JP2012256695A - Defect inspection method and defect inspection device of mask substrate, manufacturing method of photo mask, and manufacturing method of semiconductor device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a defect inspection method of a mask substrate that can adequately perform defect inspection of the mask substrate.SOLUTION: A defect inspection method according to an embodiment comprises: a step S12 of obtaining an asynchronous image of a mask substrate in an asynchronous state in which scan of a stage with the mask substrate placed on it is not synchronized with charge transfer from a TDI camera to obtain an image of the mask substrate; steps S13, S14, S15 of obtaining, on the basis of the asynchronous image, the number of positions having image intensity higher than a certain value; a step S16 of determining an image intensity threshold on the basis of the number of the positions; a step S17 of obtaining a synchronous image of the mask substrate in a synchronous state in which scan of the stage with the mask substrate placed on it is synchronized with charge transfer from the TDI camera to obtain an image of the mask substrate; and a step 18 of determining, on the basis of the synchronous image, that a position having an image intensity equal to or higher than the threshold is a defect.

Description

  Embodiments described herein relate generally to a mask substrate defect inspection method and the like.

  If there is a defect in the mask blank, it is difficult to accurately transfer the mask pattern formed on the photomask onto the semiconductor substrate. Therefore, it is important to inspect the mask blank in advance before producing the photomask.

  However, there are cases where a noise signal is erroneously detected as a defect signal during defect inspection. In order to detect a defect signal, a threshold is usually set, and a signal having a level higher than the threshold is determined as a defect signal. However, if the threshold level is too low, a defect signal with a low intensity can be detected, but the frequency (number of detections) of erroneous detection of the noise signal increases. On the other hand, if the threshold level is too high, the number of detected noise signals is reduced, but it becomes impossible to detect a defect signal having an intensity equal to or lower than the threshold.

  Therefore, in the past, when performing a defect inspection of a mask substrate such as a mask blank, it is impossible to accurately grasp the frequency of erroneous detection of a noise signal, while minimizing the number of detected noise signals, It was difficult to detect more defect signals.

Japanese Patent No. 3728495

  Provided is a mask substrate defect inspection method capable of accurately performing a defect inspection of a mask substrate.

  In the mask substrate defect inspection method according to the embodiment, an asynchronous image of the mask substrate is obtained in an asynchronous state in which the scanning of the stage on which the mask substrate is placed and the charge transfer of the TDI camera that acquires the image of the mask substrate are not synchronized. Obtaining, a step of obtaining the number of locations having an image intensity equal to or greater than a certain value based on the asynchronous image, a step of determining a threshold value of the image intensity based on the number of locations, and the mask substrate. Acquiring a synchronized image of the mask substrate in a synchronized state in which scanning of the stage mounted and charge transfer of a TDI camera for acquiring an image of the mask substrate are synchronized; and based on the synchronized image, the threshold value And a step of determining a portion having the above image intensity as a defect.

It is the figure which showed schematic structure of the defect inspection apparatus of the mask substrate which concerns on embodiment. It is the flowchart which showed the defect inspection method of the mask substrate which concerns on embodiment. It is the figure shown about the synchronous image. It is the figure shown about the asynchronous image. It is the figure which showed the relationship between the charge transfer speed and direction of a TDI camera, and the moving speed and direction of a mask board | substrate on the imaging surface of a TDI camera by a vector. It is the figure shown about the cumulative pixel number distribution of image intensity. It is the functional block diagram which showed the structure for performing the various processes of embodiment. It is the figure shown about the example of a change of embodiment. 5 is a flowchart showing a photomask manufacturing method and a semiconductor device manufacturing method.

  Hereinafter, embodiments will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating a schematic configuration of a defect inspection apparatus for a mask substrate according to an embodiment.

  EUV (extreme ultra violet) light from the light source 11 is applied to the mask substrate 14 via the elliptical mirror 12 and the plane mirror 13. In this embodiment, a mask blank is used as the mask substrate 14. This mask blank 14 is a mask blank of a reflective photomask for EUV exposure. Specifically, the mask blank 14 has a multilayer reflective film formed on the glass substrate and an absorption layer formed on the multilayer reflective film. The mask blank 14 is placed on a stage 15 that can move in the X, Y, and Z directions.

  The light irradiated on the mask blank 14 is scattered on the surface of the mask blank 14. Scattered light having a radiation angle smaller than a predetermined angle is shielded by a shielding part (convex mirror) 16. Scattered light having a radiation angle larger than a predetermined angle is collected by the concave mirror 17 and enters the shielding portion (convex mirror) 16. Light from the shielding unit (convex mirror) 16 is imaged on the imaging surface of a TDI (time delay integration) camera 18. The optical system as described above is a dark field optical system, and the TDI camera 18 captures a dark field image.

  A signal based on the image obtained by the TDI camera 18 is sent to a personal computer 19 serving as a calculation unit, and the personal computer 19 performs calculation for defect determination.

  FIG. 2 is a flowchart illustrating a mask substrate defect inspection method according to the embodiment. Hereinafter, the defect inspection method will be described with reference to the flowchart of FIG.

  First, the mask blank 14 is placed on the stage 15 (S11). Next, an asynchronous image in a partial region of the mask blank 14 is acquired (S12). Here, the partial area is an area on the pattern processing surface of the mask blank 14.

  Here, the asynchronous image refers to an image obtained in a state (asynchronous state) in which the synchronous operation of the TDI (time delay integration) method is not performed. The TDI method is a method for continuously acquiring images while scanning an imaging target. In the TDI method, scanning of a stage on which a mask substrate is placed and charge transfer of a TDI camera are usually synchronized. That is, the charge transfer operation (charge transfer direction and charge transfer speed) of the TDI camera is synchronized (matched) with the movement operation (movement direction and speed) of the mask substrate on the imaging surface of the TDI camera. An image obtained in such a synchronized state is a synchronized image. Therefore, an asynchronous image is an image obtained in a state (asynchronous state) that is not such a synchronous state. That is, a state in which the charge transfer direction of the TDI camera and the movement direction of the mask substrate on the imaging surface of the TDI camera do not match is defined as state X, and the charge transfer speed of the TDI camera and the mask on the imaging surface of the TDI camera Assuming that the state where the moving speed of the substrate does not match is the state Y, the state becomes an asynchronous state when it is at least one of the state X and the state Y.

  FIG. 3 is a diagram showing a synchronous image. As shown in FIG. 3A, when the synchronization operation is performed, the defect is obtained as a point image. Therefore, as shown in FIG. 3B, a sharp peak is observed in the image intensity distribution of the synchronous image.

  FIG. 4 is a diagram showing an asynchronous image. Here, an asynchronous image is shown when the charge transfer direction of the TDI camera and the movement direction of the mask substrate (mask blank) on the imaging surface of the TDI camera are opposite to each other. As shown in FIG. 4A, when an asynchronous operation is performed, the defect is stretched and a slender image is obtained. Therefore, as shown in FIG. 4B, no sharp peak is observed in the image intensity distribution of the asynchronous image.

  FIG. 5 is a diagram showing the relationship between the charge transfer speed and direction of the TDI camera and the movement speed and direction of the mask substrate on the imaging surface of the TDI camera in vectors. The former is indicated by a vector V1, and the latter is indicated by a vector V2. The length of the image obtained by extending the defect is proportional to the absolute value of the difference vector V3 between the vector V1 and the vector V2.

  In step S12, a sufficiently stretched image is obtained so that the asynchronous image of the defect cannot be recognized as a defect. That is, the charge transfer speed of the TDI camera 18 and the moving speed of the stage 15 on which the mask blank 14 is placed are adjusted so that such a sufficiently stretched image is obtained. Since the defect is not recognized in the asynchronous image thus obtained, all signals showing sharp image intensity peaks in the asynchronous image are noise signals.

  Next, a cumulative pixel number distribution of image intensity is obtained for an asynchronous image acquired from a partial region on the surface of the mask blank 14 (S13). FIG. 6 is a diagram showing the cumulative pixel number distribution of image intensity. The plot P1 in FIG. 6 shows the cumulative pixel number distribution of the image intensity for the partial area. That is, the plot P1 in FIG. 6 shows the number of pixels having an image intensity equal to or higher than each image intensity on the horizontal axis in FIG. Generally speaking, the plot P1 in FIG. 6 shows the number of locations having an image intensity greater than a certain value. In the example of FIG. 6, the number of pixels hardly changes until the image intensity reaches I1. Therefore, in the example of FIG. 6, it is considered that most of the pixels in the partial area have an image intensity equal to or higher than I1.

  Next, a cumulative pixel number distribution of image intensity is predicted for a desired inspection region on the surface of the mask blank 14 (S14). Specifically, the area ratio of the desired inspection area to the partial area is obtained, and the plot P1 in FIG. 6 is multiplied by the area ratio. Thereby, as shown in the plot P2 of FIG. 6, a cumulative pixel number distribution of the image intensity for the desired inspection region is obtained.

  Next, fitting approximation is performed using an arbitrary function for the plot P2 obtained in step S14. By this fitting approximation, an approximate curve C1 of FIG. 6 is created (S15).

  Next, based on the approximate curve C1 in FIG. 6, the number of pixels on the vertical axis in FIG. 6 within the desired inspection region of the mask blank 14 (hereinafter sometimes referred to as the number of pixels of pseudo defects for convenience) is a predetermined number. The image intensity (signal intensity) when it is less than Nth is determined as the threshold value Ith (S16). As indicated by the approximate curve C1 in FIG. 6, the number of pixels on the vertical axis decreases rapidly from the vicinity of the image intensity I1. That is, the number of pixels having a high image intensity has decreased rapidly from the vicinity of the image intensity I1. Therefore, for example, it is considered that pixels having an image intensity I1 or higher are affected by noise, and the image intensity when the number of pixels on the vertical axis in FIG. It is defined as Ith. That is, a part (pixel) having an image intensity equal to or less than a certain value is regarded as a noise part. Here, the predetermined number Nth is an expected value of the allowable number of noise detections. For example, if one noise detection is allowed in 10 inspections, the predetermined number Nth is 0.1. In this way, the number of locations having an image intensity equal to or greater than a certain value is acquired based on the asynchronous image, and the threshold value Ith of the image intensity is determined based on the acquired number of locations.

  Next, a synchronous image in the desired inspection area on the surface of the mask blank 14 is acquired (S17). That is, a synchronized image from the desired inspection region of the mask blank 14 is acquired in a state where the scanning of the stage 15 on which the mask blank 14 is placed and the charge transfer of the TDI camera 18 are synchronized. Here, the desired inspection area is an area on the pattern processing surface of the mask blank 14.

  Next, based on the acquired synchronous image, a portion having an image intensity (signal intensity) equal to or higher than the threshold value Ith is determined as a defect (S18). As described above, when the synchronization operation is performed, a peak appears in the image intensity distribution (signal intensity distribution) of the synchronized image at the defective portion. However, if the threshold value of image intensity is not appropriate, accurate defect determination cannot be performed. Therefore, a portion having an image intensity equal to or higher than the threshold value Ith obtained in step S16 is determined as a defect.

  Next, the coordinates of the part determined as a defect in step S18 are output (S19). If there is no portion determined as a defect in step S18, information indicating that the defect is zero is output.

  FIG. 7 is a functional block diagram showing a configuration for executing the various processes described above. The various processes described above are mainly performed by the computer 19.

  The configuration illustrated in FIG. 7 includes an asynchronous image acquisition unit 21, a location number acquisition unit 22, a threshold value determination unit 23, a synchronous image acquisition unit 24, and a defect determination unit 25. The asynchronous image acquisition unit 21 acquires an asynchronous image from the surface of the mask substrate in an asynchronous state where the scanning of the stage and the charge transfer of the TDI camera are not synchronized. The location number acquisition unit 22 acquires the number of locations having an image intensity equal to or greater than a certain value based on the asynchronous image. The threshold value determination unit 23 determines a threshold value of the image intensity based on the number of the locations. The synchronized image acquisition unit 24 acquires a synchronized image from the surface of the mask substrate in a synchronized state in which the scanning of the stage and the charge transfer of the TDI camera are synchronized. The defect determination unit 25 determines a portion having an image intensity equal to or higher than the threshold value as a defect based on the acquired synchronous image. The function and operation of each part are as described above.

  As described above, in the present embodiment, based on the asynchronous image, the number of locations having an image intensity equal to or greater than a certain value is acquired, and the threshold value Ith of the image intensity is determined based on the number of acquired locations, A portion having an image intensity equal to or higher than the threshold value Ith is determined as a defect based on the synchronized image. As described above, by determining the threshold value of the image intensity using the asynchronous image, the optimum threshold value can be obtained, and accurate defect determination can be performed. As described above, if the threshold level is too low, a defect signal with low intensity can be detected, but the number of detected noise signals increases. If the threshold level is too high, the number of detected noise signals decreases, but the threshold value It becomes impossible to detect a defect signal having the following intensity. In the present embodiment, by optimizing the threshold level, it is possible to detect more defects with the minimum necessary number of noise detections.

  In the above-described embodiment, the defect is determined for each pixel, that is, the image intensity (signal intensity) is detected for each pixel, but a plurality of pixels are imaged as one defect determination location. The intensity (signal intensity) may be detected. FIG. 8 is a diagram showing an example thereof. In the example of FIG. 8, the 3 × 3 pixel area A1 is set as one defect determination portion.

  Further, when the intensity distribution of illumination light cannot be ignored and a certain level of intensity is detected even in a non-defect area, the peripheral area may be considered. Specifically, as shown in FIG. 8, the average intensity of the area A2 excluding the inner 5 × 5 pixel area from the 9 × 9 pixel area is set as the background level, and the center 3 × 3 pixel area A1 The intensity obtained by subtracting the background level of the surrounding area A2 from the intensity may be the true image intensity of the 3 × 3 pixel area A1. By doing so, the influence of the intensity distribution of the illumination light can be reduced.

  In the above-described embodiment, a reflective photomask mask blank for EUV exposure is used. However, the method of the above-described embodiment can also be applied to a transmissive photomask mask blank. However, the above-described method is more effective for a mask blank of a reflective photomask for EUV exposure. That is, a mask blank for a reflective photomask has a multilayer reflective film in which two types of films (layers) having different refractive indexes are alternately laminated, and the phases of reflected light from each layer are aligned. High reflectivity is obtained. Therefore, when the phase of the reflected light is disturbed by a defect, a phase defect occurs. By using the method of this embodiment, it is possible to detect such a phase defect with the minimum necessary number of noise detections.

  In the embodiment described above, a mask blank is used as the mask substrate. However, the method described above can be applied to a mask substrate other than the mask blank. For example, the above-described method can be applied to a photomask having a pattern with a constant period such as a line and space pattern. In the case of a fine line and space pattern with a constant period, since there are many line patterns and space patterns in one pixel of the TDI camera, it is considered that the intensity in each pixel is constant in the absence of defects. Therefore, it is possible to detect a defect in the photomask by using the method described above.

  Note that the method of the above-described embodiment can be applied to a photomask manufacturing method and a semiconductor device manufacturing method.

  FIG. 9 is a flowchart showing a photomask manufacturing method and a semiconductor device manufacturing method. First, the mask blank is inspected by the method described above (S21). Next, a photomask is manufactured using the inspected mask blank (S22). Further, a semiconductor device is manufactured using the manufactured photomask (S23). Specifically, the mask pattern (circuit pattern) formed on the photomask is transferred to the photoresist on the semiconductor substrate. Subsequently, the photoresist is developed to form a photoresist pattern. Further, the conductive film, the insulating film, the semiconductor film, or the like is etched using the photoresist pattern as a mask.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 11 ... Light source 12 ... Elliptical mirror 13 ... Plane mirror 14 ... Mask blank 15 ... Stage 16 ... Shield part (convex mirror)
DESCRIPTION OF SYMBOLS 17 ... Concave mirror 18 ... TDI camera 19 ... Personal computer 21 ... Asynchronous image acquisition part 22 ... Location number acquisition part 23 ... Threshold value determination part 24 ... Synchronous image acquisition part 25 ... Defect determination part

Claims (8)

Acquiring an asynchronous image of the mask substrate in an asynchronous state in which the scanning of the stage on which the mask substrate is placed and the charge transfer of the TDI camera for acquiring the image of the mask substrate are not synchronized;
Obtaining a number of locations having an image intensity greater than a certain value based on the asynchronous image;
Determining an image intensity threshold based on the number of locations;
Obtaining a synchronized image of the mask substrate in a synchronized state in which the scanning of the stage on which the mask substrate is placed and the charge transfer of the TDI camera that obtains the image of the mask substrate are synchronized;
A step of determining a defect having an image intensity equal to or higher than the threshold based on the synchronous image as a defect;
A defect inspection method for a mask substrate, comprising: 2. The defect inspection method for a mask substrate according to claim 1, wherein the threshold value of the image intensity is obtained as an image intensity when the number of the locations in a certain area of the mask substrate is smaller than a predetermined number. . The defect inspection method for a mask substrate according to claim 1, wherein the mask substrate is a mask blank. The defect inspection method for a mask substrate according to claim 3, wherein the mask substrate is a mask blank for EUV exposure. The defect inspection method for a mask substrate according to claim 1, wherein the synchronous image and the asynchronous image are acquired using a dark field optical system. A stage on which the mask substrate is placed;
A TDI camera for acquiring an image of a mask substrate placed on the stage;
An asynchronous image acquisition unit for acquiring an asynchronous image of the mask substrate in an asynchronous state in which the scanning of the stage and the charge transfer of the TDI camera are not synchronized;
Based on the asynchronous image, a location number acquisition unit that acquires the number of locations having an image intensity equal to or greater than a certain value;
A threshold value determination unit for determining a threshold value of the image intensity based on the number of the points;
A synchronized image acquisition unit that acquires a synchronized image of the mask substrate in a synchronized state in which scanning of the stage and charge transfer of the TDI camera are synchronized;
A defect determination unit that determines, as a defect, a portion having an image intensity equal to or higher than the threshold based on the synchronous image;
A defect inspection apparatus for a mask substrate, comprising: A photomask is manufactured using the mask blank inspected by the method according to claim 3. A semiconductor device is manufactured using the photomask manufactured by the method of Claim 7. The manufacturing method of the semiconductor device characterized by the above-mentioned.

Images