​Step S60: The control computer 22 connects the needle 18 to the specimen piece Q. Here, the control computer 22 uses a deposited film for the connection.

[0190] Step S70: The control computer 22 processes and separates the specimen S from the specimen piece Q. Here, Figure 30 The figure showing the processing and separation shows the cutting position T1 of the support portion Qa of the specimen S and the specimen piece Q in the SIM image data of the embodiment of the present invention.

[0191] Step S80: The control computer 22 retracts the needle 18. Here, the control computer 22 detects the position of the tip of the needle 18 in the same manner as the movement process of the needle 18 in step S50, and moves the needle 18 for retraction.

[0192] Step S90: The control computer 22 moves the specimen stage 12. Here, the control computer 22 moves the specimen stage 12 through the stage drive mechanism 13 so that the specific columnar portion 44 registered in the above step S20 enters the observation field region based on the charged particle beam.

[0193] (Specimen piece mounting process)

[0194] Figure 31 The figure shows an example of the specimen piece mounting process of the present embodiment. Here, the specimen piece mounting process refers to the process of transferring the extracted specimen piece Q to the specimen piece holder P. Step S100: The control computer 22 determines the transfer position of the specimen piece Q. Here, the control computer 22 determines the specific columnar portion 44 registered in the above step S20 as the transfer position.

[0195] Step S110: The control computer 22 detects the position of the needle 18. Here, the control computer 22 detects the position of the tip of the needle 18 in the same manner as in the above step S520.

[0196] Step S120: The control computer 22 moves the needle 18. Here, the control computer 22 moves the needle 18 to the transfer position of the specimen piece Q determined in step S100 through the needle drive mechanism 19. The control computer 22 stops the needle 18 with a predetermined gap between the columnar portion 44 and the specimen piece Q.

[0197] Step S130: The control computer 22 connects the specimen piece Q connected to the needle 18 to the columnar portion 44. Step S140: The control computer 22 separates the needle 18 from the specimen piece Q. Here, the control computer 22 separates by cutting the deposited film DM2 connecting the needle 18 and the specimen piece Q.

[0198] Step S150: The control computer 22 retracts the needle 18. Here, the control computer 22 separates the needle 18 from the specimen piece Q by a predetermined distance through the needle driving mechanism 19.

[0199] Step S160: The control computer 22 determines whether to perform the next sampling. Here, performing the next sampling means continuing sampling from different locations of the same specimen S. The number of samplings to be performed is set in advance in Step S10. Therefore, the control computer 22 confirms this data and determines whether to perform the next sampling. If it is determined to perform the next sampling, the control computer 22 returns to Step S50 and continues the subsequent steps as described above to perform the sampling operation. On the other hand, when the control computer 22 determines not to perform the next sampling, it ends the series of processes of the automatic MS.

[0200] In addition, in the present embodiment, an example in which the learning data is a set of a learning image and information indicating the position of an object in the learning image is described, but it is not limited thereto. In the learning data, in addition to the learning image, it may also include the following parameter information, which is information indicating the type of specimen, scanning parameters (such as the acceleration voltage of the focused ion beam irradiation optical system 14 and the electron beam irradiation optical system 15), the number of uses since the start of cleaning the needle 18, whether there is a foreign object attached to the tip of the needle 18, and the like.

[0201] In this case, the machine learning model M1 is generated by performing machine learning based on the learning image and the parameter information. In addition, the determination unit 304 obtains the parameter information from the control computer 22 in addition to the image data of the SIM image and the SEM image, and determines the position of the object in the image based on the image data, the parameter information, and the machine learning model M1.

[0202] In addition, the above-described direction information may also be included in the parameter information. When the learning data includes the direction information, the relationship between the learning object and the direction of observing the object (the direction based on the specimen stage 12) is learned to generate the machine learning model M1. Therefore, the determination unit 304 does not need to use the direction information in the determination of the position of the object.

[0203] In addition, as described above, the computer (the control computer 22 in the present embodiment) performs control of the position related to the second object (the columnar portion 44, the needle 18, and the specimen piece Q in the present embodiment) based on the result obtained by determining the position related to the second object (the columnar portion 44, the needle 18, and the specimen piece Q in the present embodiment) according to the model based on machine learning (the machine learning model M1 in the present embodiment) by the image processing computer 30 and the second information including the second image (the SIM image and the SEM image of the columnar portion 44, the needle 18, and the specimen piece Q in the present embodiment). In addition, the image processing computer 30 and the control computer 22 may be configured integrally in the charged particle beam apparatus 10.

[0204] (Second Embodiment)

[0205] Hereinafter, the second embodiment of the present invention will be described in detail with reference to the drawings.

[0206] In the present embodiment, the following case will be described: as the learning image, a pseudo-image generated according to the type of the object or a machine learning model selected according to the type of the object is used.

[0207] The charged particle beam apparatus 10 of the present embodiment is referred to as the charged particle beam apparatus 10a, and the image processing computer 30 is referred to as the image processing computer 30a.

[0208] Figure 32 FIG. is an example showing the structure of the image processing computer 30a of the present embodiment. When comparing the image processing computer 30a of the present embodiment ( Figure 32 ) with the image processing computer 30 of the first embodiment ( Figure 6 ), the learning image generation unit 306a, the classification unit 307a, the machine learning model M1a, and the classification learning model M2a are different. Here, the functions of other structural elements are the same as those in the first embodiment. The description of the functions the same as those in the first embodiment is omitted, and in the second embodiment, the description will be centered on the parts different from the first embodiment.

[0209] The control unit 300a includes, in addition to the learning data acquisition unit 301, the learning unit 302, the determination image acquisition unit 303, and the determination unit 304, a learning image generation unit 306a and a classification unit 307a.

[0210] The learning image generation unit 306a generates a pseudo-image PI as a learning image. In the present embodiment, the pseudo-image PI refers to an image generated based on a SIM image and an SEM image obtained in advance by irradiating an object with a charged particle beam. As an example, the learning image generation unit 306a generates a pseudo-image PI based on a bareware BW and a pattern image PT.

[0211] The bareware BW is an image that shows the shape of the object by removing the surface pattern from the object. The bareware BW is preferably a plurality of images that show the shapes of a plurality of objects with different sizes, contrasts, focuses, etc. The bareware BW is an image obtained by depicting using image software with different SIM images and SEM images.

[0212] The pattern image PT is an image that shows a pattern corresponding to the internal structure of the object. The pattern image PT can be a SIM image or an SEM image obtained by irradiating with a charged particle beam, or an image obtained by depicting using image software.

[0213] The learning image generation unit 306a uses a pseudo-image generation algorithm to apply random noise to the pattern corresponding to the internal structure of the object shown in the pattern image PT, and generates a pseudo-image PI by overlapping it with the bareware BW.

[0214] In the present embodiment, as an example, a case where the learning image generation unit 306a generates a pseudo-image PI as a learning image of the specimen piece Q is described, but it is not limited thereto. The learning image generation unit 306a can also generate a pseudo-image PI as a learning image of the needle 18 and the columnar portion 44. In addition, the learning image generation unit 306a can also generate a pseudo-image PI in the case where a part of the specimen piece Q is attached to the tip of the needle 18 as a learning image of the above abnormal situation.

[0215] In addition, the learning image generation unit 306a can also include a SIM image and an SEM image obtained in advance by irradiating the object with a charged particle beam in the first embodiment in the learning image. That is, the learning image generation unit 306a can use only the pseudo-image PI as a learning image, or can use the pseudo-image PI, the SIM image, and the SEM image in combination as a learning image.

[0216] In machine learning, the learning unit 302 extracts the shape of the surface of the object and the pattern of the internal structure from the learning image generated by the learning image generation unit 306a as feature amounts, and generates a machine learning model M1a.

[0217] Here, with reference to Figures 33 to 35 , a method for generating the pseudo-image PI will be described.

[0218] Figure 33This is a diagram showing an example of the bare part BW of the present embodiment. In Figure 33 the bare part BWs, the bare part BW1, the bare part BW2, and the bare part BW3 are shown as the bare part BW of the sample piece Q. The bare part BW1, the bare part BW2, and the bare part BW3 are images that imitate the shapes of the sample pieces Q of multiple sizes. In addition, images corresponding to the needles 18 are respectively included in the bare part BW1, the bare part BW2, and the bare part BW3 as information indicating the picking positions.

[0219] Figure 34 This is a diagram showing an example of the pattern image PT of the present embodiment. In Figure 34 the pattern image PT, the user sample U1 is shown. The user sample U1 is an image prepared in advance according to the type of the sample piece Q to be processed by the user of the charged particle beam device 10a. In the user sample U1, patterns corresponding to the types of substances constituting the multiple layers are depicted for the sample piece composed of multiple layers.

[0220] Figure 35 This is a diagram showing an example of the pseudo image PI of the present embodiment. In Figure 35 the pseudo image PI, the bare part BWs of Figure 33 and the pseudo images PI1, PI2, and PI3 generated based on the user sample U1 of Figure 34 are shown. The pseudo images PI1, PI2, and PI3 have the patterns of the internal structure shown in the user sample U1 overlapping on the shapes of the sample pieces Q of multiple sizes.

[0221] Return Figure 32 , and continue with the description of the structure of the image processing computer 30a.

[0222] The classification unit 307a classifies the determination image obtained by the determination image acquisition unit 303 based on the classification learning model M2a. The classification learning model M2a is a model for selecting, according to the type of the object, the model for the determination unit 304 to use from among the multiple models included in the machine learning model M1a. Here, the multiple models included in the machine learning model M1a are distinguished not only by the set of learning data used for generating the model but also by the machine learning algorithm.

[0223] The classification learning model M2a, for example, associates the type of each sample piece Q to be processed by the user with the models included in the machine learning model M1a. The classification learning model M2a is pre-generated by machine learning and stored in the storage unit 305.

[0224] Next, with reference to Figure 36, to describe the process of detecting the pickup position of the specimen wafer Q as the operation of the automatic MS of the charged particle beam apparatus 10a using the classification learning model M2a.

[0225] Figure 36 It is a diagram showing an example of the detection process of the pickup position in the present embodiment.

[0226] Step S310: The classification unit 307a classifies the determination image acquired by the determination image acquisition unit 303 based on the classification learning model M2a.

[0227] Step S320: Based on the classification result, the classification unit 307a selects the machine learning model used by the determination unit 304 for determination from the multiple models included in the machine learning model M1a. Additionally, the classification unit 307a may select template matching as the algorithm used by the determination unit 304 for determination according to the classification result.

[0228] Step S330: The determination unit 304 determines the pickup position of the specimen wafer Q included in the determination image acquired by the determination image acquisition unit 303 based on the machine learning model selected by the classification unit 307a. Here, in step S330, the determination unit 304 performs the above-mentioned Figure 23 pickup position determination process.

[0229] Here, the classification by the classification unit 307a is an example of the third determination for selecting the type of determination. Instead of Figure 23 the template matching in step S5310, any one of the multiple models included in the machine learning model M1a can be used to perform the machine learning-based determination. Therefore, according to the result of the third determination for selecting the type (algorithm) of determination, the classification unit 307a selects the type (algorithm) of determination for at least one of the first determination (as an example, the determination in step S5310) and the second determination (as an example, the determination in step S5340).

[0230] Step S340: The determination unit 304 determines whether the pickup position of the specimen wafer Q can be determined. When the determination unit 304 determines that the pickup position can be determined (step S340; YES), it outputs the position information indicating the determined pickup position to the control computer 22 and ends the pickup position determination process. On the other hand, when the determination unit 304 determines that the pickup position cannot be determined (step S340; NO), it performs the process of step S350.

[0231] Step S350: The determination unit 304 stops the automatic MS of the control computer 22. Here, the determination unit 304 outputs a stop signal for stopping the automatic MS to the control computer 22. After that, the determination unit 304 ends the pickup position determination process.

[0232] In addition, in the above-described embodiment, an example of a case where the determination unit 304 performs a second determination when the first determination fails in the determination of the position of the object is described, but it is not limited thereto. The determination unit 304 may also continue to perform the second determination even when the first determination is successful, and determine the position of the object based on both the result of the first determination and the result of the second determination.

[0233] For example, the determination unit 304 may determine the position of the object based on template matching, and then determine the position of the object based on machine learning. When the positions indicated by the respective determination results match, the position indicated by the determination result is determined as the position of the object.

[0234] In addition, the determination unit 304 may also control the position of the object according to the result of the fourth determination selected based on at least one of the result of the first determination and the result of the second determination. A specific example of this case will be described below.

[0235] For example, for the determination of the position of the object, the determination unit 304 may also select the next determination method based on the result of the determination performed last time. When the next determination method is selected based on the result of the determination performed last time for the determination of the position of the object, for example, the determination unit 304 may also perform the second determination first in the next determination when the accuracy of the first determination is lower than the accuracy of the second determination, based on the result of the first determination performed last time and the result of the second determination performed last time.

[0236] In addition, the determination unit 304 may also select the type of determination to be used next based on the type of the first determination and the type of the second determination, based on the type of determination performed last time.

[0237] In addition, the determination unit 304 may also select the type of the first determination to be used for the next determination based on the result of the second determination performed based on the result of the first determination performed last time. For example, the determination unit 304 may also select the type of template to be used in template matching according to the accuracy of the determination based on machine learning performed when template matching fails.

[0238] In addition, the determination unit 304 may also select the type of the second determination for the next determination based on the result of the previous second determination. For example, the determination unit 304 may continue to use a certain type among the types of the second determination in the determinations after the next time until the accuracy of the second determination becomes equal to or lower than a specified value, and change the type of the second determination when the accuracy of the second determination becomes equal to or lower than the specified value. In this case, for example, the determination unit 304 may continue to use a certain model among the multiple machine learning models in the determinations after the next time until the accuracy of the determination based on this model becomes equal to or lower than the specified value, and change the machine learning model when the accuracy of the determination based on this model becomes equal to or lower than the specified value.

[0239] In this way, the determination unit 304 may also control the position of the object based on the result of the fourth determination selected based on at least one of the result of the first determination and the result of the second determination, and information including an image obtained by irradiation with a charged particle beam.

[0240] In addition, an example of the following situation has been described: Figure 36 In the detection process of the shown pick-up position, in step S320, an example of the case of selecting a machine learning model for determination from among the multiple models included in the machine learning model M1a according to the result of classifying the determination image, but it is not limited thereto. The machine learning model for determination may also be selected based on the score calculated for the result of classifying the determination image.

[0241] For example, the classification unit 307a classifies the determination image obtained by the determination image acquisition unit 303 based on the classification learning model M2a, and then calculates a score (referred to as a classification score) for the classification result. The classification unit 307a calculates the classification score, for example, by calculating the posterior probability for the classification result. The classification unit 307a calculates the classification score as a numerical value having a value within a specified range such as 0 to 100 points. When the calculated classification score is equal to or higher than the specified value, the classification unit 307a selects, according to the classification result, the machine learning model for the determination unit 304 to use for determination from among the multiple models included in the machine learning model M1a.

[0242] On the other hand, when the calculated classification score is less than a specified value, the classification unit 307a selects, from among the multiple models included in the machine learning model M1a, not only the machine learning model corresponding to the classification result but also the machine learning model corresponding to a classification similar to the classification result. That is, the classification unit 307a selects multiple machine learning models from among the multiple models included in the machine learning model M1a. The determination unit 304 determines the position of the object included in the determination image for each of the multiple machine learning models selected by the classification unit 307a. The determination unit 304 compares the determination results among the multiple machine learning models. For example, the determination unit 304 calculates a score (referred to as a position determination score) for the determined result and selects the result with the highest position determination score as the determination result of the position of the object. The position determination score is a score for the determination of the position of the object based on the machine learning model M.

[0243] In addition, in order to enable the determination unit 304 to determine the position of the object, a threshold value of the position determination score may be set in advance. At a time before the determination unit 304 determines the position of the object, the user of the charged particle beam apparatus 10 sets the threshold value of the position determination score in advance. In this case, when determining whether the position of the object can be determined, the determination unit 304 determines not only whether the position of the object can be determined but also whether the position determination score is equal to or greater than the threshold value. When the determination unit 304 determines that the position of the object can be determined and the position determination score is equal to or greater than the threshold value, it determines that the position of the object can be determined. Even when the determination unit 304 determines that the position of the object can be determined, when the position determination score is less than the threshold value, it determines that the position of the object cannot be determined.

[0244] Furthermore, when it is known in advance that the position of the object is included in a certain range, a limit may be set on the range of the coordinates indicating the position of the object in the determination result of the position of the object. At a time before the determination unit 304 determines the position of the object, the user of the charged particle beam apparatus 10 sets the range of the coordinates in advance. In this case, when determining whether the position of the object can be determined, the determination unit 304 determines that the position of the object can be determined only when it determines that the position of the object can be determined, the position determination score is equal to or greater than the threshold value, and the coordinates indicating the position of the object are within the range set in advance. Even when the determination unit 304 determines that the position of the object can be determined, when the position determination score is less than the threshold value or the coordinates indicating the position of the object are outside the range set in advance, it determines that the position of the object cannot be determined.

[0245] In addition, in the above-described embodiment, an example in which the charged particle beam apparatuses 10 and 10a include two charged particle beam irradiation optical systems, i.e., a convergent ion beam irradiation optical system 14 and an electron beam irradiation optical system 15, is described, but the present invention is not limited thereto. The charged particle beam apparatus may include one charged particle beam irradiation optical system. In this case, it is preferable that, in the determination image obtained by irradiating with a charged particle beam through the charged particle beam irradiation optical system, for example, in addition to the object being reflected, the shadow of the object is also reflected. Further, in this case, the object is the needle 18.

[0246] The shadow of the needle 18 refers to a phenomenon that occurs when the needle 18 approaches the surface of the specimen piece Q in an inclined direction inclined by a predetermined angle from the vertical direction with respect to the specimen stage 12, and the secondary electrons (or secondary ions) generated from the surface of the specimen piece Q near the needle 18 are blocked from reaching the detector 16. The closer the distance between the needle 18 and the surface of the specimen piece Q, the more significant this phenomenon becomes. Therefore, the closer the distance between the needle 18 and the surface of the specimen piece Q, the higher the brightness value of the shadow in the determination image.

[0247] In addition to determining the position of the tip of the needle 18 as two-dimensional coordinates in the determination image, the image processing computer 30 calculates the distance between the tip of the needle 18 and the surface of the specimen piece Q based on the brightness value of the shadow of the needle 18. Thus, the image processing computer 30 determines the position of the tip of the needle 18 as a value of three-dimensional coordinates based on the determination image.

[0248] In addition, a part of the control computer 22, the image processing computers 30 and 30a in the above-described embodiments can also be implemented by a computer. For example, the learning data acquisition unit 301, the learning unit 302, the determination image acquisition unit 303, the determination unit 304, the learning image generation unit 306a, and the classification unit 307a can be implemented. In this case, it can also be implemented by recording a program for implementing this control function on a computer-readable recording medium and causing a computer system to read and execute the program recorded on this recording medium. In addition, the "computer system" mentioned here is a computer system built in the control computer 22, the image processing computers 30 and 30a, and includes hardware such as an OS and peripheral devices. In addition, the "computer-readable recording medium" refers to a removable medium such as a floppy disk, an optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in a computer system. In addition, the "computer-readable recording medium" can also include: a recording medium that dynamically holds a program for a short time, such as a communication line in the case of transmitting a program via a network such as the Internet or a communication line such as a telephone line; a recording medium that holds a program for a certain period of time, such as a volatile memory inside a computer system that becomes a server or a client in this case. In addition, the above program can be a program for implementing a part of the above functions, and can also be a program that can implement the above functions by combining with a program already recorded in a computer system.

[0249] In addition, a part or all of the control computer 22, the image processing computers 30 and 30a in the above-described embodiments can be implemented as an integrated circuit such as an LSI (Large Scale Integration). Each functional block of the control computer 22, the image processing computers 30 and 30a can form a processor individually, or a part or all of them can be integrated to form a processor. In addition, the method of forming an integrated circuit is not limited to LSI, and can also be implemented by a dedicated circuit or a general-purpose processor. In addition, in the case where an integrated circuit technology that replaces LSI appears due to the progress of semiconductor technology, an integrated circuit based on this technology can also be used.

[0250] As described above, one embodiment of the present invention has been described in detail with reference to the drawings, but the specific structure is not limited to the above structure, and various design changes and the like can be made without departing from the gist of the present invention.

Claims

1. A charged particle beam device that automatically produces a specimen piece from a specimen, wherein, the charged particle beam device includes: a charged particle beam irradiation optical system that irradiates a charged particle beam; a specimen stage that mounts and moves the specimen; a specimen piece transfer unit that holds and conveys the specimen piece separated and extracted from the specimen; a holder fixing stage that holds a specimen piece holder for transferring the specimen piece; and a computer that controls the position based on the result of a second determination for the position made according to the result of a first determination for a position related to an object, and information including an image obtained by irradiation with the charged particle beam, wherein the first determination is made based on template matching using a template for the object, and the second determination is made based on a machine learning model learned from second information including a second image of a second object, or the first determination is made based on the machine learning model, and the second determination is made based on the template matching.

2. The charged particle beam device according to claim 1, wherein, the computer selects the type for at least one of the first determination and the second determination according to the result of a third determination for selecting the type of determination.

3. The charged particle beam device according to claim 1 or 2, wherein, the computer controls the position based on the result of a fourth determination selected based on at least one of the result of the first determination and the result of the second determination, and information including an image obtained by irradiation with the charged particle beam.

Images