JP2005129546A - Method and apparatus for inspecting pattern defect - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a high-resolution and high-speed defect inspection apparatus for inspecting the defect of a pattern on a wafer, foreign particles, residues, steps or the like by an electron beam in a process for manufacturing a semiconductor device. <P>SOLUTION: The defect inspection apparatus is constructed as to store images of a plurality of areas of a semiconductor surface to be inspected by forming images by a image forming lens 11 by reflecting an electron beam (a planar electron beam) at the near surface of a semiconductor to be inspected. The electron beam contains an energy component that can not reach the surface of the semiconductor to be inspected caused by a decelerating electronic field formed for decelerating the electron beam on the surface of the semiconductor 7 to be inspected. The defect inspection apparatus is constituted to detect the presence or absence of defects and positions of the defects in the area by comparing stored images in the plurality of areas. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method and apparatus for inspecting the surface state of a sample (semiconductor device or the like).

  As a method for detecting a defect of a circuit pattern formed on a wafer by a comparative inspection of an image in a manufacturing process of a semiconductor device, a so-called SEM pattern comparative inspection method that scans a point-shaped electron beam is “ JP-A-59-192943 "," J. Vac. Sci. Tech. B, Vol. 9, No. 6, pp. 3005-3009 (1991) "," J. Vac. Sci. Tech. B, Vol. 10, No. 6, pp. 2804-2808 (1992) ”,“ SPIE Vol. 2439, pp. 174-183 ”,“ Japanese Patent Laid-Open No. 05-258703 ”, and the like.

  These techniques are used to detect electrical defects such as minute etching residue and minute pattern defects that are less than the resolution of an optical microscope, and non-opening defects of minute conduction holes. In these techniques, it is necessary to acquire a pattern image at a very high speed in order to obtain a practical inspection speed. In order to secure an S / N ratio of an image acquired at high speed, a beam current that is 100 times or more (10 nA or more) that of a normal scanning electron microscope is used.

  Further, in “JP-A-7-249393”, “JP-A-10-194627”, JP-A-2000-340160, “JP-A-11-108864”, etc., a rectangular electron beam is used. An apparatus for inspecting a semiconductor wafer at high speed by a so-called projection method is described, such as forming an image of reflected electrons, secondary electrons, or electrons reflected without being irradiated on the wafer by forming a reverse electric field by a lens. .

  In addition, a technique for obtaining an image of the outermost surface of the specimen by applying a potential to the specimen and creating a situation in which an electron beam is not irradiated and reflected by the electric field near the surface is known as a mirror microscope (for example, “Rheinhold Godehardt, ADVANCES IN IMAGING AND ELECTRON PHYSICS, VOL.94, p.81-150 ").

JP 59-192943 A JP 05-258703 A JP 7-249393 A Japanese Patent Laid-Open No. 10-1974627 JP 2000-340160 A JP-A-11-108864 J. Vac. Sci. Tech. B, Vol. 9, No. 6, pp. 3005-3009 (1991) J. Vac. Sci. Tech. B, Vol.10, No.6, pp.2804-2808 (1992) SPIE Vol.2439, pp.174-183 Rheinhold Godehardt, ADVANCES IN IMAGING AND ELECTRON PHYSICS, VOL.94, p.81-150

  In the conventional inspection technique using the above-described electron beam, the following problems remain in both the SEM method, the reflected electron and secondary electron projection methods.

  First, with respect to the SEM method, in order to form an image that maintains an inspectable S / N ratio, the electron beam has a larger current than a general SEM. However, since the electron beam is focused into a point and this “point beam” is scanned in a plane (two-dimensionally) on the sample surface, there is a limit to speeding up (reducing the inspection time).

Further, there is a limit to increasing the current of the electron beam due to the brightness of the electron source employed, the space charge effect, and the like. For example, when trying to obtain a resolution of about 0.1 μm, the electron beam current has a theoretical limit of about several hundred nA, and in practice, only about 100 nA can be used. The S / N ratio of an image is determined by the number of electrons used to form an image, that is, the product of the beam current value and the time required for image acquisition. In order to ensure an S / N ratio of an image in which image processing can be normally performed, when the beam current value is 0.1 μm at 100 nA, 100 sec or more is required to inspect an area of 1 cm ² on the sample surface.

  On the other hand, in the projection method, it is possible to irradiate an electron beam having a larger current than in the SEM method at a time, and it is possible to form an image at a time. However, the emission angle distribution of secondary electrons is spread over a wide angle, and the energy is also spread with about 1 to 10 eV. It is stated in "LSI Testing Symposium / 1999 Conference Proceedings, P142" that sufficient resolution cannot be obtained unless most of the secondary electrons are cut when forming an enlarged image of the sample by imaging such electrons. It can be easily judged from the described FIG. This shows the negative sample applied voltage for accelerating the secondary electrons emitted from the sample and the imaging resolution of the secondary electrons. According to this, when the sample applied voltage is -5 kV, the resolution is approximately 0.2 μm.

  Not all of the emitted secondary electrons can be used for image formation. For example, in the calculation of the cited document, a beam having an opening angle of 1.1 mrad or less is used on the image plane after passing through the objective lens. . The total number of secondary electrons within the range of the opening angle is about 10%. Furthermore, the energy width of the secondary electrons used for imaging is calculated at 1 eV. However, the energy width of the emitted secondary electrons is actually emitted with a width of several eV or more. The side skirt exists up to about 50 eV. Of the secondary electrons having such a wide energy distribution, when only those having an energy width of at most 1 eV are extracted, it becomes a fraction.

  As described above, it is difficult to ensure the S / N ratio of the image because the ratio of electrons that can actually contribute to the image formation is low even if an electron beam is an area beam and a large current is applied to form an image collectively. As a result, it is impossible to shorten the inspection time as expected. Similarly, in the case of using reflected electrons, only an emission amount two orders of magnitude smaller than the irradiation beam current can be obtained, and it is difficult to achieve both high resolution and high speed as in the case of secondary electrons.

  In addition, regarding the above-described mirror microscope technique, there has been no report on a technique and an apparatus for applying this technique to inspection of a semiconductor wafer. In addition, there is no restriction on the direction of the trajectory of the electron beam directed at the sample, and the electron beam is irradiated onto the sample with a wide angle. Therefore, the resolution is about sub-μm, which is insufficient to observe the current semiconductor. It was.

  The present invention has been made paying attention to the above points, and an object thereof is to provide a defect inspection method and a defect inspection apparatus for detecting a defect portion in a pattern formed on a wafer with high resolution and at high speed. And

  The object of the present invention can be achieved by the following method.

  That is, a plurality of irradiation regions (area regions) on the sample surface are sequentially irradiated as an “area beam” (planar electron beam) having a two-dimensional spread instead of “point beam”. A negative potential is applied to the wafer. This negative potential is set to such a value that most of the electron beam is returned near the outermost surface of the wafer. Specifically, the negative potential is 0.5 V to 5 V higher than the potential of the electron source. As a result, the electron withdrawn is imaged (hereinafter, the electron withdrawn by the electric field without colliding with the sample will be referred to as “retracted electron” or “mirror electron”). Then, enlarged images of the plurality of irradiation regions are sequentially formed, the enlarged images of the plurality of irradiation regions are converted into electrical image signals, and the image signals for the plurality of irradiation regions are compared with each other. Thereby, it becomes possible to detect the pattern defect about each said irradiation area | region with high resolution and high speed.

  A typical configuration example of the present invention will be described. First, a pattern defect inspection method according to the present invention applies a planar electron beam having a two-dimensional extent to a plurality of irradiations of a sample surface to which a negative potential is applied. Sequentially irradiating a region, forming an image of electrons drawn back near the surface of the sample without colliding with the sample, sequentially forming enlarged electron images of the plurality of irradiated regions, and forming the plurality of irradiations formed It is configured to detect pattern defects formed in the sample by converting an enlarged electronic image of the region into an electrical image signal and comparing the image signals for the plurality of irradiation regions. To do. Here, the planar electron beam is configured to irradiate each of the plurality of irradiation regions so that the traveling direction is aligned substantially parallel to the sample surface and is incident substantially perpendicularly.

  The pattern defect inspection apparatus according to the present invention includes a first electron optical system that irradiates a plurality of irradiation regions on a sample surface with an electron beam from an electron source as a planar electron beam having a two-dimensional spread, Means for generating an electric field such that electrons forming the electron beam irradiated toward the sample are pulled back near the outermost surface of the sample, and images the electrons pulled back from the plurality of irradiation regions to form an image A second electron optical system for forming a magnified image of the plurality of irradiation areas, an image signal detection means for detecting each of the magnified images of the plurality of irradiation areas by converting them into electrical image signals, and Image signal processing means for detecting pattern defects in each irradiation region by comparing image signals of a plurality of irradiation regions on the sample surface, and detecting pattern defects formed in the sample Characterized by being configured.

  The pattern defect inspection apparatus according to the present invention includes an electron beam irradiation optical system that irradiates a plurality of irradiation regions of a sample surface with a two-dimensionally spread electron beam from an electron source, and the surface. Means for generating an electric field such that all or a part of the electrons forming the electron beam are drawn back immediately before they collide with the surface of the sample, and images the electrons drawn back from the plurality of irradiation regions to form an image. An imaging optical system for forming magnified images of a plurality of irradiation regions, and a sample for moving the sample so that the sample is placed and the electron beam is irradiated to each of the plurality of irradiation regions A moving stage, image signal detection means for detecting each of the enlarged images of the plurality of irradiation areas by converting them into electrical image signals, and image signals of the plurality of irradiation areas detected on the sample surface Compare to, and characterized by having an image signal processing means for detecting a pattern defect by determining whether differences exist above a predetermined threshold at each irradiation area.

  The pattern defect inspection apparatus according to the present invention is characterized in that an electron optical system for acquiring an SEM image of a sample is provided at a predetermined position in the moving direction of the sample moving stage.

  The first electron optical system or the electron beam irradiation optical system is disposed between an electron gun that generates an electron beam, a condenser lens that focuses the electron beam emitted from the electron gun, and the condenser lens and the sample. An objective lens and an electron beam deflection mechanism, the condenser lens is used to place the focal point of the electron beam on the focal plane on the electron source side of the objective lens, and the traveling direction is substantially parallel to the sample surface. Then, a planar electron beam incident substantially perpendicularly is formed and irradiated to a plurality of irradiation regions on the sample surface.

  The means for generating an electric field such that electrons irradiated toward the sample are pulled back on the outermost surface of the sample is constituted by, for example, a power source that applies a predetermined negative potential to a conductive sample holder that holds the sample. In addition, in a sample such as a semiconductor having an insulating film on the surface, a pre-charge control is provided in which a second electron gun is provided at a position away from the optical axis of the electron beam for image formation and the sample is irradiated with electrons before image formation. Means. Furthermore, a grid capable of applying a potential between the sample and the second electron gun is provided, and means for irradiating the sample with electrons using the second electron gun while controlling a voltage applied to the grid is provided.

  Further, the sample moving stage can be set to move the sample continuously at substantially the same speed, so that the defect inspection can be speeded up. In this case, it is needless to say that it is necessary to control the electron beam irradiation area on the sample surface to be the same position for a predetermined time by monitoring the position of the sample moving stage. That is, a stage position measuring mechanism that measures the position of the sample moving stage that continuously moves in real time is measured, and a position variation that occurs during the stage movement with continuous movement is measured by the stage position measuring mechanism, and the electronic Feedback is provided to the electron beam deflection mechanism of the beam irradiation optical system so that the positional relationship between the electron beam and the sample moves at a substantially constant speed in a predetermined direction.

  Further, the image signal detection means converts an enlarged electronic image of an irradiation area formed by the second electron optical system or the imaging optical system onto a fluorescent plate to convert it into an optical image, and this optical image Is imaged on the optical image detection element via an optical lens or an optical fiber. Alternatively, the enlarged electron image formed by the optical system may be directly formed on an image sensing element having electron sensitivity. As the image detection element, a charge-coupled element (CCD sensor) or an element (TDI sensor) that integrates and outputs an optical signal input with a time delay can be used. In addition, reading of detection signals from the image detection element is performed in a multi-channel reading method in parallel.

  According to the present invention, a fine pattern defect on a sample surface of a semiconductor device or the like and an electrical defect such as an open, short, and leak are imaged and inspected with high sensitivity, high resolution, and high speed using an electron beam. An inspection method and an inspection apparatus capable of performing the above are realized.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(Example 1)
FIG. 1 shows the minimum components necessary for explaining the operating principle of the present invention. The electron beam emitted from the electron source 1 is converged by the condenser lens 2 and forms a crossover around the beam separator 3 and on the front focal plane of the objective lens. The electron beam is deflected to the optical axis perpendicular to the wafer 7 by the beam separator 3. The beam separator 3 has a deflection action only with respect to the electron beam from above. For example, an ExB deflector in which an electric field and a magnetic field are orthogonal is used. The electron beam deflected by the beam separator 3 is formed by the objective lens 6 into a planar electron beam aligned in a direction perpendicular to the sample (wafer) surface.

  A negative potential that is substantially equal to or slightly higher than the acceleration voltage of the electron beam is applied to the sample (wafer) 7 by the power source 9, and the shape of the semiconductor pattern formed on the surface of the wafer 7 and the state of charging are changed. A reflected electric field is formed. By this electric field, most of the planar electron beam is pulled back immediately before it collides with the wafer 7 and rises with a direction and intensity reflecting the pattern information of the wafer 7.

  The pulled back electron beam is converged by the objective lens 6, and the beam separator 3 has no deflection effect on the electron beam traveling from below, so it rises vertically, and the imaging lens 11 causes the image detection unit to move upward. An image of the surface of the wafer 7 is formed on 103. As a result, a local change in charging potential on the surface of the wafer 7 and a structural difference such as unevenness are formed as an image. This image is converted into an electrical signal and sent to the image processing unit 104.

  In order to detect defects in the semiconductor pattern formed on the wafer 7, the image processing unit 104 compares the image with the surrounding same shape pattern unit or the image of the defect-free unit acquired in advance. Remember where you are. The wafer 7 is placed on a stage (not shown), and the stage is continuously moved by a stage control system 30. The stage control system 30 and the beam control system 28 are linked to move the image acquisition region continuously while finely adjusting the position of the electron beam by a deflector (not shown) as the stage moves.

  When an insulator is present on the surface of the wafer 7, the surface potential of the wafer 7 cannot be determined only by the power source 9. For this purpose, a preliminary charging control device 32 is provided as a function for charging the wafer surface potential to a desired potential. The apparatus controls the charged potential of the surface of the wafer 7 by irradiating the wafer 7 with an electron beam while applying a voltage to the grid electrode adjacent to the surface of the wafer 7. The operation principle and structure will be described later. When the surface of the wafer 7 is charged in advance before the inspection, the image is acquired by passing under the preliminary charging control device 32 to set the inspection area to a desired charging voltage and then directly under the objective lens 6. To do.

  Next, the principle of imaging defects on the wafer surface under the condition that the electron beam is pulled back on the surface of the wafer 7 will be described.

  FIG. 2 schematically shows how the electron beam 201 perpendicularly incident on the equipotential line 205 near the outermost surface of the wafer 7 is pulled back. Due to the defects 202 existing on the surface of the wafer 7, the equipotential lines 205 have a non-uniform shape where the defects exist. Accordingly, the electron beam vertically incident thereon is not pulled back vertically, but is pulled back at an angle as shown and enters the lens 204. The lens 204 shows the operation of the objective lens 6 and the imaging lens 11 with a single equivalent lens. It can be seen from FIG. 2 that when an image is formed on the imaging surface 203 by this lens, the electron beam from the defect 202 portion is concentrated at one place on the imaging surface and the portion becomes brighter than the surroundings. It is possible to detect the presence of a defect and the position of the defect from this image.

  FIG. 3 shows the result of numerical simulation of the equipotential lines 305 near the wafer surface and the electron beam trajectory 306 in the present invention. The pattern cross-sectional portion 304 includes a conductive material portion 302 (white) and an insulating film portion 301 (shaded portion) having a size of 70 nm, and only the central conductive material portion (conductive portion) 303 is 1 V, and other conductive material portions. Was assumed to be 0V. That is, it was assumed that only the central conductive material portion was charged with 1V positive compared with the surroundings because of insufficient conduction with the substrate.

  The energy of the electron beam was set to −1 eV with respect to the wafer potential. That is, when the potential applied to the wafer is −5000V, the energy of the electron beam is calculated to be 4999eV. In this case, it is shown that the electron beam is strongly influenced by the disturbance of the equipotential lines formed by the patterns having different central potentials, and the vertically incident beam is reflected with a large angle.

  FIG. 4 shows the density of the electron beam drawn back by the electric field on the wafer surface. As shown in the “Structure” column in the drawing, the portion arranged in 3 × 3 is a conductive material, and only the central potential is set to a potential different from the surrounding by 1V. The figure below plots the electrons returning from the wafer surface after performing the electron beam trajectory calculation with the energy width (ΔE) of the electron beam being 2 eV in this case. Where the density of dots is high, the electron density is high. It is shown that electrons are concentrated in the center and there is a part with a high electron density, and it is possible to detect a change of 1V of a 70 nm fine pattern by the principle explained in FIG. ing.

  Here, it will be described that the present invention improves the inspection speed by a difference compared to the method of scanning a point beam and the method of imaging secondary electrons. In the case of an image forming apparatus using an electron beam, what limits the speed of image acquisition ultimately results in the S / N ratio required for the image. The S / N ratio of the image is determined by the number of electrons used to form the image. The required S / N ratio is determined by the magnitude of contrast that the defect to be detected causes in the image. That is, if the defect contrast is signal C, the noise needs to be smaller than that. Noise N is defined by the 3σ value of the signal. The σ value is determined by shot noise of the number of irradiated electrons, and is a square root (√S) of the number of electrons S irradiated per pixel. Therefore, the noise N is 3√S.

In the case where the electron beam is made to collide with the wafer 7 as in the prior art and the secondary electrons generated at that time are detected, there is a stochastic process of secondary electron emission from the sample. Assuming that this is a process, the noise N is N = (3√2) √S. For example, if the defect contrast C is 5% of the average signal amount S, C = 0.05 × S, and the noise N needs to satisfy N ≦ 0.05 × S, so that S ≧ 7200. Become. Based on this concept, the inspection time T per 1 cm ² is obtained as follows.
T = (0.01 / x) ² · t
= (1.6e-19 ・ 0.01 ²・ (3√2) ² ) / (I ・ η ・ C ²・ Pix ² ) (1)
T = ((1.6e-19 · (3√2) ² ) / (I · η · C ² )) · (x ² / Pix ² )
Here, t is the time that the electron beam must stay in the same place in order to make the noise smaller than the contrast C. That is, in the case of the SEM type, it is the time for the electron beam probe to irradiate one pixel. In the case of surface beam irradiation, it is time that the beam needs to be directed to a certain point, and this time is called shot time.

Pix is the required resolution, x is the length of one side of the area beam (in the case of SEM, the pixel size, ie, the same as Pix), I is the beam current, and η is the efficiency of electrons that can be used for image formation. On the other hand, in the present invention, since the electron beam does not impinge on the surface of the wafer 7 and is only scattered by the electric field, there is no stochastic process associated with secondary electron emission. Therefore, √2 in the equation (1) is not necessary, and the equation (2) is obtained.
T = (0.01 / x) ² · t
= (1.6e-19 ・ 0.01 ²・ (3) ² ) / (I ・ η ・ C ²・ Pix ² ) (2)
T = ((1.6e-19 · (3) ² ) / (I · η · C ² )) · (x ² / Pix ² )
Here, η and C in each method are estimated. In the case of the SEM type, almost the same number of secondary electrons as the irradiated electron beam are emitted, and almost 100% of them can be taken into the detector, so that η is almost 1.

  On the other hand, in the secondary electron projection method in which the wafer 7 is irradiated with a planar electron beam and the generated secondary electrons are imaged, secondary electrons having a limited vertical direction component among the emitted secondary electrons. If only the image is not formed, the resolution is deteriorated. This will be described with reference to FIGS.

  FIG. 5 shows the resolution of an image with respect to the emission half angle β of secondary electrons or reflected electrons contributing to imaging. That is, the half-open angle of electrons taken into the imaging system. For example, it is shown that the resolution when an image is formed by secondary electrons having an emission angle of 100 mrad or less is about 100 nm. As calculation conditions, the energy of electrons irradiated on the wafer is 500 eV, the wafer surface is under a strong electric field of 5 kV / mm, and the energy width of secondary electrons is 5 eV. The energy distribution of the secondary electrons is spread over 10 eV or more, but only a component of ± 2.5 eV centered on the emission energy 2 eV is used for imaging. This corresponds to about 1/2 of all secondary electrons.

  Further, considering only the elastic scattered electrons as the reflected electrons, the energy width was set to 1 eV. From these figures, it is necessary to set the emission angle to 25 mrad for the secondary electron to have a resolution of, for example, 40 nm. In this case, the probability that the secondary electron is scattered within the emission angle β is about 0.1%. . Assuming that the emission efficiency of secondary electrons (ratio of the number of secondary electrons to the number of irradiated electrons) is about 1, η in the case of the secondary electron imaging type is 1/2 × 0.001 × 1 = 0.0005. .

  On the other hand, in order to obtain a resolution of 40 nm in the case of reflected electrons, the probability of the presence of reflected electrons within the sample emission angle (β) of 80 mrad is 0.2% from FIG. According to the reference document “Image Formation in Low-Voltage Scanning Electron Microscopy, SPIE, Bellingham, p.43, p67, 1993”, the emission efficiency of reflected electrons is as follows. It is about 0.02 to 0.03. Therefore, η in the backscattered electron imaging type is a very small value of 0.002 × 0.025 = 5e−5.

  On the other hand, in the present invention, since the electron beam is bounced vertically upward as it is on a flat wafer surface, the opening angle of the beam is very small (several mrad), equivalent to the angular variation of the irradiation beam. FIG. 7 is an explanatory diagram for further understanding the above description.

  As shown in the left diagram in FIG. 7, secondary electrons are emitted from the sample into the vacuum with a spread of 180 degrees, whereas in the present invention, all the electrons are almost directly above as shown in the right diagram. Therefore, the irradiated electrons can be effectively used as an image. On the other hand, when unevenness or potential distribution exists on the surface, it proceeds upward with a certain angle, not in the vertical direction. In that case, the proportion of electrons that directly contribute to imaging decreases, but the change in angle itself is a factor that forms an image on the wafer surface, so the contrast increases. That is, it is equivalent to increasing the defect contrast C, which is advantageous for defect detection.

  After all, in the present invention, η and C are linked, and if η is limited, C increases accordingly. Strictly speaking, the type of pattern differs depending on the type of defect, but here, the signal that can be detected as an image in the defective portion is 1/2 of all electrons. The remaining half of the signal contributes to the contrast. Therefore, η = 0.5 and C = 0.5.

  The above is summarized as a table shown in FIG. In this case, FIG. 8B shows the relationship between the beam current and the inspection time. This relationship was calculated with Pix = 40 nm. It can be seen that the present invention can be inspected in an overwhelmingly short time compared to other methods.

  Next, the configuration of an embodiment of the present invention will be described in detail. FIG. 9 shows the configuration of an inspection apparatus according to an embodiment of the present invention. The inspection apparatus according to the present embodiment is roughly composed of an electron optical system 101, a sample chamber 102, an image detection unit 103, an image processing unit 104, and a control unit 105. Each part will be described below.

  First, the electron optical system 101 will be described. The acceleration electron beam emitted from the electron source 1 to which a negative high potential is applied by the acceleration power source 23 is converged by the condenser lens 2 and irradiates the diaphragm 4 having a rectangular opening. As the electron source 1, a Zr / O / W type Schottky electron source was used. A uniform planar electron beam having a large current beam (for example, 1.5 μA) and an energy width of 1.5 eV can be stably formed. Then, it is deflected in the direction of the wafer 7 by the beam separator 3. The beam separator 3 is for separating the optical path of the incident electron beam from the electron source 1 and the mirror electron beam from the sample. The condenser lens 2 forms a crossover on the front focal plane of the objective lens 6. Further, the arrangement of the diaphragm and the lens is optimized so that an image of the diaphragm 4 is formed on the surface of the wafer 7 by the objective lens 6.

  As a result, a planar electron beam which is oriented in a direction perpendicular to the surface of the wafer 7 and whose orbits of the electrons are aligned substantially in parallel and which is shaped into the aperture shape of the diaphragm 4 is formed. The size of the rectangular aperture on the aperture 4 is, for example, 100 μm square, and this is reduced to 1/2 by the objective lens 6 so that a 50 μm square planar electron beam can be obtained on the surface of the wafer 7. . This planar electron beam can be moved (or scanned) to an arbitrary position on the surface of the wafer 7 by the irradiation system deflector 5.

  There is no problem as long as the front focal plane of the objective lens and the crossover position cannot be perfectly matched but are within a certain allowable range. Also, the size of the crossover is ideally zero, but in reality it has a finite size due to the aberration of the electron gun and condenser lens. If this size is within a certain allowable range, there is no problem. In the electron optical system in which the position of the crossover is accurately controlled and the aberrations of the electron gun and condenser lens are sufficiently reduced, the spread of the sample incident angle can be suppressed to 0.5 mrad or less. This incident angle spread is one of the factors that determine the resolution of the magnified image of the sample surface by mirror electrons, and is expressed by the following equation.

r0 = β ² · Zm (3)
Here, r0 is the resolution determined by the spread of the incident angle, β is the maximum incident half angle, and Zm is the distance in which the electric field that pulls back the electrons is generated.

  In this embodiment, β is 0.25 mrad and Zm is 5 mm. If this is substituted into the equation (3), r0 becomes 0.3 nm, and it can be seen that this embodiment does not affect the resolution. Therefore, the beam current can be further increased as necessary.

  Note that even if the resolution is about 30 nm, it is considered sufficient for detecting defects in semiconductors. Therefore, if Zm is 5 mm, β can be tolerated to 2.4 mrad. In this case, there is a considerable margin in the displacement of the position of the crossover between the front focal plane of the objective lens and the size of the crossover.

Assuming that the beam opening half-angle at the front focal plane is α, the focal length of the objective lens is f, the crossover positional deviation is Δf, and the radius of the planar electron beam is X, the following equation is established: Δf = f · β / α (4)
α = X / (2f) (5)
From Expressions (4) and (5), for example, when the focal length f of the objective lens is 10 mm and the size X of the planar beam is 40 μm, there is no problem even if the deviation Δf of the crossover position is displaced by about 10 mm. When this is converted into the beam diameter at the front focal plane, it is about 40 μm. In any case, it can be seen that sufficient resolution can be obtained by arranging the electron beam crossover in the vicinity of the front focal plane of the objective lens.

  Here, the beam separator 3 will be briefly described. The beam separator 3 deflects the electron beam emitted from the electron source 1 in the direction of the wafer 7, while the mirror electrons drawn back from the wafer 7 are not in the direction of the electron source 1 but in the direction of the imaging lens 11. To deflect. A deflector using a magnetic field is optimal for the deflector having such an action. This is because the direction of the deflection action by the magnetic field differs depending on the incident direction of electrons.

  Further, as will be described later with reference to FIG. 10 as Example 2, in the case of an optical system in which the optical axis of the imaging lens and the optical axis of the objective lens 6 are arranged in a straight line, an electric field and a magnetic field are made to go straight, The ExB deflector is used that makes the mirror electrons go straight and deflects only the electron beam from above.

  A negative potential slightly higher (absolute value) than that of the electron source 1 is applied to the wafer 7 and the wafer (sample) moving stage 8 by the power source 9. Specifically, it is preferable to set the negative potential by 0.5 to 5V. If the negative potential is too high, the resolution of the image deteriorates. In addition, if the potential is too small, a slight change in surface irregularities or potential is imaged as extremely strong contrast, making it difficult to detect only the truly necessary defects.

  The electron beam directed perpendicular to the surface of the wafer 7 is decelerated in front of the wafer 7 by the negative potential and pulled back upward by the electric field on the surface of the wafer 7. As described above, the electrons reflect the information on the surface of the wafer 7. The mirror electrons are focused by the objective lens 6 and deflected in the direction of the imaging system deflector 10 and the imaging lens 11 by the beam separator. Then, the state of the surface of the wafer 7 is formed as an electronic image by the imaging lens 11. By magnifying and projecting this electronic image onto the fluorescent plate 15 by the magnifying lenses 13 and 14, a fluorescent image (microscopic image) reflecting the pattern on the surface of the wafer 7 and the charged state can be obtained.

  In order to improve the contrast and resolution of this electronic image, a contrast aperture 12 can be inserted into the crossover surface. With this contrast aperture 12, the resolution and contrast of the image can be enhanced by removing electrons that are greatly deviated from the vertical direction when pulled back by the surface electric field of the wafer 7.

  In the image forming principle of the present invention, the sensitivity for detecting subtle differences in the charging of the wafer surface and the resolution of the image are determined by the energy width of the planar electron beam. FIG. 10 shows the result of comparison by simulation. Assuming the same pattern as in FIG. 4, the images obtained for two cases where the energy width (ΔE) of the electron beam is 2 eV and 4 eV were compared.

  According to this result, when an energy width of 4 eV is assumed, the contrast of the portion having a different potential at the center of the pattern is not recognized. Considering the progress of semiconductor miniaturization, it is necessary to detect a difference in charging potential of about 1 V as a defect in a fine pattern as shown in FIG. Therefore, it can be seen that the energy width of the electron beam used in the present invention is desirably 2 eV or less.

  As described above, in this embodiment using the Zr / O / W Schottky electron source, there is no problem because the energy width is 1.5 eV. For example, when an electron source having a larger energy width is used, an energy filter is provided on the optical path of the electron beam, and the electron is not emitted from the electron source until the final image is formed. The energy width needs to be 2 eV or less. Although it is desirable to provide the energy filter between the electron source and the wafer 7, the same effect can be obtained by performing energy filtering on the mirror electrons from the wafer 7.

  In the present invention, the electron beam does not collide with the wafer 7. Therefore, in principle, even if an insulating film exists on the surface of the wafer 7, the surface is not charged. Therefore, if inspection is performed in a state where charging is not performed, the only defects that can be detected are shape defects (those whose shape is different from the normal part).

  However, in semiconductor pattern defect inspection using an electron beam, it is called a so-called electrical defect, such as a continuity failure, a thing to be insulated is short-circuited, or a leakage current is larger than a normal part. A method has been used in which an object is charged by irradiation with an electron beam and detected by a potential contrast of an SEM image generated by a difference in potential.

  In order to detect such defects with high sensitivity, the present invention includes a preliminary charging control device that irradiates an electron beam dedicated to charging control in advance before acquiring an inspection image. If the inspection is performed after the wafer 7 is charged to a predetermined potential in advance by this apparatus, not only a shape defect but also an electrical defect such as a conduction failure portion can be detected. Hereinafter, this operation and configuration will be described.

  FIG. 11 is a diagram for explaining the operating principle of the preliminary charging control device. The electron source 41 is an electron source that emits a high-current electron beam from a surface having a certain extent (several hundred μm to several tens of mm). For example, an electron source in which carbon nanotubes are bundled, a tungsten filament thermoelectron source, a LaB6 electron source, or the like can be used. A voltage extraction voltage is applied to the extraction grid 42 by the extraction electrode 48 to emit the electron beam 43 from the electron source 41. The electron beam passes through the control grid 44 and irradiates the insulating film 46. As a result, secondary electrons 45 are emitted.

  The secondary electrons have an energy of about 2 eV with respect to the surface potential of the insulating film 46. If the surface of the insulating film is equal to the potential of the substrate 47, the irradiation energy of the electron beam is the voltage of the accelerating power supply 49, and this voltage is set to a value such that the secondary electron emission efficiency is 1 or more. In general, the insulating film material for a semiconductor device may be 500V. At this time, since the secondary electron emission efficiency is larger than 1, the surface of the insulating film is positively charged.

  Since the control power supply 50 is connected to the control grid 44 so that an arbitrary positive or negative voltage can be applied, the potential of the insulating film surface becomes more positive than the set potential of the control grid 44, and the secondary voltage When electrons are drawn back to the surface of the insulating film, the positive charging of the surface of the insulating film stops. At this time, the charging potential on the surface of the insulating film is stabilized at a positive potential slightly lower (about 2 V) than the potential of the control grid. The reason why it is not equal to the potential of the control grid is that secondary electrons have energy. Based on the principle described above, the potential on the surface of the insulating film 46 can be controlled by the potential of the control grid 44.

  FIG. 12 shows a configuration of a preliminary charging control device using a carbon nanotube electron source. The electron source 41 is held in a vacuum state by an insulator 51 so that a potential can be applied. The control grid 44 is arranged facing the wafer 7, and the extraction grid 42 extracts electrons from the electron source 51.

  FIG. 13 shows a configuration of a preliminary charging control device using a LaB6 electron source. When a LaB6 electron source is used for a microscope, a Wehnelt electrode is used to form a crossover immediately after electron emission. In this case, since the light source does not need to be small, an extraction electrode 42 'is provided instead.

  In the sample chamber 102, a wafer 7 is placed on a sample moving stage 8 that can move in a two-dimensional (X, Y) direction. As described above, most of the electron beam is placed on the wafer 7 by the power source 9. A negative potential is applied so as not to collide with. A stage position measuring device 27 is attached to the sample moving stage 8 to accurately measure the stage position in real time. This is for acquiring images while continuously moving the stage 8. For the stage position measuring device 27, for example, a laser interferometer is used.

  In addition, an optical sample height measuring device 26 is attached in order to accurately measure the height of the surface of the semiconductor sample (wafer). For example, a system in which light is incident on an area to be inspected on the wafer surface from an oblique direction and the height of the wafer surface is measured from a change in position of the reflected light can be used. In addition, the sample chamber 102 is also provided with an optical microscope 31 used for positioning the inspection region.

Next, the settling time of the sample moving stage 8 will be described. If the movement method of stage 8 is the step and repeat method, the settling time of stage 8 is msec.
Since an order is required, even if the image S / N ratio is improved and the image acquisition time is shortened, it takes time to move the stage and the inspection time cannot be shortened. Therefore, the moving method of the stage 8 is a continuous moving method in which the stage is always moving at a substantially constant speed. This eliminates the restriction of the inspection time due to the stage settling time. However, if the stage 8 is continuously moving, the stage 8 also moves during one shot, which is the time required to form an image at the same location, and the irradiation position on the sample surface changes. End up. Therefore, the irradiation electron beam is caused to follow the movement of the stage 8 by the irradiation system deflector 5 so that the irradiation position does not change during one shot. Further, when viewed from the electron optical system which is a stationary coordinate system, the electron beam irradiation position moves, so that the image 12 formed by the imaging lens 11 also moves. In order to prevent this movement, the imaging system deflector 10 is operated in conjunction with the irradiation system deflector 5.

  Next, the image detection unit 103 will be described. For image detection, a fluorescent plate 15 for converting an enlarged image of the scattered electron image 12 into an optical image and an optical image detection element (for example, a CCD element) 17 are optically coupled by an optical fiber bundle 16. Thereby, an optical image on the fluorescent plate 15 is formed on the light receiving surface of the optical image detection element 17. The optical fiber bundle 16 is a bundle of thin optical fibers equal to the number of pixels. An optical lens may be used in place of the optical fiber bundle 16, and an optical image on the fluorescent plate 15 may be formed on the light receiving surface of the optical image detection element (CCD) 17 by the optical lens. Electrodes 300 and transparent electrodes 301 are provided on both surfaces of the fluorescent plate 15, and the transparent electrode 301 side applies a positive high voltage between the two electrodes to prevent scattering of the electron beam. The optical image detection element (CCD) 17 converts the optical image formed on the light receiving surface into an electrical image signal and outputs it. The output image signal is sent to the image processing unit 104 where image signal processing is performed.

  Next, the reading time of the image detection element (CCD) will be described. In this embodiment, the charges accumulated in the CCD 17 can be read out in parallel from the 128-channel reading port at a reading speed of 8 Mlines / sec. The number of pixels per line per channel is 8, and the required readout time per line is 125 nsec. Therefore, the required readout time per pixel is 125 nsec / 8 (pixel) = 16 nsec. On the other hand, when the reading of image data from the CCD is a one-channel method, reading at a very high speed is required, which is difficult to realize.

  In this embodiment, the readout port for image data from the CCD is divided into 128 channels, and the parallel readout is performed on these 128 channels, so that the required readout time per pixel is 16 nsec, and a sufficiently realizable readout speed. It is said. This is schematically shown in FIG.

Since the number of channels for reading image data from the CCD 17 is 128 channels and there are 8 pixels × 1024 lines for each channel, the time required to read one piece of image data from the CCD is about 125 μsec. That is, the image signal of one shot area can be taken in 125 μsec. If the pixel size is 50 nm and the one shot area is 50 μm square, the time required for inspection per 1 cm ^{2 of the} sample surface area is 5 sec.

As described above, a speed increase of 80 times can be achieved as compared with the time required for inspection of about 400 sec per 1 cm ² of the sample area according to the conventional method when inspecting at a pixel size of 50 nm. In this embodiment, the time required for the inspection is determined by the signal reading speed from the CCD element. Therefore, if a higher-speed data reading method in the CCD element is realized in the future, further inspection speed is expected. it can.

  The image processing unit 104 includes image signal storage units 18 and 19, a calculation unit 20, and a defect determination unit 21. The image storage units 18 and 19 store images of adjacent portions of the same pattern, and both the images are calculated by the calculation unit 20 to detect different locations of both images. This result is determined as a defect by the defect determination unit 21 and the coordinates thereof are stored. The captured image signal is displayed on the monitor 22 as an image.

  Operation commands and operation conditions of each part of the apparatus are input / output from the control computer 29 in the control unit 105. The control computer 29 is preliminarily inputted with various conditions such as an acceleration voltage at the time of generating an electron beam, an electron beam deflection width / deflection speed, a sample stage moving speed, and an image signal capturing timing from the image detection element. The beam control system 28 receives a command from the control computer 29, generates a correction signal based on the signals from the stage position measuring device 27 and the sample height measuring device 26, and the electron beam is always irradiated to the correct position. Thus, a correction signal is sent to the objective lens power supply 25 and the scanning signal generator 24. The stage control system 30 receives a command from the control computer 29 and controls the sample moving stage 8.

  Next, an actual inspection procedure will be described. First, an alignment method using the optical microscope 31 and an electron beam image will be described. The wafer 7 is placed on a wafer moving stage (XY-θ stage) 8 and moved below the optical microscope 31. An optical microscope image of the surface of the wafer 7 is observed by the monitor 22, and an arbitrary pattern appearing in the center of the screen is stored. At this time, the pattern to be selected needs to be a pattern that can be observed on the electron beam image.

  Next, rotation correction is performed by the wafer movement stage 8 so that the circuit pattern on the surface of the wafer 7 is parallel or orthogonal to the stage movement direction using the optical microscope image. At the time of rotation correction, an optical image of an arbitrary pattern portion in an arbitrary chip of a circuit pattern on the surface of the wafer 7 at a certain stage position is captured and displayed on the monitor 22, and marking is given to an arbitrary position in the display screen. Then, the optical image signal is stored in the storage unit 18.

  Next, the stage 8 is moved in the x direction or y direction by a distance of several chips of the circuit pattern on the surface of the wafer 7, and an optical image of the same pattern portion as the tip in the new chip is captured and displayed on the monitor 22. Let Then, after marking the part corresponding to the previous marking part, the new optical image signal is stored in the storage unit 19. Next, the arithmetic unit 20 compares the optical image signals stored in the storage units 18 and 19 and calculates the amount of positional deviation of the marking portion between the two images. The rotation angle error of the wafer 7 is calculated from the positional deviation amount of the marking portion and the stage movement amount between both images, and the rotation angle is corrected by rotating the stage 8 correspondingly. The above rotation correction operation is repeated several times so that the rotation angle error becomes a predetermined value or less.

  Furthermore, the circuit pattern on the surface of the wafer 7 is observed using the optical microscope image, and the position of the chip on the wafer and the distance between the chips (for example, the repetition pitch of the repeating pattern such as a memory cell) are measured in advance. The value is input to the control computer 29. Then, the inspected chip on the surface of the wafer 7 and the inspected area in the chip are set on the optical microscope image of the monitor 22. The optical microscope image can be observed at a relatively low magnification, and even when the circuit pattern on the surface of the wafer 7 is covered with a transparent film such as a silicon oxide film, the substrate can be observed. Therefore, the layout of the in-chip circuit pattern can be easily observed, and the inspection area can be easily set.

  Next, the wafer 7 is moved under the electron optical system. Therefore, an electron beam image of an area expected to include the inspected area previously set on the optical microscope image is acquired. At this time, the region to be inspected is placed in one shot region. Also on the electron beam image, the stage 8 is moved so that the pattern of the previously marked portion appears on the same screen as that marked on the previous optical microscope image. As a result, it is possible to make a correspondence between the electron beam irradiation position and the optical microscope observation position in advance before starting the inspection, and to calibrate the image acquisition position. Then, on this electron beam image, the same operation as that previously performed on the optical microscope image is performed. As a result, it is possible to easily confirm and align the observation position using the optical microscope and adjust the electron beam irradiation position.

  Furthermore, after performing a certain degree of rotation correction, it is possible to perform more accurate rotation correction using an electron beam image that has a higher resolution than an optical microscope image and can obtain a high-magnification image. Furthermore, using this electron beam image, it is possible to confirm and correct the region to be inspected or the same pattern region with high magnification and high accuracy. However, if the entire surface (or part) of the semiconductor wafer 7 is covered with an insulator, the charge potential on the insulator surface may not be equal to the substrate potential, so call it before image acquisition. It is necessary to control the charging voltage on the surface by the charge control device 32.

  When the setting of the inspection conditions is completed, a part of the inspection area on the surface of the semiconductor wafer 7 is converted into an electron beam image under exactly the same conditions as the actual inspection conditions, and an image depending on the material and shape of the inspection area is obtained. Brightness information and its variation range are calculated and stored in a table. Then, a determination condition for determining whether or not the pattern portion in the inspection region actually imaged and detected in the subsequent inspection process is defective is determined.

  When the setting of the inspection area and the defect determination condition is completed by the above procedure, the inspection is actually started. At the time of inspection, the stage 8 on which the sample (semiconductor wafer) 7 is mounted continuously moves at a constant speed in the x direction. Meanwhile, the electron beam irradiates the same irradiation area (area area) on the surface of the wafer 7 for each shot for a certain shot time (in this embodiment, 50 μsec or more). Since the stage 8 is continuously moving, the electron beam is deflected and scanned by the irradiation system deflector 5 following the movement of the stage 8.

  The irradiation region or irradiation position of the electron beam is constantly monitored by a stage position measuring device 27, a sample height measuring device 26 and the like provided on the stage 8. The monitor information is transferred to the control computer 29 so that the amount of misalignment is grasped in detail, and the misalignment amount is accurately corrected by the beam control system 28. As a result, accurate alignment required for pattern comparison inspection can be performed at high speed and with high accuracy.

  Further, the surface height of the semiconductor wafer 7 is measured in real time by means other than the electron beam, and the focal lengths of the objective lens 6 and the imaging lens 11 for irradiating the electron beam are dynamically corrected. As means other than the electron beam, for example, there is an optical height measuring device 26 by a laser interference method, a method of measuring a change in position of reflected light, or the like. Thereby, an electron beam image always focused on the surface of the inspection area can be formed. Further, the warpage of the wafer 7 is measured in advance before the inspection, and the above focal length correction is performed based on the measurement data so that it is not necessary to measure the surface height of the wafer 7 during the actual inspection. May be.

  An electron beam is directed toward the surface of the wafer 7, and an enlarged optical image of a desired inspection area (area area) on the surface of the wafer 7 is formed on the fluorescent plate 15 by mirror electrons. The enlarged optical image is converted into an electrical image signal by the CCD element 17 and the image signal is taken into the image processing unit 104. Then, it receives an instruction from the control computer 29 and stores it in the storage unit 18 (or 19) as an electron beam image signal for the area region corresponding to the electron beam irradiation position given by the control unit 28.

  In the case of performing a pattern inspection between adjacent chips A and B having the same design pattern formed on the surface of the semiconductor wafer 7, first, an electron beam image signal for a region to be inspected in the chip A is captured. And stored in the storage unit 18. Next, an image signal for the inspection region corresponding to the above in the adjacent chip B is captured and stored in the storage unit 19, and at the same time, compared with the stored image signal in the storage unit 18. Further, an image signal for the corresponding inspection area in the next chip C is acquired and stored in the storage unit 18 by overwriting, and at the same time, the storage for the inspection area in the chip B in the storage unit 19 is stored. Compare with image signal. Such an operation is repeated, and comparison is performed while sequentially storing image signals for corresponding inspection regions in all inspection chips.

  In addition to the above method, it is also possible to adopt a method in which the electron beam image signal of a desired inspection region for a standard non-defective product (having no defect) is stored in the storage unit 18 in advance. In that case, the inspection area and the inspection condition for the non-defective sample are previously input to the control computer 29, the inspection for the non-defective sample is executed based on these input data, and the acquired image for the desired inspection area is obtained. The signal is stored in the storage unit 18. Next, the wafer 7 to be inspected is loaded on the stage 8 and the inspection is executed in the same procedure as before.

  Then, the acquired image signal for the inspection region corresponding to the above is taken into the storage unit 19, and at the same time, the image signal for the sample to be inspected and the image signal for the non-defective sample previously stored in the storage unit 18 Compare Thereby, the presence / absence of a pattern defect in the desired inspection region of the inspection object sample is detected. The standard (non-defective) sample may be a wafer that is known to be free of pattern defects in advance, different from the sample to be inspected, or may have no pattern defects in advance on the surface of the sample to be inspected. A known area (chip) may be used. For example, when a pattern is formed on the surface of a semiconductor sample (wafer), misalignment failure between the lower layer pattern and the upper layer pattern may occur over the entire wafer surface. In such a case, if the comparison object is a pattern in the same wafer or the same chip, the defect (defect) generated over the entire wafer surface is overlooked.

  However, according to this embodiment, the image signal of the area that is known to be non-defective (no defect) is stored in advance, and the stored image signal is compared with the image signal of the inspection target area. Such a defect that has occurred over the entire wafer surface can be detected with high accuracy.

  Both image signals stored in the storage units 18 and 19 are respectively taken into the calculation unit 20, where various statistics (specifically, average values of image density) are calculated based on the already determined defect determination conditions. , Statistics such as variance), difference values between neighboring pixels, and the like are calculated. Both image signals subjected to these processes are transferred to the defect determination unit 21 and compared there to extract a difference signal between the two image signals. These difference signals are compared with the already determined and stored defect determination conditions to determine the defect, and the image signal of the pattern area determined to be a defect and the image signal of the other area are separated.

  Presence or absence of pattern defects by forming an image reflecting the potential and shape information of the surface of the wafer 7 by the inspection method and inspection apparatus described so far, and comparing and inspecting the image signal for the corresponding pattern region Can be detected. As a result, the inspection can be performed at a very high speed as compared with the conventional inspection apparatus using an electron beam.

(Example 2)
In Example 1, since the area of the one-shot electron beam irradiation region is as large as 50 μm × 50 μm, there is a problem that distortion occurs in the peripheral portion of the enlarged image of the semiconductor sample, and the uniformity of the beam current density in the irradiation region. May cause problems. When image distortion or current density non-uniformity occurs in a fixed manner, it can be corrected by changing the fiber strand arrangement of the optical fiber bundle 16. Further, although correction can also be performed by weighting the acquisition sensitivity of image signals and image processing, if they fluctuate with time, it is difficult to cope with these methods.

  In this embodiment, the one-shot irradiation area is set to 5 μm square, and problems of distortion and non-uniformity of current density are prevented from occurring in the one-shot irradiation area. The irradiation electron beam current is 1 μA per shot. At this time, when the electron imaging efficiency η is 0.5, the irradiation time t per shot is 0.18 μsec from the above equation (1). After irradiating one irradiation area (5 μm square) with a shot time of 0.18 μsec, the electron beam is moved onto the next adjacent irradiation area (5 μm square) by the irradiation system deflector 5. In this way, the irradiation position is moved one after another, and the entire range of 100 μm in the x direction and 100 μm in the y direction is irradiated with 20 × 20 = 400 shots.

  At this time, an enlarged image is obtained on the CCD element 17 for each shot at a position corresponding to the electron beam irradiation position at that time, and the CCD element is moved according to the movement of the electron beam irradiation position by scanning the electron beam. The magnified image position obtained also moves. This is shown in FIG.

  A CCD element 17 having 1024 × 1024 pixels was used. One pixel on the CCD element corresponds to a 50 nm square area on the surface of the wafer 7, and therefore, one shot irradiation area (5 μm square) on the wafer 7 surface is 100 × on the light receiving surface of the CCD element. The area is 100 pixels (corresponding to 1/100 of the entire CCD light receiving surface). The entire light-receiving surface of the CCD element can cover a 50 μm square area on the sample surface. Therefore, 0.18 (μsec) × 100 (shot) = 18 (μsec) is required to obtain an enlarged image of a 50 μm square region on the sample surface.

  As described above, when an image of a 50 μm square area on the surface of the wafer 7 is formed on the CCD in 18 μsec, the image signal accumulated in the CCD is stored in the image storage unit 18 as a digital signal. In order to acquire an image signal of the next adjacent region on the sample surface, it is necessary to move the stage 8 by 50 μm. For this stage movement, as in the case of the first embodiment, the stage 8 is continuously moved at a constant speed. At that time, the irradiation system deflector 5 performs the deflection scanning by following the movement of the stage 8 by the irradiation system deflector 5 so as to make the stage 8 stand still with respect to the irradiation electron beam. As a result, the dead time generated when the stage 8 is moved and stopped is reduced to zero.

  In the follow-up scanning of the irradiation electron beam to the continuous movement of the stage 8, a deflection correction signal is calculated by referring to a signal from the stage position measuring device 27 in the beam control system 28, and this deflection correction signal is used as the irradiation system. The deflector 5 is sent to control the deflection of the irradiation electron beam. Furthermore, corrections related to distortion of the magnified specimen image due to the electron beam, position drift, and the like are also superimposed on the deflection correction signal to correct these. In addition, the imaging system deflector 10 is also operated in conjunction with the irradiation system deflector 5 so that the position of the magnified sample image on the CCD is not affected by the beam position movement due to the stage following. This eliminates wasted time due to moving the stage and enables high-speed and high-precision inspection. Note that image processing and the like for defect inspection after that are the same as those in the first embodiment.

When the inspection proceeds according to the procedure described above, the time T required to sequentially form enlarged images per 1 cm ^{2 of} the sample surface on the CCD becomes 0.72 sec. On the other hand, as in the case of the first embodiment, since 125 μsec is required to read out one image (image with respect to the sample surface of 50 μm square) from the CCD, 5 sec is required per 1 cm ^{2 of} the sample surface area. Since image formation and image signal readout in the CCD element are performed in parallel, the time required for inspection is the longer of the time required for image formation and the time required for image signal readout. In this embodiment, the time required for reading out an image signal is longer than the time required for image formation, and is 5 seconds per cm ^2. Therefore, the time required for inspection per sample surface area of 1 cm ² in this embodiment is 5 seconds.

  In the above description, the case where the one-shot electron beam irradiation region is fixed to a size of 5 μm square is illustrated, but the size of the electron beam irradiation region can be changed according to the pattern repetition pitch on the surface of the semiconductor wafer 7. You may be able to do it. As described above, in this embodiment, the one-shot electron beam irradiation area is set smaller. Therefore, even if some distortion occurs at the connection between the irradiation areas, the same distortion always occurs at the same location, and the appearance of distortion on the two images to be compared with each other is also equal. Therefore, the problem of erroneous detection due to distortion is eliminated. Thereby, highly reliable pattern defect inspection can be realized.

(Example 3)
In this embodiment, a time accumulation type CCD sensor is used as an element for converting a sample surface image into an electric signal. This element is called a TDI sensor and is generally used in an optical inspection apparatus. The rest is the same as in the case of the second embodiment. The operation concept of this TDI sensor will be described with reference to FIG.

  In the TDI sensor, the electric charge generated according to the intensity of the light received in each light receiving area is moved to the line in the x direction, and at the same time, the electric charge generated according to the intensity of the light received at the destination is changed. It works to add up sequentially. Then, when the final line of the light receiving surface is reached, it is output to the outside as an electrical signal. Accordingly, by making the movement speed of the charge in the x direction equal to the movement speed of the image on the light receiving surface in the x direction, the signals during the movement of the image on the sensor are integrated and output.

  In the present embodiment, like the CCD sensor in the first to fourth embodiments, the signal reading is divided into 128 channels and each is read in parallel, so that the reading speed is 4M lines / second. The size of the light receiving area was 64 pixels in the x direction and 2048 pixels in the y direction. The length of one line in the x direction corresponds to 50 nm on the sample surface, and the length in the y direction corresponds to about 100 μm. At this time, since an image having a length of 50 nm and a width of 100 μm is output at a speed of 4 M / sec, the continuous moving speed of the stage is set to the same speed (50 nm / 250 nsec = 200 mm / sec). As described above, the x-direction movement of the inspection region is performed by moving the stage 8.

On the other hand, since the irradiation area of one shot is 5 μm square, it is necessary to move the irradiation area in the y direction by scanning the electron beam as shown in FIG. That is, it is necessary to scan the electron beam by 100 μm in the y direction while the stage 8 moves by one shot (5 μm) in the x direction. If the time required for one shot is 1.25 μsec, 25 μsec is required to scan 100 μm (20 shots) in the y direction. On the other hand, since the moving speed of the stage 8 in the x direction is 200 mm / sec, the time required for the stage 8 to move exactly one shot (5 μm) in the x direction is 25 μsec. As described above, the time required for moving the stage for one shot (5 μm) in the x direction and the time required for the electron beam scanning for 20 shots (100 μm) in the y direction are matched to prevent the generation of dead time. . To obtain an image of the sample surface area 1 cm ² by this method, since it takes 2 × 10 ⁵ times the scan time required (25 .mu.sec) for the unit scan area of 5 [mu] m × 100 [mu] m described above, the sample surface area 1 cm ² per The inspection required time is 5 seconds.

  As described above, in this embodiment, the stage moving speed determined from the signal output speed of the TDI sensor is 200 mm / sec. Therefore, the inspection area can be sufficiently moved in the x direction by moving the stage. In addition, a sufficient time for scanning in the y direction on the inspection region of the electron beam can be secured. In this embodiment, since the inspection speed is determined by the signal output speed of the TDI sensor, if this signal output speed is improved, the inspection can be performed at a higher speed.

Example 4
In this embodiment, an electron optical system capable of acquiring an SEM image is employed. FIG. 17 shows the configuration.

  The electron source 201, the condenser lens 202, and the SEM objective lens 233 employ the elements constituting the SEM electron optical system as they are. The electron source 201 is a Zr / O / W Schottky electron source. The electrons extracted from the electron source are deflected by the beam separator 243, are subjected to the change of the alternative angle by the electrostatic sector type electron deflector 205, are guided to the beam separator 203, and are vertically incident on the objective lens 206. The electron beam forms a crossover at the front focal plane of the objective lens, and becomes a planar electron beam aligned in a direction perpendicular to the surface of the wafer 207 by the objective lens 206. The voltage applied to the wafer 207, the arrangement of the diaphragm, etc. are the same as in the first embodiment.

  The present embodiment is characterized in that a high-resolution SEM image can be observed without taking out the wafer 7 from the apparatus when it is desired to observe in detail a defect image detected after the wafer inspection. That is, if the electron beam is caused to go straight without operating the beam separator 243 and at the same time the wafer 207 is moved under the optical axis of the SEM objective lens 233 by the wafer moving stage 208, observation of an arbitrary position of the wafer 207 is executed. it can. In the figure, 211 is an imaging lens, 213 and 214 are magnifying lenses, 222 is a condenser lens for SEM, 228 is a beam control system, 232 is a preliminary charging control device, 252 is a sample chamber, and 263 is an image detection unit. .

  This function can be used not only for observing detected defects, but also for wafer pattern confirmation, inspection condition setting, alignment, etc. before inspection.

The figure which shows schematic structure of the test | inspection apparatus which becomes a 1st Example of this invention. The figure explaining the principle of this invention. The figure explaining the principle of this invention. The figure which shows the example of an image of the defect obtained by this invention. The figure explaining the efficiency of the electron which can be utilized for image formation in a prior art example. The figure explaining the efficiency of the electron which can be utilized for image formation in a prior art example. The figure explaining the difference in the efficiency of the electron which can be utilized for the image formation of a prior art example and this invention. The figure which compares the contrast (a) and inspection time (b) of a prior art example and this invention. The figure explaining the structure of the 1st Example of this invention. Diagram explaining the relationship between electron energy width and defect detection sensitivity FIG. 3 is a diagram illustrating the principle of a preliminary charging control device according to the first embodiment of the present invention. 1 is a diagram illustrating a configuration of a preliminary charging control device according to a first embodiment of the present invention. FIG. The figure explaining other structures of the preliminary charging control apparatus in the 1st Example of the present invention. The figure explaining the image signal detection means in 1st Example of this invention. The figure explaining the image signal detection means in 2nd Example of this invention. The figure explaining the image signal detection means in the 3rd Example of this invention. The figure explaining the structure of the electron optical system in the 4th Example of this invention.

Explanation of symbols

1: electron source
2: Condenser lens,
3: Beam separator,
4: Aperture,
5: Irradiation system deflector,
6: Objective lens,
7: Sample (wafer),
8: Wafer moving stage,
9: Power supply,
10: Imaging system deflector,
11: Imaging lens,
12: Contrast aperture,
13: Magnifying lens,
14: Magnifying lens,
15: fluorescent plate, 16: optical fiber bundle,
17: CCD,
18: Image storage unit
19: Image storage unit
20: arithmetic unit,
21: Defect determination unit,
22: Monitor,
23: Acceleration power supply,
24: Scanning signal generator,
25: Objective lens power supply, 26: Sample height measuring device,
27: Stage position measuring device, 28: Beam control system,
29: control computer, 30: stage control system, 31: optical microscope, 32: preliminary charging control device,
101: Electron optical system,
102: Sample chamber,
103: Image detection unit,
104: an image processing unit,
105: control unit, 201: electron gun, 202: condenser lens, 203: beam separator, 205: deflector, 206: objective lens, 207: sample (wafer), 208: wafer moving stage, 211: imaging lens, 213 , 214: magnifying lens, 222: SEM condenser lens, 228: beam control system, 232: preliminary charging control device, 233: objective lens for SEM, 243: beam separator, 252: sample chamber, 263: image detector.

Claims (1)

  An electron optical system that irradiates the sample with an electron beam, a sample stage that holds the sample, and a voltage at which a part or all of the electron beam irradiated on the sample is reflected without entering the sample. Means for applying to the sample stage or the sample, an image detection unit for detecting the reflected electrons to form an image, an image processing unit for analyzing the obtained image and inspecting the sample for defects, A defect inspection apparatus comprising: a preliminary charging control device for charging the sample.

Images