JP6309221B2 - Ultra-high-speed review device and ultra-high-speed review method - Google Patents

Description

  The present invention relates to an ultra-high-speed review apparatus and an ultra-high-speed review method that automatically and rapidly acquires images of a plurality of measurement objects on a sample.

  Semiconductor devices are shrinking every year according to Moore's Law, and the minimum feature size is about 20 nm in the latest devices. In order to realize a small feature size, a technique capable of exposing a smaller pattern is required. Conventionally, a laser beam with a wavelength of 193 nm has been used for exposure, but since the optical limit has already been exceeded, in recent years, an exposure technique using EUV light with a wavelength of 13.5 nm has been vigorously advanced.

  This technology originated in Japan and has been researched and developed for practical use by ASML for more than 10 years. The optical system is almost completed, but the current EUV exposure system is required for economical mass production. Since the light source power of 200 W cannot be obtained, the use for 20 nm generation exposure was skipped. Instead, double or triple exposure is used in actual production sites, where 193 nm exposure can be repeated multiple times to achieve even smaller feature sizes. In principle, it is possible to create a tiny pattern by repeating 193 nm lithography a plurality of times, but the alignment accuracy on the order of nm, the roughness coming from the molecular structure of the resist, etc. are limited. In a memory device having a relatively simple structure, it is used to realize 20 nm to 16 nm.

  On the other hand, the logic device also needs to reduce the pattern in order to enhance functions and reduce power consumption. In order to make a logic device by double or triple exposure, a considerably complicated pattern is required. A very complicated calculation is required to divide a pattern formed on one mask into a plurality of mask patterns, but the required result cannot be obtained due to the division calculation diverging depending on the pattern. There is. In order to avoid such complexity, a lithography facilitating technique called complementary lithography is being used. This exposure method is characterized by using a pattern obtained by reducing a complicated logic circuit to a simple L & S pattern. By doing this, the complicated logic device pattern is only the most easily exposed L & S pattern and the process of cutting the line, so no complicated calculation is required, the process is simple, and the calculation is about 8 nm. It is said that it can be realized up to exposure.

  If the life of the 193 nm exposure technique is extended in this way, the number of photomasks used for it will increase rapidly, and more measurement requirements will arise than ever for one mask. For example, since a plurality of masks are used one after another, the absolute position accuracy and alignment accuracy of the lines formed on the mask are more important than ever. Optical correction for exposing a finer pattern becomes complicated and precise. In the production of masks, it is necessary to reduce the defect density more than before, for example, defects that could not be ignored in the past affect the exposure yield, and fine particles are also the object of observation.

  Because of the above requirements, photomasks are required to have more precise defect inspections than before. However, since the photomask defect inspection apparatus uses 193 nm laser light as usual, it is very difficult to reliably find fine defects below the wavelength, and after detecting many defect candidates by forcibly increasing the sensitivity. A method of classifying and removing non-defective items later is performed. A review SEM is used for this purpose. The image acquired by the review SEM is used for automatic defect classification by measuring the length and area, performing a precise comparison with a defect-free image, and performing a printability check by optical simulation.

  The defect candidates found by the defect detection device are stored in the defect position coordinates in the form of KLARF, and the information is sent to the review SEM, and the mask is again observed at a high magnification to obtain an image necessary for defect determination to determine the printability. It is carried out.

  Since the conventional review SEM performs the stage movement of the step-and-repeat method, it takes 3 seconds or more to acquire one point image, and about 1000 points per hour are full. Since the number of defects to be distinguished is enormous, a review capability of 10,000 points or more per hour is required.

  The review SEM also requires a material analysis function. I want to analyze from which location of which equipment, such as fine particles, among the defects classified by image. 2. Description of the Related Art Conventionally, a material analysis method called EDX has been performed by detecting intrinsic X-rays obtained during electron beam irradiation.

  Due to the complexity of photomasks, photomasks are required to have more precise defect inspections than before. However, since the defect inspection apparatus uses 193 nm laser light as usual, it is very difficult to find fine defects below the wavelength with certainty, and after detecting a large number of defect candidates by forcibly increasing the sensitivity, There is a method of classifying and removing items that are not. A review SEM is used for this purpose.

  The defect candidates found by the defect detection apparatus are stored in the defect position coordinates in the form of KLARF and the information is sent to the review SEM, and the mask is again observed at a high magnification to obtain an image necessary for defect determination and printability. Discriminating.

  Since this conventional review SEM performs stage movement of the step-and-repeat method, it takes 3 seconds or more to acquire one point image, and about 1000 points per hour are full. Since the number of defects to be distinguished is enormous, there has been a problem that a review capability of 10,000 points or more per hour is required.

  In addition, the latter described above requires an nm-order analysis resolution due to the miniaturization of the analysis object, but EDX has a major problem that the nm-order analysis resolution cannot be obtained due to its principle. Furthermore, since the analysis speed of particles such as foreign matters is around one point, it takes an order of 1 minute to 1 hour, and the total throughput is determined by the material analysis time. Therefore, a semiconductor that requires analysis of many particle elements and the like. There was a problem that it was not practical at all in the industry.

  The present invention has been devised to solve the above-described problem, and in an ultra-high-speed review apparatus that automatically and rapidly acquires images of a plurality of measurement objects on a sample, Measurement target coordinate data in which the coordinates of the measurement target are set, and grouping means for grouping the measurement targets that are collectively scanned for a plurality of measurement targets on the sample based on the measurement target coordinate data to generate a measurement target group, For each measurement target group grouped by the grouping means, the measurement target group moves as fast as possible between the measurement target groups, while moving the sample within the measurement target group at a constant speed and moving the sample with an electron beam Scanning image acquisition means for acquiring images by repeatedly scanning in a direction perpendicular to each other, and measuring a plurality of measurement objects from the images of each measurement object group acquired by the scanning image acquisition means. The so that to perform respectively.

  At this time, the grouping means divides the entire region on the sample into a mesh shape with a quadrilateral (rectangle) of the scanning width that scans the sample with an electron beam, and the coordinates of the measurement target set in the measurement target coordinate data are After adding a mark to an existing mesh, a mesh provided with a mark included within the stage scanning width in a direction perpendicular to the scanning of the electron beam on the sample is set to an adjacent mesh or a predetermined mesh number N (N Is an integer greater than or equal to 1), and meshes adjacent to each other are grouped together to generate a measurement target group.

  In addition, the scanning image acquisition means sets a predetermined scanning area quadrangle around the coordinates based on the coordinates of each measurement target in the measurement target group, and calculates the scan start point, end point, and center as measurement points. Then, the sample is moved at high speed from the end point of a certain measurement target group to the start point of another measurement target group, and the sample is moved at a constant speed from the start point to the end point to acquire an image.

  In addition, in addition to the coordinates of the measurement target, the size of the measurement target is set in the measurement target coordinate data, and the scanning image acquisition unit sequentially scans each measurement target group grouped by the grouping unit. At this time, the width of a predetermined scanning area quadrilateral to be scanned with an electron beam corresponding to the size of the measurement object is determined and scanned.

  The scanning image acquisition means scans a predetermined scanning area quadrilateral with an electron beam by N times to acquire an image by scanning (oversampling), and then reduces the number of scanning lines to 1 / N. An image is generated.

  In addition, a calibration curve of a plurality of different element numbers and their intensities is calculated in advance by detecting the intensity of secondary electrons or reflected electrons when a substance having a plurality of different element numbers on the sample or reference sample is scanned with an electron beam. Set in the table, and detects the intensity of the secondary electrons or reflected electrons when the predetermined scanning quadrilateral area is scanned with the electron beam, and refers to the calibration curve in the table to determine whether the predetermined scanning quadrilateral area is within the predetermined scanning quadrilateral area. Substance estimation means for estimating the substance at the measurement point is provided.

  The present invention generates a measurement target group summarizing a plurality of measurement targets on the sample based on the measurement target coordinate data, and moves between the measurement target groups as fast as possible for each generated measurement target group, By moving the sample within the measurement target group at a constant speed and repeatedly scanning with an electron beam in the direction perpendicular to the moving direction of the sample, images are acquired and images of many measurement points on the sample are acquired. It becomes possible to acquire at ultra-high speed.

  In the present invention, the calibration curve of the element number and the intensity of secondary electrons or reflected electrons is set in a table in advance, and the intensity of secondary electrons or reflected electrons when a predetermined scanning quadrilateral region is scanned with an electron beam is determined. By detecting and referring to the calibration curve in the table and estimating the substance at the measurement point in the predetermined scanning quadrilateral area, the measurement point on the sample or a substance such as a foreign substance in the minute area in the vicinity can be identified. It is possible to estimate without reducing the throughput.

  The details are as follows.

  (1) Review images can be acquired at high speed, and semiconductor manufacturing productivity is improved. A review image having an SNR similar to that of the stored image can be obtained at an extremely high speed. The total review time can be greatly shortened by acquiring the image by selectively moving the stage only where there is a defect. Since the stage is not stopped, it becomes strong against disturbances that cause image vibration such as vibration and magnetic field. Since the stage is not stopped, the stage does not require a high degree of static accuracy, and costs can be reduced. Since the stage is moved in one direction as much as possible, there is no overload on the stage and the life of the stage is extended.

(2) Surface analysis, which has taken time with conventional EDX analyzers, can be performed instantaneously.
The irradiation current used at the time of analysis is very small on the order of pA and the acceleration voltage is as low as several KV, so that it is possible to prevent the sample from being damaged or contaminated. Since the secondary electron or reflected electron generation efficiency is calibrated on the spot using EDX, the secondary electron or reflected electron generation efficiency and the element can be linked one-to-one regardless of the measurement conditions and columns. Like the secondary electrons, the reflected electrons have a very high spatial resolution and can be analyzed in nm order.

  The present invention generates a measurement target group summarizing a plurality of measurement targets on the sample based on the measurement target coordinate data, and moves between the measurement target groups as fast as possible for each generated measurement target group, Within the measurement target group, the sample is moved at a constant speed and repeatedly scanned with an electron beam in a direction perpendicular to the moving direction of the sample. Achieved high-speed acquisition.

  In the present invention, the calibration curve of the element number and the intensity of secondary electrons or reflected electrons is set in a table in advance, and the intensity of secondary electrons or reflected electrons when a predetermined scanning quadrilateral region is scanned with an electron beam is determined. Detect and refer to the calibration curve in the table to estimate the substance at the measurement point in the predetermined scanning quadrilateral area, and specify the measurement point on the sample or a substance such as a foreign substance in the minute area near it. Realized the estimation without deteriorating.

  FIG. 1 shows a block diagram of an embodiment of the present invention. FIG. 1 irradiates a sample 11 with an electron beam 3 and detects generated secondary electrons and reflected electrons with an electron detector 4 to acquire an image.

  In FIG. 1, an electron gun 1 generates an electron beam 3, which is a so-called thermal emitter in which a cathode made of, for example, W or LaB6 is heated, a ZrO / W TFE emitter, or a cold cathode such as W or CNT. Various emitters such as cathode emitters known in the world are used. There is a high voltage power source 9 for applying a high voltage to the electron gun 1, and an electron beam 3 having a desired energy (energy corresponding to the acceleration voltage) can be taken out.

  The high voltage power source 9 is a known one that generates and applies a high voltage for accelerating electrons emitted from the electron gun 1.

  The electron beam column 2 focuses the electron beam 3 emitted from the electron gun 1 with a focusing lens (not shown) constituting the electron beam column, and further deflects the focused beam on the surface of the sample 11 with the objective lens 10. This is a known device that scans the surface of the narrowed electron beam 3 by means of a detector (scanning coil) 6.

  The electron beam 3 is an electron beam emitted from the electron gun 1.

  The electron detector 4 detects and amplifies secondary electrons and reflected backscattered electrons that are emitted while scanning the surface of the sample 11 while scanning the surface of the sample 11 with a finely focused electron beam 3. An electronic detection device such as MCP.

  The amplifier 5 amplifies (preamplifies) the signal detected by the electron detector 4.

  The deflector 6 is for scanning the surface of the electron beam 3 on the sample 11.

  The electron beam deflecting means 7 supplies a predetermined scanning current to the deflector 6 and controls the deflector 6 so as to scan the surface of the electron beam 3 on the sample 11.

  The electron beam column control means 8 controls the electron beam deflection means 7, the high voltage power source 9, and the like.

  The high-voltage power supply 9 supplies a high voltage to the electron gun 1.

  The objective lens 10 narrows down the electron beam 3 on the sample 11.

  The sample 11 is a sample to be subjected to surface scanning by narrowing down the electron beam 3, and detecting the emitted secondary electrons, reflected electrons, and the like to obtain an enlarged image. For example, a photomask, a semiconductor wafer, a glass substrate Etc.

  The stage 12 is a known one that can fix the sample 11 and precisely control its movement (X direction and Y direction) by a laser interferometer (not shown).

  The vacuum chamber 13 is a room (sealed space) in which the sample 11, the stage 12, and the like are placed in a vacuum.

  The stage controller 14 controls the stage 12, and here, based on a signal from a laser interferometer (not shown), a predetermined coordinate position (for example, a predetermined coordinate on a coordinate system of CAD data, or The movement is controlled to a predetermined coordinate position on the stage coordinate system. The sample is placed on the 11 stage 12, and the movement of the stage 12 means the movement of the sample 11.

  The inspection device 31 is a known device that inspects defects on the sample (photomask) 11 and outputs a defect coordinate file 32 (KLARF data in FIG. 9B described later) here.

  The defect coordinate file 32 is a file in which the coordinates of the defect of the pattern on the sample (photomask) 11 are set, for example, KLARF data in FIG.

  The PC 21 is a personal computer, and here performs various controls, calculations, processes, etc. according to a program. The defect coordinate capturing means 22, grouping means 23, scanning order determining means 24, scanned image acquiring means 25, And material estimation means 26 and the like.

  The defect coordinate capturing means 22 inspects the pattern on the sample (photomask) 11 with the inspection device 31 (for example, inspects a defect by comparing an image acquired by an optical microscope, an electron microscope, etc. with a pattern on CAD data). Then, the defect coordinates are output to the defect coordinate file 32 (see KLARF data in FIG. 9B described later).

  The grouping means 23 groups the coordinates of defect points (measurement points) based on the defect coordinate file (for example, KLARF data in FIG. 9B) acquired by the defect coordinate acquisition means 22 ( This will be described later with reference to FIGS.

  The scanning order determining unit 24 determines the scanning order of the coordinate group (measurement target group) grouped by the grouping unit 23 (described later with reference to FIGS. 9 to 11, for example).

  The substance estimation means 26 estimates the substance of the foreign substance near the measurement point in the scanning region (which will be described later with reference to FIGS. 14 to 18).

  Here, the schematic operation of the configuration of FIG. 1 will be described below.

  (1) The configuration of FIG. 1 includes a vacuum chamber 13, an electron beam column 2 for irradiating an electron beam 3, a high-voltage power source 9, an electron detector 4, and a sample 11 as seen in a normal scanning electron microscope. A stage (XY stage) 12 and stage control means 14 for moving, and a mirror and laser interferometer (not shown) for measuring the stage position in real time are provided. As a drive system of the stage 12, an ultrasonic motor, DC, AC servo motor, stepping motor, linear motor, or the like can be used. A piezoelectric element for performing fine movement may be incorporated as required.

  (2) The ultra-high speed image acquisition disclosed in the present invention is performed by a combination of constant speed travel of the stage 12 and scanning of the electron beam 3 performed in a direction substantially orthogonal thereto.

  For example, when the XY stage 12 moves at a constant speed in the Y-axis direction, the electron beam 3 is scanned at a constant speed in the X-axis direction orthogonal to the traveling direction of the stage 12. In practice, after scanning in a certain direction, a ramp waveform scan is performed so that the original position is rapidly returned. When returning, blanking is performed by a blanking deflector (not shown) so that the surface of the sample 11 is not irradiated with the electron beam 3. The movement of the stage 12 and the scanning of the electron beam 3 are configured to be synchronized, and the surface of the sample 11 is scanned two-dimensionally with the electron beam 3 without any problem by these two continuous operations. The spot size of the electron beam 3 is optimized according to the target pixel size. Before scanning, the intensity of the objective lens 10 is adjusted so that the surface of the sample 11 is irradiated with the electron beam 3 with a desired spot size from optical or image information. During the scanning period, the irradiation current of the electron beam 3 is controlled to be constant. The amount of irradiation current can be determined by collecting all the irradiation currents with a Faraday cup (not shown) and measuring with a precision ammeter.

  Secondary electrons or reflected electrons generated when scanning the surface of the sample 11 are detected by an electron detector 4 such as a scintillator, an electron multiplier, an MCP or a semiconductor detection device, and after current-voltage conversion, the image processing device. An image of the surface of the sample 11 is obtained by converting to the luminance of the image.

  (3) A proportional relationship is established between the moving speed of the stage 12 during image acquisition and the scanning speed of the electron beam 3. For example, when an image of 1000 pixels is acquired in the X-axis direction, the stage 12 is moved in the Y-axis direction at a speed that is 1/1000 of the X-axis scanning speed. When operated in this way, pixels having the same size as X and Y are realized. By changing the proportional coefficient, the X and Y pixel sizes can be changed. The absolute value of the speed of the stage 12 and the scanning speed of the electron beam 3 is determined by the speed at which one pixel is acquired, and is determined by the pixel clock. For example, when an image of 1000 × 1000 pixels is acquired in one second, a 1 MHz clock may be used.

  In order to acquire an image, since the electron beam 3 needs to be narrowed down on the sample surface, autofocus is necessary. In order not to interrupt the continuous image acquisition, when the sample is introduced into the chamber in advance, the value to be just focused on a plurality of representative points on the mask surface is acquired and recorded (prefocus, mapping), and the value is It is desirable to use autofocus by estimating the focus value of the measurement point. For example, the focus value of the intermediate coordinates may be obtained by linear approximation using the just focus value at the start point and end point of the range where the stage is moved once.

  Next, the calculation of the scanning area will be described in detail with reference to FIG. 2 and FIG. 3 based on the defect coordinate data.

  FIG. 2 is a flowchart for explaining the operation of the present invention (scanning area calculation).

  In FIG. 2, S1 mesh-divides the entire mask area with the electron beam scanning width. As shown in FIG. 3, which will be described later, the entire mask region (the entire measurement target region of the sample (photomask) 11) is mesh-divided by a scanning width of the electron beam, for example, a 10-micrometer square (rectangle).

  S2 marks the mesh where the measurement point exists. This is done by marking a mark (marked with a circle in the figure) on the mesh of FIG. 3 where the defect coordinate file 32, for example, the defect coordinate file (KLARF data) 32 of FIG. Set.

  In S3, a mark area (for example, an inscribed quadrilateral area) included in the stage scanning width is obtained. For example, a quadrilateral including a portion (mesh) marked with a circle in FIG. 3 is obtained as illustrated. Here, a case where the marked meshes are adjacent to each other, and a predetermined number of meshes N (N is an integer equal to or greater than 1), where N = 2 are separated from each other, are summarized, and the inscribed inside of the group of meshes is inscribed. The tangent quadrilateral is calculated as a mark area and grouped.

  As described above, the entire measurement area of the sample (photomask) 11 is mesh-divided with a quadrilateral (for example, 10 microns) of the scanning width of the electron beam 3 as shown in FIG. 3, and the defect coordinates in FIG. Marks (marked with circles in FIG. 3) are given to meshes having defect coordinates (measurement points) in the data, and then within the stage scanning width (for example, 100 microns), these marked meshes are adjacent. The meshes and meshes that are adjacent to each other by a predetermined number of meshes N (N = 2 in the case of FIG. 3) are gathered, an inscribed quadrilateral inscribed in the gathered meshes is calculated as a mark area, and defect coordinates ( Measurement points) can be grouped. By setting N to an arbitrary integer, it becomes possible to group further distant meshes (mesh having defect coordinates) together to reduce the number of groups. If N is increased too much, a portion having no defect coordinates (measurement points) is scanned with an electron beam, so the value of N that maximizes the throughput is experimentally determined and set.

  FIG. 3 is an explanatory diagram of the present invention (grouping of measurement points (defect coordinates)).

  FIG. 3 shows a part of the measurement region of the sample (photomask) 11 of FIG. The small mesh is a quadrilateral having a scanning width (for example, 10 microns) scanned with the electron beam 3, and shows a state where the measurement region is divided into meshes. The horizontal width in FIG. 3 indicates the scanning width scanned by the stage 12 and is, for example, 100 microns.

  In FIG. 3, when the defect coordinates in FIG. 9 (b) are the meshes indicated by ○, as described above, the meshes adjacent to the meshes indicated by ○ are separated by N = 2 in this case. When adjacent meshes are also gathered and inscribed quadrilaterals inscribed in these gathered mesh groups are calculated, the grouped mesh groups (defect coordinate group, measurement target group) shown in the figure are obtained. Then, as will be described later, it is possible to scan with the electron beam 3 while moving the stage 12 only within the inscribed quadrilateral, and to collectively acquire images in the quadrilateral area shown in the figure. The other portions move at high speed, and move to the next grouped inscribed quadrilateral, so that it is possible to acquire images of a large number of defect coordinates (measurement points) required at ultra high speed.

  FIG. 4 shows a high-speed measurement flowchart of the present invention.

  In FIG. 4, S11 receives KLARF. The defect coordinate capturing means 22 in FIG. 1 captures (receives) KRARF data (defect coordinate file 32) in FIG. 9B described later.

In step S12, target coordinates are calculated. This is based on, for example, the first defect coordinates (measurement point coordinates) (X11, Y11) of defect number A in FIG. 9B in the KLARF data received in S11. First target coordinates P01 (X01, Y01), P02 (X02, Y02)
・ P11 (X11, Y11), P12 (X12, Y12) of the second target coordinates
(Details will be described later with reference to FIGS. 5 and 6).

  In step S13, the first target coordinate is reached. This reaches the first target coordinates of FIG. 6 by high speed traveling.

  S14 travels at a constant speed. This has reached the first target coordinate (scanning area) in FIG. 6 in S13, and therefore the stage 12 is moved at a constant speed.

  S15 reaches the second target coordinate. This means that the stage is moved at a constant speed in S14, reaches the second target coordinate at the end of the scanning area, and the scanning of the scanning area is completed.

  S16 moves to the next place at a high speed. This reaches the second target coordinates (second target coordinates in FIG. 6) in S15, and the scanning of the scanning area is completed. Therefore, high-speed movement is performed toward the next scanning area.

  As described above, as shown in FIG. 6, the coordinates of the start point of the scanning area (more precisely, the offsets 1 and 2 of FIG. 5A, which are faster (outside the scanning area)), are the first target coordinates. ) Is reached (S13), the scanning area is scanned at a constant speed (S14), and when the end point of the scanning area is reached (S15), the next place (scanning area) is shifted to high speed. It becomes possible to go to. Thus, it is possible to travel at a constant speed in the scanning area, to travel at a high speed between the scanning areas, and to acquire an image in the traveling area with a necessary measurement point (defect coordinate) as a center.

  FIG. 5 is an explanatory diagram (coordinates of the scanning area) of the present invention. This shows the calculation formulas for calculating the coordinates of the first target and the first target of the scanning region based on the target coordinates (coordinates of the measurement points) in S12 of FIG. 4 described above (FIG. 6). Schematically shown).

  In FIG. 5A, the X width and Y width (see FIG. 6) of the scanning region are obtained by the following equation.

X width: L1 = √S + offset1
Y width: L2 = √S + offset2
Here, as shown in FIG. 6, the X width and the Y width of the scanning area centered on the measurement point coordinates (X11, Y11) are the X direction width (L1) and the Y direction width (L2). ). The X direction is a width for scanning by repeatedly deflecting the electron beam, and the Y direction is a width for scanning by moving the stage in a constant direction at a constant speed. Offsets 1 and 2 are widths (running regions) for traveling (scanning) at a constant speed in the traveling region in consideration of the inertia of the stage 12 and the linearity of the starting point of the electron beam 3.

FIG. 5B shows an example in which the coordinates P01 and P02 of the first target are calculated. here,
Start point coordinate P01: X01 = X11− (L1 / 2)
Y01 = Y11− (L2 / 2)
End point coordinate P02: X02 = X11 + (L1 / 2)
Y02 = Y11− (L2 / 2)
It becomes. Here, (X11, Y11) are the coordinates of the measurement point (defect point) (see FIG. 9B). L1 and L2 are L1 and L2 calculated in FIG. The start point coordinates P01 and the end point coordinates P02 are the start point coordinates P01 of the first target coordinates and the end point coordinates P02 of the first target coordinates shown in FIG.

FIG. 5C shows an example in which the coordinates P11 and P12 of the second target are calculated. here,
Start point coordinate P11: X11 = X11− (L1 / 2)
Y11 = Y11 + (L2 / 2)
End point coordinate P12: X12 = X11 + (L1 / 2)
Y12 = Y11 + (L2 / 2)
It becomes. Here, (X11, Y11) are the coordinates of the measurement point (defect point) (see FIG. 9B). L1 and L2 are L1 and L2 calculated in FIG. The start point coordinates P11 and the end point coordinates P12 are the start point coordinates P11 of the second target coordinates and the end point coordinates P12 of the second target coordinates shown in FIG.

  FIG. 6 is an explanatory diagram of the present invention (scanning of the scanning area). FIG. 6 schematically shows a state in which a high-speed movement moves to the first target in the scanning area, travels at a constant constant speed within the scanning area and reaches the second target, and moves at a high speed. . Here, the travel area, the first target (coordinates), and the second target (coordinates) are the travel area of FIG. 5A, the first target of FIG. 5B, and the second target of FIG. 5C. Correspond to each goal. Each symbol corresponds to each symbol in FIG.

  FIG. 7 is an explanatory diagram (stage movement) of the present invention. For the stage movement, the high-speed movement (speed), the first target, the measurement point, the second target, and the high-speed movement (speed) described above with reference to FIGS. .

    -Start: Stop position (home position).

    ・ High-speed movement (speed) indicates a state where the start stop position is shifted to high-speed movement.

    The first target indicates a state in which the first target (first target in FIG. 6) in the target scanning area is reached in the state of high speed movement. The deceleration starts rapidly from this first target state.

    The constant speed movement indicates a state in which the stage is moved at a constant speed after starting to decelerate rapidly from the first target.

    The measurement point is a point (measurement point) at the center of the scanning area, and is a point at which an image is acquired and measured for measurement of defect coordinates and the like.

    The second target is the second target shown in FIG. 6 and indicates a state in which a high-speed movement is made when the end point is reached by moving at a constant speed (or in a state before a predetermined offset).

    The high-speed movement indicates a state from the time when the second target in the scanning area is reached to the time when the high-speed movement is reached and the first target in the next scanning area is reached.

  FIG. 8 is an explanatory diagram (stage speed control) of the present invention. This shows an example of parameters set for controlling the speed of the stage shown in FIG. Here, the setting is made as shown below.

    ・ At the start: Accelerate to V0 at maximum acceleration.

    ・ Decelerate to constant speed V1 at the first target.

    ・ In the second target, accelerate to V0 at the maximum acceleration.

    • Accelerate to V0 with the next first goal.

  With the above parameter setting, when the stage is moved as described above with reference to FIG. 6, the movement of the stage is controlled as shown in FIG.

  FIG. 9 is an explanatory diagram of the present invention (scanning region scanning 2).

  FIG. 9A schematically shows a state in which the four scanning areas are sequentially scanned in the direction of the arrow shown here, in this case, the moving direction of the stage, and FIG. 9B is an example of defect coordinate data. Indicates.

  As shown in FIG. 9B, when the four defect coordinate data of defect numbers A, B, C, and D are set as shown, A, B, C, and D in FIG. In order of D, the stage performs high-speed movement between scanning areas and constant constant-speed movement within the scanning areas.

  As described above, as indicated by the arrows in FIG. 9A, the respective scanning regions centered on the defect coordinates of defect numbers A, B, C, and D, which are the defect coordinate data in FIG. By repeatedly moving the stage at a high speed and moving at a constant speed, it is possible to acquire images of the scanning areas of all defect numbers A, B, C, and D at high speed.

  Note that the defect position coordinates output from the inspection apparatus (defect inspection apparatus) 31 are composed of a plurality of points and scattered throughout the wafer and the mask surface. Defects are concentrated in some places and there are no defects in some places. Alternatively, when an abnormality occurs in a specific process, a macro pattern may be exhibited as a whole. Since an inertial force is acting on the stage 12 that is continuously moving in a certain direction, it takes a lot of energy and time to change the moving direction of the stage 12 greatly. This may take longer than stopping the stage 12.

  Therefore, when the stage 12 is used to travel in a certain direction for as long a straight time as possible, there is a high possibility that the total measurement time can be shortened. For this reason, instead of measuring a large number of defect coordinates output from the defect inspection apparatus 31 as a KLARF file in the same order, the stage 12 passes through as many measurement points as possible in one direction. Therefore, continuous measurement is desirable.

  For this reason, as shown in FIG. 9, when the stage moves in the Y-axis direction, first, among the defect position coordinates described in the KLARF data, the Y-axis coordinates (the progression axis of the stage 12) are in order of size. Then, a list is created by selecting from the entire KLARF data a list in which the X-axis coordinates are within a predetermined range (for example, a range of 100 microns), and measured in the order shown in the figure.

  FIG. 10 is an explanatory diagram of the present invention (scanning region scanning 3). FIG. 10 shows an example in which the measurement target range is divided into a plurality of strips for high-speed automatic review (image acquisition) of the entire mask.

  In FIG. 10, in order to reduce the number of movements of the stage 12 as much as possible and to enable review (image acquisition) as soon as possible, the scanning range in the X-axis direction depends on the defect density. A range on the order of 100 microns is selected. This value is determined in consideration of the limit at which the stage 12 can go straight without sudden movement speed change. When the defect coordinate interval is very long, it may be a wider range exceeding 100 microns.

  First, the entire mask area where the pattern exists is divided into 100-micron units in the X-axis direction to form band-like areas, and defects included in the KLARF data (see FIG. 9B) for each band-like area. By creating a list in which the coordinates are sorted by the size of the Y coordinate, a list can be created in which defect coordinates existing in the mask are divided, for example, every 100 microns.

  For example, the size of a typical photomask is 15 cm square, but the area size where the pattern is actually written is 10 cm square, so if the stage 12 is scanned a maximum of 1000 times, an image of the entire mask can be acquired. Can do. Since one stage scan scan takes about 1 second, the entire image acquisition of one mask (photomask) is completed in a maximum of about 1000 seconds = about 17 minutes.

  In the above, a method for easily shortening the stage moving time has been disclosed. However, as a method for obtaining a trajectory for acquiring all defect images in a minimum time, in addition to the above method, the speed of the stage is used as a parameter. In addition, a method for solving the shortest path problem such as Dijkstra method, Bellman Ford method, or Warshall Floyd method can be used in consideration of inertia force, stop time, acceleration time, number of groups, and the like.

  FIG. 11 is an explanatory diagram of the present invention (scanning area scanning 4). Here, when the number of defects to be observed is smaller than 1000, for example, there are also strip-shaped regions where no defects exist. In that case, it is inefficient to perform stage scanning of the entire mask 11 uniformly.

  In such a case, a skip is performed for a band-like region not including a defect (image acquisition is not performed). For example, if there are only 300 defects in the entire mask, the maximum number of strip-like regions required is 300. In general, there is a possibility that a plurality of defects are included in one strip-shaped region, and therefore the number of strip-shaped regions including defects is smaller than 300. For example, if only 300 strip-shaped regions including defects are scanned on the stage 12, one strip-shaped region can be scanned in about 1 second, and image acquisition of 300 defects is completed in about 5 minutes. To do. This is a very fast review (image acquisition) of 1 second per defect.

In order to acquire an image at a higher speed, it is only necessary to avoid the stage 12 from moving in a place that does not include a defect in one band-like region. For example, when the stage 12 is moved, the defect coordinates are rearranged in advance. At this time, the defect having the maximum Y-axis value or the defect coordinate having the minimum Y-axis value can be extracted. This value is always smaller than the maximum and minimum Y coordinates representing the edges of the pattern existing on the mask. Since there is no defect at a location beyond this coordinate, it is useless to move the stage 12 with respect to a coordinate greater than or less than that. Therefore, when the position of the stage 12 exceeds the coordinates of the extreme end of the defect coordinates included in one belt-like area, the moving direction of the stage 12 is changed to change the edge of the edge included in the next belt-like area. By moving toward the defect coordinates, useless stage movement can be prevented. When the total number of defects affecting the exposure is small, there may be a case where only one point or several points are acquired in one band-like region. In such a case, the movement of the stage 12 is started from the top of the place where the defect exists, and the moving direction of the stage is changed at the bottom. The stage 12 is moved to the next lowest part, and the stage 12 is moved to the highest part. In this way, the image acquisition time can be shortened by selecting only the place where the defect exists and moving the stage 12.
Here, as the driving mode, there are the following, and they are designated as appropriate.

Mode 1: Mode with little folding Mode 2: Shortest traveling distance Mode 3: Shortest time Mode 1 is a mode in which folding of the stage in the scanning direction is small when moving between the groups (measurement target groups) in FIG. Mode 2 is a mode in which the high-speed moving distance between the groups (measurement target groups) in FIG. 11 is the shortest. In mode 3, the fast movement time between the groups (measurement target groups) in FIG. 11 is the shortest mode. Any one of the modes may be designated, combined or designated by actual measurement.

  FIG. 12 is an explanatory diagram (oversampling scan) of the present invention. In image acquisition, for example, in the case of an exposure mask using a wavelength of 193 nm, 1/20 or less of the wavelength can be ignored, so if there is a resolution of 10 nm per pixel with a field of view of 10 microns, the role as a review image acquisition is sufficient. Can be fulfilled. However, depending on the sample, the SNR of the image obtained from the continuously moving stage may not reach the value necessary for image acquisition. In such a case, it is possible to increase the SNR using an oversampling technique in which the number of pixels per unit area is increased N times to acquire an image. For example, when the goal is to finally obtain an image having a size of 10 nm, the stage 12 advances in the Y direction by 2 nm, which is one-fifth the length of a pixel, during one X-axis scanning period. To control. The image obtained in this way is an image having different pixel densities in the vertical direction and the horizontal direction. Finally, image processing such as averaging processing is performed, and five pixels are converted into one pixel along the Y axis. To be. By this image processing, an image having the same X-axis and Y-axis pixel sizes is finally obtained. In this way, for example, an image equivalent to an image that has been accumulated five times can be obtained, and the SNR of the review image can be increased. In other words, when performing one X-axis electron beam scan, the Y-axis is adjusted so that the Y-axis moves by a pixel size of 1 / N, and the finally obtained image has N pixels in the Y-axis direction. If the averaging process is performed so that becomes one pixel, an image having the same quality as the image obtained by performing N accumulations can be obtained even though the image is acquired while moving the stage. Become.

  FIG. 13 is an explanatory diagram (beam size) of the present invention.

  FIG. 13A schematically shows an example of mode 1, and FIG. 13B schematically shows an example of mode 2.

  In the case of mode 1 in FIG. 13A, the beam size is scanned N times fine without changing the beam size, and in the case of mode 2 in FIG. 13B, the beam size is scanned finely by 1 / N. As described with reference to FIG. 12, the images obtained in these modes 1 and 2 are acquired while moving the stage 12 by performing averaging processing so that N pixels become one pixel. Nevertheless, it is possible to obtain an image having the same quality as an image obtained by performing N accumulations, and SNR can be improved.

  FIG. 14 shows another explanatory diagram of the present invention (calibration curve of secondary electron and reflected electron generation efficiency and atomic number). In FIG. 14, the horizontal axis represents the atomic number, and the vertical axis represents the secondary potential and the generation efficiency (reflection efficiency) of the reflected electrons. The solid line is for secondary electrons, and the dotted line is for reflected electrons. As can be seen from the curve in FIG. 14 (calibration curve), the generation efficiency (reflection efficiency) of secondary electrons and reflected electrons increases as the atomic number increases, as indicated by the curves in the figure. Therefore, by preparing a reference sample with a plurality of atomic numbers for the secondary electrons and reflected electrons shown in the figure in advance, and measuring the generation efficiency (reflection efficiency) from each of them and calibrating the curve (calibration curve) shown in the figure, It is possible to change the substance (atomic number) such as a foreign substance in the sample. Details will be sequentially described below.

  FIG. 15 shows another explanatory diagram of the present invention (estimation of unknown substance).

  FIG. 15A shows an example of a photomask, and FIG. 15B shows an estimation example of an unknown substance.

  On the photomask of FIG. 15 (a), there are two strip-like patterns in the longitudinal direction of Cr as shown in the figure. This is because the characteristic X-rays are measured by EDX and the Cr strips are formed from the characteristic X-rays. The pattern is specified, and the base portion is similarly specified as silicon Si by EDX. As another method, a reference sample of a known substance measured in advance next to the photomask (here, Si, Fe, Cr, etc. is placed as described on the left side.

  In FIG. 15B, in the illustrated table, the horizontal axis represents the atomic number, and the vertical axis represents the reflection (secondary) electron generation efficiency (reflection efficiency). Here, in the longitudinal direction specified by EDX in FIG. 15A, the A-shaped rectangular region of Cr in the vertical direction, the rectangular region of B of the base silicon, and the reflection of the rectangular region of unknown foreign matter C (2 Next) Measure the electron generation efficiency, plot A and B in the table, and connect them with a straight line (a curve when there are three or more) to obtain a calibration curve. Next, for the unknown foreign matter C, when the position of the atomic number corresponding to the position shown in the figure corresponding to the reflection (secondary) electron generation efficiency on the calibration curve corresponds to Fe here, (a ) C rectangular region can be estimated as iron Fe.

  Further, as another calibration method, the reflection (secondary) electron generation efficiency in the rectangular region is measured for the reference sample (Si, Fe, Cr, etc.) in FIG. Si, Fe. A calibration curve as shown in the figure is created by plotting the points of Cr, and the reflection (secondary) electron generation efficiency of the rectangular region C in FIG. 15A is measured based on the created calibration curve. Based on this measured value, the atomic number (here, iron) at the corresponding position on the calibration curve created from the reference sample of FIG. 15B is obtained, and the foreign substance in the rectangular region of C is estimated. You may do it.

  Next, the procedure for calculating the calibration curve will be described in detail according to the order of the flowchart of FIG.

  FIG. 16 is a flowchart for explaining another operation of the present invention.

  In FIG. 16, in step S21, mask loading is performed. This is done by loading and fixing the above-described photomask of FIG. 15A on the stage 2 of FIG.

  In S22, EDX measurement is performed on a plurality of types of known materials on the mask. This is because EDX measurement of a plurality of types of known materials on a photomask fixed to the stage 12 of FIG. 1, for example, the photomask of FIG. 15A, that is, a vertical created by Cr of FIG. A characteristic X-ray is measured (energy analysis) with a non-illustrated EDX for the rectangular region A on the band-shaped pattern in the direction and the rectangular region B of the base portion made of silicon, and the rectangular region A is Cr, The rectangular area of B is specified as Si. Note that the rectangular regions A and B do not need to be known materials, and the materials may be specified by measuring characteristic X-rays using EDX.

  S23 measures the reflected (secondary) electron intensity at the same location. This is the same location where the characteristic X-rays were measured in S22. In the example of FIG. 15A, the reflected (secondary) electron intensity of each rectangular area was measured for the rectangular area A and the rectangular area B. To do.

  S24 obtains a calibration curve. This is because, for example, the rectangular region A on the photomask in FIG. 15A is identified as EDX by Cr and the material B by EDX in the rectangular region B, and the material is Si by EDX in each rectangular region of A and B. Since the reflected (secondary) electron intensity was measured, a line plotted in the above-described table of FIG. 15B (the horizontal axis is the atomic number and the vertical axis is the reflected (secondary) electron intensity). The calibration curve shown in the figure is created by connecting the minutes (a curve when there are three or more).

  As described above, for example, the reflected (secondary) electron intensities of rectangular regions of a plurality of substances (Si, Cr, etc.) are measured on the photomask of FIG. 15A, and based on these, (b) of FIG. ) Calibration curve can be calculated. Although the intensity is measured here, the intensity varies depending on the amount of irradiation current, energy, and the like. Therefore, it is better to use a value normalized by the irradiation current.

  In S22, the substances in the rectangular areas A and B on the photomask are known or specified (confirmed) by EDX. However, the present invention is not limited to this, and Si, as a reference sample, is placed next to the photomask in FIG. If a sample composed of Fe, Cr or the like is arranged in advance (or loaded as necessary), a calibration curve can be created as necessary, and the accuracy can be improved. In particular, when identifying a foreign substance that is an unknown substance, a reference sample (rectangular area for a plurality of substances) for a rectangular area of the same size under the same conditions (the same conditions such as electron beam intensity, size, and image acquisition magnification) Measure the reflected (secondary) electron intensity from the same size rectangular area on the foreign material that is an unknown substance on the photomask, create a calibration curve from the reference sample, and based on this calibration curve It is possible to improve the accuracy by estimating which substance (atomic number) from the foreign material.

  FIG. 17 is a flowchart (part 2) illustrating another operation of the present invention.

  In FIG. 17, S31 loads a standard sample. This loads the reference sample shown in FIG. 15A on the stage 12 shown in FIG.

  In S32, the reflected (secondary) electron intensity of a plurality of types of known materials on the mask is measured. This is because reflection (2) of the rectangular region A (rectangular region on Cr) and the rectangular region B (rectangular region on Si) of the known material on the photomass shown in FIG. Next) Measure the electron intensity.

  In S33, the standard sample is unloaded.

  S34 loads a sample. This is done by loading a photomask on which an unknown foreign substance, for example, FIG.

  In S35, the reflected (secondary) electron intensity of the sample is measured. For example, the reflected (secondary) electron intensity is actually measured for the rectangular region C on the unknown foreign matter on the photomask of FIG.

  In S36, elemental analysis is performed. This creates a calibration curve of (b) of FIG. 15 based on the reflected (secondary) electron intensity measured for a rectangular region on a known sample (or the reference sample of (a) of FIG. 15) in S32, Next, an atomic number corresponding to the position of the corresponding reflection (secondary) electron generation efficiency on the calibration curve is obtained based on the reflection (secondary) electron intensity measured in the rectangular region on the unknown foreign substance in S35. Estimate by elemental analysis with the substance of the atomic number (for example, iron).

  As described above, the calibration curve of FIG. 15B is created based on the standard sample (or the reference sample of FIG. 15A), and the reflected (secondary) electrons of the rectangular area on the unknown foreign substance of the mask are created. The atomic number (substance) of the foreign substance can be estimated from the calibration curve based on the intensity. At this time, by selecting and determining in advance an atomic number (substance) that can exist as a foreign substance, any one of the substances selected in advance based on the reflected (secondary) electron intensity of the rectangular region on the foreign substance is selected. It is possible to accurately determine whether the substance is a foreign substance and more accurately estimate the foreign substance.

  FIG. 18 shows another explanatory diagram (high-speed analysis function) of the present invention. Here, the reflected electron image and the secondary electron image have a spatial resolution on the order of nm, unlike the case of specifying the foreign element by EDX. Furthermore, the amount of current required for obtaining the secondary electron image and the reflected electron image is very small compared with the pA order and EDX, and the sample is not damaged or soiled. In addition, the image acquisition speed is extremely high, and one second is not required to acquire one screen. This is a big difference from the fact that EDX requires a large current of nA or more and requires several tens of minutes to one hour to perform a surface analysis of several hundred pixel squares. Since the calibration curve calibrated by the method of FIG. 17 described above can directly estimate the atomic number from the backscattered electron and secondary electron generation efficiency, the elemental surface analysis (instantaneous element surface analysis (secondary electron image or backscattered electron image) ( Estimation) can be performed. The generation efficiency regions corresponding to different elements can be expressed by distinguishing the elements by displaying different colors on the display. For example, as shown in FIG. ) Can be performed with a resolution of nm order.

  The image shown in FIG. 18 is a reflected (secondary) electronic image, and the size thereof is 500 nm square in the vertical and horizontal directions as shown in the figure. On this image, Cr. In the regions indicated by Mo and Fe, the reflected (secondary) electron intensity is measured for the rectangular region of each part on the reflected (secondary) electron image of FIG. 18 based on the calibration curve according to the flowchart of FIG. (Alternatively, the intensity in the rectangular area of the foreign matter is averaged on the image), and the elements are estimated based on the calibration curve. As described above, elements that can exist as foreign substances are determined in advance, and by selecting (estimating) which element is closest to the reflected (secondary) electron intensity, accuracy is ensured. It becomes possible.

  As described above, it is possible to estimate the substance of the minute foreign matter existing on the mask with good accuracy.

1 is a configuration diagram of one embodiment of the present invention. It is an operation explanation flowchart (scanning area calculation) of the present invention. It is explanatory drawing (grouping of a measurement point) of this invention. It is a high-speed measurement flowchart of the present invention. It is explanatory drawing (coordinate of a scanning area | region) of this invention. It is explanatory drawing (scanning of a scanning area | region) of this invention. It is explanatory drawing (stage movement) of this invention. It is explanatory drawing (stage speed control) of this invention. It is explanatory drawing (scanning 2 of a scanning area | region) of this invention. It is explanatory drawing (scan 3 of a scanning area | region) of this invention. It is explanatory drawing (scanning 4 of a scanning area | region) of this invention. It is explanatory drawing (oversampling scan) of this invention. It is explanatory drawing (beam size) of this invention. It is explanatory drawing (relational curve of the generation efficiency of a secondary electron and a reflected electron, and an atomic number) of this invention. It is another explanatory drawing (estimation of an unknown substance) of the present invention. It is another operation | movement description flowchart of this invention. It is another operation | movement description flowchart (the 2) of this invention. It is another explanatory drawing (high-speed analysis function) of this invention.

1: electron gun 2: electron beam column 3: electron beam 4: electron detector 5: amplifier 6: deflector 7: electron beam deflecting means 8: electron beam column control means 9: high voltage power supply 10: objective lens 11: sample ( Photomask)
12: Stage 13: Vacuum chamber 14: Stage control means 21: PC (personal computer)
22: Defect coordinate capturing means 23: Grouping means 24: Scanning order determining means 25: Scanned image acquiring means 26: Material estimating means 31: Inspection apparatus (defect inspection apparatus)
32: Defect coordinate file

Claims (7)

In an ultra-high-speed review device that automatically acquires images of multiple measurement objects on a sample,
Measurement object coordinate data in which coordinates of a plurality of measurement objects on the sample are set; and
Grouping means for grouping a plurality of measurement objects on the sample based on the measurement object coordinate data into one or a plurality of area units that can be scanned with an electron beam by one stage movement to generate a measurement object group When,
For each measurement object group grouped by the grouping means, the measurement object group is moved as fast as possible, and the sample is moved at a constant speed in a scanning region where the measurement object is scanned, And scanning image acquisition means for repeatedly scanning in the direction perpendicular to the moving direction of the sample with an electron beam and acquiring an image,
An ultra-high-speed review device that causes the plurality of measurement objects to be measured from the images of the measurement object groups acquired by the scanning image acquisition unit.   The grouping means virtually divides the entire area on the sample into a mesh shape with a quadrilateral of scanning width that scans the sample with an electron beam, and the coordinates of the measurement target set in the measurement target coordinate data After a mark is added to a mesh on which the mark exists, the mesh added with the mark included in the stage scanning width in the direction perpendicular to the scanning of the electron beam on the sample is adjacent to the mesh, or further, a preset mesh distance 2. The ultra-high-speed review device according to claim 1, wherein meshes adjacent to each other within N (N is an integer of 1 or more) are grouped together to generate a measurement target group.   The scanning image acquisition means sets the scanning region around the coordinates based on the coordinates of each measurement target in the measurement target group, and uses the start point, end point, and center of the electron beam scan as measurement points. The sample is calculated, the sample is moved at a high speed from the calculated end point of the measurement target to the start point of another measurement target, and the sample is moved at a constant speed from the start point to the end point to acquire an image. The super-high-speed review apparatus of Claim 1 or Claim 2. In the measurement object coordinate data, in addition to the coordinates of the measurement object, set the size of the measurement object,
The scanning image acquisition unit determines a width of the scanning region to be scanned with an electron beam corresponding to the size of the measurement target when sequentially scanning each measurement target group grouped by the grouping unit. The ultra-high-speed review device according to claim 1, wherein the ultra-high-speed review device performs scanning.   The scanning image acquisition unit scans the scanning region with an electron beam to obtain N times the number of scanning lines, acquires an image, and then generates an image with the number of scanning lines reduced to 1 / N. The ultrahigh-speed review device according to any one of claims 1 to 4. Calculate the calibration curves for the different element numbers and their intensities in advance by detecting the intensity of secondary or reflected electrons generated when scanning the electron beam with a substance with different element numbers on the sample or reference sample. And set it in the table,
Substance estimation means for detecting the intensity of secondary electrons or reflected electrons generated when the scanning area is scanned with an electron beam and referring to the calibration curve in the table to estimate the substance at the measurement point in the scanning area The ultrahigh-speed review device according to any one of claims 1 to 5, further comprising: In an ultra-high-speed review method that automatically and rapidly acquires images of multiple measurement objects on a sample,
Based on measurement object coordinate data in which coordinates of a plurality of measurement objects on the sample are set, a plurality of measurement objects on the sample are converted into one or a plurality of area units that can be scanned with an electron beam by one stage movement. A grouping step for generating a group of objects to be measured by grouping;
For each measurement target group grouped by the grouping step, the measurement target group moves as fast as possible between the measurement target groups, while moving the sample at a constant speed in a scanning region scanning the measurement target, And scanning with an electron beam in a direction perpendicular to the moving direction of the sample, and a scanning image acquisition step for acquiring an image,
An ultra-high-speed review method characterized by causing each of the plurality of measurement objects to be measured from the image of each measurement object group acquired by the scanning image acquisition step.