JP2002353279A - Circuit-pattern inspection method and its apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect failures of a plurality of kinds of circuit-patterns and the like on a wafer at the same time, simplify an operation for setting inspection conditions, and improve inspection functions. SOLUTION: Images 106, 107 in regions 104, 105 of a pattern 101 to be inspected and a comparison pattern 102 are obtained under one inspection condition and images 106', 107' in regions 104, 105 are obtained under another inspection condition. The compared result 108 between the image 106 and the image 107 and the compared result 109 between the image 106' and the image 107' are compared. The image 106 and the image 107 are also compared.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for inspecting a circuit pattern, and more particularly to inspecting a semiconductor device such as a memory having a fine circuit pattern, an LSI, a liquid crystal, a photomask and the like using a charged particle beam. The present invention relates to a circuit pattern inspection method and apparatus suitable for the above.

[0002]

2. Description of the Related Art A semiconductor device is manufactured by repeating a process of transferring a circuit pattern formed on a substrate such as a semiconductor wafer or a photomask by lithography and etching. The generation of foreign substances in the manufacturing process and the generation of defects in each step deteriorate the yield of semiconductor devices. Therefore, the formed pattern is inspected for each process in the manufacturing process in order to detect the occurrence of abnormality or defect in each process early or in advance.

As an example of an apparatus for inspecting a defect existing in a pattern on a semiconductor wafer, a semiconductor wafer is irradiated with white light, and an optical image is used to compare repeated patterns such as a plurality of LSI patterns. Defect inspection equipment is known (Semiconductor World (Semiconductor W
orld), August, Vol. 17, No. 13, pp. 96-99, 1995).
However, with the miniaturization of circuit patterns, the complexity of circuit pattern shapes, and the diversification of materials, it has become difficult to detect defects using optical images. As defects that are difficult to detect with an optical inspection device, residues and defects of a pattern having a silicon oxide film or a photosensitive photoresist material as a light transmitting material on the surface,
Examples of such defects include an etching residue that is lower than the resolution of the optical system, a non-opening defect of a minute conduction hole, and a defect generated at a step bottom of a wiring pattern.

In order to detect a defect or the like which is difficult to detect with such an optical inspection apparatus, a circuit using a charged particle beam, particularly an image obtained by scanning with an electron beam, which is desired to have a higher resolution than an optical image, is required. A method for comparing and inspecting patterns has been proposed (for example, Japanese Patent Application Laid-Open No. Sho 59-192943, Journal of Vacuum Science Technology B, J. Vac. Sci. Tech. B), Vol. .6,
pp. 3005-3009 (1991), J. Vac. Sci. Tech. B, Vo
l. 10, No. 6, pp. 2804-2808 (1992), and JP-A-5-258703, and U.S. Pat.
No. 06, etc.).

In order to obtain a practical throughput with an inspection apparatus using an electron beam, it is necessary to acquire an image at a very high speed. Then, in order to secure the SN ratio of the image acquired at high speed, a normal scanning electron microscope (Scanning Electron Microscope) is used.
n Microscopy (hereinafter abbreviated as SEM) 100 times or more (1
(0 nA or more), and the S / N ratio of the image is secured while maintaining a practical inspection speed. At this time,
The beam diameter is much wider than normal SEM,
It is about 0.05 μm to 0.2 μm. this is,
This value is limited by the brightness of the electron gun and the Coulomb effect because the beam current is large. Then, an image signal obtained by such an electron optical system is sent to an image processing system and
One image is obtained for one region, and a comparison inspection is performed between adjacent images of the same pattern portion. If there is a different brightness portion when comparing the images, the portion is regarded as a defect and the coordinates are stored. With this configuration, a defect having a size of about 0.05 to 0.1 μm is automatically detected.

In order to accurately detect defects automatically,
It is important to adjust various parameters such as the electron beam irradiation energy and the detection sensitivity in consideration of the relationship between charge-up on the sample surface, secondary electron emission efficiency, potential contrast, and the like. Charge-up is a phenomenon in which the surface of a sample is positively or negatively charged by electron beam irradiation, and occurs remarkably when the sample is an insulating material. When the charge-up occurs, the secondary electron emission efficiency changes, so that the quality of the obtained electron beam image, for example, the brightness and contrast of the pattern of the electron beam image changes. In the pattern inspection system using an electron beam, the sample is charged up and a potential contrast is generated on the sample surface, as described in JP-A-11-121561, in order to efficiently detect the defect to be detected. May be caused. In addition, excessive charge-up causes distortion of an image, and uneven charge-up may destabilize image quality.
As described in Japanese Patent Publication No. 684, a method of obtaining an electron beam image by always controlling the charge-up of a sample to the same state has been devised.

[0007] When parameters such as settings related to the charge-up control are defective, the number of erroneous detections increases and the defect detection rate decreases. In a pattern comparison inspection device using an electron beam, such parameter tuning,
After the completion of the main inspection, an image of the detected defect coordinate part is obtained, and a review operation is performed by which the operator confirms what kind of defect has been detected. The function required for the review work is a function of imaging a specific narrow visual field and visually observing and confirming it, as in a normal SEM.
In the review function, the above-mentioned 0.
A high-resolution SEM that can confirm the type and shape of a defect of about 0.05 to 0.1 μm is required. JP-A-2000-1
Japanese Patent Publication No. 88075 discloses a method in which an electronic optical system for inspection and an electronic optical system for review are provided on the same sample chamber, and the visual inspection of defects detected by an inspection device is efficiently performed.

[0008]

As described above, in the conventional pattern inspection apparatus using charged particles, parameters are always set before inspection. These parameters determine the operating conditions of the inspection apparatus at the time of inspection (hereinafter, referred to as inspection conditions), and include, for example, irradiation energy such as acceleration voltage, beam current, pixel size, sensitivity of the secondary electron signal detection system, and the like. There are filter constants, charge-up control values, scan sequences, and the like. At this time, since the optimum value differs depending on the process to be inspected (material of the inspection sample, etc.), the pattern size, and the type of defect to be detected, a trial inspection is performed before the inspection, and an image of the detected defect coordinates is displayed.
The optimum values of the parameters are set while checking whether the defect to be detected is detected. In this defect confirmation, that is, in the review, it is determined whether the detected defect is a true defect or an erroneous detection due to a setting defect. However, as described above, in order to obtain high detection accuracy, a height different from that for inspection is used. In many cases, a review SEM for resolution is prepared.

In order to derive the optimum value of the parameter,
Trial testing and review had to be repeated over and over. In addition, when using a different SEM for review than for inspection to obtain high detection accuracy, if they are composed of different devices, it is necessary to unload and load the sample, which are in the same sample chamber. Even in such a case, there is a problem that the sample stage needs to be moved, which requires a lot of time and complicated work.

In order to reduce the optimum value of the parameter to one, for example, the insulating material residue on the sample surface is deteriorated in detection sensitivity when charged up, but the residue at the bottom of the contact hole is improved in detection sensitivity by charging up. As described above, if the detection sensitivity of a specific type of defect is improved, the detection sensitivity of another type of defect may deteriorate, and it has been difficult to detect a plurality of types of defects at the same time.

In addition to the review after the trial inspection,
There is also a need to obtain an image of the defect coordinate portion after the main inspection, confirm what kind of defect has been detected by a high-resolution review SEM, and save the image. At this time, for example, the residue at the bottom of the contact hole can be detected by the inspection device, but the difference in conditions such as the amount of charge-up at the time of inspection and at the time of review is such that it may not be visible with a normal SEM. As a result, there is a problem in that a defect cannot be confirmed and an erroneous determination is made as erroneous detection.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to compare image information between identical design patterns such as circuit patterns on a wafer in the process of manufacturing a semiconductor device, to determine a defect, In a method and an apparatus for inspecting foreign matter, residue and the like by a charged particle beam, simultaneous detection of a plurality of types of defects is performed, and inspection condition setting work is simplified by sharing inspection condition information between inspection and review,
It is an object of the present invention to provide a circuit pattern inspection method and apparatus capable of improving the inspection function and improving the reliability of the inspection.

[0013]

In order to achieve the above object, the present invention irradiates a charged particle beam onto a surface of a substrate on which a circuit pattern is formed, and detects a signal generated from the circuit pattern by the irradiation. In a circuit pattern inspection method for detecting a defect in the circuit pattern using an image signal obtained by performing, a plurality of inspection conditions including at least the irradiation conditions of the charged particle beam or the charging conditions of the substrate or comparison conditions, A circuit pattern inspection method is provided, wherein a defect of the circuit pattern is extracted by comparing image signals obtained for each of the inspection conditions. This enables simultaneous detection of defects with completely different features,
The defect detection function and the defect detection efficiency can be improved.

Further, according to the present invention, in the above-described circuit pattern inspection method, first and second inspection conditions are determined, and the inspection target area on the circuit pattern obtained under the first and second inspection conditions is determined. Are the first and second images, and the images of the comparison target pattern of the inspection target region obtained under the first and second inspection conditions are the third and fourth images. A first comparison image is generated by comparing the third image, a second comparison image is generated by comparing the second image and the fourth image, and the first comparison image and the second image are further generated. And extracting a defect in the inspection target area by comparing the inspection pattern with a comparison image. This makes it possible to improve the reliability of the defect feature classification and at the same time facilitate the work of optimizing the defect inspection conditions to be detected.

Further, according to the present invention, in the above-described circuit pattern inspection method, first and second inspection conditions are determined, and an inspection target area on the circuit pattern obtained under the first and second inspection conditions is determined. The first and second images are used as the first and second images, and the first and second images are compared to extract a defect in the inspection target area, thereby providing a circuit pattern inspection method. As a result, even in a state where the detection sensitivity is increased and a small difference between the images is detected, it is possible to reduce the erroneous detection of a defect in which a portion having no defect is determined as a defect. And the reliability of the inspection can be improved at the same time.

Further, according to the present invention, in the circuit pattern inspection method, when the detected defect is reviewed by using an apparatus, the observation of the observation apparatus is performed with reference to the inspection condition at the time of detecting the defect. A circuit pattern inspection method is provided, wherein a condition is determined and a review is performed under the condition. As a result, the task of confirming defects or optimizing inspection conditions, which has conventionally required skill, and was complicated and time-consuming, has been simplified and assured. High reliability of inspection can be achieved.

Further, the present invention is a circuit pattern inspection apparatus for inspecting a circuit pattern using the circuit pattern inspection method, wherein an inspection condition setting means for setting the plurality of inspection conditions, First control means for controlling the irradiation of the charged particle beam to the circuit pattern for each of the inspection conditions set by this means, and an image signal obtained for each of the inspection conditions corresponding to the inspection condition And a second control means for controlling the circuit pattern to be stored in a memory and performing a comparison process. This makes it possible to realize a circuit pattern inspection apparatus which is capable of simultaneously inspecting defects having completely different characteristics and which has high inspection efficiency and is easy to use.

[0018]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(Circuit Pattern Inspection Method of the Present Invention) FIG.
FIG. 3 is an explanatory view illustrating a circuit pattern inspection method of the present invention. The circuit pattern in the present invention is a memory or LS
1 shows a pattern formed of materials having different electrical characteristics, such as a semiconductor device such as I, a liquid crystal, and a photomask. Inspection of these patterns acquires a pattern image to be inspected,
The comparison is performed by comparing the image with the design data, or when the pattern is repeatedly present, by comparing the pattern image to be inspected with the pattern image at another position which is repeatedly present. The example of FIG. 1 is the latter case, in which a comparison inspection of the inspection target pattern 101 and another pattern 102 that is repeatedly present is performed. In the comparison of the two images, only the difference is detected, and it is not possible to identify which pattern is defective. However, here, the comparison of the two images is shown for the sake of simplicity. In an actual inspection, there are a plurality of these patterns, and the defect position is specified from the history of the difference data detected as a result of the comparison.

In FIG. 1, inspection conditions including various parameters such as the amount of charge-up on the sample surface, electron beam irradiation energy, and detection sensitivity are considered in consideration of the relationship between secondary electron emission efficiency, potential contrast, and defect detection efficiency. When the irradiation area 103 is irradiated, the area 104 of the inspection target pattern 101 and the area 1 of another pattern 102
It is assumed that the image No. 05 is obtained. Conventionally, a trial inspection and a review are repeated to determine an optimum condition for a target defect inspection, and the above-described regions 104 and 10 are determined under the optimum condition.
The pattern 5 is acquired once or plural times. Then, at one time, the obtained images are compared. In addition, in the case of plural times, a method is used in which these are averaged and then compared. In the present invention, as will be described in detail below, the irradiation area 103 stores an image obtained under different inspection conditions or similar inspection conditions or an image partially processed by averaging or the like, Compare in various ways. Some examples of this comparison method are described below.

As an example of a method of comparing a plurality of images, there is a method of comparing the comparison results under each inspection condition. In the example of FIG. 1, the comparison result 108 between the image 106 and the image 107 and the image 1
06 ′ and the comparison result 109 of the image 107 ′ are compared. At this time, the inspection condition J1 when the images 106 and 107 are obtained, and the inspection condition J when the images 106 'and 107' are obtained.
2, and necessary ones of the acquired images and the images of the comparison results are stored in association with each other. By doing so, it is possible to improve the reliability of defect feature classification and to facilitate the work of optimizing the defect inspection conditions, which is the purpose of detection.

As another example of a method of comparing a plurality of images, there is a method of comparing a plurality of images (106 and 106 'in FIG. 1) obtained in the same area. When the inspection conditions are largely different, the obtained images are often completely different. In this case, it is preferable to use an image in which the inspection conditions are almost the same or not so much different. For example, when comparing a plurality of images obtained under the same inspection conditions, it is effective to remove false detections that occur with a low probability such as noise. The reliability of the erroneous detection is improved as compared with the conventional addition method. Further, as an example of using an image in which the inspection conditions are changed little by little, in an inspection apparatus that obtains an image of the area 103 by scanning a charged particle beam, a beam shape is circular, a vertical ellipse having a predetermined ratio, May be a horizontally long ellipse. In this case, an image is acquired by changing the detection sensitivity of the vertical edge or the horizontal edge portion of the circuit pattern, and it is possible to detect erroneous detection due to a stochastic light-dark difference occurring at the edge portion of the circuit pattern. Become. The obtained plurality of comparison results (comparison result 110 in FIG. 1) can be used for judging an erroneously detected defect and a true defect. As a result, erroneous detection of a defect can be reduced, and the detection sensitivity can be reduced. Can be set high, and the defect detection efficiency and inspection reliability are improved. This comparison method can also be used when performing noise evaluation.

When all the acquired images are stored as in a test inspection described later, a comparison method can be selected after the images are acquired, and the comparison results can be analyzed. However, in the present inspection for inspecting a wide area, an enormous storage device is required if all the images are stored, so that only the final result is often stored. In this case, it is necessary to specify which inspection method should be used for the acquired image, which comparison method should be used, which comparison result should be compared, or which comparison result should be stored. However, the image of the representative defect having the predetermined characteristic and the degree of the defect image which cannot be determined for some reason may be stored in the main inspection.

In the defect review, since the defect is minute in the review function of the inspection apparatus, its existence, shape, characteristic, and the like may not be confirmed visually. In such a state, it cannot be determined whether the detected defect is an erroneous detection or a true defect. Therefore, the review may use a high-resolution observation device different from the inspection device. In such a case, it is necessary for the observation apparatus to set the observation conditions including the irradiation condition of the charged particle beam or the charging condition of the substrate. Further, even when a review is performed using the review function of the inspection apparatus, a technique is required for setting conditions for image acquisition such as a defect that cannot be seen due to charge-up. Defects detected in the inspection device,
Despite the fact that a defect actually exists, there is a case where the condition setting is not optimized so that it cannot be visually confirmed in the review. Conventionally, erroneous detection is determined even in such a case, which affects the reliability of the inspection. In the present invention, first, the inspection apparatus stores the detected defect and the data of the inspection condition in association with each other. At this time, it is preferable to use the inspection condition with the largest number of detected defects. Next, the observation condition of the observation device is set from the inspection condition of the inspection device corresponding to the defect to be reviewed. In the case where the inspection apparatus and the observation apparatus are different as described above, the method of setting conditions is often different because the structure of the electron optical system and the apparatus structure such as the charge-up control method are different. For the first time or when the optimal correspondence between inspection conditions and observation conditions is not known, it is necessary to make adjustments by trial and error, but once the correspondence is known, record and record multiple correspondences between inspection conditions and observation conditions. . This correspondence may be derived from defect information detected by another type of inspection device. If the inspection conditions are automatically converted into observation conditions based on this record, the operation can be further simplified. The corresponding observation conditions can be further fine-tuned to obtain and store clear defect images. In this way, the defects detected by the inspection device and the defects that actually exist can be visually confirmed by the review without fail, simplifying the delicate and advanced skill and the time-consuming review work, and erroneous detection. High reliability of the judgment is realized, and high reliability of the inspection becomes possible.

(Outline of Configuration of Circuit Pattern Inspection Apparatus of the Present Invention) FIG. 2 shows an example of the configuration of a circuit pattern inspection apparatus of the present invention for performing inspection using the above-described pattern inspection method. In the example shown here, an electron beam is used in particular, but the inspection method of the present invention is applicable to charged particles. In the example shown in FIG. 2, a semiconductor wafer or a mask,
Although a reticle pattern is assumed, the present invention is also applicable to a circuit pattern including a liquid crystal or the like.

The circuit pattern inspection apparatus 1 is roughly divided into an inspection room 2 comprising an electron optical system 3, an optical microscope unit 4, and a sample room 8, an image processing unit 5, a control unit 6, a secondary electron detection It comprises a unit 7 and a spare room (not shown). The preliminary chamber includes a substrate 9 to be inspected in the inspection chamber 2 where the chamber is evacuated.
In order to transport the vacuum chamber, the vacuum chamber can be evacuated independently of the inspection room 2.

The electron optical system 3 includes an electron gun 10, an electron beam extraction electrode 11, a condenser lens 12, a blanking deflector 13, an aperture 14, a scanning deflector 15, an objective lens 16, a reflection plate 17, and an ExB (E A (cross-B) deflector 18, and an electron beam 19 generated by the electron gun 10 and extracted by the extraction electrode 11 is irradiated to the sample 9 through the condenser lens 12, the aperture 14, and the objective lens 16. You. The electron beam 19 is a narrowed beam, and scans the sample 9 by the scanning deflector 15. The secondary electrons 51 generated from the sample 9 by this scanning are Ex
The orbit is bent by the B deflector 18 to irradiate the reflecting plate 17 to generate second secondary electrons 52, and the secondary electron detector 2
0 is detected. When the irradiation of the sample 9 with the electron beam 19 is restricted, the blanking deflector 13 directs the electron beam 19 to the outside of the opening of the stop 14.

The sample chamber 8 includes a sample stage 30, an X stage 3,
1, a Y stage 32, a rotation stage 33, a position monitor length measuring device 34, and a sample height measuring device 35. As the position monitor length measuring device 34, a length measuring device based on laser interference was used here. The positions of the X stage 31, the Y stage 32, and the rotary stage 33 can be monitored in real time,
The location information can be sent to Although not shown, data such as the number of rotations of the motors of the X stage 31, the Y stage 32, and the rotation stage 33 are similarly transmitted from the respective drivers to the control unit 6. The control unit 6 controls the electron beam 19 based on these data.
The region and position where the electron beam 19 is irradiated can be accurately grasped, and the positional deviation of the irradiation position of the electron beam 19 can be corrected in real time using the correction control circuit 43 as needed. . Further, even if the sample 9 is changed, the area irradiated with the electron beam can be stored for each sample.
As the sample height measuring device 35, an optical measuring device using a measuring method other than the electron beam, for example, a laser interferometer or a reflected light measuring device for measuring a change in the position of reflected light is used.
The height of the sample 9 mounted on the stage 31 and the Y stage 32 can be measured in real time. In this example, a method of irradiating the sample 9 with elongated white light passing through a slit through a transparent window, detecting the position of reflected light with a position detection monitor, and calculating the amount of change in height from the change in position. Using. Based on the measurement data of the sample height measuring device 35, the focal length of the objective lens 16 for narrowing down the electron beam 19 and the positional deviation of the irradiation position of the electron beam 19 are dynamically corrected, and the focus is always focused on the inspection area. Matched electron beam 19
Can be irradiated. Further, it is also possible to configure so that the warp and the height distortion of the sample 9 are measured in advance before the electron beam irradiation, and the correction conditions for each inspection area of the objective lens 16 are set based on the data. .

Operation commands and operation conditions for each section of the apparatus are input from the control section 6. In the control unit 6, conditions such as an acceleration voltage at the time of generation of an electron beam, an electron beam deflection width, a deflection speed, a signal capture timing of a secondary electron detection device, and a sample stage moving speed are arbitrarily or selected according to the purpose. It has been entered so that it can be set. The control unit 6 monitors the displacement of the position and the height from the signals of the position monitor length measuring device 34 and the inspected substrate height measuring device 35 by using the correction control circuit 43,
A correction signal is generated from the result, and the correction signal is sent to the objective lens power supply 45 and the scanning signal generator 44 so that the electron beam 19 is always irradiated to a correct position. The sample 9 is mounted on the X stage 31 and the Y stage 32. The X stage 31 and the Y stage 32 are stationary when the inspection is performed, and the electron beam 19 is two-dimensionally scanned when the inspection is performed. , The Y stage 32 is continuously moved in the Y direction at a constant speed, and the electron beam 19 can be linearly scanned in the X direction. The former method is effective for inspecting a specific relatively small area, and the latter method is effective for inspecting a relatively large area.

The optical microscope unit 4 is installed near the electron optical system 3 in the room of the inspection room 2 and separated from the electron optical system 3 so as not to affect each other. The distance between 4 is known. Then, the X stage 31 or the Y stage 32 reciprocates a known distance between the electron optical system 3 and the optical microscope unit 4. Optical microscope 4
Is composed of a white light source 40, an optical lens 41, and a CCD camera 42. Although not shown, an acquired image is sent to the image processing unit 5 in the same manner as an electron beam image described later. The captured optical image is displayed on the monitor 50.

A sample 9 is irradiated with a narrowed electron beam 19,
An image of the sample 9 is acquired by generating secondary electrons 51 and detecting them in synchronization with the scanning of the electron beam 19 and the movement of the X stage 31 and the Y stage 32. In the automatic inspection targeted by the present invention, a high inspection speed is essential. Therefore, unlike the ordinary SEM, the electron beam having the electron beam current of the pA order is not scanned at a low speed. In applications where high-speed performance is required rather than accuracy, or in applications where charging of insulating materials is suppressed, multiple scans and superposition of each image are not performed, so at least one high-speed electron beam scan is required. The device must be capable of acquiring images. Therefore, in this embodiment, a high current electron beam of about 100 times or more, for example, 10 to 100 nA compared with the normal SEM is used, and the SN of the image is maintained while maintaining a practical inspection speed.
Is secured. The beam diameter is limited by the brightness and Coulomb effect of the electron gun due to the large beam current, and is considerably wider than that of a normal SEM.
It is about 0.2 μm. The standard value of the scan width for forming an image is 100 μm, and the standard value of one pixel is 0.05 μm.
m to 0.1 μm, and the standard time for one scan is about 10 μm.
s. However, these values can be changed by setting.

The electron gun 10 uses a diffusion-supply type thermal field emission electron source. By using the electron gun 10, a stable electron beam current can be secured as compared with a conventional tungsten (W) filament electron source or a cold field emission electron source, for example. An image is obtained. With this electron gun 10, the electron beam current can be set large. The electron beam 19 is extracted from the electron gun 10 by applying a voltage between the electron gun 10 and the extraction electrode 11. The acceleration of the electron beam 19 is performed by applying a high negative potential to the electron gun 10. As a result, the electron beam 19 advances toward the sample stage 30 with energy corresponding to the potential, and the condenser lens 1
The light is converged at 2 and is further narrowed down by the objective lens 16 to irradiate the sample 9 mounted on the X stage 31, the Y stage 32, and the rotating stage 33 on the sample stage 30.
A scanning signal generator 44 for generating a scanning signal and a blanking signal is connected to the blanking deflector 13, and an objective lens power supply 45 is connected to the condenser lens 12 and the objective lens 16. A negative voltage can be applied to the sample 9 by the retarding power supply 36. By adjusting the voltage of the retarding power supply 36, the primary electron beam can be decelerated, and the irradiation energy of the electron beam to the sample 9 can be adjusted to an optimum value without changing the potential of the electron gun 10. When the electron beam 19 needs to be blanked, the electron beam 19 is blanked by the blanking deflector 13.
Is deflected so that the electron beam 19 does not pass through the stop 14.

The secondary electrons 51 generated by irradiating the sample 9 with the electron beam 19 are accelerated by the negative voltage applied to the sample 9. The ExB deflector 18 is disposed above the sample 9, and the accelerated secondary electrons 51 are deflected in a predetermined direction. The amount of deflection can be adjusted by the voltage applied to the ExB deflector 18 and the strength of the magnetic field. Further, this electromagnetic field can be changed in conjunction with a negative voltage applied to the sample. The secondary electrons 51 deflected by the ExB deflector 18 are reflected by the reflector 1 under predetermined conditions.
Collision 7 This reflector 17 has a conical shape,
The electron beam 19 passes through an opening provided at the center.
When the accelerated secondary electrons 51 collide with the reflector 17, a second secondary electron 52 having an energy of several V to 50 eV is generated from the reflector 17.

In the secondary electron detector 7, the secondary electron detector 20 is disposed above the objective lens 16 in the inspection room 2, detects the second secondary electrons 52, and
Output signal from the preamplifier 2 installed outside the inspection room 2
The signal is amplified by 1 and converted into digital data by the AD converter 22. The digital data is transmitted from the optical conversion unit 23 to the electric conversion unit 25 of the image processing unit 5 by the optical transmission unit 24. As described above, since the detected analog signal is digitized and transmitted immediately after the detection, a high-speed signal having a high SN ratio can be obtained. When the reflection plate 17 is not provided, the secondary electrons 51 are used instead of the second secondary electrons 52 instead of the secondary electrons 52.
It may be detected as 0.

The high voltage power supply 26 includes a preamplifier drive power supply 27 for driving the preamplifier 21, an AD converter drive power supply 28 for driving the AD converter 22, and a secondary electron detector 20 for attracting a second secondary electron. Power is supplied to a reverse bias power supply 29 that supplies a voltage to be applied to the power supply. Reverse bias power supply 2
The supply of 9 generates an attractive electric field in the secondary electron detector 20, and this electric field guides the second secondary electrons to the secondary electron detector 20. The secondary electron detector 20 has an electron beam 19.
The secondary electrons 51 generated during the irradiation of the sample 9 are then accelerated and collide with the reflector 17 to generate the second secondary electrons 52 in conjunction with the scanning timing of the electron beam 19. It is configured to detect. Note that FIG.
In the above description, the secondary electron detector 20 is applied with the reverse bias voltage from the reverse bias power supply 29. However, the configuration may be such that the reverse bias voltage is not applied. As the secondary electron detector 20, a PIN semiconductor detector or another type of semiconductor detector, for example, a Schottky semiconductor detector, an avalanche semiconductor detector, or the like may be used. Moreover, if conditions such as responsiveness and sensitivity are satisfied, an MCP (micro channel plate) can be used as a detector.

The image processing section 5 includes a first storage section 46, a second storage section 47, a calculation section 48, a defect determination section 49, and a monitor 50. The image signal of the sample 9 transmitted by the optical fiber 24 is stored in the first storage unit 46 or the second storage unit 47 after being converted into an electric signal again by the electric conversion unit 25. The arithmetic unit 48 performs various image processings on the stored image signal to align with the image signal of the other storage unit, standardize the signal level, and remove the noise signal. Perform a comparison operation on the signals. The defect judging unit 49 compares the absolute value of the difference image signal calculated by the operation unit 48 with a predetermined threshold value. If the difference image signal level is larger than the predetermined threshold value, the pixel is determined to be defective. The position is determined as a candidate, and the number of defects and the like are displayed on the monitor 50. Further, a function of displaying the captured electron beam image or optical image on the monitor 50 as it is is also provided.

Although not shown, in the present invention, the first storage unit 46 and the second storage unit 47 are used to acquire a plurality of images.
Is composed of several blocks. One block stores a pattern image obtained under predetermined inspection conditions. The image signal input to the first storage unit 46 or the second storage unit 47 is block-selected and stored according to the designation from the control unit 6 in advance, so that a plurality of images for each inspection condition are simultaneously obtained. can do. At this time, the operation of each unit accompanying the inspection is performed by the control unit 6 designating data based on the inspection condition to the control circuit 43. Therefore, the control unit 6 manages the relationship between the image signal and the inspection condition. are doing. Further, in accordance with designation from the control unit 6, comparison of these stored images, that is, comparison of an image of a certain block with another block is performed. The resulting data is stored in the defect determination unit 49. The relationship between the result data, the inspection condition, and the image is managed by the control unit 6. With the configuration shown in one embodiment of the present invention,
The present invention is implemented.

(Operation Procedure of Circuit Pattern Inspection) FIG. 3 is a diagram simply showing the operation procedure of the circuit pattern inspection. FIG.
An operation procedure when a semiconductor wafer subjected to pattern processing in a manufacturing process is inspected by the circuit pattern inspection apparatus 1 shown in FIG. 2 will be described with reference to FIG. 3 and FIG.

First, after loading a wafer, rough alignment is performed by an optical microscope (steps 301 and 30).
2). This is because the scanning range is narrow in an electron beam scanning image, and there is a possibility that the alignment mark does not fall within the scanning range of the electron beam only by mechanical positioning during loading. Therefore, alignment is performed with an accuracy of approximately several tens of μm using a low-magnification optical microscope having a wide field of view.

The details of the procedures 301 and 302 are as follows. First, the sample 9 is loaded into a sample exchange chamber (not shown). After the sample 9 is mounted and fixed on a sample holder (not shown), the sample exchange chamber is evacuated, and when the sample exchange chamber reaches a certain degree of vacuum, it is transferred to the inspection chamber 2. In the inspection room 2, the sample holder is placed on the sample table 30 via the X stage 31, the Y stage 32, and the rotating stage 33, and held and fixed. The sample 9 is placed on the X stage 31 and the Y stage 3 based on predetermined inspection conditions registered in advance.
2 is moved to the predetermined first coordinates below the optical microscope unit 4 by the movement in the X and Y directions.
An optical microscope image of the circuit pattern formed on is observed. Then, it is compared with an equivalent circuit pattern image stored in advance for position rotation correction, and a position correction value of the first coordinate is calculated. Next, it moves away from the first coordinate by a certain distance, moves to the second coordinate where a circuit pattern equivalent to the first coordinate exists, and similarly observes the optical microscope image and stores it for position rotation correction. The position correction value of the second coordinates and the amount of rotation deviation with respect to the first coordinates are calculated by comparing the corrected circuit pattern image with the calculated circuit pattern image. The rotation stage 33 is rotated by the calculated rotation shift amount to correct the rotation. Although the rotation amount is corrected by the rotation of the rotation stage 33, a method in which the rotation stage 33 is not provided and the scanning deflection position of the electron beam is corrected based on the calculated rotation amount is also used. The amount can be corrected. Next, for future position correction, the first coordinate, the positional deviation amount of the first circuit pattern by optical microscope image observation, the second coordinate, the positional deviation amount of the second circuit pattern by optical microscope image observation Is sent to the control unit 6 and stored. Next, a circuit pattern formed on the sample 9 is observed by an optical microscope, and the position of a chip where the circuit pattern is located, the distance between chips, or the repetition pitch of a repetitive pattern such as a memory cell is measured in advance. The measured value is input to the control unit 6. Further, a chip to be inspected and a region to be inspected in the chip are designated and input to the control unit 6.
The image of the optical microscope can be observed at a relatively low magnification, and when the surface of the sample 9 is covered with, for example, a silicon oxide film, the image can be observed through the base. In addition, the arrangement of the chips and the layout of the circuit patterns in the chips can be easily observed, and the region to be inspected can be easily set. Preparatory work such as predetermined correction work and inspection area setting by the optical microscope unit 4 is performed as described above.

Next, the conditions of the electron optical system, which is a part of the inspection conditions, are set (step 303). First, the irradiation energy of the electron beam, the pixel size, the beam current, and the like are set according to the type of the pattern of the wafer to be inspected. Here, the optimum irradiation energy of the electron beam differs depending on the material of the pattern. Normally, the conductive material is set to a higher energy of several keV or more so as to obtain high resolution, and is preferably set to 1.5 keV or less for a pattern containing an insulator in order to prevent charging. In FIG. 2, a negative voltage for decelerating the primary electrons of the electron beam 19 is applied to the sample 9 by the retarding power supply 36, and the irradiation energy of the electron beam 19 can be appropriately adjusted by adjusting this voltage. Make up. As for the pixel size, a size is set according to the defect size of particular interest. Insignificantly reducing the pixel size increases the time required for inspection. Although the default value of the beam current is unique to the apparatus, it is preferable to set the beam current to a smaller value especially for a wafer which is easily charged. When it is known in advance that the contrast of a defect to be detected is large, it is possible to narrow the electron beam by reducing the current, so that high-resolution inspection can be performed. Subsequent to the above setting, an image of the wafer to be inspected by the electron beam is displayed, and the focus and astigmatism are corrected, and the detection system gain is set. This can be automated by capturing an image and using a computer or a dedicated image processing device.

After the above setting is completed, next, the image is finely aligned by the electron beam (step 30).
4). As a result, the coordinate system of the stage and the coordinate system of the wafer accurately match, and a desired position on the wafer can be inspected by measuring the coordinates of the stage with a laser length measuring device or the like. In this procedure 304, first, the X stage 31
The sample 9 is moved below the electron optical system 3 by the movement of the Y stage 32. Based on the coordinate values stored and corrected by the alignment based on the optical microscope image, the electron beam 19 is two-dimensionally scanned in the X and Y directions by the scanning deflector 15 in the same circuit pattern as observed by the optical microscope unit 4. Is scanned. Due to the two-dimensional scanning of the electron beam, secondary electrons 51 are generated from the observed part, and the second secondary electrons 52 generated by the reflection plate 17 are detected by the secondary electron detector 20 to obtain an electron beam image. Is done. Using this image, procedure 302
The same operation as the correction operation by the optical microscope unit 4 and the setting of the inspection area in the above is performed by the electron beam image. In this work, simple inspection position confirmation and position alignment and position adjustment have already been performed using the optical microscope image, and rotation correction has also been performed in advance, so no large adjustment is required.
In addition, the electron beam image has higher resolution than the optical image, and can perform positioning, position correction, and rotation correction with high magnification and high accuracy.

Next, an area to be inspected is inputted and set from a control screen such as a display of the control section 6 (step 30).
5) Then, a test inspection is performed (procedure 306). In this test inspection, since images are acquired under a plurality of electron optical system conditions, various parameters of a plurality of inspection conditions to be tested are set. These inspection conditions include scanning conditions,
Details of the scanning conditions will be described later.

The test inspection is performed to determine whether the various parameters set above are appropriate. First, a part of the pattern to be inspected is specified. This is the test inspection area designation. Next, an image is detected by electron beam scanning in the same manner as in the actual inspection, but in the trial inspection, the acquired image is entirely taken into the memory without passing through the image processing circuit. Next, image processing conditions are set.
This is also included in the inspection conditions, and when the image processing apparatus compares two images of the same adjacent pattern,
This is a setting of a threshold value and a filter of a brightness and a positional deviation for determining whether or not a defect is present. When these thresholds are set low, the defect detection sensitivity is improved, but the possibility that non-defect portions are regarded as defects increases. On the other hand, when the threshold value is increased, the detection sensitivity decreases. Here, the brightness threshold value is a threshold value of a difference between luminance signals. Since the image formed by the electron beam is rougher than the optical microscope image (that is, the S / N ratio of the image is smaller), even if the inspection objects to be compared are the same and the average brightness is the same, it is slightly different for each pixel. There is light-dark variation. Therefore, a light and dark threshold value is provided, and images are compared to detect only a portion where a light and dark difference is greater than the threshold value as a defect, so that the roughness of the image is not recognized as a defect. Further, the positional deviation threshold value is provided so that the incomplete alignment of the two images to be compared is not erroneously recognized as a defect. It is considered that there is an image positional deviation about the threshold set here, and only when a difference larger than that occurs, is judged as a defect. A filter is a filter applied to an acquired image,
It is used to reduce the roughness of the detected image or to emphasize the edges. In this method, an optimum one is selected by trial and error according to a type of a pattern and a defect to be detected.

When the above-described image processing conditions are set, a comparative inspection is performed using a series of images that have already been acquired using the image processing conditions and loaded into the memory. This inspection detects a defect, stores its coordinates, and displays the location of the defect on the control unit screen. If no defect is detected at this time, the sensitivity is increased by lowering the threshold value of the image processing condition, and the comparative inspection is performed again using the image on the memory.

Next, an image of the coordinates detected in the test inspection is observed. This is a review (procedure 307). Since the image of that portion has already been stored in the memory, the image is first displayed on the display of the control section 6, a display dedicated to image display, or the like, and what kind of defect is detected is visually observed. Here, for a considerably large defect, the shape, type, and the like of the defect can be easily found visually. When the shape and type of the defect are found in the review of the test inspection, appropriate values such as the optical system condition, the scanning condition, the comparison method, and the inspection sensitivity as the inspection condition for the main inspection are derived. From the results,
By repeatedly narrowing down the parameters and repeating the review, it is possible to find an inspection condition that can reliably detect a defect to be detected.

However, in many cases, fine defects close to the performance limit of the apparatus and defects having a small contrast cannot be determined sufficiently visually or even the presence or absence of a defect cannot be determined. Therefore, a high-resolution SEM dedicated to review may be used as an observation device. In addition, if the sample is visually observed for a long time, the sample is irradiated with a large amount of an electron beam, and the sample may be destroyed or an image may not be observed due to charging. As described above, unlike the inspection conditions, the observation conditions of the observation device (corresponding to the inspection conditions of the inspection device) at the time of review do not require high-speed image acquisition, and can be clearly observed with higher resolution (larger magnification). Must be a condition. Conventionally, one of the causes of failure to determine a defect is a setting failure of an observation condition, and in this condition setting work, trial and error has to be repeated in trial inspection and review, which requires a great deal of labor and time. . In the present invention, the relationship between the detected defect and the optimal inspection condition can be analyzed by various types of comparison methods using images acquired under a plurality of conditions for one inspection region,
In addition, by acquiring and recording the relationship between the inspection conditions and the observation conditions by the tendency obtained by accumulating the past data and the detailed adjustment during the trial inspection, it is possible to simplify the condition optimization work,
High reliability can be realized. In this way, in this inspection (actual automatic inspection), the observation conditions for the detected defect are known in advance, and the review can be easily performed.

It is determined whether or not the desired inspection sensitivity and accuracy have been obtained by the above-described review, and if insufficient, the procedure returns to step 306 to change the inspection conditions of the trial inspection and repeat the following ( Step 308). Then, when the desired result is obtained, the main inspection is started (procedure 309). This inspection is performed automatically and unattended. When this inspection is completed, a review is performed again (step 310). As a result, when it is determined that sufficient inspection sensitivity and accuracy are not realized, the test inspection of step 306 is repeated (step 311).
If a sufficient result is obtained, the wafer to be inspected is unloaded (step 312), and if necessary, the wafer is moved to another analysis / inspection apparatus to perform various inspection / analysis (step 313).

(Relationship Between Defect Image and Inspection Condition) Next, a detailed comparison will be made on how the defect image obtained by the inspection condition is affected, and a comparison method will be examined. There are many types of circuit pattern defects. First, there is a difference depending on the material. In addition, there are differences in defect positions and shapes even with the same material. For example, the appearance of an observed image differs between a defect on the sample surface and a defect at the bottom of a contact hole having a large aspect ratio.
As an example of the types of the defects, the defects of the circuit pattern formed on the wafer include those called short circuit, disconnection, isolated defect, defective opening, uneven thickness of oxide film, and the like.

FIG. 4A shows that each material is charged.
The relationship between the irradiation energy E and the secondary electron generation rate δ when irradiating the material 111, 112, 113 with an electron beam when not performed is shown. The higher the secondary electron generation rate δ, the more secondary electrons are generated, and thus the acquired image is displayed brighter. FIG. 4B shows the energy E of FIG.
The brightness L1 to L3 of the acquired image when each of the materials 111 to 113 is irradiated with 1, E2, and E3 is shown. However, this brightness indicates a relative relationship with respect to the same irradiation energy, and L1 (brightest)> L2 (second brightest)> L3 (darkest). When the energy is different, the same brightness is not always the same even with the same L2. As can be seen from FIG. 4B, the materials 111 and 112 cannot be identified in the case of the irradiation energy E1, and the materials 111 and 113 cannot be identified in the case of the irradiation energy E3.
In this case, to detect three materials, the contrast is greatest when the electron beam is irradiated under the condition of the irradiation energy E2. However, the farther the secondary electron generation rate δ is from 1, the more charge-up occurs. Secondary electron generation rate δ
Is larger than 1, the amount of emitted electrons is larger than the amount of incident electrons, and is positively charged. If smaller than 1, it is negatively charged. Therefore, when the electron beam is irradiated under the condition of the irradiation energy E2, the brightness of the image is not stabilized, and the difference between the three materials finally disappears (contrast).
The light and dark are reversed depending on the surface potential. Therefore, in the case of such a sample, in the present invention, images can be obtained under the respective conditions of the irradiation energies E1 and E3 at which charge-up hardly occurs, and stable detection can be performed by comparing the plurality of images. Becomes

FIG. 5A is a schematic diagram showing the influence of charge-up when the sample has irregularities. If there is no charge in the vicinity, it goes straight like secondary electrons 115, is absorbed like secondary electrons 114 in the vicinity of positively charged material 501, and conversely, is near negatively charged material 502. It is returned like secondary electrons 116. Since the brightness of the image at the portion where the secondary electrons do not travel straight is displayed dark, the charge-up affects the shape and dimensions. In some cases, charge-up charges flow on the sample surface, causing unevenness in the acquired image. When an image is obtained by scanning an electron beam, a portion to which the electron beam is first applied (including a portion which has been irradiated immediately before) is being charged. Therefore, depending on the scanning direction of the electron beam and the scanning interval (distance, time). The acquired images may be different. However, the brightnesses L1 to L5 in FIG.
2> L3 (L4)> L5.

FIG. 5B shows a case where the sample is formed of the insulating material 503 and a case where a positive charge is properly applied only to the surface. In this case, 2 from the bottom 121 of the hole having a large aspect ratio such as a contact hole.
The secondary electrons 122 are accelerated by the positive charges of the insulating material 503, and the detection efficiency is improved (brighter). When there is no charge, the material is absorbed by the material 503, so that the detection efficiency is poor. Therefore, in the contact hole inspection, it is more convenient to detect defects if charging is performed properly. Further, in this case, a potential is generated on the surface due to the dielectric constant between the charged material surface and the substrate 123. When the conductor 504 exists inside the insulating material 503 as shown in FIG.
The potential of the surface 119 and the potential of the surface 120 are different. Since the amount of secondary electrons generated also varies depending on the potential difference (potential contrast), in this case, a defect of the conductor 504 inside the insulating material 503 can be secondarily observed. However, the brightnesses L1 to L3 in FIG. 5B have a relationship of L1>L2> L3.

As described above, the optimum inspection conditions are completely different depending on the material and the shape of the sample and the defect. The inspection conditions here include the amount of electron beam irradiation per unit area, the electron beam current value, the scanning speed of the electron beam, and the irradiation energy of the electron beam irradiating the sample. Although the above described the physical phenomena in the image detection process very simply, actually more complex phenomena have occurred, and the determination of the optimal conditions for the above parameters requires a great deal of labor and time. It is necessary to determine the optimum values of these parameters for each circuit pattern shape and material. Since it is necessary to freely adjust and control the irradiation energy of the electron beam irradiating the sample, as described above, in the apparatus shown in FIG. 2, a negative voltage for decelerating the primary electrons is applied to the sample by the retarding power supply, and this is applied. The configuration is such that the irradiation energy of the primary electron beam can be appropriately adjusted by adjusting the voltage. In addition, since the types of defects that can be detected are limited by one type of inspection condition, the present invention obtains a plurality of images having different contrasts and brightnesses by combining several inspection conditions, and performs comparison and evaluation. It enables good defect detection. Further, in the present invention, the observation condition is determined based on the inspection condition with the highest detection efficiency of the detected defect, so that the defect can be visually observed efficiently and reliably. At this time, it is desirable that the observation device is provided with an observation condition control function such as charge-up such as the inspection device or more.

(Outline of Scanning Conditions) Next, the scanning method will be described in detail. The method of irradiating the sample with the charged particles includes a batch exposure method for a certain area (in this case, an image is obtained by imaging secondary electrons distributed two-dimensionally), and a charging method that is narrowed down as described above. There is a method of scanning a particle beam (beam).
Although the present invention is not limited thereto, it is assumed here that the sample is scanned by a charged particle beam. Since the scanning control of the circuit pattern inspection apparatus shown in FIG. 1 described above is a completely digital method, any scanning method can be realized by scanning parameters or program description. The images that can be obtained are completely different due to the influence of the above. Therefore, in addition to the condition of the electron optical system, the condition of the scanning method also becomes one factor of the inspection condition, and it is necessary to determine the condition according to the user's requirements, the electrical characteristics of the sample, and the required accuracy.

FIGS. 6 (a) and 6 (b) show examples of differences in scanning method by beam deflection, and FIGS. 6 (c) and 6 (d) show examples of differences in scanning method by stage movement. FIG. 6A shows a method of scanning in one direction, in which a dotted line indicates blanking so as not to irradiate the wafer with a beam in a retrace line. The scanning condition parameters at this time include a scanning speed and an interval (distance, time). Further, as described above, the image that can be obtained differs depending on the scanning direction such as the X direction, the Y direction, and the oblique direction depending on the circuit pattern of the sample. Also, do not blank this return line,
There is also a scanning method in which a retrace position, beam shape and size are controlled, and the next scan area or the area to be scanned is charged. The same position on the sample may be repeatedly scanned several times to acquire a plurality of images whose charge gradually increases. At this time, the management of the images in the storage units 46 and 47 of the image processing unit 5 is such that the image obtained in the first scan is in block 1, the image obtained in the second scan is in block 2, and so on. At the time of storage, addresses (storage blocks) are sorted and stored. However, the image management method is merely an example, and the image may be stored in the memory simply in the order of input at the time of storage, and the address may be converted when the image is taken out. Further, as for these stored images, an image on which an addition process, a difference process, a filter process, and the like have been performed may be stored again in the storage device and used as a comparison image.

FIG. 6B shows a method of scanning in a reciprocating direction.
Although there is no need for blanking and blanking,
Suitable for high-speed operation. However, since there is non-uniformity between the forward path and the return path, the image accuracy is inferior to that of the scanning method shown in FIG. The return path may be used as a dedicated scan for charge control.
In this case, the charge control can be performed with higher accuracy than the charge control using the retrace of the scanning method in FIG.

FIG. 6C shows the trajectory when the same scanning method as that of FIG.
FIG. 6D shows the locus of scanning in the step-and-repeat method. In the step-and-repeat method, an image for the deflection area is acquired by beam scanning while the stage is stopped at one time, and the stage is moved to the next inspection position by a step operation, and an image for the next deflection area is acquired by beam scanning. By repeating the above operation, images of the plurality of deflection areas are joined to obtain an image of the wafer. The continuous stage movement method obtains a continuous image by controlling the beam movement on the sample and the continuous movement of the stage. FIG.
As shown in (c) and (d), the beam scanning direction is performed in the reciprocating direction of the stage, and it is possible to inspect the entire surface of the sample having a wide area. The scanning area or the scanning operation on the outward or return path of the stage movement is called a stripe. Although the stage continuous movement method without the step operation time of the stage enables faster and continuous inspection,
Deflection control or stage control is required so that the target position does not deviate from the deflection area. The stage moving direction and the beam scanning direction may be any directions as long as the comparative inspection can be performed. However, since the pattern of the chip is rectangular, it is preferable to scan in accordance with the chip pattern direction. In this case, the stage movement direction is basically substantially orthogonal to the beam scanning direction. From the viewpoint of stage movement accuracy, it is better to operate by one axis alone than to operate by two axes. For this reason, in an actual inspection device,
In some cases, the chip direction, that is, the direction of the wafer is aligned with the stage axis.

As a method of acquiring a plurality of images having different inspection conditions by a scanning method involving stage movement, it is possible to scan the same area a plurality of times in one stripe to obtain an image. The inspection conditions may be changed for each step. This inspection condition change includes a charge-up operation by beam scanning. Considering these scanning methods and image acquisition methods in terms of the required memory amount, an enormous amount of memory is required to store all the images in a full inspection such as the main inspection. 1 scan is 1
000 pixels, 1 pixel of 0.1 μm, the memory capacity of one stripe, when the sample is a 300 mm wafer,
About a few gigabytes. Several tens in case of multiple image acquisition
A gigabyte capacity is sufficient, and such a capacity can be easily realized. However, in the case of a full image, this about 10
A capacity of 00 times is required. Therefore, from the viewpoint of the amount of memory, it is good to set the comparison image of about one stripe as one unit. However, when there is no time limit, using an off-line dedicated recording device makes it impossible to store the image of the entire inspection or compare the images. Not possible. The present invention is implemented by the above-described scanning conditions and image storage means.

(Operation Panel) Finally, with respect to the inspection apparatus shown in FIG. 1, the characteristic features of the operation for implementing the present invention will be described. The present invention does not limit the details of the operation panel, such as the display of the inspection apparatus, or the method of designating the inspection operation. However, there is a difference from the conventional case where only one type of inspection is performed. First, the method of specifying the inspection operation becomes somewhat complicated, and a main menu screen or a file for specifying the procedure (flow) of the inspection operation is added. In addition, a selection of a comparison image linked to the inspection condition, a comparison method selection, and an image display selection screen are added. It may be an additional function, but important functions include managing the correlation between inspection conditions and types of detected defects, sample conditions (process conditions, constituent material conditions), and the correlation between those conditions and observation conditions. Management functions are required. Record these data automatically or as a flow operation,
By making use of the analysis, it is possible to reduce the labor required for the user and to manage the reliability and reproducibility of the inspection device.

[0060]

The main effects of the present invention are as follows. 1) Simultaneous detection of a plurality of types of defects having completely different features becomes possible, and the defect detection function and defect detection efficiency can be improved. 2) Simultaneously with the improvement of the reliability of the feature classification of the defect, it is possible to easily perform the work of optimizing the inspection condition of the defect which is the detection purpose. 3) Since false detection of defects can be reduced, and thus the detection sensitivity can be set high, it is possible to increase the defect detection efficiency and at the same time improve the reliability of inspection. 4) The defect checking operation or the inspection condition optimizing operation can be simplified and ensured, the erroneous detection judgment can be made more reliable, and the inspection can be made more reliable.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram illustrating a circuit pattern inspection method according to the present invention;

FIG. 2 is a diagram showing a configuration example of a circuit pattern inspection device of the present invention.

FIG. 3 is a diagram illustrating an operation procedure of a circuit pattern inspection method according to the present invention.

FIG. 4 is an explanatory diagram of a relationship between electron beam irradiation energy and an acquired image.

FIG. 5 is an explanatory diagram of the influence of charge-up.

FIG. 6 is a diagram showing an example of electron beam scanning.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 circuit pattern inspection device 2 inspection room 3 electron optical system 4 optical microscope unit 5 image processing unit 6 control unit 7 secondary electron detection unit 8 sample room 9 substrate to be inspected 46 first image storage unit 47 second image storage unit 48 operation Unit 49 defect determination unit 50 monitor 101, 102 circuit pattern 103 irradiation area 104, 105 area 106, 107 image 108 to 110 comparison result

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification FI FI Theme Court II (Reference) H01J 37/28 G01R 31/28 L (72) Inventor Yasuhiro Gunji 882- Address Ichimo, Ichiki, Hitachinaka City, Ibaraki Pref. (72) Inventor Hiroyuki Shinada 1-280 Higashi Koigakubo, Kokubunji-shi, Tokyo Inside the Hitachi, Ltd.Central Research Laboratories Co., Ltd. F-term in the Measuring Instruments Group of the Works (Reference) PA12 2G132 AA00 AD15 AF12 AL09 AL11 4M106 AA01 BA02 CA38 CA70 DB05 DH33 DJ04 DJ06 DJ07 DJ17 DJ 18 DJ20 DJ21 5C033 UU05 UU06

Claims (5)

[Claims] A surface of a substrate on which a circuit pattern is formed is irradiated with a charged particle beam, and a defect generated in the circuit pattern is detected using an image signal obtained by detecting a signal generated from the circuit pattern by the irradiation. In the method of inspecting a circuit pattern to be detected, a plurality of inspection conditions including at least a charged particle beam irradiation condition or a charging condition of the substrate or a comparison condition are determined,
A circuit pattern inspection method, wherein a defect of the circuit pattern is extracted by comparing image signals obtained for each of the inspection conditions. 2. The circuit pattern inspection method according to claim 1, wherein first and second inspection conditions are determined, and an image of an inspection target region on the circuit pattern obtained under the first and second inspection conditions. Are the first and second images, the images of the comparison target pattern of the inspection target area obtained under the first and second inspection conditions are the third and fourth images, and the first and second images are the same.
A first comparison image is generated by comparing the image and the third image, and a second comparison image is generated by comparing the second image and the fourth image.
A circuit pattern inspection method, comprising: generating a comparison image of (1), and extracting a defect in the inspection target area by comparing the first comparison image with the second comparison image. 3. The circuit pattern inspection method according to claim 1, wherein first and second inspection conditions are determined, and an image of an inspection target area on the circuit pattern obtained under the first and second inspection conditions. A first and a second image, and comparing the first image with the second image to extract a defect in the inspection target area. 4. The circuit pattern inspection method according to claim 1, wherein, when the detected defect is reviewed using an observation device, the observation condition of the observation device is referred to by referring to the inspection condition at the time of detecting the defect. And performing a review under the conditions. 5. A circuit pattern inspection apparatus for inspecting a circuit pattern using the circuit pattern inspection method according to claim 1, wherein: an inspection condition setting unit for setting the plurality of inspection conditions; Control means for controlling the irradiation of the charged particle beam to the circuit pattern for each of the inspection conditions set by (1), and associating the image signal obtained for each of the inspection conditions with the inspection condition A second control means for performing control so as to store the data in a memory and perform a comparison process.