JP5474924B2 - Inspection apparatus using charged particle beam and device manufacturing method using the inspection apparatus - Google Patents

Description

  The present invention relates to an inspection apparatus for inspecting a defect or the like of a pattern formed on a surface to be inspected using an electron beam, and more specifically, inspecting an electron beam as in the case of detecting a wafer defect in a semiconductor manufacturing process. Inspection that irradiates the target and captures secondary electrons that change according to the surface properties to form image data, and inspects the pattern formed on the surface of the inspection target with high throughput based on the image data The present invention relates to an apparatus and a device manufacturing method for manufacturing a device with high yield using such an inspection apparatus. More specifically, the present invention relates to a detection apparatus using a projection method using a surface beam and a device manufacturing method using the apparatus.

In the semiconductor process, the design rule is about to reach the age of 100 nm, and the production form is changed from mass production of small varieties represented by DRAM to SOC (Silicon on
chip), and is shifting to high-mix low-volume production. Along with this, the number of manufacturing processes increases, and it is essential to improve the yield for each process, and defect inspection due to the process becomes important. The present invention relates to an apparatus used for inspecting a wafer or the like after each step in a semiconductor process, and relates to an inspection method and apparatus using an electron beam or a device manufacturing method using the same.

  With the high integration of semiconductor devices and the miniaturization of patterns, inspection apparatuses with high resolution and high throughput are required. In order to investigate defects on a wafer substrate of 100 nm design rule, it is necessary to look at pattern defects and particle / via defects in wirings having a line width of 100 nm or less and their electrical defects, and therefore a resolution of 100 nm or less is required. In addition, since the amount of inspection increases due to an increase in manufacturing process due to high integration of devices, high throughput is required. In addition, as the number of devices increases, the inspection apparatus is also required to have a function of detecting a contact failure (electrical defect) of a via that connects wirings between layers. At present, optical defect inspection equipment is mainly used. However, in terms of resolution and contact defect inspection, instead of optical defect inspection equipment, defect inspection equipment using electron beams will be used in the future. Expected to become mainstream. However, the electron beam type defect inspection apparatus has a weak point, which is inferior to the optical method in terms of throughput.

  Therefore, development of an inspection apparatus capable of detecting electric defects with high resolution and high throughput is required. The resolution in the optical system is said to be limited to 1/2 of the wavelength of the light to be used, and is about 0.2 μm in the case of visible light that has been put into practical use.

  On the other hand, in a method using an electron beam, a scanning electron beam method (SEM method) is usually put into practical use, with a resolution of 0.1 μm and an inspection time of 8 hours / piece (200 mm wafer). The electron beam method is also characterized by being capable of inspecting electrical defects (wire disconnection, conduction failure, via conduction failure, etc.), but a defect inspection device with a very low inspection speed and a high inspection speed. Development is expected.

  In general, an inspection apparatus is expensive and has a low throughput as compared with other process apparatuses. Therefore, after an important process at present, for example, after etching, film formation, or CMP (chemical mechanical polishing) planarization process, etc. It is used.

A scanning (SEM) type inspection apparatus using an electron beam will be described. The SEM inspection apparatus narrows the electron beam (this beam diameter corresponds to the resolution), scans it, and irradiates the sample in a line shape. On the other hand, by moving the stage in the direction perpendicular to the scanning direction of the electron beam, the observation region is irradiated with the electron beam in a planar shape. The scanning width of the electron beam is generally several 100 μm. A secondary electron from a sample generated by irradiation of the finely focused electron beam (referred to as a primary electron beam) is detected by a detector (scintillator + photomultiplier (photomultiplier tube) or semiconductor type detector (PIN diode type). Etc.). The coordinates of the irradiation position and the amount of secondary electrons (signal intensity) are combined and imaged and stored in a storage device, or an image is output on a CRT (CRT). The above is the principle of SEM (scanning electron microscope), and a defect in a semiconductor (usually Si) wafer in the process is detected from an image obtained by this method. The inspection speed (corresponding to the throughput) is determined by the amount of primary electron beam (current value), the beam diameter, and the response speed of the detector. The beam diameter is 0.1 μm (which may be considered to be the same as the resolution), the current value is 100 nA, the detector response speed is 100 MHz, and the inspection speed is said to be about 8 hours per 20 cm diameter wafer in this case. ing. A serious problem is that the inspection speed is extremely slow (1/20 or less) compared to the optical system. In particular, a device pattern with a design rule of 100 nm or less formed on a wafer, that is, detection of a shape defect such as a line width of 100 nm or less or a via having a diameter of 100 nm or less, or an electrical defect, and a high speed of dust of 100 nm or less. Detection is required.

  In the SEM inspection apparatus described above, the above inspection speed is considered to be almost the limit, and a new method is necessary to further increase the speed, that is, to increase the throughput.

To meet these needs, the present invention provides:
A method for inspecting a substrate having a pattern formed in a matrix of dies using electrons,
Placing the substrate on a stage with a specified direction;
Selecting a reference die to be a positioning reference, and obtaining a pattern matching template image including the coordinates of the feature points of the reference die;
Performing a pattern match on any die in a row or column containing the reference die using the template image to obtain the coordinates of the feature points of the any die;
Calculating a deviation angle formed by a row or column including the reference die and a direction in which the electrons scan the substrate based on the feature point coordinates of the arbitrary die and the reference die;
Moving the stage to correct the deviation angle and aligning the substrate;
Irradiating the electrons toward the substrate, and providing an inspection method.

  The electron that has obtained information on the surface of the sample is at least one of secondary electrons, reflected electrons, and backscattered electrons generated from the sample, or mirror electrons reflected near the surface of the sample. desirable.

  By the inspection method of the present invention, it becomes possible to inspect defects of a substrate such as a wafer having a wiring having a line width of 100 nm or less.

It is a figure which shows the whole structure of a semiconductor inspection apparatus. It is a figure which shows the whole structure of the apparatus of FIG. It is the figure which looked at the whole structure of the apparatus of FIG. 1 from the function. It is a figure which shows the main components of the test | inspection part of the apparatus of FIG. It is a figure which shows the main components of the test | inspection part of the apparatus of FIG. It is a figure which shows the main components of the test | inspection part of the apparatus of FIG. It is a figure which shows the main components of the test | inspection part of the apparatus of FIG. It is a figure which shows the main components of the test | inspection part of the apparatus of FIG. It is a figure which shows the main components of the test | inspection part of the apparatus of FIG. It is a figure which shows the main components of the test | inspection part of the apparatus of FIG. It is a figure which shows the test | inspection part exterior of the apparatus of FIG. It is a figure which shows the test | inspection part exterior of the apparatus of FIG. 1 is an elevation view showing main components of a semiconductor inspection apparatus according to the present invention. It is a front view which shows the main components of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows an example of a structure of the cassette holder of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the structure of the mini environment apparatus of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the structure of the loader housing of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the structure of the loader housing of the semiconductor inspection apparatus which concerns on this invention. (A) And (B) is a figure explaining the electrostatic chuck used for the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the electrostatic chuck used for the semiconductor inspection apparatus which concerns on this invention. (A) And (B) is a figure explaining the other example of the electrostatic chuck used for the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the bridge tool used for the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the other example of the bridge tool used for the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the structure and operation | movement procedure (A)-(C) of the elevator mechanism in the load lock room of FIG. It is a figure explaining the structure and operation | movement procedure (D)-(F) of the elevator mechanism in the load lock room of FIG. It is a figure which shows the modification of the support method of the main housing in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the modification of the support method of the main housing in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the structure of the electron optical system of the mapping projection type | formula electron beam inspection apparatus among the semiconductor inspection apparatuses which concern on this invention. It is a figure which shows the structure of the electron optical system of the scanning electron beam inspection apparatus among the semiconductor inspection apparatuses which concern on this invention. It is a figure which shows roughly the structure of an example of the detector rotation mechanism of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows roughly the structure of an example of the detector rotation mechanism of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows roughly the structure of an example of the detector rotation mechanism of the semiconductor inspection apparatus which concerns on this invention. 1 is a diagram showing a first embodiment of a semiconductor inspection apparatus according to the present invention. (1)-(5)-is a figure explaining the shape of a sample irradiation beam. (1-1)-(4) is a figure for demonstrating the irradiation shape of a linear beam. It is a figure explaining taking-out of the secondary electron from the lens-barrel in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows 2nd Embodiment of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows 3rd Embodiment of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows 4th Embodiment of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows 5th Embodiment of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the irradiation area | region which covers an observation area | region. It is a figure explaining an irradiation shape and irradiation efficiency. It is a figure which shows 6th Embodiment of the semiconductor inspection apparatus which concerns on this invention, and is a figure which shows the structure of the detection system using a relay lens. It is a figure which shows 6th Embodiment of the semiconductor inspection apparatus which concerns on this invention, and is a figure which shows the structure of the detection system using FOP. (A) And (B) is a figure which shows 8th Embodiment of the semiconductor inspection apparatus based on this invention. It is a graph which shows the opening part diameter dependence of the transmittance | permeability. It is a figure which shows the specific structural example of the electron detection system in the apparatus of FIG. (A) And (B) is a figure explaining the requirements for operating the electron detection system in the apparatus of FIG. 37 in three modes. It is a figure which shows the structure of the E * B unit of the semiconductor inspection apparatus which concerns on this invention. It is sectional drawing which follows the line A of FIG. It is a figure which shows 9th Embodiment of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the simulation of electric field distribution. It is a figure which shows the structure of the power supply part of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the circuit system which generate | occur | produces the DC voltage of the power supply part shown in FIG. It is a figure which shows an example of the circuit structure of the static bipolar power supply of the power supply part shown in FIG. It is a figure which shows the special power supply in the power supply part shown in FIG. It is a figure which shows the special power supply in the power supply part shown in FIG. It is a figure which shows the special power supply in the power supply part shown in FIG. It is a figure which shows an example of the power supply circuit for the retarding chuck | zipper in the power supply part shown in FIG. FIG. 46 is a diagram illustrating an example of a hardware configuration of an EO correction deflection voltage in the power supply unit illustrated in FIG. 45. It is a figure which shows an example of the circuit structure of the octopole conversion part in the power supply part shown in FIG. (A) shows an example of the circuit configuration of the high-speed high-voltage amplifier in the power supply section shown in FIG. 45, and (B) is a diagram showing the output. It is a figure which shows 1st Embodiment of the pre-charge unit of the semiconductor inspection apparatus shown in FIG. It is a figure which shows 2nd Embodiment of the pre-charge unit of the semiconductor inspection apparatus shown in FIG. It is a figure which shows 3rd Embodiment of the pre-charge unit of the semiconductor inspection apparatus shown in FIG. It is a figure which shows 4th Embodiment of the precharge unit of the semiconductor inspection apparatus shown in FIG. It is a figure which shows the imaging device provided with the pre-charge unit shown in FIGS. FIG. 61 is a diagram for explaining the operation of the apparatus in FIG. 60. It is a figure which shows the other structural example of the defect inspection apparatus provided with the precharge unit. FIG. 62 is a diagram showing a device for converting a secondary electron image signal into an electric signal in the device shown in FIG. 61. FIG. 62 is a flowchart for explaining the operation of the apparatus shown in FIG. 61. (A), (b), (c) is a figure explaining the defect detection method in the flowchart of FIG. It is a figure which shows the other structural example of the defect inspection apparatus provided with the precharge unit. It is a figure which shows the further another structural example of the defect inspection apparatus provided with the precharge unit. It is a figure explaining operation | movement of the control system of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining operation | movement of the control system of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining operation | movement of the control system of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining operation | movement of the control system of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining operation | movement of the control system of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining operation | movement of the control system of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining operation | movement of the control system of the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the alignment procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the alignment procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the alignment procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the defect inspection procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the defect inspection procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the defect inspection procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the defect inspection procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the defect inspection procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the defect inspection procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the defect inspection procedure in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the structure of the control system in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the structure of the user interface in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the structure of the user interface in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining the other function and structure of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the electrode in the other function and structure of the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the electrode in the other function and structure of the semiconductor inspection apparatus which concerns on this invention. It is a graph which shows the voltage distribution between a wafer and an objective lens. It is a flowchart explaining the secondary electron detection operation | movement in the other function and structure of the semiconductor inspection apparatus which concerns on this invention. FIG. 92 is a diagram showing a potential application mechanism in the apparatus shown in FIG. 91. (A) And (B) is a figure explaining the electron beam calibration method in the apparatus shown in FIG. (A) And (B) is a figure explaining the alignment control method in the apparatus shown in FIG. (A) And (B) is a figure explaining the concept of EO correction | amendment in the apparatus shown in FIG. FIG. 92 is a diagram illustrating a specific device configuration for EO correction in the apparatus shown in FIG. 91. (A) And (B) is a figure explaining EO correction | amendment in the apparatus shown in FIG. It is a figure explaining EO correction | amendment in the apparatus shown in FIG. It is a figure explaining EO correction | amendment in the apparatus shown in FIG. It is a figure explaining EO correction | amendment in the apparatus shown in FIG. It is a figure explaining the idea of a TDI transfer clock. It is a figure explaining the idea of a TDI transfer clock. FIG. 103 is a timing chart illustrating operation of the circuit in FIG. 102. It is a figure which shows the modification of the defect inspection apparatus which concerns on this invention. FIG. 105 is a flowchart for explaining the operation of the apparatus shown in FIG. 104. FIG. 105 is a flowchart for explaining the operation of the apparatus shown in FIG. 104. FIG. 105 is a flowchart for explaining the operation of the apparatus shown in FIG. 104. FIG. 105 is a flowchart for explaining the operation of the apparatus shown in FIG. 104. FIG. 105 is a flowchart for explaining the operation of the apparatus shown in FIG. 104. It is a figure explaining the semiconductor device manufacturing method concerning the present invention. It is a figure explaining the semiconductor device manufacturing method concerning the present invention. It is a figure explaining the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the basic flow of the inspection procedure of the semiconductor device manufacturing method concerning the present invention. It is a figure which shows the setting of inspection object die | dye. It is a figure explaining the setting of the inspection area | region inside die | dye. It is a figure explaining the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. (A) And (B) is a figure explaining the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure which shows the scanning example in the test | inspection procedure of the semiconductor device manufacturing method which concerns on this invention, when there exists one test | inspection die. It is a figure which shows an example of an inspection die. It is a figure explaining the production | generation method of a reference image in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the adjacent die comparison method in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the adjacent die comparison method in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the reference | standard die comparison method in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the reference | standard die comparison method in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the reference | standard die comparison method in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the focus mapping in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the focus mapping in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the focus mapping in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the focus mapping in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the focus mapping in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the focus mapping in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the litho margin measurement in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the litho margin measurement in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the litho margin measurement in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the litho margin measurement in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the litho margin measurement in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the litho margin measurement in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure explaining the litho margin measurement in the test | inspection procedure of the semiconductor device manufacturing method concerning this invention. It is a figure which shows an example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows an example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows an example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the other example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the other example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows the further another example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows another example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows another example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows another example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows another example of the stage apparatus in the semiconductor inspection apparatus which concerns on this invention. (A) And (B) is a figure which shows the conventional stage apparatus. It is a figure which shows the optical system and detector in the semiconductor inspection apparatus which concern on this invention. (A) And (b) is a figure which shows other embodiment of the semiconductor inspection apparatus based on this invention. It is a figure which shows the electron beam apparatus of FIG. 150 in detail. It is a figure which shows the primary electron irradiation method in the semiconductor inspection apparatus which concerns on this invention. It is a figure which shows embodiment of the semiconductor inspection apparatus which concerns on this invention, and is equipped with the electrode structure which prevents a dielectric breakdown. 153 is a table for explaining the operation of the apparatus in FIG. 153. FIG. 157 is a diagram showing a structure of an electrode in the device of FIG. 153. FIG. 157 is a diagram showing a structure of an electrode in the device of FIG. 153. FIG. 157 is a diagram showing a structure of an electrode in the device of FIG. 153. FIG. 157 is a diagram showing a structure of an electrode in the device of FIG. 153. It is a figure which shows embodiment of the semiconductor inspection apparatus which concerns on this invention, and is equipped with the damping device. (A)-(c) is a figure explaining the apparatus of FIG. It is a figure explaining the apparatus of FIG. It is a figure explaining the apparatus of FIG. It is a figure explaining the apparatus of FIG. FIGS. 9A to 9C are diagrams for explaining a pattern matching method in the apparatus of FIG. It is a figure explaining holding | maintenance of the wafer in the semiconductor inspection apparatus which concerns on this invention. It is a figure explaining holding | maintenance of the wafer in the semiconductor inspection apparatus which concerns on this invention. (A) And (b) is a figure explaining holding | maintenance of the wafer in the semiconductor inspection apparatus which concerns on this invention. 167 is a diagram showing an electron beam apparatus including the chuck described in FIG. 166. FIG. FIG. 168 is a diagram showing an E × B separator in the apparatus shown in FIG. 168. FIG. 168 is a diagram showing an E × B separator in the apparatus shown in FIG. 168. It is a figure which shows embodiment which connected the inspection apparatus which concerns on this invention to the production line. (A) is a diagram schematically showing an embodiment of a mapping projection type electron beam apparatus that can be used by switching between secondary electrons and reflected electrons, and (B) shows the configuration of the secondary optical system. It is a figure shown roughly. 172 is a diagram showing a specific configuration of the secondary electron detection system in FIG. (A) And (B) is a figure explaining the different operation mode of the defect inspection apparatus shown to (A) of FIG. 172 is a diagram illustrating a specific configuration of a lens of the secondary optical system of the defect inspection apparatus illustrated in FIG. 172A. FIG. (A) is a figure which shows schematically the structure of the modification of the mapping projection type electron beam apparatus shown to (A) of FIG. 172, (B) is a figure explaining the scanning method. (A) is a figure which shows roughly the structure of the other modification of the mapping projection type electron beam apparatus shown to (A) of FIG. 172, (B) is a figure explaining the scanning method. 172 is a diagram showing the structure of a vacuum chamber and an XY stage of the mapping projection type electron beam apparatus shown in FIG. 172A and an inert gas circulation piping system therefor. FIG. 178 is a diagram showing an example of the differential exhaust mechanism in FIG. 178. FIG. It is a figure showing roughly the composition of the whole inspection system.

Hereinafter, embodiments of a semiconductor inspection apparatus according to the present invention will be described in detail in the following order with reference to the drawings.
Notes 1. Overall configuration 1-1) Main chamber, stage, vacuum transfer system exterior 1-1-1) Active vibration isolation table 1-1-2) Main chamber 1-1-3) XY stage 1-2) Laser interference measurement system 1 -3) Inspection unit exterior 2. Embodiment 2-1) Conveying system 2-1-1) Cassette holder 2-1-2) Mini-environment device 2-1-3) Main housing 2-1-4) Loader housing 2-1-5) Loader 2-1-6) Stage device 2-1-7) Wafer chucking mechanism 2-1-7-1) Basic structure of electrostatic chuck 2-1-7-2) Chucking for 200/300 bridge tool Mechanism 2-1-7-3) Wafer chucking procedure 2-1-8) Apparatus configuration for 200/300 bridge tool 2-2) Wafer transfer method 2-3) Electro-optical system 2-3-1) Outline 2-3-2) Configuration details 2-3-2-1) Electron gun (electron beam source)
2-3-2-2) Primary optical system 2-3-2-3) Secondary optical system 2-3-3) E × B unit (Wien filter)
2-3-4) Detector 2-3-5) Power supply 2-4) Precharge unit 2-5) Vacuum exhaust system 2-6) Control system 2-6-1) Configuration and function 2-6-2) Alignment procedure 2-6-3) Defect inspection 2-6-4) Control system configuration 2-6-5) User interface configuration 2-7) Description of other functions and configuration 2-7-1) Control electrode 2-7 -2) Potential application method 2-7-3) Electron beam calibration method 2-7-4) Electrode cleaning 2-7-5) Alignment control method 2-7-6) EO correction 2-7-7) Image Comparison Method 2-7-8) Device Manufacturing Method 2-7-9) Inspection 2-8) Inspection Method 2-8-1) Overview 2-8-2) Inspection Algorithm 2-8-2-1) Array Inspection 2 -8-2-2) Random inspection 2-8-2-3) Focus mapping 2-8- -4) lithography margin measurement 3. Other Embodiments 3-1) Modified Examples of Stage Device 3-2) Other Embodiments of Electron Beam Device 3-2-1) Electron Gun (Electron Beam Source)
3-2-2) Structure of electrode 3-3) Embodiment related to vibration control device 3-4) Embodiment related to wafer holding 3-5) Embodiment of E × B separator 3-6) Production line Embodiments of 3-7) Embodiments using other electrons 3-8) Embodiments using secondary electrons and reflected electrons.

1. Overall Configuration First, the overall configuration of the semiconductor inspection apparatus will be described.
The overall configuration of the apparatus will be described with reference to FIG. The apparatus includes an inspection apparatus main body, a power supply rack, a control rack, an image processing unit, a film forming apparatus, an etching apparatus, and the like. A roughing pump such as a dry pump is placed outside a clean room. As shown in FIG. 2, the main part inside the inspection apparatus main body is composed of an electron beam optical barrel, a vacuum transfer system, a main housing that houses a stage, a vibration isolation table, a turbo molecular pump, and the like.

  The control system includes two CRTs and a command input function (such as a keyboard). FIG. 3 shows the configuration in terms of function. The electron beam column mainly includes an electron optical system, a detection system, an optical microscope, and the like. The electron optical system includes an electron gun, a lens, and the like, and the transfer system includes a vacuum transfer robot, an atmospheric transfer robot, a cassette loader, various position sensors, and the like.

  Here, the film forming apparatus, the etching apparatus, and the cleaning apparatus (not shown) are arranged side by side near the inspection apparatus main body, but these may be incorporated in the inspection apparatus main body. These are used, for example, for suppressing charge of the sample or for cleaning the sample surface. When the sputtering method is used, both functions of controlling film and etching can be provided by one unit.

  Although not shown, depending on the intended use, the related devices may be installed side by side near the inspection apparatus main body, or the related apparatuses may be incorporated into the inspection apparatus main body and used. Or you may incorporate an inspection apparatus in those related apparatuses. For example, a chemical mechanical polishing apparatus (CMP) and a cleaning apparatus may be incorporated in the inspection apparatus main body, or a CVD (chemical vapor deposition) apparatus may be incorporated in the inspection apparatus. Advantages such as saving the number of units for conveyance and shortening the conveyance time can be obtained.

Similarly, the inspection apparatus main body may be incorporated in a film forming apparatus such as a plating apparatus. It can also be used in combination with a lithographic apparatus.
1-1) Main chamber, stage, vacuum transfer system exterior In FIGS. 4, 5, and 6, main components of the inspection unit of the semiconductor inspection apparatus are shown. The inspection unit of the semiconductor inspection apparatus includes an active vibration isolation table 4. 1 for blocking vibrations from the external environment, a main chamber 4. 2 serving as an inspection room, and an electro-optical device 4. 3, an XY stage 5, 1 for wafer scanning mounted inside the main chamber, a laser interference measurement system 5, 2 for controlling the operation of the XY stage, and a vacuum transfer system 4, 4 attached to the main chamber, They are arranged in a positional relationship as shown in FIGS. The inspection unit of the semiconductor inspection apparatus further includes an exterior 6. 1 for enabling environmental control and maintenance of the inspection unit, and is arranged in a positional relationship as shown in FIG. 6.

1-1-1) Active anti-vibration table Active anti-vibration table 4.1 is equipped with an active anti-vibration unit 5, 3 on which a welding surface plate 5, 4 is mounted, and is an inspection room on this welding surface plate. The main chambers 4 and 2, the electron optical devices 4 and 3 installed on the upper part of the main chamber, and the vacuum transfer systems 4 and 4 associated with the main chamber are held. Thereby, the vibration from the external environment in a test | inspection part can be suppressed now. In this embodiment, the natural frequency is within ± 25% with respect to the X direction 5 Hz, the Y direction 5 Hz, and the Z direction 7.6 Hz, and the control performance is 0 dB or less at 1 Hz in the transfer characteristics of each axis. It is -6.4 dB or less at 7.6 Hz, -8.6 dB or less at 10 Hz, and -17.9 dB or less at 20 Hz (above, no load on the surface plate). In another structure of the active vibration isolation table, the main chamber, the electron optical device, and the like are suspended and held. In still another structure, a stone surface plate is mounted to hold the main chamber and the like.

1-1-2) Main chamber The main chambers 4 and 2 hold the turbo molecular pumps 7 and 2 directly in the lower part in order to realize a vacuum degree (10 ⁻⁴ Pa or less) as an inspection environment, and perform wafer scanning. A high-precision XY stage 5. 1 is provided on the inside to shield the magnetism from the outside. In this embodiment, the following structure is used to improve the flatness of the installation surface of the high-precision XY stage as much as possible. The lower plates 7 and 3 of the main chamber are installed and fixed on the portions 7 and 4 having a particularly good flatness (in this embodiment, flatness of 5 μm or less) prepared on the welding surface plate. Further, an intermediate plate is provided inside the main chamber as a stage installation surface. The middle plate is supported at three points with respect to the lower plate of the main chamber, and is not directly affected by the flatness of the lower plate. In this embodiment, the support portion is constituted by spherical seats 7 and 6. The intermediate plate can achieve a stage installation surface with a flatness of 5 μm or less when subjected to its own weight and stage weight. In addition, in order to suppress the influence of the main chamber deformation on the stage mounting surface due to the internal pressure change (vacuum degree 10 ⁻⁴ Pa or less from atmospheric pressure), the area near the three-point support part of the middle plate of the lower plate is directly on the welding surface It is fixed.

  In order to control the XY stage with high accuracy, a stage position measurement system using a laser interferometer is installed. The interferometer 8 · 1 is arranged in a vacuum in order to suppress measurement errors, and in this embodiment, the chamber wall 7 having high rigidity is used in order to make the vibration of the interferometer itself, which directly causes measurement errors, zero.・ It is fixed directly to 7. Further, in order to eliminate the error between the measurement position and the inspection position, the extension line of the measurement part by the interferometer is made to coincide with the inspection part as much as possible. The motors 8 and 2 for performing the XY operation of the stage are held by the chamber walls 7 and 7 in this embodiment. However, when it is necessary to further suppress the influence of the motor vibration on the main chamber. It is directly held by the welding surface plate 7. 1 and attached to the main chamber by a structure that does not transmit vibrations such as bellows.

  The main chambers 4 and 2 are made of a material having high magnetic permeability in order to block the influence of an external magnetic field on the inspection portion. In this embodiment, permalloy and iron such as SS400 are plated with Ni as a rust-proof coating. In other embodiments, permale, supermalloy, electromagnetic soft iron, pure iron or the like is used. Furthermore, it is also effective as a magnetic shielding effect to directly cover the periphery of the inspection portion inside the chamber with a material having high magnetic permeability.

1-1-3) XY stage The XY stage 5.1 can scan a wafer with high accuracy in a vacuum. The strokes of X and Y are, for example, 200 mm to 300 mm for 200 mm wafers and 300 mm to 600 mm for 300 mm wafers, respectively. The driving of the XY stage in this embodiment is performed by motors 8 and 2 for driving the X and Y axes fixed to the main chamber wall and ball screws 8 and 5 attached to these via magnetic fluid seals 8 and 3, respectively. Is done. In order to perform the XY operation in a state where the ball screws for driving X and Y are fixed to the chamber wall, in this embodiment, the stage structure is as follows.

First, on the lower stage, Y stages 7 and 10 are arranged, and ball screws 7 and 8 and cross roller guides 7 and 11 for driving are installed. An X stage 7, 13 is mounted on the upper part of the Y stage via an intermediate stage 7, 12 on which ball screws 7, 14 for driving the X axis are installed. The intermediate stage, the Y stage, and the X stage are connected by a cross roller guide in the Y-axis direction. As a result, when the Y-axis is moved, the X stage is moved by the Y stage and the connecting portions 7 and 14, and the intermediate stage remains fixed. In another embodiment, the intermediate stage has a two-stage structure arranged side by side with the upper stage shaft. In the XY stage according to another embodiment, the XY stage itself is driven by a linear motor. Furthermore, high-precision mirrors 8 and 4 (in this embodiment, flatness λ / 20 or less, and the material is aluminum deposition on synthetic quartz) are provided so that measurement by a laser interferometer can be performed over the entire stroke.

  In order to perform wafer alignment in vacuum, θ stages 7 and 15 are installed on the XY stage. In the θ stage in this embodiment, two ultrasonic motors for driving and a linear scale for position control are arranged. Various cables connected to the movable part that performs the X, Y, and θ operations are clamped by cable bearers respectively held by the X stage and the Y stage, and then to the outside of the main chamber through feedthroughs installed on the chamber walls. Connected with.

  Tables 1 and 2 show specifications of the present embodiment having the above structure.

１−２）レーザ干渉測定系
レーザ干渉測定系は、Ｘ軸およびＹ軸に平行で、その延長線上が検査位置に相当する光軸を有するレーザ光学系と、その間に配された干渉計８・１により構成されている。本実施の形態における光学系の配置は、図９、図１０に示されるような位置関係で配置されている。溶接定盤上に設置されたレーザ９・１より発射されたレーザ光は、ベンダ９・２により垂直に立ち上げられたのちにベンダ１０・１により測定面と平行に曲げられる。さらに、スプリッタ９・４によりＸ軸測定用とＹ軸測定用に分配された後に、ベンダ１０・３およびベンダ９・６によりそれぞれＹ軸およびＸ軸に平行に曲げられ、メインチャンバ内部へと導入される。

1-2) Laser interferometry system The laser interferometer system is a laser optical system having an optical axis parallel to the X axis and the Y axis and whose extension line corresponds to the inspection position, and an interferometer 8. 1. The arrangement of the optical system in the present embodiment is arranged in a positional relationship as shown in FIGS. The laser beam emitted from the laser 9 · 1 installed on the welding surface plate is raised vertically by the vendor 9 · 2 and then bent parallel to the measurement surface by the vendor 10 · 1. Furthermore, after being distributed for X-axis measurement and Y-axis measurement by splitters 9 and 4, they are bent parallel to Y-axis and X-axis by vendors 10 and 3 and 9 and 6, respectively, and introduced into the main chamber. Is done.

  An adjustment method at the time of starting up the optical system will be described below. First, the laser light emitted from the laser is adjusted to be bent vertically by the vendor 9 • 2 and horizontally by the vendor 10 • 1. After that, the optical beam is bent by the vendor 10.3, and reflected back to the mirror 8.4, which is placed vertically with high accuracy with respect to the Y-axis, so that the optical axis returning completely coincides with the incident optical axis. Adjust. By checking the optical axis immediately after the laser in a state where the interferometer is removed so as not to disturb the reflected light, highly accurate adjustment is possible. The optical axis adjustment of the X axis can be performed independently by the splitters 9 and 4 and the vendors 9 and 6 after the optical axis adjustment of the Y axis. The adjustment procedure is the same as that for the Y-axis. Furthermore, after adjusting the axes of incident light and reflected light on the X-axis and Y-axis, it is necessary to make the intersection of each optical axis (in the case where there is no mirror) coincide with the wafer inspection position. For this reason, the brackets fixing the vendors 10 and 3 are made perpendicular to the Y axis, and the brackets fixing the vendors 9 and 6 are made perpendicular to the X axis so that the incident light and the reflected light coincide with each other. It can be moved as it is. Furthermore, it is desirable that the vendors 10 and 1, the splitters 9 and 4, the vendors 10 and 3, and the vendors 9 and 6 can move up and down while maintaining their positional relationships.

  In addition, an optical axis adjustment method associated with laser replacement in this apparatus during operation after startup will be described below. In an apparatus in which the inside of the main chamber during operation is maintained in a vacuum, the optical axis and the like from which the interferometer is removed are difficult. Therefore, a jig is prepared in which several targets 10.2 are installed in the optical path outside the main chamber, and the optical path at the time of start-up can be determined only outside the main chamber. After the laser replacement, the adjustment performed at the time of start-up can be reproduced by adjusting the optical axis with respect to the targets 10 and 2 only by the adjustment function provided on the laser mounting base.

1-3) Inspection unit exterior The inspection unit exteriors 4 and 7 are configured to have a function as a frame structure for maintenance. In the present embodiment, a retractable double-supported crane 11.1 is mounted on the upper part. The crane 11.1 is attached to the traverse rail 11.2, and the traverse rail is further installed on the traveling rail (vertical) 11.3. While the traveling rail is normally stored as shown in FIG. 11, it is raised as shown in FIG. 12 during maintenance, and the vertical stroke of the crane can be increased. Thereby, at the time of maintenance, the electron optical devices 4 and 3, the main chamber top plate, and the XY stage 5.1 can be attached to and detached from the back of the device by a crane built in the exterior. In another embodiment of the crane built in the exterior, the crane structure has a rotatable cantilever shaft.

  Further, the exterior of the inspection unit can also have a function as an environmental chamber. This has a magnetic shielding effect as well as temperature and humidity management as required.

2. Embodiment Hereinafter, with reference to the drawings, a preferred embodiment of the present invention will be described as a semiconductor inspection apparatus for inspecting a substrate, ie, a wafer, on which a pattern is formed as an inspection object.

2-1) Transport System FIGS. 13 and 14 show the main components of the semiconductor inspection apparatus according to the present invention in an elevation view and a plan view. The semiconductor inspection apparatus 13. 1 includes a cassette holder 13. 2 for holding a cassette containing a plurality of wafers, a mini-environment apparatus 13. 3, a loader housing 13. 5 constituting a working chamber, and a wafer. A loader 13 and 7 for loading from a cassette holder 13 and 2 onto a stage device 13 and 6 disposed in the main housing 13 and 4 and an electro-optical device 13 and 8 attached to a vacuum housing are shown. 13 and the positional relationship as shown in FIG.

  The semiconductor inspection apparatus 13. 1 further includes a precharge unit 13. 9 disposed in a vacuum main housing 13. 4, a potential application mechanism for applying a potential to the wafer, an electron beam calibration mechanism, and a stage apparatus. The optical microscopes 13 and 11 constituting the alignment control devices 13 and 10 for positioning the wafer are provided.

2-1-1) Cassette Holder The cassette holder 13.2 is a cassette 13, 12 (for example, SMIF manufactured by Assist Co., Ltd.) in which a plurality of (for example, 25) wafers are stored in a state of being arranged in parallel in the vertical direction , A closed cassette such as FOUP) is held (two in this embodiment). The cassette holder 13.2 has a structure suitable for the case where the cassette is transported by a robot or the like and automatically loaded into the cassette holder 13.2, and is suitable for the case where the cassette is manually loaded. The open cassette structure can be selected and installed arbitrarily. In this embodiment, the cassette holders 13 and 2 are automatically loaded with the cassettes 13 and 12. For example, the cassette holders 13 and 12 and an elevating mechanism 13 that moves the elevating tables 13 and 13 up and down are provided. 14 and the cassettes 13 and 12 can be automatically set on the lifting tables 13 and 13 in the state shown by the chain line in FIG. 14, and after setting, the cassettes 13 and 12 are automatically rotated to the state shown by the solid line in FIG. Directed to the axis of rotation of the first transport unit in the mini-environment device.

Further, the elevating tables 13 and 13 are lowered to a state indicated by a chain line in FIG. As described above, the cassette holder used for automatic loading or the cassette holder used for manual loading may be any known structure, so that a detailed description of the structure and functions thereof is possible. Is omitted.

  In another embodiment, as shown in FIG. 15, a plurality of 300 mm substrates are accommodated in a grooved pocket (not shown) fixed inside the box body 15. Is what you do. The substrate transport box 15.2 is connected to a rectangular tube-shaped box body 15.1 and a substrate loading / unloading door automatic opening / closing device, and a substrate loading / unloading door capable of opening and closing a side opening of the box body 15.1 by a machine. 15.3, lids 15 and 4 that are located on the opposite side of the opening and cover the opening for attaching and detaching the filters and fan motor, and a grooved pocket for holding the substrate W (FIG. 13) (Not shown), ULPA filters 15 and 5, chemical filters 15 and 6, and fan motors 15 and 7. In this embodiment, the substrate is loaded and unloaded by the robot-type first transfer units 15 and 7 of the loaders 13 and 7.

  The substrates or wafers stored in the cassettes 13 and 12 are wafers to be inspected, and such inspections are performed after or during the process of processing the wafers during the semiconductor manufacturing process. Specifically, a substrate that has been subjected to a film formation process, CMP, ion implantation, or the like, that is, a wafer having a wiring pattern formed on the surface, or a wafer on which a wiring pattern has not yet been formed is housed in a cassette. Since a large number of wafers accommodated in the cassettes 12 and 12 are arranged in parallel and spaced apart in the vertical direction, the first wafer can be held by a wafer at an arbitrary position and a first transfer unit described later. The arm of the transfer unit can be moved up and down. The cassette is also provided with a function for controlling moisture in the cassette in order to prevent oxidation of the wafer surface after the process. For example, a dehumidifying agent such as silica gel is placed in the cassette. In this case, any dehumidifying effect can be used.

2-1-2) Mini-Environment Device In FIGS. 13 to 16, the mini-environment device 13.3 includes a housing 16.2, which constitutes a mini-environment space 16.1, the atmosphere of which is controlled. , A gas circulation device 16.3 for controlling the atmosphere by circulating a gas such as clean air in the mini-environment space 16.1, and a part of the air supplied in the mini-environment space 16.1. It includes a discharge device 16.4 for collecting and discharging, and a pre-aligner 16.5 disposed in the mini-environment space 16.1 for roughly positioning a substrate, i.e., a wafer, to be inspected.

  The housing 16. 2 has a top wall 16. 6, a bottom wall 16. 7 and a peripheral wall 16. 8 surrounding the four circumferences, and has a structure for blocking the mini-environment space 16. In order to control the atmosphere of the mini-environment space 16.1, the gas circulation device 16.3 is attached to the top wall 16.6 in the mini-environment space 16.1 as shown in FIG. A gas supply unit 16.9 which cleans the gas (air in this embodiment) and flows clean air in a laminar flow through one or more gas outlets (not shown); A recovery duct 16, 10 which is disposed on the bottom wall 16, 7 in the environment space 16, 1 and recovers the air flowing down toward the bottom, the recovery duct 16, 10 and the gas supply unit 16, 9 and conduits 16 and 11 for returning the recovered air to the gas supply units 16 and 9.

In this embodiment, the gas supply units 16 and 9 are designed to take in and clean about 20% of the supplied air from the outside of the housing 16 and 2. However, the ratio of the gas taken in from the outside is arbitrary. Can be selected. The gas supply units 16.9 are equipped with a known structure HEPA or ULPA filter for producing clean air. The laminar flow of the clean air, that is, the downward flow, is mainly supplied so as to flow through a conveyance surface by a first conveyance unit, which will be described later, disposed in the mini-environment space 16.1. This prevents dust that may be generated due to the above from adhering to the wafer. Therefore, it is not always necessary that the downflow jet outlet is located close to the top wall as shown in the drawing, and it is sufficient if it is above the transport surface of the transport unit. Further, it is not necessary to flow over the entire mini-environment space 16.1.

  In some cases, cleanliness can be ensured by using ion wind as clean air. In addition, a sensor for observing the cleanliness can be provided in the mini-environment space 16.1, and the apparatus can be shut down when the cleanliness deteriorates.

  Entrances 13 and 15 are formed in portions of the peripheral walls 16 and 8 of the housings 16 and 2 adjacent to the cassette holders 13 and 2. A shutter device having a known structure may be provided near the entrances 13 and 15 so that the entrances 13 and 15 are closed from the mini-environment device side. The laminar flow downflow created near the wafer may be, for example, a flow rate of 0.3 to 0.4 m / sec. The gas supply unit 16.9 may be provided outside the mini-environment space 16.1.

  The discharge device 16.4 includes suction ducts 16 and 12 disposed at a lower portion of the transfer unit at a position below the wafer transfer surface of the transfer unit, and blowers 16 and 13 disposed outside the housing 16.2. , Conduits 16 and 14 for connecting the suction ducts 16 and 12 and the blowers 16 and 13 are provided. The discharge device 16.4 flows around the transport unit and sucks the gas containing dust that may be generated by the transport unit through the suction ducts 16 and 12 and draws the conduits 16 and 14 and the blowers 16 and 13 together. To the outside of the housing 16. In this case, the air may be discharged into an exhaust pipe (not shown) drawn near the housing 16.

  The pre-aligner 16.5 disposed in the mini-environment space 16.1 is an orientation flat formed on the wafer (referred to as a flat portion formed on the outer periphery of a circular wafer, hereinafter referred to as an orientation flat), One or more V-shaped notches or notches formed on the outer periphery of the wafer are optically or mechanically detected to determine a rotational position about the wafer axis OO of about ± 1 degree. Positioning is performed in advance with accuracy. The pre-aligner 16.5 constitutes a part of the mechanism for determining the coordinates of the inspection object, and is responsible for the rough positioning of the inspection object. Since the pre-aligner 16.5 itself may have a known structure, the description of the structure and operation is omitted.

  Although not shown, a recovery duct for a discharge device may be provided below the pre-aligner 16.5, and the air containing dust discharged from the pre-aligner 16.5 may be discharged to the outside.

2-1-3) Main Housing In FIGS. 13 to 15, the main housings 13 and 4 constituting the working chambers 13 and 16 include housing main bodies 13 and 17, and the housing main bodies 13 and 17 include the base frame 13. 18 is supported by a housing support device 13/20 mounted on a vibration isolator 13 or 19 on a vibration isolator 13/19. The housing support devices 13 and 20 include frame structures 13 and 21 assembled in a rectangular shape. The housing main bodies 13 and 17 are disposed and fixed on the frame structures 13 and 21, and the bottom walls 13 and 22, the top walls 13 and 23, the bottom walls 13 and 22 and the top walls 13 and 22 mounted on the frame structure. The working chambers 13 and 16 are isolated from the outside by being provided with peripheral walls 13 and 24 which are connected to the walls 13 and 23 and surround the four circumferences. In this embodiment, the bottom walls 13 and 22 are made of a relatively thick steel plate so as not to be distorted by weighting by a device such as a stage device mounted thereon. It may be.

  In this embodiment, the housing body and the housing support devices 13 and 20 are assembled in a rigid structure, and vibrations from the floor on which the base frames 13 and 18 are installed are prevented from being transmitted to the rigid structure. It is blocked by the vibration device 13/19. Outer / outlet ports 14 for loading and unloading wafers are formed on the peripheral walls 13 and 24 of the housing main bodies 13 and 17 adjacent to the loader housing described later.

In addition, the vibration isolator 13 * 19 may be an active type having an air spring, a magnetic bearing, or the like, or a passive type having these. Since any of them may have a known structure, description of its own structure and function is omitted. The working chambers 13 and 16 are maintained in a vacuum atmosphere by a known vacuum device (not shown). A control device 2 for controlling the operation of the entire device is disposed under the base frames 13 and 18. The pressure of the main housing is normally kept at 10 ⁻⁴ to 10 ⁻⁶ Pa.

2-1-4) Loader Housing In FIG. 13 to FIG. 15 and FIG. 17, the loader housings 13 and 5 are housing main bodies 14 and 2 constituting first loading chambers 14 and 2 and second loading chambers 14 and 3, respectively. 4 is provided. The housing main body 14. 4 is a partition that divides the bottom wall 17. 1, the top wall 17. 2, the peripheral wall 17. 3 surrounding the four circumferences, and the first loading chamber 14. 2 and the second loading chamber 14. Walls 14 and 5 are provided so that both loading chambers can be isolated from the outside. The partition walls 14 and 5 are formed with openings or entrances 17.4 for exchanging wafers between the loading chambers. Moreover, the entrance / exit 14 * 6 and 14 * 7 are formed in the part adjacent to the mini-environment apparatus and main housing of the surrounding wall 17.3.

  The housing main bodies 14 and 4 of the loader housings 13 and 5 are placed on and supported by the frame structures 13 and 21 of the housing support devices 13 and 20. Accordingly, the vibration of the floor is not transmitted to the loader housings 13.5. The entrances 14 and 6 of the loader housings 13 and 5 and the entrances 13 and 25 of the housings 16 and 2 of the mini-environment device 13 and 3 are aligned with each other, and there are the mini-environment space 16 and the first loading. Shutter devices 14 and 8 that selectively block communication with the chambers 14 and 2 are provided.

  The shutter devices 14 and 8 surround the doorways 13 and 25 and 14 and 6 so as to be in close contact with the side walls 17 and 3 and are fixed in contact with the side walls 17 and 3. There are doors 13 and 27 for preventing the air from passing through, and drive devices 13 and 28 for moving the doors. Further, the entrances 14 and 7 of the loader housings 13 and 5 are aligned with the entrances and exits 14 and 1 of the housing main bodies 13 and 17, and the second loading chambers 14 and 3 and the working chambers 13 and 16 are arranged there. Shutter devices 13 and 29 for selectively preventing communication are provided. The shutter devices 13 and 29 have sealing members 13 and 30 and sealing members 13 and 30 which are fixed to the side walls 17 and 3 and 24 in close contact with the periphery of the entrances 14 and 7 and 14 and 1. And doors 14 and 9 that prevent air from flowing through the doorway, and drive devices 13 and 31 that move the doors.

  Furthermore, the opening formed in the partition walls 14 and 5 is provided with a shutter device 14 and 10 that closes the door by a door and selectively blocks communication between the first and second loading chambers. These shutter devices 14, 8, 13, 29, and 414, 10 are adapted to hermetically seal each chamber when in the closed state. Since these shutter devices may be known ones, detailed description of their structure and operation will be omitted.

  Note that the support method of the housings 16 and 2 of the mini-environment devices 13 and 3 is different from the support method of the loader housing, and vibrations from the floor via the mini-environment devices 13 and 3 are caused by the load from the loader housings 13 and 5 and the main housing 13. In order to prevent transmission to 4, a vibration-proof cushioning material may be arranged between the housings 16, 2 and the loader housings 13, 5 so as to surround the doorway in an airtight manner. .

  In the first loading chambers 14 and 2, wafer racks 14 and 11 are disposed that support a plurality of (two in this embodiment) wafers in a horizontal state with a vertical separation. As shown in FIG. 18, the wafer racks 14 and 11 are provided with support pillars 18 and 2 fixed in an upright state at four corners of a rectangular substrate 18 and 1 and each of the support pillars 18 and 2 has two stages. The support portions 18.3 and 18.4 are formed, and the periphery of the wafer W is placed on the support portion and held. Then, the tips of arms of first and second transfer units, which will be described later, are brought close to the wafer from between adjacent columns, and the wafer is held by the arm.

The loading chambers 14 · 2 and 14 · 3 are controlled in an atmosphere to a high vacuum state (the degree of vacuum is 10 ⁻⁴ to 10 ⁻⁶ Pa) by a known vacuum exhaust device (not shown) including a vacuum pump (not shown). To be able to be. In this case, the first loading chamber 14. 2 is kept in a low vacuum atmosphere as a low vacuum chamber, and the second loading chamber 14. 3 is kept in a high vacuum atmosphere as a high vacuum chamber to effectively prevent contamination of the wafer. You can also. By adopting such a structure, the wafer accommodated in the loading chamber and subsequently inspected for defects can be transferred into the working chamber without delay. By adopting such a loading chamber, the throughput of defect inspection is improved together with the multi-beam electronic device principle described later, and the degree of vacuum around the electron source that is required to be kept in a high vacuum state is further increased. The vacuum can be as high as possible.

  The first and second loading chambers 14, 2 and 14, 3 are respectively connected to a vacuum exhaust pipe and a vent pipe (not shown) for an inert gas (for example, dry pure nitrogen). Thereby, the atmospheric pressure state in each loading chamber is achieved by an inert gas vent (injecting an inert gas to prevent oxygen gas other than the inert gas from adhering to the surface). Since the apparatus for performing such an inert gas vent itself may have a known structure, a detailed description thereof will be omitted.

In the inspection apparatus of the present invention using an electron beam, a typical lanthanum hexaboride (L _a B ₆ ) used as an electron source of an electron optical system described later has a high temperature to such an extent that it emits thermal electrons once. In the case of being heated to a state, it is important not to make contact with oxygen or the like as much as possible in order not to shorten the lifetime, but in the stage before carrying the wafer into the working chamber where the electron optical system is arranged, By performing the atmosphere control as described above, it can be executed more reliably.

2-1-5) Loader The loaders 13 and 7 include robot-type first transfer units 16 and 14 disposed in the housings 16 and 2 of the mini-environment devices 13 and 3, and second loading chambers 14 and 14. 3 is a robot-type second transfer unit 14.

The first transport units 16 and 14 have multi-node arms 16 and 16 that are rotatable about the axis O ₁ -O ₁ with respect to the drive units 16 and 15. As the multi-node arm, an arbitrary structure can be used, but in this embodiment, the multi-node arm has three portions which are rotatably attached to each other.

One part of the arms 16 and 16 of the first transport unit 16 and 14, that is, the first part closest to the drive unit 16 and 15 is a drive mechanism of a known structure provided in the drive unit 16 and 15 (see FIG. (Not shown) are attached to the rotatable shafts 16 and 17. The arms 16 and 16 can be rotated around the axis O ₁ -O ₁ by the shafts 16 and 17 and can be expanded and contracted in the radial direction with respect to the axis O ₁ -O ₁ as a whole by relative rotation between the parts. A gripping device 14 or 13 for gripping a wafer such as a mechanical chuck or an electrostatic chuck having a known structure is provided at the tip of the third portion farthest from the axes 16 and 17 of the arms 16 and 16. The drive units 16 and 15 can be moved in the vertical direction by the lifting mechanisms 16 and 18 having a known structure.

  The first transfer unit 16, 14 has an arm extending in one direction M1 or M2 of the two cassettes in which the arms 16, 16 are held by the cassette holder. The sheet is placed on one arm or gripped by a chuck (not shown) attached to the tip of the arm and taken out. Thereafter, the arm contracts (as shown in FIG. 14), and the arm rotates to a position where it can extend in the direction M3 of the pre-aligner 16.5, and stops at that position. Then, the arm is extended again and the wafer held by the arm is placed on the pre-aligner 16.5. After receiving the wafer in reverse from the pre-aligner 16.5, the arm further rotates and stops at a position where it can extend toward the second loading chamber 14.2 (direction M4). -Deliver the wafer to the wafer receiver in 2. When the wafer is mechanically gripped, the peripheral edge of the wafer (in the range of about 5 mm from the peripheral edge) is gripped. This is because a device (circuit wiring) is formed on the entire surface of the wafer except for the peripheral portion, and if this portion is gripped, the device is broken or a defect is generated.

  The second transfer units 14 and 12 are basically the same in structure as the first transfer unit, and are different only in that the wafer is transferred between the wafer rack and the mounting surface of the stage apparatus. Therefore, detailed description is omitted.

  In the loader 13, the first and second transport units 16, 14, 14, 12 are moved from the cassette held in the cassette holder onto the stage device 13, 6 disposed in the working chamber 13, 16. When the wafer is transported in a substantially horizontal state, the arm of the transport unit moves up and down simply by taking out the wafer from the cassette and inserting it into the wafer rack, and placing the wafer on the wafer rack. And the removal from the wafer and the placement of the wafer on the stage device and the removal from the wafer. Therefore, a large wafer, for example, a wafer having a diameter of 300 mm can be moved smoothly.

  The stage has a mechanism to reverse bias the wafer, so when the arm puts the wafer on the stage or picks it up, the arm is at the same or close potential as the stage, or the arm is at a floating potential. It has a mechanism that avoids problems such as electric discharge due to potential short circuit.

2-1-6) Stage device The stage devices 13 and 6 include a fixed table 13 and 32 disposed on the bottom walls 13 and 22 of the main housings 13 and 4 and a Y direction (on the paper surface in FIG. 1) on the fixed table. Y tables 13 and 33 that move in the vertical direction), X tables 13 and 34 that move in the X direction (left and right in FIG. 1) on the Y table, and rotary tables 13 and 35 that can rotate on the X table, And holders 13 and 36 disposed on the rotary tables 13 and 35, respectively. The wafer is releasably held on the wafer placement surfaces 14 and 14 of the holders 13 and 36. The holders 13 and 36 may have a known structure capable of releasably gripping the wafer mechanically or by an electrostatic chuck method. The stage devices 13 and 6 are held by the holders on the mounting surfaces 14 and 14 by operating a plurality of tables as described above using a servo motor, an encoder, and various sensors (not shown). High accuracy in the X, Y, and Z directions (up and down in FIG. 13) with respect to the electron beam emitted from the electron optical device, and in the direction around the axis perpendicular to the support surface of the wafer (θ direction) It can be positioned with.

  For positioning in the Z direction, for example, the position of the mounting surface on the holder may be finely adjusted in the Z direction. In this case, the reference position of the mounting surface is detected by a position measuring device (laser interference distance measuring device using the principle of an interferometer) using a fine-diameter laser, and the position is controlled by a feedback circuit (not shown). Instead, the position of the notch or orientation flat of the wafer is measured to detect the planar position and rotation position of the wafer with respect to the electron beam, and the rotation table is rotated by a stepping motor capable of controlling a minute angle.

  In order to prevent dust generation in the working chamber as much as possible, the servomotors 14, 15, 14, 16 for the stage device and the encoders 14, 17, 14, 18 are arranged outside the main housing 13, 4. Yes. The stage devices 13 and 6 may be of a known structure used in, for example, a stepper and the like, and a detailed description of the structure and operation is omitted. Also, since the laser interference distance measuring device may have a known structure, detailed description of the structure and operation is omitted.

  It is also possible to standardize a signal obtained by inputting the rotational position of the wafer with respect to the electron beam and the X and Y positions in advance to a signal detection system or an image processing system described later. Further, the wafer chuck mechanism provided in the holder is adapted to apply a voltage for chucking the wafer to the electrode of the electrostatic chuck, and is arranged at three points (preferably equally spaced in the circumferential direction) of the outer periphery of the wafer. It is designed to press and hold (separated). The wafer chuck mechanism includes two fixed positioning pins and one pressing crank pin. The clamp pin can realize automatic chucking and automatic release, and constitutes a conduction point for voltage application.

  In this embodiment, the table that moves in the left-right direction in FIG. 14 is the X table and the table that moves in the up-down direction is the Y table, but the table that moves in the left-right direction in FIG. The table that moves to X may be the X table.

2-1-7) Wafer chucking mechanism
2-1-7-1) Basic Structure of Electrostatic Chuck In order to accurately focus the electron optical system on the sample surface in a short time, it is preferable to make the unevenness of the sample surface, that is, the wafer surface as small as possible. Therefore, the wafer is attracted to the surface of the manufactured electrostatic chuck with good flatness (preferably flatness of 5 μm or less).

  The electrode structure of the electrostatic chuck includes a monopolar type and a bipolar type. The monopolar type is a method in which the wafer is attracted by conducting the wafer in advance and applying a high voltage (generally about several tens to several hundreds V) to one electrostatic chuck electrode. In this case, it is not necessary to conduct the wafer, and the wafer can be attracted only by applying positive and negative voltages to the two electrostatic chuck electrodes. However, in general, in order to obtain a stable adsorption condition, it is necessary to form two electrodes in a comb tooth shape, and the electrode shape becomes complicated.

On the other hand, for inspection of a sample, a predetermined voltage (retarding voltage) is applied to the wafer in order to obtain the imaging conditions of the electron optical system or to make the state of the sample surface easy to observe with electrons. There is a need to. In order to apply this retarding voltage to the wafer and to stabilize the potential on the wafer surface, the electrostatic chuck needs to be of the above-mentioned monopolar type. (However, it is necessary to make the electrostatic chuck function as a bipolar type until it is connected to the wafer with a conductive needle, as will be described later. Therefore, the electrostatic chuck has a switchable structure between a monopolar type and a bipolar type. Yes.)
Therefore, it must be brought into electrical contact with the wafer mechanically. However, the requirement for preventing contamination of the wafer has become stricter, and it is required to avoid mechanical contact with the wafer as much as possible, and contact with the edge of the wafer may not be permitted. If this is the case. Conduction must be established on the backside of the wafer.

  Usually, a silicon oxide film is formed on the back surface of the wafer, and conduction cannot be obtained as it is. Therefore, by bringing two or more needles into contact with the back surface of the wafer and applying a voltage between the needles, the oxide film can be locally broken and conductive with the silicon of the wafer base material. The voltage applied to the needle is a DC voltage or an AC voltage of about several hundred volts. Further, the needle material is required to be non-magnetic, wear-resistant, and a high melting point material, and tungsten or the like can be considered. It is also effective to coat the surface with TiN or diamond in order to further enhance durability or prevent contamination of the wafer. It is also effective to apply a voltage between the needles and measure the current in order to confirm that conduction with the wafer has been achieved.

  A chucking mechanism as shown in FIG. 19 is made from the above background. In the electrostatic chuck, electrodes 19. 1, 19. 2, which are preferably in a comb-teeth shape for stably adsorbing the wafer W, and three pusher pins 19. Two or more conductive needles 19 and 4 for wafer application are provided. Further, a correction ring 19.5 and a wafer dropping mechanism 19.6 are disposed around the electrostatic chuck.

  The pusher pins 19.3 protrude in advance from the surface of the electrostatic chuck when the wafer W is transferred by the robot hand. When the wafer W is placed on the wafer W by the operation of the robot hand, the pusher pins 19.3 are slowly lowered. The wafer W is placed on the electrostatic chuck. When the wafer is taken out from the electrostatic chuck, it performs the reverse operation to deliver the wafer W to the robot hand. For pusher pins 19.3, the surface material must be selected so that the wafer position is not shifted or contaminated, and silicon rubber, fluorine rubber, ceramics such as SiC and alumina, resin such as Teflon and polyimide, etc. are used. It is desirable to do.

There are several methods for driving the pusher pins 19.3. One is a method of installing a non-magnetic actuator below the electrostatic chuck. This may be a method in which the pusher pin is directly linearly driven by an ultrasonic linear motor, or a method in which the pusher pin is linearly driven by a combination of a rotary ultrasonic motor and a ball screw or rack and pinion gear. This method is on the XY stage mounting the electrostatic chuck table, while the pusher mechanism is compactly, wirings such as the actuator and limit sensor becomes very large. These wires are connected from the XY-operating table to the sample chamber (main chamber or main housing) wall surface, but bend as the stage moves, so it is necessary to arrange them with a large bend R and take up space. End up. Moreover, since it becomes a particle generation source and the wiring needs to be replaced periodically, the number of use should be minimized.

  Therefore, as another method, there is a method of supplying driving force from the outside. When the stage is moved to the position where the wafer W is attached or detached, the shaft protruding into the vacuum via the bellows is driven by the air cylinder provided outside the chamber, and the shaft of the pusher drive mechanism provided below the electrostatic chuck It is supposed to press. The shaft is connected to the rack and pinion or link mechanism inside the pusher drive mechanism, and the reciprocating movement of the shaft is interlocked with the vertical movement of the pusher pin. When the wafer W is transferred to and from the robot hand, the pusher pins 19.3 are raised by adjusting the speed to an appropriate level with a controller and pushing the shaft into a vacuum with an air cylinder.

  The drive source of the shaft from the outside is not limited to the air cylinder, but may be a combination of a servo motor, a rack and pinion, and a ball screw. It is also possible to use an external drive source as the rotation axis. In that case, the rotary shaft is provided with a vacuum seal mechanism such as a magnetic fluid seal, and the pusher drive mechanism has a built-in mechanism for converting the rotation into a linear motion of the pusher.

  The correction ring 19.5 has an action to keep the electric field distribution at the edge of the wafer uniform, and basically applies the same potential as the wafer. However, a potential slightly different from the wafer end potential may be applied in order to cancel out the influence of a minute gap between the wafer and the compensation ring and a minute difference between the wafer and the compensation ring surface height. The correction ring has a width of about 10 to 30 mm in the radial direction of the wafer, and a nonmagnetic and conductive material such as titanium, phosphor bronze, TiN or TiC coated aluminum can be used.

  The conductive needles 19 and 4 are supported by springs 19 and 7. When the wafer is mounted on the electrostatic chuck, it is lightly pressed against the back surface of the wafer by the spring force. In this state, electrical continuity with the wafer W is obtained by applying a voltage as described above.

The electrostatic chuck main body is made up of non-magnetic planar electrodes 19, 19, 2, such as tungsten, and a dielectric formed thereon. As the dielectric material, alumina, aluminum nitride, polyimide, or the like can be used. In general, ceramics such as alumina are a perfect insulator having a volume resistivity of about 10 ¹⁴ Ωcm, so that no charge transfer occurs inside the material and a Coulomb force acts as an adsorption force. On the other hand, the volume resistivity can be set to about 10 ¹⁰ Ωcm by slightly adjusting the ceramic composition, and this causes the movement of electric charge inside the material, so that the wafer adsorption force is stronger than the Coulomb force. A so-called Johnson-Rahbek force acts. If the attractive force is strong, the applied voltage can be lowered accordingly, a margin for dielectric breakdown can be increased, and a stable attractive force can be easily obtained. Further, by processing the electrostatic chuck surface into, for example, a dimple shape, even if particles adhere to the surface of the electrostatic chuck, the particles may fall into the valley portion of the dimple, which affects the flatness of the wafer. The effect of reducing the possibility can be expected.

As described above, the electrostatic chuck material is made of aluminum nitride or alumina ceramic whose volume resistivity is adjusted to about 10 ¹⁰ Ωcm, and the surface has dimple-like irregularities formed on the surface, and the flatness of the surface formed by the aggregate of the convex surfaces. What is processed into about 5 μm is practical.

2-1-7-2) Chucking mechanism for 200/300 bridge tool The apparatus may be required to inspect two types of wafers of 200 mm and 300 mm without mechanical modification. In this case, the electrostatic chuck must chuck two types of wafers and place a correction ring in accordance with the wafer size on the periphery of the wafer. FIGS. 19A, 19B, and 20 show structures for this purpose.

  FIG. 19A shows a state where a 300 mm wafer W is mounted on the electrostatic chuck. A correction ring 19. 1 having an inner diameter slightly larger than the size of the wafer W (a gap of about 0.5 mm) is positioned and mounted on a metallic ring-shaped part on the outer periphery of the electrostatic chuck with an inlay. The correction ring 19.1 is provided with three wafer dropping mechanisms 19.2. The wafer dropping mechanism 19.2 is driven by a vertical drive mechanism that is linked to the drive mechanism of the pusher pins 19.3, and is supported so as to be rotatable around a rotation axis provided in the correction ring 19.1.

When receiving the wafer W from the robot hand, the pusher pin drive mechanism operates to push the pusher pins 19.3 up. At the same time, correct ring 19.
As shown in FIG. 19B, the wafer dropping mechanism 19.2 provided in 1 also rotates upon receiving a driving force. Then, the wafer dropping mechanism 19.2 forms a tapered surface that guides the wafer W to the center of the electrostatic chuck. Next, after the wafer W is placed on the pushed-up pusher pins 19.3, the pusher pins 19.3 are lowered. By appropriately adjusting the timing of the driving force applied to the wafer dropping mechanism 19.2 with the lowering of the pusher pins 19.3, the wafer W is electrostatically chucked while its position is corrected by the tapered surface of the dropping mechanism 19.2. The center of the wafer W and the center of the electrostatic chuck are placed on the top.

  It is desirable to form a low friction material such as Teflon, preferably a conductive low friction material (for example, conductive Teflon, conductive diamond-like carbon, TIN coating) on the tapered surface of the dropping mechanism 19 or 2. Reference numerals A, B, C, D, and E in the figure are terminals (described later) for applying a voltage, and 19.4 is a wafer for detecting that the wafer W is placed on the electrostatic chuck. It is a conduction needle and is pushed up by a spring 19.5.

  FIG. 20 shows a state in which a 200 mm wafer W is mounted on the same electrostatic chuck. Since the surface of the electrostatic chuck is exposed because the wafer diameter is smaller than that of the electrostatic chuck, the correction ring 20. 1 having a size that completely hides the electrostatic chuck is mounted. The positioning of the correction ring 20.1 is the same as in the case of the 300 mm correction ring.

  A step is provided in the inner peripheral portion of the correction ring 20. 1 so as to be accommodated in the ring-shaped groove 20. 2 on the electrostatic chuck side. This is a structure for concealing the electrostatic chuck surface from the gap between the inner periphery of the correction ring 20.1 and the outer periphery of the wafer W with a conductor (correction ring 20.1) when a 200 mm wafer is mounted. It is. If the electrostatic chuck surface is visible, when the electron beam is applied, the electrostatic chuck surface is charged and the sample surface potential is disturbed.

  To replace the correction ring 20.1, a correction ring replacement place is provided at a predetermined position in the vacuum chamber, and a correction ring having a required size is transported by the robot and attached to the electrostatic chuck (into the inlay portion). Insert).

  The 200 mm correction ring is also provided with a wafer dropping mechanism 20. 2 as in the case of 300 mm. A relief is formed on the electrostatic chuck side so as not to interfere with the wafer dropping mechanism 20. The method of mounting the wafer on the electrostatic chuck is exactly the same as in the case of 300 mm. Reference numerals A, B, C, D and E are terminals for applying a voltage, 20 and 3 are push pins similar to the push pins 19 and 3, and 20 and 4 are similar to the wafer conduction needles 19.4. Wafer conduction needle.

  FIGS. 20A and 20B are diagrams schematically showing a configuration of an electrostatic chuck capable of supporting both a 300 mm wafer and a 200 mm wafer, and FIG. (B) shows a state in which a 200 mm wafer is placed. As understood from (A) of FIG. 20-1, the electrostatic chuck has a size capable of mounting a 300 mm wafer, and as shown in (B) of FIG. This portion is wide enough to mount a 200 mm wafer, and a groove 20. 6 into which the inner periphery of the correction ring 20. 1 is fitted is provided so as to surround it. Reference signs A, B, C, D, and E are terminals for applying a voltage.

In the case of the electrostatic chuck shown in FIGS. 20A and 20B, there is a correction ring whether the wafer is placed on the electrostatic chuck, whether the wafer is correctly placed on the electrostatic chuck, or not. Whether or not is detected optically. For example, an optical sensor is installed above the electrostatic chuck, and the wafer is placed horizontally by detecting the optical path length when the light emitted from the optical sensor is reflected by the wafer and returns to the optical sensor again. It can be detected whether it is mounted or tilted. The presence or absence of the correction ring is determined by providing an optical transmitter that obliquely illuminates an appropriate point in the place where the correction ring should be placed and an optical receiver that receives the reflected light from the correction ring. Can be detected. Furthermore, a combination of an optical transmitter for obliquely irradiating an appropriate point where a correction ring for a 200 mm wafer is placed, an optical receiver for receiving reflected light from the correction ring, and a correction ring for a 300 mm wafer A combination of an optical transmitter that obliquely illuminates an appropriate point on the place where the light is placed and an optical receiver that receives reflected light from the correction ring, and which optical receiver receives the reflected light It is possible to detect which one of the 200 mm wafer correction ring and the 300 mm wafer correction ring is placed on the electrostatic chuck.

2-1-7-3) Wafer Chucking Procedure The wafer chucking mechanism having the above-described structure chucks a wafer according to the following procedure.

(1) A correction ring suitable for the wafer size is transported by a robot and mounted on an electrostatic chuck.
(2) The wafer is placed on the electrostatic chuck by the wafer transfer by the robot hand and the vertical movement of the pusher pin.

(3) An electrostatic chuck is applied in a bipolar manner (positive and negative voltages are applied to the terminals C and D) to attract the wafer.
(4) A predetermined voltage is applied to the conduction needle to destroy the insulating film (oxide film) on the back surface of the wafer.

(5) The current between the terminals A and B is measured, and it is confirmed whether or not electrical connection with the wafer is obtained.
(6) The electrostatic chuck is shifted to monopolar adsorption. (Apply the same voltage to terminals A and B as GRD and terminals C and D)
(7) The voltage at the terminal A (, B) is lowered while maintaining the potential difference between the terminal A (, B) and the terminal C (, D), and a predetermined retarding voltage is applied to the wafer.

2-1-8) Device configuration for 200/300 bridge tool FIGS. 21 and 22 show a configuration for making an apparatus capable of inspecting both 200 mm wafers and 300 mm wafers without mechanical modification. Hereinafter, differences from the 200 mm wafer or 300 mm wafer dedicated apparatus will be described.

  It is possible to install wafer cassettes according to the wafer size and wafer cassette type determined by the user specifications at the wafer cassette installation location 21.1 which is exchanged for each specification such as 200 / 300mm wafer, FOUP, SMIF, and open cassette. It has become. Atmospheric transfer robots 21 and 2 are equipped with hands that can handle different wafer sizes. In other words, a plurality of wafer drop-in portions are provided in accordance with the wafer size, and are mounted on the hand at locations that match the wafer size. It has come to be. The atmospheric transfer robot 21. 2 sends the wafer from the installation location 21. 1 to the pre-aligner 21. 3 to adjust the orientation of the wafer, then takes the wafer out of the pre-aligner 21. 3 and sends it into the load lock chamber 21.

The wafer racks inside the load lock chambers 21 and 4 have the same structure, and a plurality of drop portions corresponding to the wafer size are formed on the wafer support portion of the wafer rack. The height of the robot hand is adjusted so that the wafer mounted on the wafer is loaded in the drop-in part that matches the size, the wafer is inserted into the wafer rack, and then the wafer is supported by lowering the robot hand. The wafer is placed on a predetermined drop of the part.

  The wafers placed on the wafer racks in the load lock chambers 21 and 4 are then taken out from the load lock chambers 21 and 3 by the vacuum transfer robots 21 and 6 installed in the transfer chambers 21 and 5 and then the sample chamber 21. -It is conveyed on stage 21 * 8 in 7. Similarly to the atmospheric transfer robots 21 and 2, the hands of the vacuum transfer robots 21 and 6 also have a plurality of drop portions that match the wafer size. The wafer mounted on the predetermined drop portion of the robot hand is placed on the electrostatic chuck on which the correction rings 21 and 9 suitable for the wafer size are mounted in advance on the stages 21 and 8, and is attracted and fixed by the electrostatic chuck. Is done. The correction rings 21 and 9 are placed on correction ring racks 21 and 10 provided in the transfer chambers 21 and 5. Therefore, the vacuum transfer robots 21 and 6 take out the correction rings 21 and 9 corresponding to the wafer size from the correction ring racks 21 and 10 and transfer them onto the electrostatic chuck, and positioning inlay portions formed on the outer periphery of the electrostatic chuck. After the correction rings 21 and 9 are fitted into the wafer, the wafer is placed on the electrostatic chuck.

  When replacing the correction ring, the reverse operation is performed. That is, the correction rings 21 and 9 are removed from the electrostatic chuck by the robots 21 and 6, the correction rings are returned to the correction ring racks 21 and 10 in the transfer chambers 21 and 5, and the correction ring suitable for the wafer size to be inspected is corrected. It is conveyed from the ring racks 21 and 10 to the electrostatic chuck.

  In the inspection apparatus shown in FIG. 21, since the pre-aligners 21 and 3 are disposed near the load lock chambers 22 and 4, the wafer is not aligned properly, so that the correction ring cannot be mounted in the load lock chamber. In addition, it is easy to return the wafer to the pre-aligner for realignment, and there is an advantage that time loss in the process can be reduced.

  FIG. 22 shows an example in which the location of the correction ring is changed, and the correction ring racks 21 and 10 are omitted. In the load lock chamber 22.1, wafer racks and correction ring racks are formed in a hierarchy, and these can be installed in an elevator and moved up and down. First, in order to install a correction ring suitable for the wafer size to be inspected in the electrostatic chuck, the elevator of the load lock chamber 22.1 is moved to a position where the vacuum transfer robots 21 and 6 can take out the correction ring. Then, when the correction ring is installed on the electrostatic chuck by the vacuum transfer robots 21 and 6, the elevator is operated so that the wafer to be inspected can be transferred, and the wafers are taken out from the wafer rack by the vacuum transfer robots 21 and 6. After that, it is placed on an electrostatic chuck. In the case of this configuration, an elevator is required for the load lock chamber 22.1, but the vacuum transfer chamber 21.5 can be formed small, which is effective in reducing the footprint of the apparatus.

  In addition, it is desirable that the sensor for detecting whether or not a wafer exists on the electrostatic chuck should be installed at a position that can accommodate both different wafer sizes. A plurality of sensors that function may be arranged for each wafer size.

  In the inspection apparatus described with reference to FIG. 21, a procedure is adopted in which a correction ring is placed on the electrostatic chuck and the wafer is positioned so as to be applied to the inner diameter of the correction ring. Therefore, in the inspection apparatus shown in FIG. 22, a correction ring is attached to the wafer in the load lock chamber 22. 1, the wafer with the correction ring attached is conveyed along with the correction ring and introduced into the sample chamber 21. The procedure of attaching to the upper electrostatic chuck is taken. As a mechanism for realizing this, there is an elevator mechanism shown in FIGS. 22-1 and 22-2 for moving the elevator up and down to pass the wafer from the atmospheric transfer robot to the vacuum transfer robot. Hereinafter, a procedure for transporting a wafer using this mechanism will be described.

  As shown in FIG. 22-1 (A), the elevator mechanism provided in the load lock chamber has a plurality of stages (two stages in the figure) of correction ring support bases provided so as to be movable in the vertical direction. The upper correction ring support bases 22 and 2 and the lower correction ring support bases 22 and 3 are fixed to the first bases 22 and 5 that are moved up and down by the rotation of the first motors 22 and 4, thereby the first correction ring support bases 22 and 3. The rotation of the motors 22 and 4 causes the first bases 22 and 5 and the upper and lower correction ring support bases 22 and 22 and 3 to move upward or downward.

  On each correction ring support, correction rings 22 and 6 having an inner diameter corresponding to the size of the wafer are placed. Two types of correction rings 22.6 for the 200 mm wafer and for the 300 mm wafer are prepared with different inner diameters, and the outer diameters of these correction rings are the same. Thus, by using a correction ring having the same outer diameter, mutual compatibility is created, and 200 mm and 300 mm can be placed in any combination in the load lock chamber. In other words, regarding the line in which the 200 mm wafer and the 300 mm wafer are mixed and flowed, the upper stage is for 300 mm and the lower stage is for 200 mm, so that it is possible to flexibly cope with the inspection regardless of which wafer flows. In addition, if the same size wafer flows, the upper and lower tiers are for 200 mm or 300 mm, and the upper and lower tier wafers can be inspected alternately, so that the throughput can be improved.

  The second motors 22 and 7 are mounted on the first bases 22 and 5, and the second bases 22 and 8 are attached to the second motors 22 and 7 so as to be movable up and down. Upper wafer support tables 22 and 9 and lower wafer support tables 22 and 10 are fixed to the second tables 22 and 8. As a result, when the second motors 22 and 7 are rotated, the second bases 22 and 8 and the upper and lower wafer support bases 22 and 9 and 22 and 10 are integrally moved upward or downward.

  Therefore, as shown in FIG. 22-1 (A), the wafer W is loaded on the hand of the atmospheric transfer robot 21 · 2 and loaded into the load lock chamber 22 · 1, and then, as shown in FIG. The second motors 22 and 7 are rotated in the first direction to move the wafer support tables 22 and 9 and 22 and 10 upward to place the wafer W on the upper wafer support tables 22 and 9. As a result, the wafer W is moved from the atmospheric transfer robots 21 and 2 to the wafer support tables 22 and 9. Thereafter, the atmospheric transfer robots 21 and 2 are moved backward as shown in (C), and when the backward movement of the atmospheric transfer robots 21 and 2 is completed, as shown in (D), the second motors 22 and 7 are moved to the first. The wafer support bases 22, 9, 22, 10 are moved downward by rotating in the direction opposite to the direction of. As a result, the wafer W is placed on the upper correction ring 22.

  Next, as shown in (E), the hands of the vacuum transfer robots 21 and 6 are put into the load lock chambers 22 and 1 and stopped below the correction rings 22 and 6. In this state, the first motor 22.4 is rotated, and as shown in FIG. 2F, the first base 22.5, the upper and lower correction ring support bases 22.2, 22-3, the second motor 22 7 and the upper and lower wafer support tables 22, 9, 22, 10 are moved downward, whereby the correction rings 21, 6 and the wafer W placed on the upper wafer support tables 22, 9 are transferred to the vacuum transfer robot 21.・ It can be placed on the hand 6 and carried into the sample chambers 21 and 7.

The operation of returning the wafer, which has been inspected in the sample chambers 21 and 7, to the load lock chambers 21 and 4 is performed in the reverse procedure to the above, and is loaded onto the wafer support table by the vacuum transfer robot together with the correction ring. The wafer is transferred to the correction ring support table, then to the wafer support table, and finally placed on the atmospheric transfer robot. In FIGS. 22-1 and 22-2, the wafer transfer operation in the upper stage has been described. However, by adjusting the heights of the hands of the atmospheric transfer robots 21 and 2 and the vacuum transfer robots 21 and 6, in the lower stage. The same operation is possible. Thus, by appropriately switching the heights of the hands of the atmospheric transfer robots 21 and 2 and the vacuum transfer robots 21 and 6, an uninspected wafer is carried into the sample chamber from one stage, and then the inspected wafer is sampled. Unloading from the chamber to the other stage can be performed alternately.

2-2) Wafer transfer method Next, wafer transfer from the cassettes 13 and 12 supported by the cassette holders 13 and 2 to the stage devices 13 and 6 disposed in the working chambers 13 and 16 is performed in order. This will be described (see FIGS. 14 to 16).

  As described above, the cassette holders 13 and 2 have a structure suitable for manually setting a cassette, and a structure suitable for automatically setting a cassette. In this embodiment, when the cassettes 13 and 12 are set on the elevating tables 13 and 13 of the cassette holders 13 and 2, the elevating tables 13 and 13 are lowered by the elevating mechanisms 13 and 14, and the cassettes 13 and 12 are moved in and out. 13 and 15 are matched. When the cassette is aligned with the entrances 13 and 15, a cover (not shown) provided on the cassette is opened, and a cylindrical cover is formed between the cassette and the entrances 13 and 15 of the mini-environment device 13 and 3. Arranged to block the inside of the cassette and the mini-environment space from the outside. Since these structures are publicly known, detailed description of the structure and operation is omitted. In addition, when the shutter apparatus which opens and closes the entrance / exit 13 * 15 is provided in the mini-environment apparatus 13 * 3 side, the shutter apparatus operate | moves and opens the entrance / exit 13 * 15.

  On the other hand, the arms 16 and 16 of the first transport units 16 and 14 are stopped in a state facing in either the direction M1 or M2 (in this description, the direction of M1). One of the wafers stored in the cassette is received at the leading end. In this embodiment, the vertical position adjustment of the arm and the wafer to be taken out from the cassette is performed by the vertical movement of the driving units 16 and 15 of the first transfer unit 16 and 14 and the arms 16 and 16. The lifting / lowering table of the cassette holder may be moved up and down or both.

When the reception of the wafer by the arms 16 and 16 is completed, the arm contracts and operates the shutter device to close the entrance / exit (if there is a shutter device), and then the arms 16 and 16 rotate around the axis O ₁ -O _1. It will be in a state where it can move and extend in the direction M3. Then, the arm is extended and placed on the tip or held by the chuck, the wafer is placed on the pre-aligner 16.5, and the pre-aligner 16.5 is rotated in the direction of rotation of the wafer (the central axis perpendicular to the wafer plane). Rotation direction) is positioned within a predetermined range. When the positioning is completed, the transfer units 16 and 14 receive the wafer from the pre-aligner 16.5 at the tip of the arm and then contract the arm so that the arm can be extended in the direction M4. Then, the doors 13 and 27 of the shutter device 14 and 8 are moved to open the entrances 13 and 25 and 13 and 37, and the arms 16 and 16 are extended so that the wafers are moved to the wafer racks 14 and 11 in the first loading chamber 14 and 2. Place on the upper or lower side. Before opening the shutter devices 14 and 8 and delivering the wafers to the wafer racks 14 and 11 as described above, the openings 17.4 formed in the partition walls 14 and 5 are the doors 14 of the shutter devices 14 and 10. -It is closed in an airtight state by 19.

In the wafer transfer process by the first transfer units 16 and 14, clean air flows in a laminar flow from the gas supply units 16 and 9 provided on the housing of the mini-environment devices 13 and 3 (downflow). ) To prevent dust from adhering to the upper surface of the wafer during transfer. Part of the air around the transport unit (in this embodiment, air that is mainly contaminated with about 20% of the air supplied from the supply unit) is sucked from the suction ducts 16 and 12 of the discharge devices 16 and 4 to be removed from the housing. To be discharged. The remaining air is recovered through the recovery ducts 16 and 10 provided at the bottom of the housing and returned to the gas supply units 16 and 9 again.

  When a wafer is placed on the wafer racks 14 and 11 in the first loading chambers 14 and 2 of the loader housings 13 and 5 by the first transfer unit 16 and 14, the shutter devices 14 and 8 are closed, and the loading chamber 14 and 14 are closed. -Seal the inside of 2. Then, after the inert gas is expelled in the first loading chamber 14. 2 and the air is expelled, the inert gas is also discharged and the inside of the loading chamber 14. The vacuum atmosphere of the first loading chamber 14.2 may be a low degree of vacuum. When the degree of vacuum in the loading chambers 14 and 2 is obtained to some extent, the shutter devices 14 and 10 are operated to open the shutters 14 and 5 of the entrances 17 and 4 that have been sealed by the doors 14 and 19, and the second transport unit The arms 14 and 20 of the arms 14 and 12 are extended to receive one wafer from the wafer receivers 14 and 11 by the gripping device at the tip (mounted on the tip or gripped by a chuck attached to the tip). When the receipt of the wafer is completed, the arm is contracted, and the shutter devices 14 and 10 are operated again to close the entrances 17.4 by the doors 14-19.

  Before the shutter devices 14 and 10 are opened, the arms 14 and 20 can be extended in advance in the direction N1 of the wafer racks 14 and 11. Also, as described above, the doors 14 and 9 of the shutter devices 13 and 29 are closed at the doors 14 and 9 before the shutter devices 14 and 10 are opened. Communication with the chambers 13 and 16 is blocked in an airtight state, and the second loading chambers 14 and 3 are evacuated.

When the shutter devices 14 and 10 close the entrances 17.4, the second loading chambers 14 and 3 are evacuated again to a higher vacuum than in the first loading chambers 14 and 2. Meanwhile, the arms of the second transfer units 16 and 14 are rotated to a position where they can extend toward the stage devices 13 and 6 in the working chambers 13 and 16. Meanwhile the stage device 13, 6 in the working chamber 13 · 16, Y table 13 · 33, X table 13 · 34 centerline _X 0 -X ₀ is the rotation axis _{O 2} of the second transfer unit 14, 12 of 14 moves upward in FIG. 14 to a position substantially coincident with the X-axis line X ₁ -X ₁ passing through −O ₂ , and the X tables 13 and 34 move to a position approaching the leftmost position in FIG. Waiting in this state. When the second loading chambers 14 and 3 become substantially the same as the vacuum state of the working chambers 13 and 16, the doors 14 and 9 of the shutter devices 13 and 29 move to open the entrances 14 and 7 and 14 and the arms extend. Then, the tip of the arm holding the wafer approaches the stage device 13 · 6 in the working chamber 13 · 16. Then, the wafer is placed on the placement surfaces 14 and 14 of the stage devices 13 and 6. When the placement of the wafer is completed, the arm contracts, and the shutter devices 13 and 29 close the entrances 14 and 7 and 14.

  Since the stage has a mechanism that applies a reverse bias potential (retarding potential) to the wafer, when the arm goes to or gets the wafer from the stage, the arm is at the same potential as or close to the stage, or the arm is at a floating potential. Thus, a mechanism for avoiding problems such as discharge due to potential short-circuiting is provided. As another embodiment, when the wafer is transferred onto the stage device, the bias potential to the wafer may be turned off.

When controlling the bias potential, the potential may be turned off until the wafer is transported to the stage, and the bias potential may be applied after the wafer is transported and placed on the stage. When the bias potential is applied, the tact time may be set in advance, and may be applied according to the preset time. Alternatively, the sensor detects that the wafer is placed on the stage and applies the detection signal as a trigger. You may make it do. Alternatively, it may be detected that the shutter devices 13 and 29 have closed the entrances 14 and 7 and 14 and the detection signal is applied as a trigger. Further, in the case of using the electrostatic chuck, it confirms that it is attracted to the electrostatic chuck, it may be applied a bias voltage as a trigger.

  The operation until the wafers in the cassettes 13 and 12 are transferred onto the stage apparatus has been described above. However, the wafers that have been placed on the stage apparatuses 13 and 6 and completed processing are transferred from the stage apparatuses 13 and 6 to the cassettes 13 and 12. To return to the inside, the reverse operation is performed. Further, since a plurality of wafers are placed on the wafer racks 14 and 11, the wafers are transferred between the wafer racks 14 and 11 and the stage devices 13 and 6 by the second transfer units 14 and 12. The first transfer units 16 and 14 can transfer wafers between the cassette and the wafer racks 14 and 11, and can efficiently perform inspection processing.

  Specifically, when there are already processed wafers A and unprocessed wafers B in the wafer racks 14 and 11, first, the unprocessed wafers B are moved to the stage devices 13 and 6. During this time, the processed wafer A is moved from the wafer rack to the cassettes 13 and 12 by the arm, and the unprocessed wafer C is extracted from the cassettes 13 and 12 by the arm and positioned by the pre-aligner 16.5, and then the loading chamber 14 Move to the second wafer rack 14.

  In this way, in the wafer racks 14 and 11, the processed wafer A can be replaced with the unprocessed wafer C while the wafer B is being processed. Further, depending on how to use such an apparatus for performing inspection and evaluation, a plurality of stage apparatuses 13 and 6 are arranged in parallel, and a plurality of stage apparatuses are moved by moving wafers from one wafer rack 14 or 11 to each apparatus. One wafer can be processed in the same way.

  FIG. 23 shows a modification of the method for supporting the main housings 13 and 4. In the modification shown in FIG. 23, the housing support device 23. 1 is formed of a thick and rectangular steel plate 23. 2, and the housing main body 23. 3 is placed on the steel plate. Therefore, the bottom walls 23 and 4 of the housing main bodies 23 and 1 have a thin structure as compared with the bottom wall of the embodiment. In the modification shown in FIG. 24, the housing main body 24, 3 and the loader housing 24, 4 are suspended and supported by the frame structure 24, 2 of the housing support device 24, 1.

  The lower ends of the plurality of vertical frames 24 and 5 fixed to the frame structure 24 and 2 are fixed to the four corners of the bottom walls 24 and 6 of the housing main body 24 and 3 so that the peripheral wall and the top wall are supported by the bottom walls. It has become. The vibration isolators 24 and 7 are disposed between the frame structures 24 and 2 and the base frames 24 and 8. The loader housings 24 and 4 are also suspended by suspension members 24 and 9 fixed to the frame structures 24 and 2. In the modification shown in FIG. 24 of the housing main body 24. 3, it is supported in a suspended manner, so that it is possible to lower the center of gravity of the main housing and the various devices provided therein. In the main housing and loader housing support methods including the above-described modifications, vibrations from the floor are not transmitted to the main housing and the loader housing.

  In another variant, not shown, only the housing body of the main housing is supported from below by the housing support device, and the loader housing can be placed on the floor in the same way as the adjacent mini-environment device 13.3. In still another modification not shown, only the housing body of the main housings 13 and 4 is supported in a suspended manner on the frame structure, and the loader housing can be arranged on the floor in the same manner as the adjacent mini-environment device. .

According to the above embodiment, the following effects can be obtained.
(1) An overall configuration of a mapping projection type inspection apparatus using an electron beam is obtained, and an inspection object can be processed with high throughput.
(2) Inspecting the inspection object while monitoring the dust in the space by providing a sensor for observing the cleanliness by supplying a clean gas to the inspection object in the mini-environment space to prevent the adhesion of dust. Can do.
(3) Since the loading chamber and the working chamber are integrally supported via the vibration preventing device, it is possible to supply and inspect the inspection target to the stage device without being affected by the external environment.

2-3) Electron optical system
2-3-1) Overview The electron optical systems 13 and 8 are provided in the lens barrels 13 and 38 fixed to the housing main bodies 13 and 17, and the primary electron optical system schematically illustrated in FIG. An electron optical system including 25.1 (hereinafter simply referred to as a primary optical system) and a secondary electron optical system (hereinafter simply referred to as a secondary optical system) 25.2 and a detection system 25.3 are provided. The primary optical system 25. 1 is an optical system that irradiates the surface of the wafer W to be inspected with an electron beam. The electron gun 25. 4 that emits an electron beam and the primary electron beam emitted from the electron gun 25. A lens system 25.5 comprising an electrostatic lens for focusing the light, a Wien filter or E.times.B separator 25.6, and an objective lens system 25.7, as shown in FIG. The electron guns 25 and 4 are arranged in order with the uppermost part. The lenses constituting the objective lens systems 25 and 7 of this embodiment are decelerating electric field type objective lenses. In this embodiment, the optical axis of the primary electron beam emitted from the electron guns 25 and 4 is oblique with respect to the irradiation optical axis (perpendicular to the wafer surface) irradiated to the wafer W to be inspected. It has become. Electrodes 25 and 8 are arranged between the objective lens systems 25 and 7 and the wafer W to be inspected. The electrodes 25 and 8 have an axisymmetric shape with respect to the irradiation optical axis of the primary electron beam, and the voltage is controlled by the power sources 25 and 9.

  The secondary optical system 25. 2 includes lens systems 25 and 10 including electrostatic lenses that pass secondary electrons separated from the primary optical system by the E × B deflector 25 and 6. The lens systems 25 and 10 function as a magnifying lens that magnifies the secondary electron image.

The detection systems 25 and 3 include detectors 25 and 11 and image processing units 25 and 12 arranged on the image planes of the lens systems 25 and 10.
The incident direction of the primary beam is usually the E direction of the E × B filter (the reverse direction of the electric field), and this direction is the same as the integration direction of the integration type line sensor (TDI: time delay integration). The TDI integration direction may be different from the primary beam direction.

The electron beam optical system barrel includes the following components.
(1) Column magnetic shield Preferably, a nickel alloy such as permalloy or a magnetic material such as iron is preferably used as a member constituting the lens barrel, and an effect of suppressing the influence of magnetic disturbance can be expected.
(2) Detector rotating mechanism In order to make the scanning axis direction of the stage coincide with the scanning direction of the detector, the inside of the lens barrels 13 and 38 is kept in a vacuum state above the lens barrels 13 and 38. It has a detector rotation mechanism that allows the detectors 25 and 11 such as TDI to be rotated about ± several degrees around the optical axis so as to eliminate the deviation in the scanning direction caused by the assembly of the apparatus. In this mechanism, rotational resolution and rotational position reproducibility are required for about 5 to 40 seconds. This arises from the necessity of keeping the deviation between the scanning direction of the stage and the scanning direction of the detector within about 1/10 of one pixel while scanning an image for one frame. According to the detector rotation mechanism, the angle error between the moving direction of the stage and the integration direction of TDI can be adjusted to 10 mrad or less, preferably 1 mrad or less, more preferably 0.2 mrad or less.

  Hereinafter, an example of the configuration of the detector rotation mechanism will be described with reference to FIGS. 25-3 to 25-5. FIG. 25-3 is a diagram illustrating the entire configuration of the detector rotating mechanism provided on the upper part of the lens barrels 13 and 38, and FIG. 25-4 is a schematic diagram of a mechanism for rotating the upper lens barrel. FIG. 25-5 shows a mechanism for sealing the upper and lower lens barrels.

  25-3, the upper ends of the lens barrels 13 and 38 are composed of upper lens barrels 25 and 20 to which the detectors 25 and 11 are attached, and lower lens barrels 25 and 21 fixed to the main housings 13 and 4. . The upper lens barrels 25 and 20 are supported by bearings 25 and 22 with respect to the lower lens barrels 25 and 21 and can rotate around the optical axis of the secondary optical system. Sealing portions 25 and 23 are provided between the lower barrels 25 and 21 in order to keep the inside of the barrels 13 and 38 in a vacuum. Specifically, the seal portions 25 and 23 are installed between the lower ends of the upper lens barrels 25 and 20 and the upper ends of the lower lens barrels 25 and 21, and the upper lens barrel 25 The flanges 25 and 24 are provided so as to surround 20, and the bearings 25 and 22 are installed between the flanges 25 and 24 and the side surfaces of the upper lens barrels 25 and 20.

  Bearing holders 25, 25, 25, and 26 for holding the bearings 25 and 22 are screwed to the upper and lower barrels 25 and 20, respectively. Furthermore, in order to rotate the upper lens barrels 25 and 20 with respect to the lower lens barrels 25 and 21, a drive mechanism shown in FIG. 25-4 is provided. That is, protrusions 25 and 27 are provided on a part of the bearing holders 25 and 26 provided at the upper ends of the flange portions 25 and 24, and on the other hand, a mounting member (bracket) 25 protruding from the upper lens barrels 25 and 20 is provided. The actuators 25 and 29 are fixed to 28. The shafts 25 and 30 of the actuators 25 and 29 are in contact with the projections 25 and 27, and the projections 25 and 27 are provided between the flange portions 25 and 24 and the mounting members (brackets) 25 and 28 to which the actuators 29 and 29 are fixed. There are provided preload springs 25 and 31 to which a pulling force is applied. As a result, the actuators 25 and 29 are actuated to change the lengths of the shafts 25 and 30 protruding from the actuators 25 and 29, so that the upper lens barrels 25 and 20 are moved at a desired angle with respect to the lower lens barrels 25 and 21. It can be rotated in a desired direction.

  For the above-mentioned rotation accuracy, the movement resolution of the actuators 25 and 29 is preferably about 5 to 10 μm. Further, as the actuators 25 and 29, piezoelectric actuators or micrometers may be motor-driven. In addition, it is desirable to measure the rotational position of the detectors 25 and 11 by attaching a sensor capable of measuring the relative distance between the brackets 25 and 28 for fixing the actuators 25 and 29 and the protrusions 25 and 27. As the sensor, a linear scale, a potentiometer, a laser displacement meter, a strain gauge, or the like can be used.

  In order to keep the inside of the lens barrels 13 and 38 in a vacuum, as shown in FIG. It is installed so that a slight gap 25, 32 (FIG. 25-5) is formed between the surfaces. The seal portions 25 and 23 are provided with partition rings 25 and 33 fixed to the center portion and two elastic seals 25 and 34, 25 and 35, and between the lip portions of the respective elastic seals 25 and 34, 25 and 35. Are provided with springs 25, 36, 25, and 37 for ensuring the surface pressure of the sealing surface and improving the sealing performance, respectively. In the center of the partition rings 25 and 33, exhaust ports 25 and 39 that are continuous with the exhaust passages 25 and 38 formed in the lower lens barrels 25 and 21 are provided. The elastic seals 25, 34, 25, and 35 are preferably made of a material having an extremely small friction coefficient and excellent slidability. For example, an omni seal manufactured by Huron, USA can be used.

In this way, by arranging the elastic seals in a double manner and evacuating the spaces 25 and 40 between them, the upper lens barrels 25 and 20 rotate and some leaks are generated in the elastic seals 25 and 35 on the atmosphere side. Even if it occurs, the leaked air is exhausted through the exhaust passages 25 and 38, and the pressure in the spaces 25 and 40 does not increase so much. Therefore, there is no leakage from the elastic seals 25 and 34 into the lens barrel, and the vacuum in the lens barrel is not deteriorated. The spaces 25 and 40 may be continuously evacuated, but can be evacuated only when the detector rotation mechanism is operated. This is because leakage is likely to occur during rotation, and when it is not rotated, sufficient sealing can be achieved with the surface pressure between the elastic seals 25, 34, 25, 35 and the lower ends of the upper lens barrels 25, 20.

  It is important to appropriately set the surface pressure between the elastic seals 25, 34, 25, 35 and the upper and lower surfaces, and this can be realized by adjusting the size of the gaps 25, 32. The clearances 25 and 32 can be adjusted by inserting shims 25 and 41 between the bearings 25 and 22 and the upper end surfaces of the lower lens barrels 25 and 21. By inserting shims 25 and 41 here, the height of the bearings 25 and 22 with respect to the lower lens barrels 25 and 21 can be changed. On the other hand, since the bearings 25 and 22 are sandwiched by the pressers 25 and 25 and 25 and 26 in the upper barrels 25 and 20, the bearings 25 and 22 are structured so as to move up and down together with the upper barrels 25 and 20. The gaps 25 and 32 between the upper lens barrel 25 and 20 and the lower lens barrel 25 and 21 are changed by the thickness of the shims 25 and 41, respectively.

  Depending on the specifications of the lens barrel, as shown in FIG. 25-5, there is a case where it is sufficient not to provide a double seal but to perform vacuum evacuation between the seals using only a single seal. However, the double seal is more reliable and it is easier to obtain a high vacuum. In the above description, the springs 25, 36, 25, 37 are provided inside the elastic seals 25, 34, 25, 35. However, the elastic seals 25, 34, 25, The springs 25, 36, 25, and 37 may be omitted when 25 is sufficiently pressed against the upper and lower surfaces, or when the elastic seals 25, 34, 25, 35 themselves have a sufficient repulsive force.

  In order to align the direction of the detector and the stage by the rotation mechanism configured as described above, the detectors 25 and 11 are rotated by a small amount, and scanning and imaging of the detectors 25 and 11 are performed each time, and the sharpest What is necessary is just to match | combine the angle of the detector 25 * 11 with the angle when an image was acquired. Hereinafter, the specific method will be described.

In the rotation movable range of the detector rotation mechanism, the detectors 25 and 11 are rotated by a small angle to perform scanning imaging of the detectors 25 and 11, and image processing is performed on the obtained images, thereby obtaining an image such as contrast. Find a number that can evaluate quality. By repeating this, the relationship between the rotational position of the detectors 25 and 11 and the image quality is obtained, and the rotational position of the detectors 25 and 11 when the image quality is the best is obtained. Therefore, the positioning operation of the detectors 25 and 11 is completed by rotating the detectors 25 and 11 to that position.

  The allowable value of the positional deviation between the stage and the detectors 25 and 11 is one in which the deviation between the scanning direction of the stage and the scanning direction of the detector is 1 pixel while scanning the image of one frame of the detectors 25 and 11. Decided from the need to be within about / 10. Therefore, the allowable angle deviation when the pixels are arranged in about 500 stages in the scanning direction is about 40 seconds.

In order to adjust the angle deviation between the stage and the detector to 40 seconds or less, the relationship between the position of the detector and the image quality is quantified by a method such as polynomial approximation, and detection is performed when the image quality is the best. The method of obtaining the position of the detectors 25 and 11, or by first rotating the detectors 25 and 11 to perform imaging, obtaining an approximate relationship between the position of the detector and the image quality, and the detector having the best image quality A method can be used in which the range of positions is narrowed down and the detector is rotated by a small amount within the range and the same operation is performed to obtain a detector position with the best image quality with high accuracy. In this way, it is effective to provide a lock mechanism in order to prevent an angle shift after the angle alignment between the stage and the detector is completed. For example, a plate-shaped component may be passed between the bearing retainers 25, 25, 25, and 26, and the plate-shaped component and the bearing retainers 25, 25, 25, and 26 may be fixed with bolts.

(3) The NA moving mechanism NA is held by a mechanism that can move about several centimeters in the optical axis direction or in a direction orthogonal to the optical axis, and adjusts the NA to an optically optimal position in conjunction with the change in magnification. It is possible to do. It is desirable that multiple NAs be attached to the NA holder, and if it is desired to change the NA deterioration or transmittance by adding such a mechanism, it is possible to replace the NA while keeping the inside of the lens barrel in a vacuum. become.
Further, the NA holding unit is preferably provided with a heater unit, and holding the NA at a high temperature has an effect of making the NA difficult to deteriorate. It is also effective to install a reactive gas pipe, and it is possible to clean the NA while keeping the inside of the lens barrel in a vacuum.

(4) Isolation valve The lens barrel is preferably provided with a valve for dividing the inside of the lens barrel into a plurality of spaces. Specifically, it is effective to install a valve so that the space of the MCP part and the electron gun part can be separated from the space of the stage part. With such a configuration, it is possible to perform maintenance around the stage and the like while keeping the MCP unit and the electron gun unit in a vacuum. Conversely, maintenance of the MCP unit and the electron gun unit can be performed while the stage unit and the like are held in a vacuum.

(5) Optical axis shield tube It is desirable that the periphery of the optical axis is surrounded by a cylindrical member that is grounded to the earth. With such a configuration, an effect of suppressing the influence of electrical disturbance can be expected. .

(6) Orifice in front of the MCP An orifice-shaped or elongated cylindrical member is installed between the series of electron optical systems and the MCP portion so that the conductance of the path connecting the two spaces becomes small. Thus, it becomes easy to maintain the pressure of the MCP portion about 1/5, preferably about 1/10, more preferably about 1/100 lower than that of the electron optical system.

(7) Integration of electrodes and high precision For parts that need to be arranged on the concentric shaft with an accuracy of several μm or less on an electro-optical basis, it is desirable to assemble them by methods such as aligning members or cold fitting. Good to have been.

(8) Optical microscope A sample image at a low magnification or an image when viewed with light is provided with an optical microscope for comparison with an electron beam image. The magnification is about 1/10 to 1/5000 of the electron beam image, preferably 1/20 to 1/1000, and more preferably about 1/20 to 1/100. The image of light from the sample surface is detected by a two-dimensional solid-state image sensor (CCD)
It can be displayed on the CRT. It can also be stored in memory.

(9) A non-vibration type vacuum pumping system such as a coaxial ion pump ion pump is arranged in a rotationally symmetrical manner around the optical axis in the vicinity of the electron gun section and MCP section, thereby affecting the influence of charged particles, magnetic fields, etc. by the pumping system itself. The effect of holding the part in a high vacuum can be expected while offsetting. This is because the pipe conductance is reduced when the ion pump is connected to an electron gun unit or the like through a pipe and exhausted.

Hereinafter, specific embodiments will be described.
(1) Embodiment 1
It is an example of an inspection apparatus mainly composed of a vacuum chamber, a vacuum exhaust system, a primary optical system, a secondary optical system, a detector, an image processor, and a control computer. An example is shown in FIG.

  A secondary optical system for guiding a primary optical system 26. 1 for irradiating a sample with an electron beam and electrons emitted from the sample surface, for example, secondary electrons, reflected electrons, backscattered electrons, etc., to a detector. There is a system 26.2. The secondary optical system is a mapping projection optical system. In order to separate the primary system and the secondary system, an E × B beam separator 26.3 is used. The electronic image signals detected by the detectors 26 and 4 are converted into optical signals and / or electric signals and processed by the image processors 26 and 5. At this time, an image can be satisfactorily formed even if the number of electrons incident on the detector is 200 or less in an area corresponding to one pixel. Of course, it is needless to say that an image can be satisfactorily formed even when there are 200 or more pixels in one pixel region.

The electron gun 26,6, which is a component of the primary optical system, uses LaB ₆ as a heat filament, and draws electrons from the cathode by Wehnelt and extraction electrodes 26,7. Thereafter, the beam is converged to the apertures 26 and 9 by the two-stage A lens (Einzel lens) 26 and 8 to form a crossover. Thereafter, the light passes through the two-stage aligners 26 and 10, the apertures 26 and 11, the three-stage quadrupole lenses 26 and 12, and the three-stage aligners 26 and 13, enters the beam separator, and is deflected toward the sample surface. 26 and 14 and a secondary system P lens (objective lens) 16 and 16 are irradiated almost perpendicularly to the sample surface.

  The apertures 26 and 9 pass through a beam region having high uniformity at crossover and high brightness, and the apertures 26 and 11 define an incident angle of the beam to the quadrupole lens 26. 10 is used for adjustment for making the beam incident on the center of the optical axis of the apertures 26 and 11 and the quadrupole lenses 26 and 12. The quadrupole lenses 26 and 12 are used to change the shape of the beam by changing the trajectory of the beam in two directions, for example, the X and Y directions. For example, in the sample irradiation beam shape, it is possible to change the ratio of the shape in the x and y directions of a circle, an ellipse, a rectangle, a rectangle and an ellipse (see FIG. 27). After passing through the quadrupole lens, it is adjusted by the aligners 26 and 14 so as to pass through the centers of the apertures 26 and 15 and the P lenses (objective lenses) 26 and 16, and enters the sample surface. At this time, the shape of the irradiation beam can be symmetrically formed with respect to at least one of the two axes. The beam shape may be asymmetric. The energy of the beam applied to the sample surface is finally determined by the voltage difference between the cathode and the sample surface. For example, when the cathode is −5.0 kV and the sample surface is −4 kV, the irradiation beam energy is 1 keV (see FIG. 26).

  In this case, the voltage error is ± 10 V and the energy error is ± 20 eV. When secondary electrons are used as detection electrons, when the beam irradiation energy is 1.5 keV ± 10 eV to 5 keV ± 10 eV, the sample is in a negatively charged state, and secondary electrons from that state are emitted from the sample. The image is enlarged and imaged by the secondary system and guided to the detection system. When the irradiation energy is 50 ± 10 eV to 1500 eV ± 10 eV, the sample surface becomes positively charged, and the emitted secondary electrons. Guided to the detection system. Positive charging can operate with relatively low damage, but is more susceptible to charge-up effects or non-uniform surface potential due to charge-up. In the operation with negative charging, it is easy to obtain a stable image, and the distortion of the image due to the effect of charge-up or uneven surface potential due to charge-up can be made smaller than that of positive charge.

Further, there are cases where the positions of the apertures 26 and 15 are operated by shifting the positions of the crossover between the secondary system and the primary system. For example, the secondary system forms a crossover of secondary electrons on the center of the optical axis of the secondary system, and the crossover of the primary system is shifted by 50 to 500 μm from the center of the optical axis of the secondary system (X, Y may be formed and operated. As a result, the two crossovers of the primary system and the secondary system in the apertures 26 and 15 do not overlap, and the current density can be relaxed. Therefore, the blur due to the space charge effect when the beam current amount is large can be increased. It becomes possible to suppress. This is effective, for example, when the primary system irradiation beam current density is 1 × 10 ⁻³ A / cm ² or more. When the current density is lower than that, there is little influence even if the optical axis centers are the same.

  One or more kinds of secondary electrons, reflected electrons, and backscattered electrons are used as the emitted electrons from the sample surface. For example, when the incident beam energy is 1000 eV ± 10 eV, the emission energy from the sample surface is approximately 0 to 10 eV, 1000 eV ± 10 eV, and 10 to 1000 eV, respectively. In addition, electrons transmitted through a thin sample or a sample having a hole (for example, a slangyl mask) are also used. In this case, in the former thin sample, the incident energy decreases by the thickness, and in the holed sample, the incident energy becomes the same energy.

  A focused ion beam (FIB) may be used instead of the electron beam. As the FIB source, a liquid metal Ga ion source is generally used, but other liquid metal ion sources using a metal that is easily liquefied, ion sources of different systems, for example, a duoplasmatron using a discharge, and the like can be used.

  As a sample, various samples are used from a chip of about 10 × 10 mm to a 2, 4, 6, 8, 12 inch wafer. In particular, it is effective for detecting a wiring pattern having a line width of 100 nm or less, a via defect having a diameter of 100 nm or less, and dust, and is convenient for detecting these electrical defects. As the sample, a Si wafer, a semiconductor device wafer obtained by processing Si, a wafer subjected to micromachining, a liquid crystal display substrate, a hard disk head processed wafer, or the like is used.

  In the secondary system 26.2, mapping projection optics for imaging and guiding the emitted electrons from the sample, for example, secondary electrons, reflected electrons, backscattered electrons and transmitted electrons to the detection system at a magnification. An example in which the system is used will be described. Examples of the lens configuration of the column include P lenses (objective lenses) 26 and 16, apertures 26 and 15, aligners 26 and 14, beam separators 26 and 3, T lenses (intermediate lenses) 26 and 17, and aligners 26 and 18. , Apertures 26 and 19, P lenses (projection lenses) 26 and 20, aligners 26 and 21, and a microchannel plate (MCP) unit. Hermetic quartz glass is installed on the top flange of the column. A relay lens and a two-dimensional charge coupled device (2D-CCD) are installed on the top thereof, and an image formed on the phosphor screen is formed on the 2D-CCD sensor.

The emitted electrons from the sample surface form a crossover with apertures 26 and 15 by P lenses (objective lenses) 26 and 16, and form an image at the center of the beam separator 26 and 3. It is effective to operate under the condition that the image is formed at the center of the beam separator because the influence of the aberration of the secondary beam generated in the beam separators 26 and 3 can be reduced. This, for example, when passing the beam at E × B, to come different deflection amount, aberrations image height, by imaging, because it is possible to minimize the aberration incurred the imaging component It is. The same can be said for the primary system. Therefore, in the primary system, not only the imaging conditions are formed on the sample, but also the imaging point is formed near the center of the beam separator. This is effective in reducing the aberration of the primary beam and minimizing uneven current density on the sample.

The aligners 26 and 14 are used to adjust the beam to the center of the P lens (intermediate lens) 26 and 17 at the top. In order to adjust the beam to the center of the P lens (projection lens) 26/20 located upstream thereof, aligners 26/18 are used. There are aligners 26 and 21 to adjust the beam to the center of the MCP at the top. The magnification of the P lens (objective lens) 26 and 16 is 1.5 to 3 times, the magnification of the P lens (intermediate lens) 26 and 17 is 1.5 to 3, and the magnification of the P lens (projection lens) 26 and 20 is 30. ~ 50. In order to achieve these magnifications, adjustment is performed by applying a voltage corresponding to each magnification to each lens. In order to finely adjust the focus, a dedicated focus correction lens is incorporated in a P lens (objective lens) system, and focusing is achieved by fine adjustment of the voltage applied to the electrode. Further, at the positions of the apertures 26 and 15 and the apertures 26 and 19, when both form a crossover, the apertures 26 and 15 are used for noise cut, and the apertures 26 and 19 determine the aberration and contrast. It can also be used to fulfill

  As the size, for example, the apertures 26 and 15 and the apertures 26 and 19 can be used in a range of φ30 to φ2000 μm, preferably φ30 to φ1000 μm, more preferably φ30 to φ500 μm. At this time, when aberrations, transmittance, and contrast characteristics are mainly determined by the apertures 26 and 15, the apertures 26 and 15 are, for example, φ30 to φ500 μm, and the apertures 26 and 19 are used at φ1000 to φ2000 μm. When the aberration, transmittance, and contrast characteristics are mainly determined by the apertures 26 and 19, for example, the apertures 26 and 19 are used at φ30 to φ500 μm, and the apertures 26 and 15 are used at φ1000 to φ2000 μm.

In some cases, stig electrodes are provided above and below the P lenses (intermediate lenses) 27 and 17. This is used to correct astigmatism generated by the beam separator 26. For example, it is possible to use a stig having an electrode configuration of 4, 6, and 8 poles. For example, different voltages can be applied to the respective electrodes of the 8 poles to be used for correcting astigmatism and spherical aberration.

  Further, in the lens operation when the backscattered electron image and the backscattered electron are used, if the P lens (projection lens) 26 and 20 at the final stage uses a deceleration lens (a negative voltage application lens), noise reduction of secondary electrons is performed. It becomes effective. Usually, the amount of secondary electrons is about 10 to 1000 times larger than the amount of reflected electrons, which is particularly effective when imaging using reflected electrons and backscattered electrons is performed. For example, when the cathode voltage of the primary electron source is −4 kV and the sample potential is −3 kV, the reflected electron energy from the sample is 1 keV, and the detector voltage is the installation potential. The energy difference between electrons is 1 keV. At this time, in the negative voltage lens operation of the P lens (projection lens) electrode, it is possible to use a condition in which the center voltage passes the reflected electrons and cuts off the secondary electrons. These conditions can be obtained by simulation.

  In the beam separators 26 and 3, separators that use only E × B in which the electrodes and the magnetic poles are orthogonal or a magnetic field B are used. In the example of E × B, an E electrode that forms an electric field distribution and a magnetic pole that has a magnetic pole surface orthogonal to the E electrode and forms a magnetic flux density distribution in an orthogonal direction. For example, when the optical axis of the secondary system is perpendicular to the sample surface, the incident beam of the primary system can be set at 10 to 90 degrees with respect to the axis of the secondary system. At this time, the primary system is deflected by E × B and can be perpendicularly incident on the sample surface, and the emitted electrons from the sample surface are guided by E × B in the optical axis direction, that is, the vertical direction from the sample surface. This is achieved by the voltage applied to the E electrode and the magnetic flux density formed on the B electrode. For example, a magnetic flux density distribution is formed in parallel to a pair of E electrodes from ± 2 kV ± 1 V and a pair of B electrodes. For example, a magnetic flux density in the magnetic pole direction of 1 to 60 G ± 1 G is generated at the center of E × B. (See FIG. 26).

  E × B can also be applied when the deflection relationship between the primary system and the secondary system is reversed. That is, a primary system incident beam source is provided directly above the sample, and a secondary system detector is provided in a direction that forms an angle of 10 to 80 degrees with the primary system axis. The system beam can be incident perpendicularly to the sample without applying a deflection force, and can be guided toward the detector by applying a deflection force to the electrons (secondary system beam) emitted from the sample.

  In the detectors 26 and 4, signal electrons are introduced into the electron multiplier tubes 28 and 1 such as MCP, and the amplified electrons are irradiated onto the fluorescent screen to form a fluorescent image. The fluorescent screen is one in which a fluorescent material is coated on one surface of a glass plate 28.2 such as quartz glass. This fluorescent image is picked up by the relay lens system 28.3 and the two-dimensional CCD 28.4. The relay lens system and CCD are installed at the top of the column. Hermetic glass 28.6 is installed on the top flange of the column, separating the vacuum environment inside the column from the external atmospheric environment, and forming a fluorescent image on the CCD with reduced distortion and contrast degradation. The fluorescent image can be taken efficiently.

  An integrated line image sensor (TDI-CCD) camera can be used instead of the CCD. In this case, TDI imaging can be performed while moving the stage on the stage, for example, in the E electrode direction or the B magnetic pole direction. For example, when the number of TDI integrated stages is 256, the number of 2048 pixels / stage per stage, the element size is 15 × 15 μm, and the MCP imaging magnification with respect to the sample surface is 300 times, the line / space is 0.1 / 0. When the thickness is 1 μm, the sample surface size is 30/30 μm on the MCP surface. When the relay lens magnification is 1 ×, 30 μm is imaged with two element sizes. At this time, electrons emitted from a sample position corresponding to one element, that is, a sample size of 0.05 × 0.05 μm, are accumulated during the stage movement by the number of stages of 256 elements, and the total acquired light quantity can be increased and imaged. This is particularly effective when the stage speed is high, such as for a line rate of 100 kHz to 600 kHz. This is because when the line rate is high, the number of acquired electrons per element, that is, the acquired light intensity per element of the TDI sensor decreases, so that integration is performed to increase the final acquired light intensity, and contrast and S This is because / N can be increased. The line rate is 0.5 kHz to 100 MHz, preferably 1 kHz to 50 MHz, more preferably 20 kHz to 10 MHz. Correspondingly, the video rate is also used at 1 to 120 MHz / tap, preferably 10 to 50 MHz / tap, more preferably 10 to 40 MHz / tap per tap. The number of taps is 1 or more and 520 or less, preferably 4 or more and 256 or less, and more preferably 32 or more and 128 or less (see FIGS. 28 and 29).

CCDs and TDI sensors / cameras having low noise and high sensitivity characteristics are used. For example, it can be set at 100 to 100000 DN / (nJ / cm ² ), but among these, it is efficient when used at 1000 to 50000 DN / (nJ / cm ² ). Further, when used at 10,000 to 50,000 DN / (nJ / cm ² ), a high-quality image can be obtained with good S / N even at a high line rate.

In addition, when an image is acquired using a CCD or TDI sensor, the area of the number of pixels × the number of stages of these sensors can be used in a state where they substantially coincide with the irradiation area of the primary beam, which is efficient. In addition, noise is reduced. Noise, electrons from higher image height portion other than the area used in the image as noise, there is reaching the detector. In order to reduce them, it is effective to reduce the beam irradiation of parts other than the effective visual field. Image information acquired by the CCD and TDI sensors is converted into an electrical signal and processed by an image processor. By this image processing, cell-to-cell, Die to Die, Die to Any Die image comparison is performed, and defect inspection can be performed. For example, pattern defects, particle defects, potential contrast defects (for example, wiring or plating electrical connection defects).

Stages 26 and 22 are X, Y, Z,

A stage installed by a combination of one or more of the moving mechanisms is used. In such an electron beam inspection apparatus, the following equipment elements can be used as the above elements.
Primary system <br/> electron source W _filament, L a _{B 6} filament, TFE, FE
Lens Made of metal or ceramic, phosphor bronze, Ti, Al as metal
Einzel lens, quadrupole lens aligner 4 pole, 6 pole, 8 pole lens aperture Material, Mo, Ta, Ti, phosphor bronze
Secondary lens Lens Made of metal or ceramic, phosphor bronze, Ti, Al as metal
Ceramic electrode is treated with Au plating
Einzel lens, 4-pole lens aligner 4-pole, 6-pole, 8-pole lens aperture Material, MO (molybdenum), Ta, Ti, phosphor bronze
Electron beam separator E electrode Metal or ceramic, phosphor bronze, Ti, Al
The ceramic electrode is a B magnetic pole processed with Au plating or the like, Permalloy B, Permalloy C, or the like, such as a material having a high saturation magnetic flux density and high permeability (for example, 10 ³ to 10 ⁷ , preferably 10 ⁴ to 10 ⁷ , more preferably Is 10 ⁵ to 10 ⁷ )
Sample Si wafer, 3-5 group compound semiconductor wafer, liquid crystal substrate, hard disk head processed wafer,
2, 4, 6, 8, 12 inch wafers are used
Detector MCP / Fluorescent screen / Relay lens / CCD
MCP / Fluorescent screen / Relay lens / TDI
MCP / Fluorescent Plate / FOP (Fiber Optic Plate) / TDI
Photomulti-photomal can be used in combination as described above. The MCP has a function of amplifying incoming electrons, and the emitted electrons are converted into light by a fluorescent screen. If the incident electron charge is sufficiently large and does not need to be multiplied, the operation can be performed without MCP. It is also possible to use a scintillator instead of the fluorescent screen. This light signal (or image signal) is transmitted to the TDI at a predetermined magnification in the case of a relay lens, and is transmitted to the TDI or formed as an image in the case of FOP by a factor of 1 (transmitting the optical signal on a one-to-one basis). Photomal amplifies an optical signal and converts it into an electrical signal. Multiphotomal is a series of multiple photomals.
Image processor It has functions such as image comparison, defect detection, defect classification, and image data recording.

  In the electron beam inspection apparatus described above, the irradiation beam shape of the primary beam can be used such that the irradiation beam shape is symmetric with respect to the X and Y axes with respect to at least one axis. This makes it possible to form an acquired image with low aberration and low distortion on the electron incident surface of the detector by a beam centered on the optical axis.

  Further, when a CCD or TDI is used as a detector, an S / N sufficient for an incident amount of electrons of 200 or less per pixel area in the formation of one pixel on an area corresponding to one pixel, for example, MCP. And can be used for image processing and defect detection. This is because, for example, in a projection optical system, by defining the size of the apertures 26.15 or 26.19, noise reduction and aberration reduction effects can be generated. For example, an aperture having a diameter of 30 .mu.m to 1000 .mu.m is installed. As a result, an S / N improvement can be realized, and a high-quality image with high resolution can be acquired in the area of 200 electrons / 1 pixel.

TDI performs integration for the number of stages along the moving direction of the stage. In the case of this embodiment, 25
Integration is performed for 6 stages, and the number of integration stages is 114 to 8192, preferably 114 to 4096, and more preferably 512 to 4096. Even if the illuminance unevenness of the primary beam is slightly in the integration direction and the signal electrons from the sample are also uneven, the unevenness is averaged by the effect of integration, and the detected electronic information is constant and stable. Become. Accordingly, the moving direction of the stage can be determined in consideration of the direction in which the illuminance unevenness of the primary electron beam is likely to occur, so that the direction in which the unevenness in illuminance easily occurs coincides with the integration direction of TDI. Although the use of TDI enables continuous image acquisition, the CCD may be used to scan the stage in a step-and-repeat manner, and image acquisition may be performed. That is, the stage is stopped at a specific place to acquire an image, and the process is moved to the next place, where the stage is stopped and image acquisition is repeated. The same can be done using TDI. That is, after using the TDI still mode (stop image acquisition mode, the stage is stopped), or acquiring an image of a certain area (for example, 2048 pixels × 2048 pixels) by the TDI normal image acquisition method, Move to the next place (no image is acquired during this movement), and image acquisition is performed similarly. Therefore, in this case, the inspection is performed without stopping the stage movement.

  When the surface of the sample is enlarged by electrons and an image is formed on the detector, if the resolution of the image is about one pixel of CCD or TDI, the aberration or blur of the secondary optical system should be within one pixel. Is desirable. When the signal electrons are deflected at E × B, aberration and blur increase. Therefore, in the present embodiment, in the secondary optical system, signal electrons such as secondary electrons, reflected electrons, and backscattered electrons are expressed as E ×. In B, it is set so as to go straight without applying a deflection force. That is, the central axis of the secondary optical system is a straight line that passes through the center of the visual field of the sample, the E × B center, and the center of the detector.

It should be noted that it is sufficient that the image of the secondary optical system is not blurred even if it is other than the above-described embodiment, and it goes without saying that the present invention includes it.
(2) Embodiment 2
In the inspection apparatus similar to that in Embodiment 1, when a TDI sensor / camera is used as a detector, the number of pixels / stage is 2048 or more and 4096 or less, the number of taps is 32 or more and 128 or less, and the sensitivity is 10,000 to 40000 DN / (nJ / Cm ² ), it is possible to acquire images more efficiently at high speed. At this time, it can be used at a line rate of 100 to 400 kHz and a video rate of 10 mHz to 40 MHz. At this time, with an 8-inch Si wafer, for example, an LSI device wafer, with a resolution of 0.1 μm / pixel, the inspection time per piece can be executed in 1/8 to 2 hours.

  At this time, when the resolution is 0.1 μm / pixel, in the sample observation and defect inspection, even when the pattern shape is, for example, L / S: 0.2 / 0.2 μm, a contrast of 3 to 30% is achieved. It can be used for defect detection. A defect having a shape other than L / S can be detected as long as it has a size of one pixel or more by comparison with a contrast change. Contrast is realized at 5 to 30%, and observation and defect inspection are possible by image processing. Further, the LSI device wafer can detect defects below the design rule. In the memory, it is possible to detect a defect corresponding to the half pitch of the wiring width, and in the logic, the gate length.

When defect detection is performed using a TDI sensor / camera and an image processing mechanism, images can be continuously formed by a TDI operation, and inspection can be continuously performed. At this time, the sample is placed on the stage, and an image is obtained by performing a continuous operation in the same manner. The speed of the stage is basically determined by v = f × D. However,
v: Stage speed,
f: line frequency,

D: Size corresponding to sensor pixel on sample (determined by projection magnification)
It is. For example, when f: 300 kHz and D: 0.1 μm, v = 30 mm / s.

  FIG. 29 shows an example of a detection system having a configuration different from that of the first embodiment shown in FIG. In this case, the MCP 29.2, FOP 29.3, TDI sensor / package 29.4, connection pins 29.5 and feedthrough flange 29.6 are provided in the vacuum of the column 29.1, and the TDI sensor 29.4 The output is received by the TDI camera 29.7 via the field through flange 29.6. The FOP 29 · 3 is coated with a fluorescent material, and a fluorescent image is formed by electrons from the MCP 29 · 2. This fluorescent image is transmitted to the TDI sensor 29.4 by the FOP 29.3. The image signals of the TDI sensors 29 and 4 are transmitted to the TDI cameras 29 and 7 via the connection pins 29 and 5 and the feedthrough flanges 29 and 6. At this time, if FOP29.3 is used, optical signal transmission loss can be reduced. For example, the transmittance is improved by about 5 to 20 times compared to the relay lens. This is particularly effective when performing a TDI operation. This is because the acquired optical signal intensity can be high, so that it can be operated at a higher speed, and the fiber-shaped signal non-uniformity becomes minute due to the integration of TDI and can be ignored. Here, the connection pins 29 and 5 for connecting the pins of the TDI sensors 29 and 4 and the feedthrough flanges 29 and 6 are required. For example, one of the connection pins 29 and 5 is fixedly connected by contact with the mesh (for example, the pin side of the feedthrough), and the pin side of the TDI sensor / package is contacted by an elastic force such as a spring (not shown). )

  As a result, the pins of the feedthrough flanges 29 and 6 and the pins of the TDI sensor / package 29.4 can be installed with low pressing force, parallel position, and low impedance. A high-speed operation sensor has a large number of pins, for example, a number of pins exceeding 100 is required. If the number of pins is large, the installation pressure (pressing force) increases, and the TDI sensor / package 29.4 may be damaged. Overcoming these points, it can be installed.

  As shown in FIG. 28, the CCD or TDI is usually installed on the atmosphere side, and the MCP and fluorescent screen are installed in a vacuum. However, the relay optical system such as FOP can be shortened by placing the CCD or TDI in a vacuum. The transmission efficiency can be increased.

(3) Embodiment 3
In the inspection apparatus similar to the first and second embodiments, an EB-CCD or EB-TDI is used as a detector (see FIG. 30). EB is an electron beam, and EB-CCD or EB-TDI directly inputs an electron beam and amplifies it into an electric signal (not detecting an optical signal).

  When an EB-TDI sensor / camera is used, electrons can be directly incident on the pixel portion of the sensor and charges can be accumulated. In this case, it is not necessary to use a fluorescent plate, a relay lens system, or a hermetic glass used in a normal detector, and they can be omitted. That is, since it is possible to obtain an electrical signal directly from an electronic signal without once converting the electronic signal image into an optical signal image, the loss caused thereby can be greatly reduced. That is, adverse effects such as image distortion, contrast deterioration, and magnification fluctuation due to the fluorescent plate, hermetic glass, and relay lens system can be greatly reduced. Further, the reduction in the number of components enables downsizing, low cost, and high speed operation. This is because in high-speed operation, it is possible to reduce signal transmission speed loss and image formation speed loss.

FIG. 30 shows an example of an EB-TDI unit. See Embodiment 1 for the optical system. The surface of the TDI sensor 30. 3 is installed at the image forming point above the secondary system column, that is, above the P lens (projection lens). It consists of a TDI sensor / package 30,3, connection pins 30,4, feedthrough 30/5, TDI camera 30-1, image processor 30-6, and control PC 30-7. Electrons emitted from the sample surface (secondary electrons, reflected electrons, or backscattered electrons) are imaged by the secondary system and enter the surface of the TDI sensor 30. Charges are accumulated corresponding to the amount of electrons, and an electric signal for image formation is formed by the TDI camera 30.

  The pins of the sensor / package 30 · 3 and the pins of the feedthrough flanges 30 · 5 are connected by the connection pins 30 · 4. This is the same as in the second embodiment. At this time, since the electronic image signal is directly converted into an electric signal by the TDI sensors 30 and 3 in comparison with the detection systems of the first and second embodiments, the number of components and parts is reduced and the transmission path is shortened. It becomes possible. This makes it possible to improve the S / N, reduce the noise, increase the speed, reduce the size, and reduce the cost.

  In this embodiment, the EB-TDI 30 • 1 is used, but the EB-CCD can be used similarly. In particular, such a configuration is effective when the number of pixels is large or the number of necessary pins exceeds 100 in order to perform high-speed operation. Field-through pins and package connection pins are required. One side (for example, the package side) of this connection pin is constituted by a spring material and a contact plate, and the contact width can be reduced. If the number of contact pins increases to 100 or more, the pushing force at the time of connection increases, and if the total force exceeds 5 kg, the problem of destruction of the package occurs. For this reason, a connection pin that provides a pressing force of 50 to 10 g / piece by adjusting the spring force is used.

  When the number of incident electrons is insufficient when using EB-CCD or EB-TDI, it is also possible to use MCP, which is an electron augmentation view. The same conditions as in the first and second embodiments can be used for the number of pixels / number of stages, the number of stages, the number of taps, the line rate, and the video rate. Sensitivity can be used at 0.1 to 10,000 DN / electron.

(4) Embodiment 4
In the inspection apparatus similar to the first, second, and third embodiments, as shown in FIG. 31, the primary system 31. 1 is the same, but the configuration of the secondary system 31.2 is different. In order to achieve higher resolution, a two-stage P lens (objective lens) 31/3, a two-stage P lens (intermediate lens) 31/5, and a two-stage P lens (projection lens) 31.8 are used. ing. Further, the P lens (intermediate lens) is a zoom lens. Thereby, it is possible to realize a mapping projection beam optical system having a higher resolution and a larger field of view than in the past, and it is possible to acquire an image with an arbitrary magnification in the zoom range.

2-3-2) Details of Configuration Hereinafter, the electron gun, the primary optical system, the secondary optical system, the E × B unit, the detector, and the power source of the electron optical system shown in FIGS. 25-1 to 31 will be described in detail. To do.

2-3-2-1) Electron gun (electron beam source)
A thermal electron beam source is used as the electron beam source. The electron emission (emitter) material is L _a B ₆ . Other materials can be used as long as the material has a high melting point (low vapor pressure at high temperature) and a small work function. A cone-shaped tip or a truncated cone shape with the cone tip cut off is used. The diameter of the tip of the truncated cone is about 100 μm. As another method, a field emission type electron beam source or a thermal field emission type is used, but a relatively wide area (for example, 100 × 25 to 400 × 100 μm ² ) is large as in the present invention. When irradiating with an electric current (about 1 μA), a thermoelectron source using L _a B ₆ is optimal. In the SEM method, a thermal field electron beam source (TFE type) and a Schottky type are generally used. The thermal electron beam source emits electrons by heating the electron emission material. The thermal field emission electron beam source emits electrons by applying a high electric field to the electron emission material. This is a system in which electron emission is stabilized by heating. By selecting the temperature and electric field strength in this method, it becomes possible to extract the electron beam under an efficient condition called a Schottky condition, and this method is also often used recently.

2-3-2-2) Primary optical system An electron beam irradiated from an electron gun is formed, and an electron beam or a linear electron beam having a two-dimensional cross section such as a rectangle, a circle, and an ellipse is formed on the wafer surface. The portion to be irradiated is called a primary electron optical system. The beam size and current density can be controlled by controlling the lens conditions of the primary electron optical system. The primary electron beam is incident on the wafer perpendicularly (± 5 degrees, preferably ± 3 degrees, more preferably ± 1 degree) by the E × B filter (Wien filter) of the primary / secondary electron optical system connection portion. .

The L _a B ₆ heat electrons emitted from the cathode, Wehnelt, triple anode lens or double anode and forms a crossover image on stop cancer in a single anode. An electron beam whose angle of incidence to the lens is optimized by the illumination field stop is controlled to form an image on the NA stop in a rotationally asymmetric manner by controlling the primary electrostatic lens, and then is irradiated onto the wafer surface. The subsequent stage of the primary electrostatic lens is composed of a three-stage quadrupole (QL) and a single-stage aperture aberration correcting electrode. Although the quadrupole lens has a limitation that alignment accuracy is severe, it has a characteristic that it has a stronger convergence effect than a rotationally symmetric lens, and an aperture aberration corresponding to the spherical aberration of the rotationally symmetric lens is applied to the aperture aberration correction electrode. Then, correction can be performed. Thereby, a uniform surface beam can be irradiated to a predetermined area. Further, the electron beam can be scanned by the deflector.

  The irradiation electron beam shape and area on the sample surface include a region corresponding to the shape and area of the imaging area of the TDI-CCD on the sample, and the illuminance in the irradiation region of the electron beam irradiation is uniform and the illuminance It is desirable that the unevenness is 10% or less, preferably 5% or less, more preferably 3% or less.

  The shape and area of the TDI-CCD in this embodiment is 2048 × 512 in terms of the number of pixels, and the pixel size is 16 μm × 16 μm, so that the total is a rectangle of about 32.8 mm × 8.2 mm. When the magnification of the secondary optical system is 160 times, the irradiation area on the sample surface is 1/160 of the above 32.8 mm × 8.2 mm, and thus becomes a rectangle of 205 μm × 51.2 μm.

  Accordingly, the electron beam irradiation area in this case is preferably a rectangle including a rectangle of 205 μm × 51.2 μm. However, as long as the shape and area satisfy the above conditions, as shown in FIG. It may be a rectangle, an ellipse, a circle, or the like. When the magnification of the secondary optical system is 320 times, it becomes 1/320 of 32.8 mm × 8.2 mm, so that it becomes a rectangle of 102.4 μm × 25.6 μm, and an irradiation area of 1/4 of 160 times.

  As described above, in the present invention, the sample is irradiated with a beam having a relatively wide area including the imaging area of the TDI-CCD as a detector, and the imaging area on the sample corresponds to each pixel of the TDI-CCD. Then, the electrons emitted from the imaging region on these samples are detected by simultaneously imaging on the TDI-CCD.

  The irradiation shape of the electron beam may be linear, and may be scanned to ensure the same irradiation region as the planar beam. As shown in (1-1) and (1-2) of FIG. 27-2, the linear beam 27.1 means a beam having a shape with a vertical to horizontal ratio of 1:10 or more, and is limited to a rectangle. It may be an ellipse. Further, as shown in (2) of FIG. 27-2, the linear beam 27.1 may be partially interrupted on the way. When the beam is scanned, since the time for continuously irradiating the same spot on the sample is shortened, there is an advantage that the influence of charge-up on the sample is reduced.

  FIGS. 27-2 (3) and (4) show the relationship between the TDI-CCD multi-pixel imaging region 27.3 on the inspection object 27.2 and the linear beam 27.1. Among these, in (3) of FIG. 27-2, the linear beam 27.1 is almost perpendicular to the integration direction 27.4 of the TDI-CCD or the moving direction 27.5 of the XY stage (for example, 90 ° ± 3 °, preferably Are arranged at 90 ° ± 1 °, and the scanning direction 27.6 of the beam is the same as the integration direction 28.4 of the TDI-CCD or the moving direction 27.5 of the XY stage (eg, 0 ° ± 1 °, preferably Is 0 degree ± 1 minute, more preferably 0 degree ± 1 second).

  (4) in FIG. 27-2 shows another example, and the linear beam 27.1 is almost parallel to the integration direction 27.4 of the TDI-CCD or the moving direction of the XY stage (for example, 90 degrees ± 1 degree, Preferably it is 90 degrees ± 1 minute, more preferably 90 degrees ± 3 seconds.

2-3-2-3) Secondary optical system A two-dimensional secondary electron image generated by an electron beam irradiated on the wafer is obtained at a field stop position by an electrostatic lens (CL, TL) corresponding to an objective lens. The image is formed and enlarged and projected by a lens (PL) at a later stage. This imaging projection optical system is called a secondary electron optical system. A negative bias voltage (deceleration electric field voltage) is applied to the wafer. The decelerating electric field has a decelerating effect on the irradiation beam, reduces damage to the sample, and accelerates secondary electrons generated from the sample surface due to the potential difference between CL and the wafer, thereby reducing chromatic aberration. The electrons converged by CL are imaged on FA by TL, the image is enlarged and projected by PL, and imaged on a secondary electron detector (MCP). In this optical system, an NA is arranged between the CL and TL, and an optical system capable of reducing off-axis aberrations is configured by optimizing the NA.

  An electrostatic octupole (STIG) is placed to correct the manufacturing errors of the electron optical system and the astigmatism and anisotropic magnification of the image generated by passing through the E × B filter (Wien filter). Correction is performed, and axial deviation is corrected by a deflector (OP) disposed between the lenses. As a result, a mapping optical system with a uniform resolution within the field of view can be achieved.

Hereinafter, further description will be given using some embodiments.
(1) Embodiment 5
FIG. 32 shows an electron optical system. The primary electrons emitted from the electron gun 32. 1 pass through the image forming lens 32. 2, pass through the two-stage zoom lens 32. 3, pass through the three-stage quadrupole lens 32. The sample is deflected by 35 ° by the filters 32 and 5 and irradiated on the sample surface through the objective lenses 32 and 7 in the opposite direction parallel to the optical axis of the secondary optical system 32 and 6. The quadrupole lens may be two or more multipoles, and is not limited to an even number, and may have an odd number of poles. Further, it is desirable that the quadrupole lens has 3 to 20 stages, preferably 3 to 10 stages, more preferably 3 to 5 stages.

  Secondary electrons, reflected electrons, and backscattered electrons emitted from the sample surface by the primary electron irradiation are imaged at the center of the E × B filter 32, 5 by the objective lens 32, 7 and then by the intermediate lens 32, 8. After the magnification is changed, an image is formed in front of the projection lenses 32 and 9. The image formed by the intermediate lenses 32 and 8 is magnified about 30 to 50 times by the projection lenses 32 and 9 and formed on the detector surfaces 32 and 10.

  The image forming lenses 32 and 2 enable an image to be formed in front of the zoom lenses 32 and 3 even if the acceleration voltage changes. In FIG. It may be composed of a plurality of stages of lenses.

If the acceleration voltage of primary electrons is constant, the irradiation area and shape of the primary electrons on the sample surface are almost determined by the conditions of the zoom lens 32 and the quadrupole lens 32.4. The zoom lenses 32 and 3 change the irradiation area while maintaining the beam shape. The quadrupole lenses 32 and 4 can change the beam size, but are mainly used to change the beam shape (the aspect ratio of the ellipse). In FIG. 32, each of the zoom lenses 32 and 3 has two stages and the quadrupole lenses 32 and 4 have three stages, but the number of lens stages may be increased.

  Hereinafter, a case where the size of one pixel of the detector is 16 μm square and the size of the detector is 2048 × 512 pixels will be considered. When the magnification of the secondary optical system 32.6 is 160 times, the size corresponding to one pixel on the sample is 16 μm ÷ 160 = 0.1 μm, and the observation area is 204.8 × 51.2 μm. Since the irradiation area covering it is elliptical, it varies depending on the ratio of the major axis to the minor axis. This is shown in FIG. In FIG. 33, the horizontal axis indicates the long axis position, and the vertical axis indicates the short axis position. In considering the optimum irradiation shape, there is an idea that it is not desired to irradiate the beam to a place that is not the observation region 33.1. For this purpose, it is only necessary to find an irradiation shape that maximizes the irradiation efficiency obtained by dividing the area of the observation region by the area of the irradiation region.

  FIG. 34 is a plot of the irradiation efficiency against the ratio of the long axis to the short axis of the shape of the irradiation region. From this, it can be seen that the shape with the highest irradiation efficiency is obtained when the ratio of the major axis to the minor axis of the irradiation ellipse is equal to the ratio of the major axis to the minor axis of the rectangular observation region. That is, the beam shape for irradiating the entire observation region 204.8 × 51.2 μm is 290 × 72.5 μm. Actually, the shape of the irradiation beam becomes slightly larger due to the influence of the aberration of the irradiation optical system and the luminance unevenness of the electron gun. In order to achieve this irradiation beam shape, an image in front of the quadrupole lens 32, 4 is formed into an elliptical irradiation region on the sample surface by an optical system including the quadrupole lens 32, 4 and the objective lens 32, 7. The quadrupole lenses 32.4 may be adjusted so as to form them. In this case, it is only necessary to obtain a sufficiently flat irradiation current density over the necessary irradiation region and the entire irradiation region on the sample surface, and it is not necessary to image the irradiation beam on the sample surface. The size of the image in front of the quadrupole lenses 32 and 4 is adjusted by the zoom lenses 32 and 3 so that a predetermined irradiation area can be obtained on the sample surface.

  Consider now the case where the magnification of the secondary electron optical system 32.6 is increased from 160 times to 320 times, for example. At this time, the size corresponding to one pixel on the sample surface is 16 μm ÷ 320 = 0.05 μm square, and the observation area is 102.4 × 25.6 μm. If the irradiation area remains 160 times in this state, the amount of signal reaching one pixel of the detector is proportional to the area ratio, and thus becomes 1/4 of 160 times. If an image of a signal amount corresponding to an average of 400 electrons per pixel at 160 times is viewed, the standard deviation of fluctuation due to shot noise at that time is √ (400) = 20. Therefore, the S / N ratio is 400/20 = 20. In order to obtain an image having the same S / N ratio at a magnification of 320, the same signal amount may be included in one pixel. Since the area per pixel on the sample is 1/4, it is sufficient that the secondary electron signal density is four times per unit area.

  If the landing energy expressed as the difference between the acceleration energy of the primary electrons and the potential of the sample surface is constant, the irradiation current density and the secondary electron signal density are almost proportional. Therefore, it can be seen that the irradiation current density should be quadrupled. In order to increase the irradiation current density by a factor of 4, the irradiation current may simply be increased by a factor of 4, or the irradiation area may be reduced to ¼. In order to reduce the irradiation area to 1/4, the irradiation size should be 1/2 for both the long axis and the short axis. Since both the observation area and the irradiation area are reduced to ½ in a similar shape, the observation area can be sufficiently irradiated.

  As a means for increasing the irradiation current density, the irradiation current may be increased or the irradiation area may be reduced. However, based on the idea that irradiation is not possible in areas other than the observation region, it is possible to reduce the irradiation area. desirable.

  Table 3 shows the voltages of the respective primary optical system lenses when the secondary optical system magnification is 320 times and 160 times, and the irradiation size on the obtained sample. Thus, an irradiation area that can sufficiently follow the magnification of the secondary optical system is obtained. Although not shown in Table 3, the irradiation size at a magnification of 80 times may be an elliptical shape of 620 μm × 180 μm, and at a magnification of 480 times, it may be an elliptical shape of 100 μm × 30 μm. Thus, it is desirable to change the irradiation size according to the change or switching of the magnification.

観察領域を電子線で照明する場合、上記のように矩形又は楕円で、観察領域を全て覆う広さの面積を持つ電子ビームで照明する方法のほかに、複数の、ビームの広さが観察エリアより小さい面積をもつ電子ビームを走査して照明する方法も可能である。ビーム数は２本以上１０００本以下、好ましくは２本以上１００本以下、より好ましくは４本以上４０本以下である。二本以上のビームがつながった線状のビームを走査しても良い。この場合は線の長手方向に垂直な方向に走査させることにより、１回の走査でより広い領域を検査することが可能となる。この場合も検出器にはＣＣＤ又はＴＤＩを用いて良い。線形のビームを形成するには例えばＬａＢ６の電子源を用い光学系で線形のスリットを経由させるようにすれば良い。また、電子源の先端が鋭利で細長い形状のカソードを用いて線形のビームを形成しても良い。尚ビームの走査中のステージの移動は検査領域全てを網羅するように連続的、または断続的にＸＹ平面の少なくとも１方向に行うようにする。

When illuminating the observation area with an electron beam, in addition to the method of illuminating with an electron beam that is rectangular or elliptical and has an area that covers the entire observation area as described above, a plurality of beam widths vary depending on the observation area. A method of scanning and illuminating an electron beam having a smaller area is also possible. The number of beams is 2 or more and 1000 or less, preferably 2 or more and 100 or less, more preferably 4 or more and 40 or less. A linear beam in which two or more beams are connected may be scanned. In this case, by scanning in a direction perpendicular to the longitudinal direction of the line, a wider area can be inspected by one scan. Again, a CCD or TDI may be used as the detector. In order to form a linear beam, for example, an LaB6 electron source may be used and the optical system may be made to pass through a linear slit. Alternatively, a linear beam may be formed by using a long and narrow cathode with a sharp tip of the electron source. The stage movement during the beam scanning is performed continuously or intermittently in at least one direction of the XY plane so as to cover the entire inspection region.

(2) Embodiment 6
FIG. 35 shows a configuration of a detection system using a relay lens. The secondary electrons imaged on the surface of the MCP (microchannel plate) 35.1 by the secondary optical system pass through the channels in the MCP 35.1 and the number thereof is between the electron entrance surface and the exit surface of the MCP 35.1. It is multiplied according to the voltage applied to. The structure and operation of the MCP 35.1 are known and will not be described in detail here. In this embodiment, the pixel size on the MCP 35 · 1 is set to 26 μm, and an effective area corresponding to 1024 horizontal pixels × 512 vertical pixels is used with a channel diameter of 6 μm. The electrons multiplied in the MCP 35 · 1 are emitted from the emission surface of the MCP 35 · 1 and collide with the fluorescent screen 35 · 3 coated on the opposing glass plate 35 · 2 having a thickness of about 4 mm. Fluorescence with an intensity corresponding to is generated. A thin transparent electrode is applied between the glass plate 35.2 and the phosphor screen 35.3, and a voltage of about 2 to 3 kV is applied between the MCP emission surface and the MCP and phosphor screen. The spread of electrons at the surface is suppressed as much as possible, and the blur of the image is suppressed as much as possible. At the same time, the electrons emitted from the MCP 35 • 1 collide with the phosphor screen 35 • 3 with an appropriate energy. improves. The material of the glass plate 35.

  The light intensity signal obtained by converting the electronic signal on the fluorescent screen 35.3 passes through the glass plate 35.2, passes through the optically transparent plate 35.4 that separates the vacuum from the atmosphere, and passes through the fluorescent screen 35.3. The generated light passes through a relay lens 35 and 5 that forms an image, and is incident on a light receiving surface 35 and 6 of a CCD or TDI sensor disposed at the image forming position. In the present embodiment, relay lenses 35.5 having an imaging magnification of 0.5 times and a transmittance of 2% are used.

The light incident on the light receiving surfaces 35 and 6 is converted into an electric signal by a CCD or TDI sensor, and the electric signal of the image is output to the capturing device. The TDI sensor used in this embodiment is a pixel having a pixel size of 13 μm, a horizontal effective pixel count of 2048 pixels, an integration stage count of 144, a tap count of 8 and a line rate of up to 83 kHz. As a result of such advancement, a pixel having a larger number of effective pixels in the horizontal direction and a larger number of integration stages may be used. The structure and operation of the TDI sensor are known and will not be described in detail here.

  In Table 4, in the column of the first embodiment, the secondary electron emission current density, the secondary optical system imaging magnification, the number of pixel incident electrons obtained when the TDI line rate in the present embodiment is determined, and TDI gray Scale pixel tone values and stage speed are shown.

ここで述べたグレースケール画素階調値のフルスケールは２５５ＤＮである。これは、現状のＭＣＰダイナミックレンジが２μＡ程度しかない事に起因する。ＭＣＰダイナミックレンジの画期的な向上は現状望めないので、ある程度の画素階調値を得るためには、ＴＤＩレスポンシビティ（Responsivity）を最低２００ＤＮ／（ｎＪｃｍ^２）は確保する事が重要になる。

The full scale of the gray scale pixel gradation value described here is 255 DN. This is because the current MCP dynamic range is only about 2 μA. Since a breakthrough improvement in the MCP dynamic range cannot be expected at present, it is important to ensure a TDI responsivity of at least 200 DN / (nJcm ² ) in order to obtain a certain pixel gradation value. .

(3) Embodiment 7
FIG. 36 shows a configuration of a detection system using FOP. The structure and operation up to the phosphor screen 36. 1 are the same as those in the fifth embodiment. However, the effective area of the MCP 36.2 of the present embodiment is a horizontal size of 2048 × 512 pixels with a pixel size of 16 μm. Unlike the fifth embodiment, the phosphor screen 36. 1 is applied to an FOP (fiber optic plate) 36. 3 having a thickness of about 4 mm instead of the glass plate. The light intensity signal converted from the electronic signal by the phosphor screen 36.1 passes through each fiber of the FOP 36.3. A transparent electrode is applied to the light exit surface of the FOP 36. 3, which is at ground potential. The light emitted from the FOP 36.3 passes through another FOP 36.4 having a thickness of, for example, about 3 mm without opening a gap, and is disposed on the light output surface of the FOP 36.4 via a translucent adhesive. It is incident on the light receiving surface of the CCD or TDI sensor 36.5. Since light does not diverge beyond each fiber of the FOP, if the pixel size of the CCD or TDI sensor 36.5 is sufficiently larger than the fiber diameter, the image quality is not greatly affected.

  In the present embodiment, the fiber diameter of the FOP is 6 μm, and the pixel size of the TDI sensor 36.5 is 16 μm. The magnification of the image can be changed by changing the fiber diameter on the entrance side and the exit side of the FOP. However, the distortion and displacement of the image due to this change become large, and in this embodiment, the magnification is equal. The transmittance of the present embodiment is about 40%.

  The CCD or TDI sensor 36.5 is placed in a vacuum, and the electrical signal 36,6 of the image converted from the optical signal is output to the capture device via a feedthrough 36,7 that isolates the atmosphere from the vacuum.

  It is possible to dispose the CCD or TDI sensor 36.5 in the atmosphere and isolate the atmosphere and the vacuum with FOP, but the transmittance decreases and the distortion increases as the thickness of FOP increases. In light of the facts, there is little need for active adoption.

  The TDI sensor 36.5 used in this embodiment is a pixel having a pixel size of 16 μm, a horizontal effective number of pixels of 2048 pixels, a total number of steps of 512, a tap number of 32, and a line rate of up to 300 kHz. Due to technological advancement of the sensor, a sensor having a larger number of effective pixels in the horizontal direction and a larger number of integrated stages may be used.

  In the column of Embodiment 2 in Table 4, the secondary electron emission current density, the secondary optical system imaging magnification, and the number of pixel incident electrons obtained when determining the TDI line rate in this embodiment, the TDI grayscale pixel The gradation value and stage speed are shown.

(4) Embodiment 8
FIG. 37A is a diagram schematically showing a configuration of a defect inspection apparatus EBI of a projection projection system, and FIG. 37B is a schematic diagram showing a configuration of a secondary optical system and a detection system of the defect inspection apparatus EBI. Show. In Figure 37, the electron gun 37 · 1 has a L _a B ₆ made the cathode 37 · 2 of thermionic emission type operable at a large current, the primary electrons emitted from the electron gun 37 · 1 to the first direction Passes through a primary optical system including several stages of quadrupole lenses 37 and 3 and the beam shape is adjusted, and then passes through a Wien filter 37.4. The traveling direction of the primary electrons is changed by the Wien filter 37.4 to the second direction so as to be input to the sample W to be inspected. The primary electrons exiting the Wien filter 37.4 and proceeding in the second direction have their beam diameter reduced by the NA aperture plate 37.5 and pass through the objective lens 37.6 to irradiate the sample W.

In this way, in the primary optical system, the electron gun 37. 1 is made of a high-brightness material made of _La B ₆ , and therefore has a low energy and a large current compared to the conventional scanning type defect inspection apparatus. In addition, a primary beam having a large area can be obtained. The electron gun 37.1 is made of LaB ₆ and has a truncated cone shape, a diameter of 50 μm or more, a primary electron extraction voltage of 4.5 kV, and 1 × 10 ³ A / cm ² sr or more 1 × 10 ⁸ A / cm. ^It can be used by extracting electrons with a luminance of ² sr or less. Preferably, it is 1 × 10 ⁵ A / cm ² sr to 1 × 10 ⁷ A / cm ² sr at 4.5 kV. More preferably, it is 1 × 10 ⁶ A / cm ² sr or more and 1 × 10 ⁷ A / cm ² sr or less at 10 kV. Further, the electron gun 37.1 is a Schottky type, and draws electrons with a luminance of 1 × 10 ⁶ A / cm ² sr or more and 2 × 10 ¹⁰ A / cm ² sr or less when the extraction voltage of primary electrons is 4.5 kV. It can also be used. Preferably, it is 1 × 10 ⁶ A / cm ² sr to 5 × 10 ⁹ A / cm ² sr at 10 kV. Also, a ZrO Schottky type can be used for the electron gun 37.1.

  The shape of the irradiation region where the primary electrons irradiate the sample W is approximately symmetric with respect to the other two orthogonal axes not including the optical axis of the primary electrons, and the primary electrons in the region where the primary electrons irradiate the sample. The illuminance unevenness is 10% or less, preferably 5% or less, more preferably 3% or less, and the illuminance unevenness is extremely uniform. In this case, the beam shape can be used even when the beam shape is not substantially symmetric with respect to two other orthogonal axes not including the optical axis of the primary electrons as described above.

In this embodiment, the sample W is irradiated with a surface beam formed in a rectangular shape having a cross section of, for example, 200 μm × 50 μm by the primary optical system, so that a small area of a predetermined area on the sample W can be irradiated. become able to. In order to scan the sample W with this surface beam, the sample W is placed on a high-precision XY stage (not shown) corresponding to, for example, a 300 mm wafer, and the XY stage is two-dimensionally fixed with the surface beam fixed. Move. In addition, since it is not necessary to narrow down the primary electrons to the beam spot, the surface beam has a low current density and the sample W is less damaged. For example, in the conventional beam scanning type defect inspection apparatus, the current density of the beam spot is 10 A / cm ² to 10 ⁴ A / cm ² , but in the defect inspection apparatus of FIG. Only 0001 A / cm ^{2 to} 0.1 A / cm ² . Preferably used in ^{0.001A / cm 2 ~1A / cm 2} . More preferably used in ^{0.01A / cm 2 ~1A / cm 2} . On the other hand, the dose is 1 × 10 ⁻⁵ C / cm ² in the conventional beam scanning method, whereas 1 × 10 ⁻⁶ C / cm ² to 1 × 10 ⁻¹ C / cm ^{2 in the} present method. And this method is more sensitive. It is preferably used at 1 × 10 ⁻⁴ C / cm ² to 1 × 10 ⁻¹ C / cm ² , more preferably 1 × 10 ⁻³ C / cm ² to 1 × 10 ⁻¹ C / cm ² .

  The incident direction of the primary electron beam is basically from the E direction of E × B37.4, that is, the direction of the electric field, and the TDI integration direction and the stage moving direction are aligned with this direction. The incident direction of the primary electron beam may be the B direction, that is, the direction in which a magnetic field is applied.

  Secondary electrons, reflected electrons, and backscattered electrons are generated from the region of the sample W irradiated with the primary electrons. First, detection of secondary electrons will be described. Secondary electrons emitted from the sample W are enlarged by the objective lens 37, 6 so as to travel in the second opposite direction, and the NA aperture plate 37, 5 and Vienna. After passing through the filter 37.4, it is magnified again by the intermediate lens 37-7, further magnified by the projection lens 37-8, and enters the secondary electron detection system 37-9. In the secondary optical systems 37 and 9 for guiding secondary electrons, the objective lenses 37 and 6, the intermediate lenses 37 and 7, and the projection lenses 37 and 8 are all high-precision electrostatic lenses, and the magnification of the secondary optical system is Configured to be variable. Since primary electrons are incident on the sample W almost perpendicularly (± 5 degrees or less, preferably ± 3 degrees or less, more preferably ± 1 degree or less) and secondary electrons are taken out almost perpendicularly, the unevenness on the surface of the sample W Does not cause shading.

  The Wien filter 37.4 is also called an E × B filter, has an electrode and a magnet, has a structure in which an electric field and a magnetic field are orthogonal, and bends primary electrons, for example, 35 degrees to the sample direction (direction perpendicular to the sample). On the other hand, it has a function of causing at least one of secondary electrons, reflected electrons, and backscattered electrons from the sample to go straight.

  The secondary electron detection systems 37 and 9 that receive secondary electrons from the projection lenses 37 and 8 are microchannel plates (MCP) 37 and 10 that multiply the incident secondary electrons and electrons emitted from the MCP 37 and 10. Fluorescent screens 37 and 11 that convert light, and sensor units 37 and 12 that convert light emitted from the fluorescent screens 37 and 11 into electrical signals are provided. The sensor units 37 and 12 have high-sensitivity line sensors 37 and 13 made up of a large number of two-dimensionally arranged solid-state imaging devices, and the fluorescence emitted from the fluorescent screens 37 and 11 is the line sensors 37 and 13. Is converted into an electrical signal and sent to the image processing units 37 and 14 for parallel, multistage and high-speed processing.

  While moving the sample W and sequentially irradiating and scanning individual regions on the sample W with a surface beam, the image processing units 37 and 14 sequentially store data on the XY coordinates and the image of the region including the defect. Accumulation is performed, and an inspection result file including coordinates and images of all areas to be inspected including defects for one sample is generated. In this way, inspection results can be managed collectively. When this inspection result file is read, the defect distribution of the sample and the defect detail list are displayed on the display of the image processing unit 12.

  Actually, among the various components of the defect inspection apparatus EBI, the sensor units 37 and 12 are arranged in the atmosphere, but the other components are arranged in a lens barrel kept in a vacuum. In the embodiment, a light guide is provided on an appropriate wall surface of the lens barrel, and light emitted from the fluorescent screens 37 and 11 is taken out into the atmosphere through the light guide and relayed to the line sensors 37 and 13.

Assuming that the electrons emitted from the sample W are 100%, the ratio of electrons that can reach MCP37 · 10 (hereinafter referred to as “transmittance”) is transmittance (%) = (electrons that can reach MCP37 · 10) / ( Electrons emitted from sample W) × 100
It is represented by The transmittance depends on the opening area of the NA aperture plates 37.5. As an example, FIG. 38 shows the relationship between the transmittance and the opening diameter of the NA aperture plate. Actually, at least one of secondary electrons, reflected electrons, and backscattered electrons generated from the sample reaches the electron detection system D is about 200 to 1000 per pixel.

  The center of the image magnified and projected on the detector and the center of the electrostatic lens are the common axis, and the electron beam between the polarizer and the sample has the common axis as the optical axis, and the optical axis of the electron beam. Is perpendicular to the sample.

FIG. 39 shows a specific configuration example of the electron detection systems 37 and 9 in the defect inspection apparatus EBI of FIG. A secondary electron image or a reflected electron image 39.1 is formed on the incident surface of the MCP 37.10 by the projection lenses 37.8. For example, the MCP 37 · 10 has a resolution of 6 μm, a gain of 10 ³ to 10 ⁴ , and an execution pixel of 2100 × 520. To do. As a result, fluorescent light is emitted from the portions of the fluorescent screens 37 and 11 irradiated with electrons. The The emitted fluorescence is incident on the line sensors 37 and 13 through the optical relay lenses 39 and 3. For example, the optical relay lens 39.3 has a magnification of 1/2, a transmittance of 2.3%, and a distortion of 0.4%, and the line sensors 37 and 13 have 2048 × 512 pixels. The optical relay lens 39.3 forms an optical image 39.4 corresponding to the electronic image 39.1 on the incident surface of the line sensor 37.13. An FOP (fiber optic plate) can be used in place of the light guide 39.2 and the relay lens 39.3. In this case, the magnification is 1. Further, when the power index per pixel is 500 or more, the MCP may be omitted.

  The defect inspection apparatus EBI shown in FIG. 37 adjusts the acceleration voltage of the electron gun 37. 1 and the sample voltage applied to the sample W and uses the electron detection system 37. It can operate in either charging mode or negative charging mode. Further, by adjusting the acceleration voltage of the electron gun 37. 1, the sample voltage applied to the sample W, and the objective lens conditions, the defect inspection apparatus EBI can cause the high-energy reflected electrons emitted from the sample W by irradiation with primary electrons. Can be operated in a backscattered electron imaging mode for detecting. The reflected electrons have the same energy as that when the primary electrons are incident on the sample W, and the energy is higher than that of the secondary electrons. Therefore, the reflected electrons are not easily affected by the potential due to charging of the sample surface. . As the electron detection system, an electron impact type detector such as an electron impact type CCD or an electron impact type TDI that outputs an electric signal corresponding to the intensity of secondary electrons or reflected electrons can be used. In this case, the electron impact type detector is installed at the imaging position without using the MCP 37.10, the fluorescent screen 37.11, and the relay lens 39.3 (or FOP). With such a configuration, the defect inspection apparatus EBI can operate in a mode suitable for an inspection object. For example, in order to detect a metal wiring defect, a GC wiring defect, or a resist pattern defect, a negative charging mode or a backscattered electron imaging mode may be used, and a via conduction failure or a via bottom residue after etching may be detected. For the detection, a reflected electron imaging mode may be used.

FIG. 40A is a diagram for explaining requirements for operating the defect inspection apparatus EBI of FIG. 37 in the above three modes. The acceleration voltage of the electron gun 37.1 is V _A , the sample voltage applied to the sample _W is V _W , the irradiation energy of primary electrons when irradiating the sample is E _IN , and the secondary electron detection system 37/9 is incident. Let E _{OUT be} the signal energy of the secondary electrons. The electron gun 37.1 is configured to change the acceleration voltage _VA, and a variable sample voltage _VW is applied to the sample W from an appropriate power source (not shown). Therefore, when the acceleration voltage V _A and the sample voltage V _W are adjusted and the electron detection systems 37 and 9 are used, the defect inspection apparatus EBI has a secondary electron yield of more than 1 as shown in FIG. It can operate in a positive charging mode in a large range and in a negative charging mode in a range smaller than 1. Further, by setting the acceleration voltage V _A , the sample voltage V _W and the objective lens condition, the defect inspection apparatus EBI can distinguish between the two types of electrons using the energy difference between the secondary electrons and the reflected electrons. It can operate in a backscattered electron imaging mode that detects only electrons.

Examples of values of VA, V _W , E _IN and E _OUT for operating the defect inspection apparatus EBI in the backscattered electron imaging mode, the negative charging mode and the positive charging mode are as follows:
Reflected electron imaging mode V _A = −4.0 kV ± 1 degree V (preferably ± 0.1 degree, more preferably ± 0.01 degree or less)
V _W = −2.5 kV ± 1 degree V (preferably ± 0.1 degree, more preferably ± 0.01 degree or less)
E _IN = 1.5 keV ± 1 degree V (preferably ± 0.1 degree, more preferably ± 0.01 degree or less)
E _OUT = 4 keV or less
Negative charging mode V _A = −7.0 kV ± 1 V (preferably ± 0.1 V, more preferably ± 0.01 V or less)
V _W = −4.0 kV ± 1 V (preferably ± 0.1 V, more preferably ± 0.01 V or less)
E _IN = 3.0 keV ± 1V (preferably ± 0.1V, more preferably ± 0.01V or less)
E _OUT = 4 keV + α (α: energy width of secondary electrons)
Positive charging mode V _A = −4.5 kV ± 1 V (preferably ± 0.1 V, more preferably ± 0.01 V or less)
V _W = −4.0 kV ± 1 V (preferably ± 0.1 V, more preferably ± 0.01 V or less)
E _IN = 0.5 keV ± 1 V (preferably ± 0.1 V, more preferably ± 0.01 V or less)
E _OUT = 4 keV + α (α: energy width of secondary electrons)
It becomes.

As described above, basically in the secondary electron mode, the potential V _W of the sample is 4 kV ± 10 V (preferably 4 kV ± 1 V, more preferably 4 kV ± 0 in both the positive charge mode and the negative charge mode). Apply a constant potential of .01V or less. On the other hand, setting the acceleration voltage _{V A} is the case of the reflection electron mode 4kV ± 10V (preferably 4kV ± 1V, more preferably less 4kV ± 0.01 V) and the sample potential _{V W} to an arbitrary potential of less acceleration potential 4kV And use it. In this way, secondary electrons or reflected electrons that are signals are set to enter the MCP of the detector with an optimum energy of 4 keV ± 10 eV + α (preferably 4 kV ± 1 V, more preferably 4 kV ± 0.01 V). ing.

  The above potential setting is basically a case where the energy of the signal electrons passing through the secondary optical system is 4 keV, and an electron image of the sample surface is formed on the detector. By changing this energy, the secondary electron mode is changed. By changing the set potential in the reflected electron mode, it is possible to obtain an optimal electron image corresponding to the type of sample. As the negative charging mode, it is possible to use a region of electron irradiation energy (for example, 50 eV or less) lower than the positively charged region of FIG.

  Actually, the detected amounts of secondary electrons and reflected electrons vary depending on the surface composition, pattern shape, and surface potential of the region to be inspected on the sample W. That is, the secondary electron yield and the amount of reflected electrons differ depending on the surface composition of the object to be inspected on the sample W, and the secondary electron yield and the amount of reflected electrons are larger than those on the plane at sharp points and corners of the pattern. Further, when the surface potential of the object to be inspected on the sample W is high, the amount of secondary electron emission decreases. Thus, the electron signal intensity obtained from the secondary electrons and reflected electrons detected by the detection systems 37 and 9 varies depending on the material, pattern shape, and surface potential.

2-3-3) E × B unit (Vienna filter)
The Wien filter is a unit of an electromagnetic prism optical system in which electrodes and magnetic poles are arranged in an orthogonal direction, and an electric field and a magnetic field are orthogonal. When an electromagnetic field is selectively applied, an electron beam incident on the field from one direction is deflected, and an electron beam incident from the opposite direction is a condition that cancels the influence of the force received from the electric field and the force received from the magnetic field ( Wien condition) can be created, whereby the primary electron beam is deflected and irradiates vertically onto the wafer, and the secondary electron beam can travel straight towards the detector.

  The detailed structure of the electron beam deflecting section of the E × B unit will be described with reference to FIG. 41 and FIG. 42 showing a longitudinal section along the line AA in FIG. As shown in FIG. 41, the field of the electron beam deflecting unit 41.2 of the E × B unit 41.1 is a structure in which an electric field and a magnetic field are orthogonal to each other in a plane perpendicular to the optical axis of the mapping projection optical unit. That is, an E × B structure is adopted. Here, the electric field is generated by the electrodes 41 · 3 and 41 · 4 having concave curved surfaces. The electric fields generated by the electrodes 41 · 3 and 41 · 4 are controlled by the control units 41 · 5 and 41 · 6, respectively. On the other hand, the magnetic coils 41 · 7 and 41 · 8 are arranged so as to be orthogonal to the electric field generating electrodes 41 · 3 and 41 · 4 to generate a magnetic field. The electric field generating electrodes 41 and 3 and 41 and 4 are point objects, but may be concentric circles.

  In this case, in order to improve the uniformity of the magnetic field, a magnetic path is formed by providing a pole piece having a parallel plate shape. The behavior of the electron beam in the longitudinal section along the line AA is as shown in FIG. The irradiated electron beams 42 · 1 and 42 · 2 are deflected by the electric field generated by the electrodes 41 · 3, 41 · 4 and the magnetic field generated by the electromagnetic coils 41 · 7, 41 · 8, and then the sample surface. Incident in the direction perpendicular to the top.

  Here, the incident positions and angles of the irradiation electron beams 42. 1 and 42. 2 to the electron beam deflecting units 41 and 2 are uniquely determined when the electron energy is determined. Further, the conditions of the electric field and the magnetic field, that is, the electric field generated by the electrodes 41, 3 and 41, 4 so that the secondary electrons 42, 3 and 42, 4 go straight, and the electromagnetic coil The control unit 41 · 5, 41 · 6, 41 · 9, 41 · 10 controls the magnetic field generated by 41 · 7, 41 · 8, so that secondary electrons are converted into the electron beam deflection unit 41.2. , And enter the mapping projection optical unit. Here, v is an electron velocity (m / s), B is a magnetic field (T), e is a charge amount (C), and E is an electric field (V / m).

  Here, the E × B filter 41.1 is used to separate the primary electron beam and the secondary electron, but it goes without saying that it is possible to use a magnetic field. Further, the primary electron beam and the secondary electron may be separated only by the electric field. Furthermore, it can be used for separation of primary electrons and reflected electrons.

  Here, as a ninth embodiment, a modification of the E × B filter will be described with reference to FIG. FIG. 43 is a cross-sectional view taken along a plane perpendicular to the optical axis. The four pairs of electrodes 43.1 and 43.2, 43.3 and 43.4, 43.5 and 43.6, 43.7 and 43.8 for generating an electric field are formed of nonmagnetic conductors. As a whole, it has a substantially cylindrical shape, and is fixed to the inner surface of the electrode supporting cylinders 43 and 9 made of an insulating material by screws (not shown) or the like. The axis of the electrode supporting cylinders 43 and 9 and the axis of the cylinder formed by the electrodes are made to coincide with the optical axes 43 and 10. Grooves 43 and 11 parallel to the optical axes 43 and 10 are provided on the inner surfaces of the electrode supporting cylinders 43 and 9 between the electrodes 43 and 1 to 43 and 8, respectively. And the area | region of the inner surface is coated with the conductor 43 * 12, and is set to earth potential.

  When the electric field is generated, “cos θ1” is applied to the electrodes 43, 3 and 43.5, “−cos θ1” is applied to the electrodes 43, 6 and 43.4, “cos θ2” is applied to the electrodes 43-1, 43 and 7 and the electrodes 43 When a voltage proportional to “−cos θ2” is applied to 8, 43.2, a substantially uniform parallel electric field is obtained in a region of about 60% of the inner diameter of the electrode. FIG. 44 shows a simulation result of the electric field distribution. In this example, four pairs of electrodes are used, but even with three pairs, a uniform parallel electric field can be obtained in a region of about 40% of the inner diameter.

  The magnetic field is generated by arranging two rectangular platinum alloy permanent magnets 43 and 13 and 43 and 14 in parallel outside the electrode supporting cylinders 43 and 9. Protrusions 43 and 16 made of a magnetic material are provided around the surfaces of the permanent magnets 43 and 13 and 43 and 14 on the optical axis 43 and 10 side. The protrusions 43 and 16 compensate for the fact that the magnetic lines of force on the optical axis 43 and 10 side are distorted outwardly. Its size and shape can be determined by simulation analysis.

  On the outside of the permanent magnets 43, 13, 43, 14, the passage on the side opposite to the optical axis 43, 10 of the magnetic lines of force by the permanent magnets 43, 13, 43, 14 is a cylinder coaxial with the electrode support cylinders 43, 9. As described above, yokes or magnetic circuits 43 and 15 made of a ferromagnetic material are provided.

The E × B separator as shown in FIG. 43 can be applied not only to the mapping projection electron beam inspection apparatus as shown in FIG. 25-1, but also to the scanning electron beam inspection apparatus.
An example of the scanning electron beam inspection apparatus is shown in FIG. An electron beam is irradiated from the electron gun 25/14 toward the sample 25/15. The primary electron beam passes E × B 25 · 16, but at the time of incidence, it travels straight without being applied with a deflection force, is narrowed down by the objective lens 25 · 17, and enters the sample 25 · 15 almost perpendicularly. The electrons coming out of the samples 25 and 15 are applied with a deflection force this time by E × B 25 and 16 and guided to the detectors 25 and 18. In this way, by adjusting the electric field and magnetic field of E × B25 · 16, one of the primary and secondary charged particle beams can be made to go straight, and the other can be made to go straight in any direction.

  When E × B25 · 16 is used, aberration is generated in the changed direction due to the addition of deflection force. To correct this, the electron guns 25 and 14 of the primary optical system and E × B25 · 16 may further be provided with an E × B deflector. For the same purpose, an E × B deflector may be further provided between the secondary system detectors 25 and 18 and the E × B 25 and 16.

  In a scanning electron beam inspection apparatus or a scanning electron microscope, narrowing with a primary electron beam leads to an increase in resolution, so that no excessive deflection force is given to the primary electron beam. As shown in FIG. 25-2, the primary system electron beam generally travels straight and the secondary system beam is deflected. However, conversely, if it is preferable to deflect the primary beam and move the secondary beam straight, it may be so. Similarly, in the projection type electron beam inspection apparatus, in order to properly correspond the imaging region on the sample and the pixels on the CCD of the detector, the secondary system beam has a deflection force that causes as little aberration as possible. It is generally preferable not to give it. Therefore, as shown in FIG. 25A, it is general to adopt a configuration in which the primary system beam is deflected and the secondary system beam goes straight, but the primary system beam goes straight and the secondary system beam goes straight. If it is preferable to adopt a configuration for deflecting the beam of the system, such a configuration may be adopted.

  The setting of the electric field and magnetic field strength of E × B may be changed for each mode such as the secondary electron mode and the reflected electron mode. It is possible to set the strength of the electric field and the magnetic field so that an optimal image can be obtained for each mode. Needless to say, when there is no need to change the setting, the strength can remain constant.

As is clear from the above explanation, according to this example, a region where both the electric field and the magnetic field are uniform around the optical axis can be made large, and even if the irradiation range of the primary electron beam is expanded, E × The aberration of the image passing through the B separator can be set to a value with no problem. In addition, since the projections 43 and 16 are provided in the periphery of the magnetic pole that forms the magnetic field, and the magnetic pole is provided outside the electric field generating electrode, a uniform magnetic field can be generated and the electric field distortion due to the magnetic pole can be reduced. . Further, since the magnetic field is generated using the permanent magnet, the entire E × B separator can be stored in a vacuum. Furthermore, the entire E × B separator can be miniaturized by forming the electric field generating electrode and the magnetic path forming magnetic circuit into a coaxial cylindrical shape with the optical axis as the central axis.

2-3-4) Detector The secondary electron image from the wafer imaged by the secondary optical system is first amplified by a microchannel plate (MCP) and then converted to an image of light by hitting the fluorescent screen. The principle of MCP is 1 to 100 μm in diameter and 0.2 to 10 mm in length, preferably 2 to 50 μm in diameter and 02.02 in length. Bundled millions to tens of millions of very thin conductive glass capillaries having a diameter of ˜5 mm, more preferably 6 to 25 μm, and a length of 0.24 to 1.0 mm, and shaped into a thin plate, By applying a predetermined voltage, each of the capillaries functions as an independent secondary electron amplifier, forming a secondary electron amplifier as a whole. The image converted into light by this detector is projected one-to-one on the TDI-CCD by the FOP system placed in the atmosphere through the vacuum transmission window.

  Here, the operation of the electron optical device having the above-described configuration will be described. As shown in FIG. 25A, the primary electron beam emitted from the electron gun 25.4 is focused by the lens system 25.5. The converged primary electron beam is incident on the E × B deflector 25, 6 and deflected so as to irradiate the surface of the wafer W perpendicularly, and is imaged on the surface of the wafer W by the objective lens system 25, 8. Is done.

  The secondary electrons emitted from the wafer by the irradiation of the primary electron beam are accelerated by the objective lens system 25, 8 and incident on the E × B type deflector 25, 6 and travel straight through the deflector to the secondary optical system. Are guided to the detectors 25 and 11 through the lens systems 25 and 10. Then, it is detected by the detectors 25 and 11, and the detection signal is sent to the image processing units 25 and 12. It is assumed that the objective lens systems 25 and 7 are applied with a high voltage of 10 to 20 kV and the wafer is installed.

  Here, when there are vias 25 and 13 in the wafer W and the voltage applied to the electrodes 25 and 8 is −200 V, the electric field on the electron beam irradiation surface of the wafer is 0 to −0.1 V / mm (− is the wafer). W side indicates high potential). In this state, no discharge is generated between the objective lens systems 25 and 7 and the wafer W, and the wafer W can be inspected for defects, but the secondary electron detection efficiency is slightly lowered. Accordingly, a series of operations for irradiating an electron beam and detecting secondary electrons is performed, for example, four times, and the detection results obtained for four times are subjected to processing such as cumulative addition and averaging to obtain a predetermined detection sensitivity. .

  In addition, when the wafer does not have vias 25 and 13, even if the voltage applied to the electrodes 25 and 8 is +350 V, no discharge occurs between the objective lens systems 25 and 7 and the wafer, and the defect inspection of the wafer W is performed. I did it. In this case, since the secondary electrons are focused by the voltage applied to the electrodes 25 and 8 and further focused by the objective lenses 25 and 7, the detection efficiency of the secondary electrons in the detectors 25 and 11 is improved. Therefore, the processing as a wafer defect device has also been speeded up, and inspection can be performed with high throughput.

2-3-5) Power supply The power supply unit in this apparatus is mainly composed of a DC high-voltage precision power supply having several hundred output channels for electrode control, and the supply voltage depends on the role and positional relationship of the electrodes. Although it is different, stability is required to be on the order of several hundred ppm or less, preferably 20 ppm or less, and more preferably several ppm with respect to the set value because of the requirements of image resolution and accuracy. In order to minimize aging fluctuations, temperature fluctuations, noise and ripples, etc., circuit methods, component selection, and mounting have been devised.

The types of power sources other than the electrodes include a constant current source for heating the heater, and a high-voltage high-speed amplifier for deflecting the beam two-dimensionally in order to confirm the centering of the beam near the center of the aperture electrode when the beam of the primary system is centered A constant current source for heating the heater, a constant current source for the electromagnetic coil for E × B that is an energy filter, a retarding power source for applying a bias to the wafer, and a potential for attracting the wafer to the electrostatic chuck are generated. There are a power supply, a high-voltage high-speed amplifier that performs EO correction, an MCP power supply that amplifies electrons based on the principle of a photomultiplier, and the like.

  FIG. 45 shows the overall configuration of the power supply unit. In the figure, power is supplied from the power supply rack 45.2 and the high-speed high-voltage amplifiers 45.3, 45.4 and 45.5 to the electrodes of the lens barrel 45.1 via the connection cable. The High-speed and high-voltage amplifiers 45.3 to 45.5 are wide-band amplifiers, and the frequency of signals handled is high (DC-MHz). Install in the vicinity to prevent an increase in cable capacitance. A correction signal is output from the EO correction 45.6, and the octupole conversion units 45 and 7 convert the respective electrodes of the octupole into voltages having phases and magnitudes matched to the vector values, and the high-speed high-voltage amplifier 45 -After being input to 4 and amplified, it is supplied to the electrodes included in the lens barrel.

  AP image acquisition blocks 45 and 8 generate a sawtooth wave from the AP image acquisition blocks 45 and 8 in order to confirm the centering of the beam in the vicinity of the center of the aperture electrode when the beam of the primary system is centered. By applying the beam to the deflection electrode 45.1 and deflecting the beam two-dimensionally, the magnitude of the beam current received by the aperture electrode is related to the position, and the image is displayed, so that the beam position is brought to the mechanical center position. Serves as an auxiliary function to adjust.

  From the AF control 46.9, the voltage equivalent to the best focal point measured in advance is stored in the memory, this value is read according to the stage position, converted to analog voltage by the D / A converter, and high speed Through the high-voltage amplifier 45.5, the function of applying to the focus adjusting electrode included in the lens barrel 45.1 and observing while maintaining the optimum focus position is realized.

  The power supply racks 45 and 2 store DC high-voltage precision power supplies having about several hundreds of output channels for electrode control, which are composed of power supply groups 1 to 4. The power supply racks 45 and 2 are connected to the communication cards 45 and 11 by the control communication units 45 and 10, and the optical fiber communication 45 that has electrical insulation to ensure safety and prevent occurrence of ground loops to prevent noise contamination. 12 constitutes a system capable of receiving commands from the control CPU units 45 and 13 and transmitting a status such as abnormality of the power supply device. The UPS 45/14 prevents system damage, abnormal discharge, danger to the human body, etc. due to system runaway when a control abnormality occurs due to a power failure, unexpected power interruption, or the like. The power supplies 45 and 15 are large power receiving units, and include an interlock, a current limit, and the like, and are configured to perform safety coordination as a whole defect inspection apparatus.

The communication cards 45 and 11 are connected to the data buses 45 and 16 and the address buses 45 and 17 of the control CPU units 45 and 13 and can perform real-time processing.
FIG. 46 shows an example of a circuit configuration of a static high voltage unipolar power source (for a lens) with respect to a circuit system in the case of generating a static DC voltage of several hundred to several tens of kilovolts. In FIG. 46, an AC voltage having a frequency at which the magnetic permeability of the transformer 46.2 is optimal is generated by the signal source 46.1, and after passing through the multipliers 46.3, is led to the drive circuit 46.4 and the transformer 46.2 Thus, a voltage having an amplitude several tens to several hundreds of times is generated. The Cockcraft Walton circuit 46.5 is a circuit that boosts voltage while rectifying. A desired DC voltage is obtained by the combination of the transformer 46.2 and the Cockcraft Walton circuit 46.5, and further smoothing is performed by the low-pass filter 46/6 to reduce ripple and noise. The high voltage output voltage is divided by the resistance ratio of the output voltage detection resistors 46, 7 and 46.8, and is set within a voltage range that can be handled by a normal electronic circuit. Since the stability of this resistor determines most of the voltage accuracy, elements with excellent temperature stability, long-term fluctuation, etc. are used.Particularly, the voltage division ratio is important, so a thin film can be formed on the same insulating substrate. Or, take measures such as bringing resistance elements close together so that the temperature does not change.

  The divided results are compared with the values of the D / A converters 46 and 10 for generating the reference voltage by the operational amplifiers 46 and 9. If there is an error, the outputs of the operational amplifiers 46 and 9 increase or decrease, and multiplication is performed. An AC voltage with an amplitude corresponding to the value is output from the unit 46.3 to form a negative feedback. Although not shown, the outputs of the operational amplifiers 46 and 9 are unipolar, or the response quadrants of the multipliers 46 and 3 are limited to prevent saturation. The operational amplifiers 46 and 9 require a very large amplification degree (120 dB or more), and are used in an open loop as an element, so a low noise operational amplifier is used. From the viewpoint of accuracy, the D / A converters 46 and 10 for generating the reference voltage need to have stability equal to or higher than that of the output voltage detection resistors 46 and 7 and 46 and 8. In order to generate this voltage, although not shown, a reference IC in which a constant voltage diode using a band gap is combined with a constant temperature function using a heater is often used, but a Peltier element is used instead of the heater. It is possible to make the temperature constant. In some cases, Peltier elements are used in a single stage or in multiple stages in order to make the output voltage detection resistors 46, 7 and 46, 8 constant.

  FIG. 47 shows an example of a circuit configuration of a static bipolar power supply (for an aligner or the like). The basic idea is to generate V5 and V6 with a power supply equivalent to the circuit of FIG. 46, and use this voltage to send the command value from 47.1 to the linear amplifier composed of 47.1 to 47.6. A bipolar high-voltage power supply is formed by input. In general, since the operational amplifier 47.2 operates in the vicinity of ± 12V, it is not shown in the figure. However, an amplifier circuit using a discrete element is required between 47.2 and 47.5, 47-6, and ± V Is amplified and converted to several hundred to several kilovolts. The precautions for various characteristics required for 47.1 to 47.4 are the same as those described in the circuit of FIG.

  48 to 50 show circuit examples of the special power supply, and FIG. 48 is a circuit example for the heater and gun, which is formed of 48 · 1 to 48 · 4. A voltage source 48 • 1, a resistor 48 • 3, and a power source 48 • 4 are superimposed on the bias voltage source 48 • 2. The heater power supply 48.4 is composed of a constant current source, and the value of the actually flowing current is detected by the resistor 48.3. Although not shown in the figure, after being replaced with digital, isolation is performed using an optical fiber or the like. The value is sent to the control communication unit 45. For the setting of the voltage value of the voltage source 48. 1 and the current value of the power supply 48. 4, the values from the control communication units 45 and 10 are inversely converted on the same principle, and values are set in the actual power supply setting unit. .

  FIG. 49 shows an example of a power supply circuit for MCP, from voltage sources 49.1, 49.2, relay circuits 49.3, 49.4, current detection circuits 49.5, 49.6, 49.7. Become. Since the terminal MCP1 performs measurement from several PAs by measuring the inflow value of current into the MCP, it is necessary to have a strict shield structure to prevent leakage current and noise from entering. The terminal MCP2 includes current measurement after amplification by the MCP, and the amplification degree can be calculated by the ratio of the current values flowing through the resistors 49, 6 and 49, 7. Resistor 49.5 measures the current on the phosphor screen. Measurement and setting at the overlapped portion are the same as those for the heater and gun.

  FIG. 50 shows a circuit example of an E × B constant current source for a magnetic coil formed by 50 · 1 and 50 · 2, and generally outputs a current of several hundred mA. The stability of the magnetic field as an energy filter is important, and a stability of the order of several ppm is required.

FIG. 51 shows an example of a power supply circuit for a retarding chuck.
5 · 1 to 51 · 9. 46. A power source similar to the static bipolar power source (for an aligner or the like) in FIG. 46 is superimposed on the bias power source (for retarding) 51.10. Measurement and setting in the overlapped portion are the same as those in the heater and gun (FIG. 48).

  FIG. 52 shows an example of a hardware configuration of the deflection electrode for EO correction, which is composed of 52 · 1 to 52 · 7. Correction signals are input to the octopole converters 52 and 4 from the X-axis EO correction 52 and 1 and the Y-axis EO correction 52 and 2, and the converted output is sent to the high-speed amplifiers 52 and 5. After being amplified to several tens to several hundreds V by 52.5, a voltage is applied to the EO correction electrodes 52.6 installed every 45 degrees. The ΔX correction 52.3 is an input for performing fine correction such as mirror bending, and is added to the X signal inside the 52.4.

  FIG. 53 shows an example of the circuit configuration of the octopole converter, and signals 53. 2, 53. 3, 53 are used for the electrodes 53. 1 installed at an angle shifted by 45 degrees other than the X and Y axes. -Vector calculation is performed from 4, 53.5, and voltages such as phases are generated. The calculation example in this case uses the values described in 53 · 6, 53 · 7, 53 · 8, 53 · 9. This can be realized by an analog resistance network, or by reading a table with a ROM when 53 · 6 to 53 · 9 are digital signals.

  FIG. 54 shows an example of a high-speed and high-voltage amplifier, which is composed of 54 · 1 to 54 · 11. (B) shows an example of a waveform when a rectangular wave is output. In this example, an amplifier is constructed using a power operational amplifier PA85A manufactured by APEX in the United States, a mega band, an output range of about ± 200 V, and a slew rate of less than about 1000 V / μS can be realized. Realized the characteristic.

2-4) Precharge Unit As shown in FIG. 13, the precharge units 13 and 9 are disposed adjacent to the lens barrels 13 and 38 of the electro-optical devices 13 and 8 in the working chambers 13 and 16. . Since this inspection device is a device that inspects the device pattern formed on the wafer surface by irradiating an electron beam to the substrate to be inspected, that is, the wafer, information such as secondary electrons generated by the electron beam irradiation Is the wafer surface information, but the wafer surface may be charged (charged up) depending on conditions such as the wafer material and the energy of irradiated electrons. In addition, there may be places where the wafer surface is strongly charged and weakly charged. If the charge amount on the wafer surface is uneven, the secondary electron information is also uneven, and accurate information cannot be obtained.

  Therefore, in the embodiment of FIG. 13, in order to prevent this unevenness, precharge units 13 and 9 having charged particle irradiation units 13 and 39 are provided. Before irradiating inspection electrons to a predetermined portion of the wafer to be inspected, charged particles are irradiated from the charged particle irradiating portions 13 and 39 of the precharge units 13 and 9 in order to eliminate charging unevenness, thereby eliminating uneven charging. This charge-up of the wafer surface is detected in advance by forming an image of the wafer surface that is symmetrical to the detection, and evaluating the image, and the precharge units 13 and 9 are operated based on the detection. In the precharge units 13 and 9, the primary electron beam may be defocused, that is, the beam shape may be blurred.

FIG. 55 shows a main part of the first embodiment of the precharge units 13 and 9. The charged particles 55.1 are irradiated from the charged particle irradiation source 55.2 to the sample substrate W by being accelerated at a voltage set by the bias power source 55.3. The inspected areas 55 and 4 show the places where the pre-processed charged particle irradiation has already been performed together with the areas 55 and 5, and the areas 55 and 6 show the places where the charged particle irradiation is being performed. In this figure, the sample substrate W is scanned in the direction of the arrow in the figure. However, when reciprocal scanning is performed, the other charged particle beam sources 55 and 7 are placed on the opposite side of the primary electron beam source as indicated by the dotted lines in the figure. The charged particle beam sources 55.2, 55.7 may be alternately turned on and off in synchronization with the scanning direction of the sample substrate W. In this case, if the energy of the charged particles is too high, the secondary electron yield from the insulating part of the sample substrate W will exceed 1, the surface will be positively charged, and even if the secondary electrons are generated below this, the phenomenon will occur. Since it becomes complicated and the irradiation effect decreases, it is effective to set the landing voltage to 100 eV or less (ideally 0 eV or more and 30 eV or less) at which the generation of secondary electrons is drastically reduced.

  FIG. 56 shows a second embodiment of the precharge units 13 and 9. This figure shows an irradiation source of a type that irradiates an electron beam 56.1 as a charged particle beam. The irradiation source is composed of a heat filament 56, 2, an extraction electrode 56, 3, a shield case 56, 4, a filament power source 56, 5, and an electron extraction power source 56, 6. The extraction electrode 56.3 has a thickness of 0.1 mm, a slit having a width of 0.2 mm and a length of 1.0 mm, and a positional relationship with a filament (thermoelectron emission source) 56.2 having a diameter of 0.1 mm. Is in the form of a three-electrode electron gun. The shield case 56.4 is provided with a slit having a width of 1 mm and a length of 2 mm. The shield case 56.4 has a distance of 1 mm from the extraction electrode 56.3, and is assembled so that the slit centers thereof coincide with each other. The material of the filament is tungsten (W), which is energized and heated at 2 A, and an electron current of several μA is obtained at an extraction voltage of 20 V and a bias voltage of −30 V.

  The example shown here is one example. For example, the material of the filament (thermoelectron emission source) can be a refractory metal such as Ta, Ir, Re, Triacoat W, an oxide cathode, etc. Needless to say, the filament current varies depending on the wire diameter and length. Other types of electron guns can be used as long as the electron beam irradiation region, electron current, and energy can be set to appropriate values.

  FIG. 57 shows a third embodiment of the precharge units 13 and 9. An irradiation source of a type that irradiates ions 57.1 as a charged particle beam is shown. This irradiation source is composed of a filament 57.2, a filament power source 57.3, a discharge power source 57.4, and an anode shield case 57.5, and the anode 57.6 and the shield case 57.5 have a size of 1 mm × 2 mm. The slits of the same size are opened, and the slits are assembled so that the centers of both slits coincide with each other at intervals of 1 mm. About 1 Pa of Ar gas 57.8 is introduced into the shield case 57.5 via the pipes 55.7, and the arc discharge type is operated by the hot filament 57.2. The bias voltage is set to a positive value.

  FIG. 58 shows the case of the plasma irradiation method which is the fourth embodiment of the precharge units 13 and 9. The structure is the same as in FIG. Similarly to the above, the arc discharge type is operated by the hot filament 57.2, but when the bias potential is set to 0V, the replasma 58.1 oozes out from the slit due to the gas pressure, and is irradiated to the sample substrate. In the case of plasma irradiation, both the positive and negative surface potentials of the sample substrate surface can be brought close to 0 because of the group of particles having both positive and negative charges as compared with other methods.

  The charged particle irradiation unit disposed close to the sample substrate W has the structure shown in FIGS. 55 to 58, and the difference in the surface structure of the oxide film and nitride film of the sample substrate W and the different process steps. The sample particles are irradiated with the charged particles 55.1 under appropriate conditions so that the surface potential becomes zero. After irradiating the sample substrate with the optimum irradiation conditions, that is, After the surface potential of the sample substrate W is averaged or neutralized by charged particles, an image is formed by the electron beams 55, 8 and 55, 9 to detect defects.

As described above, in the present embodiment, the measurement can be correctly measured because the measurement image distortion due to charging does not occur or is slight even if it is caused by the process immediately before the measurement by charged particle irradiation. In addition, since the stage can be scanned by irradiation with a large amount of current (for example, 1 μA or more and 20 μA, preferably 1 μA or more and 10 μA, more preferably 1 μA or more and 5 μA) that has been a problem in use in the past, Since a large amount is released from the wafer, a detection signal with a good S / N ratio (for example, 2 or more and 1000 or less, preferably 5 or more and 1000 or less, more preferably 10 or more and 100 or less) is obtained, and reliability of defect detection is obtained. Improves. In addition, since the S / N ratio is large, good image data can be produced even if the stage is scanned earlier.
Inspection throughput can be increased.

  FIG. 59 schematically shows an imaging apparatus including a precharge unit according to the present embodiment. This imaging device 59. 1 includes a primary optical system 59. 2, a secondary optical system 59. 3, a detection system 59. 4, and charge control means 59. And 5. The primary optical system 59.2 is an optical system that irradiates the surface of an inspection object (hereinafter referred to as an object) W with an electron beam, and an electron gun 59/6 that emits an electron beam and primary electrons emitted from the electron gun 59/6. An electrostatic lens 59, 8 for deflecting the beam 59, 7, a Wien filter, ie, an E × B deflector 59, 9, for deflecting the primary electron beam so that its optical axis is perpendicular to the target surface, and an electron beam As shown in FIG. 59, these lenses are provided with deflecting electrostatic lenses 59 and 10, which are arranged in order with the electron gun 59 and 6 being at the top, and the optical axes of the primary electron beams 59 and 7 emitted from the electron gun. Are arranged so as to be inclined with respect to a line perpendicular to the surface (sample surface) of the object W. The E × B deflectors 59 and 9 are composed of electrodes 59 and 11 and electromagnets 59 and 12.

  The secondary optical system 59. 3 includes electrostatic lenses 59 and 13 arranged above the E × B type deflectors 49 and 9 of the primary optical system. The detection system 59.4 includes a scintillator and microchannel plate (MCP) combination 59/15 that converts secondary electrons 59/14 into an optical signal, a CCD 59/16 that converts the optical signal into an electrical signal, and an image processing apparatus. 59.17. Since the structures and functions of the constituent elements of the primary optical system 59 • 2, the secondary optical system 59 • 3, and the detection system 59 • 4 are the same as those in the prior art, detailed description thereof will be omitted.

  In this embodiment, the charge control means 59. 5 for equalizing or reducing the charge charged to the target is the target W and the electrostatic lenses 59. 10 of the primary optical system 59. 2 closest to the target W. Between the electrodes 59 and 18 disposed close to the object W, the changeover switches 59 and 19 electrically connected to the electrodes 59 and 18, and one terminal 59 and 20 of the changeover switches 59 and 19 Electrically connected voltage generators 59 and 21 and charge detectors 59 and 23 electrically connected to the other terminals 59 and 22 of the changeover switches 59 and 19 are provided. The charge detectors 59 and 23 have a high impedance. The charge reducing means 59. 5 further includes a grid 59. 24 disposed between the electron gun 59. 6 of the primary optical system 59. 2 and the electrostatic lens 59. 8, and the grid 59. And voltage generators 59 and 25 connected to each other. Timing generators 59 and 26 include CCDs 59 and 16 and image processing devices 59 and 17 of detection systems 59 and 4, changeover switches 59 and 19 of charge reduction means 59 and 5, voltage generators 59 and 21, and charge detectors 59 and 23. And 59.25 are instructed for operation timing.

  Next, the operation of the electron beam apparatus having the above configuration will be described. The primary electron beams 59 and 7 emitted from the electron gun 59 and 6 reach the E × B deflectors 59 and 9 through the electrostatic lenses 59 and 8 of the primary optical system 59 and 2, and the E × B deflectors 59. 9 is deflected so as to be perpendicular to the surface of the target W, and further irradiates the surface (target surface) WF of the target W through the electrostatic lenses 59 and 10. Secondary electrons 59 and 14 are emitted from the surface WF of the object W according to the properties of the object. The secondary electrons 59 and 14 are sent to the combination 59 and 15 of the scintillator and MCP of the detection system 59 and 4 via the electrostatic lenses 59 and 13 of the secondary optical system 59 and 3, and converted into light by the scintillator. The light is photoelectrically converted by the CCDs 59 and 16, and the image processing devices 59 and 17 form a two-dimensional image (having gradation) by the converted electric signal. As in a normal inspection apparatus of this type, the primary electron beam applied to the object is scanned by a known deflection means (not shown) by the primary electron beam, or a table T that supports the object. Is moved in the two-dimensional directions of X and Y, or a combination thereof, and the entire necessary portion on the target surface WF is irradiated to collect data on the target surface.

Charges are generated in the vicinity of the surface of the target W by the primary electron beams 59 and 7 irradiated to the target W and are positively charged. As a result, the orbits of the secondary electrons 59 and 14 generated from the surface WF of the target W change according to the state of the charge due to the Coulomb force with the charge. As a result, distortion occurs in the images formed in the image processing apparatuses 59 and 17. Since the charging of the target surface WF changes depending on the properties of the target W, when a wafer is used as the target, the same wafer is not necessarily the same, and also changes with time. Therefore, there is a possibility that erroneous detection occurs when two patterns on the wafer are compared.

  Therefore, in this embodiment according to the present invention, the charge detectors 59 and 23 having a high impedance are used by the charge detectors 59 and 23 using the idle time after the CCDs 59 and 16 of the detection systems 59 and 4 capture one image. The charge amount of the electrodes 59 and 18 arranged in the vicinity of W is measured. Then, a voltage for irradiating electrons corresponding to the measured charge amount is generated by the voltage generators 59 and 21, and after the measurement, the changeover switches 59 and 19 are operated to connect the electrodes 59 and 18 to the voltage generators 59 and 21, By applying the voltage generated by the voltage generator to the electrodes 59 and 18, the charged charges are canceled out. As a result, the image formed in the image processing devices 59 and 17 is prevented from being distorted. Specifically, when a normal voltage is applied to the electrodes 59 and 18, the object W is irradiated with a focused electron beam. However, if another voltage is applied to the electrodes 59 and 18, the focusing condition is greatly shifted. A wide area where charging is expected is irradiated with a small current density, and by neutralizing the positive charge of a positively charged object, the voltage in a wide area where charging is expected is changed to a specific positive (negative) voltage. By making it uniform or making it uniform and reduced, a lower positive (negative) voltage (including zero volts) can be achieved. The charged charge canceling operation as described above is performed for each scan.

  The Wehnelt electrodes, that is, the grids 59 and 24, have a function of stopping the electron beam irradiated from the electron gun 59 and 6 during the idle time, and stably performing the charge amount measurement and the charge canceling operation. The timing of the above operation is instructed by the timing generators 59 and 26, and is, for example, the timing shown in the timing chart of FIG. When a wafer is used as a target, the amount of charge varies depending on the position of the wafer. Therefore, a plurality of electrodes 59, 18, selector switches 59, 19, voltage generators 59, 21, and charge detectors 59, 23 are provided in the CCD scanning direction. It is also possible to perform grouping and subdividing to perform more accurate control.

According to the present embodiment, the following effects can be obtained.
(1) Image distortion caused by charging can be reduced regardless of the property of the inspection object.
(2) Since charging is made uniform and offset using the idle time of the conventional measurement timing, the throughput is not affected at all.
(3) Since real-time processing is possible, post-processing time and memory are not required.
(4) High-speed and high-accuracy image observation and defect detection are possible.

  FIG. 61 shows a schematic configuration of a defect inspection apparatus including a precharge unit according to another embodiment of the present invention. The defect inspection apparatus includes an electron gun 59, 6 that emits a primary electron beam, an electrostatic lens 59, 8 that deflects and shapes the emitted primary electron beam, and a sample chamber 61.1 that can be evacuated to vacuum by a pump (not shown). , A stage 61. 2 disposed in the sample chamber and movable in a horizontal plane in a state where a sample such as a semiconductor wafer W is placed, a secondary electron beam emitted from the wafer W by irradiation of the primary electron beam, and / or Electrostatic lenses 59 and 13 of a projection system for projecting a reflected electron beam at a predetermined magnification to form an image, detectors 61 and 3 for detecting the formed image as a secondary electron image of the wafer, and the entire apparatus And a control unit 61.4 that executes processing for detecting defects on the wafer W based on the secondary electron image detected by the detector 61.3. The secondary electron image includes contributions from reflected electrons as well as secondary electrons, but here it will be referred to as a secondary electron image.

A UV lamp 61.5 that emits light in a wavelength region including ultraviolet light is installed above the wafer W in the sample chamber 61.1. This glass surface of the UV lamp 61 · 5, photoelectrons e due to photoelectric effect by light emitted from the UV lamp 651 - 5 ^- photoelectron emitting member 61, 6 to emit is coated. The UV lamp 61.5 can be selected from any light source that emits light in a wavelength region having a capability of emitting photoelectrons from the photoelectron emitting materials 61, 6. In general, it is advantageous in terms of cost to use a low-pressure mercury lamp that emits ultraviolet light of 254 nm. Further, the photoelectron emitting materials 61 and 6 can be selected from any metals as long as they have the ability to emit photoelectrons. For example, Au is preferable.

  The above-described photoelectrons have energy different from that of the primary electron beam, that is, energy lower than that of the primary electron beam. Here, low energy means an order of several eV to several tens eV, preferably 0 to 10 eV. The present invention can use any means for generating such low energy electrons. For example, it can also be achieved by providing a low energy electron gun (not shown) instead of the UV lamp 61.5.

  Furthermore, when controlling the energy of this electron gun, the defect inspection apparatus of the present embodiment includes power supplies 6. The negative electrodes of the power supplies 61 and 7 are connected to the photoelectron emitting materials 61 and 6, and the positive electrodes thereof are connected to the stage 61.2. Accordingly, the photoelectron emitting members 61 and 6 are in a state where a negative voltage is applied to the voltage of the stage 61.2, that is, the wafer W. The energy of the low energy electron gun can be controlled by this predetermined voltage.

  The detectors 61 and 3 can have an arbitrary configuration as long as the secondary electron images formed by the electrostatic lenses 59 and 13 can be converted into signals that can be post-processed. For example, as shown in detail in FIG. 62, the detectors 61 and 3 include a microchannel plate (MCP) 62 and 1, a phosphor screen 62 and 2, a relay optical system 62 and 3, and a number of CCD elements. The imaging sensors 62 and 4 can be configured. The microchannel plate 62. 1 has a large number of channels in the plate, and a larger number of electrons are emitted while the secondary electrons or reflected electrons imaged by the electrostatic lenses 59 and 13 pass through the channels. Is generated. That is, secondary electrons are amplified. The fluorescent screen 62.2 converts the secondary electrons into light by emitting fluorescence with the amplified secondary electrons. The relay lenses 62 and 3 guide this fluorescence to the CCD image sensor 62 and 4, and the CCD image sensor 62 and 4 converts the intensity distribution of secondary electrons on the surface of the wafer W into an electrical signal for each element, that is, digital image data. To the control unit 61.4.

  As illustrated in FIG. 61, the control units 61 and 4 can be configured by general-purpose personal computers 61 and 8. The computers 6 and 8 include a control unit main body 61 and 9 that executes various controls and arithmetic processes according to a predetermined program, a CRT 61 and 10 that displays a processing result of the main body 61 and 9, and an operator for inputting commands. The control units 61 and 4 may be configured from input units 61 and 11 such as a keyboard and a mouse, and of course, hardware dedicated to the defect inspection apparatus or a workstation.

The control unit main bodies 61 and 9 are composed of various control boards such as a CPU, RAM, ROM, hard disk, and video board (not shown). On a memory such as a RAM or a hard disk, a secondary electron image storage area for storing electrical signals received from the detectors 61 and 3, that is, digital image data of a secondary electron image of the wafer W, is allocated. In addition to a control program for controlling the entire defect inspection apparatus, secondary electron image data is read from the storage areas 61 and 12 on the hard disk, and defects on the wafer W are automatically detected according to a predetermined algorithm based on the image data. Defect detection programs 61 and 13 to be detected are stored. The defect detection programs 61 and 13 have, for example, a function of comparing the inspection location of the wafer W with another inspection location, and reporting a pattern that is different from the pattern of most other locations to the operator as a defect. Have. Further, the secondary electron images 61 and 14 may be displayed on the display units of the CRTs 61 and 10, and defects of the wafer W may be detected by visual observation by the operator.

  Next, the operation of the electron beam apparatus according to the embodiment shown in FIG. 61 will be described using the flowchart of FIG. 63 as an example. First, the wafer W to be inspected is set on the stage 61.2 (step 63.1). This may be a form in which a large number of wafers W stored in a loader (not shown) are automatically set on the stage 61.2 one by one. Next, a primary electron beam is emitted from the electron guns 59 and 6 and irradiated to a predetermined inspection region on the set wafer W surface through the electrostatic lenses 59 and 8 (step 63.2). Secondary electrons and / or reflected electrons (hereinafter referred to as “secondary electrons”) are emitted from the wafer W irradiated with the primary electron beam, and as a result, the wafer W is charged up to a positive potential.

Next, the generated secondary electron beam is imaged on the detector 61.3 at a predetermined magnification by the electrostatic lens 59.13 of the magnification projection system (step 63.3). At this time, the UV lamp 61.5 is caused to emit light in a state where a negative voltage is applied to the photoelectron emitting material 651 with respect to the stage 61.2 (step 63.4). As a result, the ultraviolet light having the frequency ν emitted from the UV lamp 61.5 causes the photoelectrons to be emitted from the photoelectron emitting material 65.1 by the energy quantum hν (h is Planck's constant). These photoelectrons e ⁻ are irradiated from the negatively charged photoelectron emitting materials 61 and 6 toward the positively charged wafer W to electrically neutralize the wafer W. Thus, the secondary electron beam is imaged on the detector 61. 3 without being substantially affected by the positive potential of the wafer W.

  The image of the secondary electron beam emitted from the wafer W thus electrically neutralized (with reduced image disturbance) is detected by the detectors 61 and 3 and converted into digital image data (step 63).・ 5). Next, the control units 6 and 4 execute defect detection processing of the wafer W based on the detected image data according to the defect detection programs 61 and 13 (steps 63 and 6). In the defect detection process, in the case of a wafer having a large number of the same dies, the control units 6 and 4 extract a defect portion by comparing the detected images of the detected dies as described above. It is also possible to automatically detect a defective portion by comparing and comparing a reference secondary electron image of a wafer having no defect stored in advance in a memory with an actually detected secondary electron beam image. At this time, the detected image may be displayed on the CRTs 61 and 10 and the portion determined to be a defective portion may be displayed as a mark, whereby the operator finally confirms whether or not the wafer W actually has a defect. Can be evaluated. A specific example of this defect detection method will be described later.

  As a result of the defect detection process in step 63.5, when it is determined that the wafer W has a defect (affirmative determination in step 63-7), the operator is warned of the presence of the defect (step 63.8). As a warning method, for example, a message notifying the presence of a defect may be displayed on the display unit of the CRT 61 or 10, or at the same time, enlarged images 61 and 14 of a pattern having a defect may be displayed. Such a defective wafer may be immediately taken out from the sample chamber 61.1 and stored in a storage place different from the wafer having no defect (steps 63 and 9).

  As a result of the defect detection process in Steps 63 and 6, when it is determined that the wafer W has no defects (No determination in Steps 63 and 7), the area to be inspected still remains for the wafer W that is currently inspected. It is determined whether or not (steps 63 and 10). When the region to be inspected remains (Yes in Steps 63 and 10), the stage 61.2 is driven, and the wafer W is moved so that the other region to be inspected now falls within the irradiation region of the primary electron beam. (Steps 63 and 11). Thereafter, the process returns to Step 63.2 to repeat the same processing for the other inspection areas.

  When there is no region to be inspected (No at Step 63 and 10), or after the defective wafer extraction process (Steps 63 and 9), the wafer W currently being inspected is the final wafer. It is determined whether there is no uninspected wafer in a loader (not shown) (steps 63 and 12). If the wafer is not the final wafer (No at Steps 63 and 12), the inspected wafer is stored in a predetermined storage location, and a new uninspected wafer is set on the stage 61.2 instead (Steps 63 and 13). Thereafter, the process returns to Step 63.2 to repeat the same processing for the wafer. If it is the last wafer (Yes at step 63/12), the inspected wafer is stored in a predetermined storage location, and all the processes are completed. Each cassette identification number and wafer identification number, such as a lot number, are also stored and managed.

  The UV photoelectron irradiation (step 63.4) can be performed at any timing and at any timing as long as positive charge-up of the wafer W is avoided and secondary electron image detection (step 63.5) can be performed in a state in which image disturbance is reduced. Can be done within the period. While the processing of FIG. 63 is continued, the UV lamps 6 and 5 may be lit at all times, but light emission and extinguishing may be repeated by setting a period for each wafer. In the latter case, as the timing of light emission, in addition to the timing shown in FIG. 63, before the execution of secondary electron beam imaging (steps 63 and 3), and further before the execution of primary electron beam irradiation (steps 63 and 2). You may start with. Although UV photoelectron irradiation is preferably continued at least during the secondary electron detection period, UV photoelectron irradiation can be performed if the wafer is sufficiently neutralized before or during detection of the secondary electron image. May be stopped.

  Specific examples of the defect detection method in steps 63 and 6 are shown in FIGS. First, FIG. 64A shows an image 64.1 of the die detected first and an image 64.2 of the other die detected second. If the third detected image of another die is determined to be the same as or similar to the first image 64. 1, it is determined that the 64 · 3 portion of the second die image 64 • 2 has a defect. The defect portion can be detected.

  FIG. 64B shows an example in which the line width of the pattern formed on the wafer is measured. When the actual pattern 64.4 on the wafer is scanned in the direction 64.5, the actual secondary electron intensity signal is 64.6, and this signal is a threshold level 64 determined in advance by calibration. The width 64.8 of the portion continuously exceeding 7 can be measured as the line width of the pattern 64.4. When the measured line width is not within the predetermined range, it can be determined that the pattern has a defect.

  FIG. 64 (c) shows an example in which the potential contrast of the pattern formed on the wafer is measured. In the configuration shown in FIG. 61, axially symmetrical electrodes 64 and 9 are provided above the wafer W, and, for example, a potential of −10 V is applied to the wafer potential of 0 V. The equipotential surface of −2 V at this time has a shape as indicated by 64 · 10. Here, it is assumed that the patterns 64 · 11 and 64 · 12 formed on the wafer have potentials of −4V and 0V, respectively. In this case, the secondary electrons emitted from the patterns 64 and 11 have an upward velocity corresponding to kinetic energy of 2 eV on the −2 V equipotential surfaces 64 and 10, so that the trajectory crosses the potential barrier 64 and 10. 64.13 and escapes from the electrode 64.9 and is detected by the detector 61.3. On the other hand, the secondary electrons emitted from the patterns 64 and 12 are not detected because they cannot pass through the potential barrier of −2 V and are driven back to the wafer surface as indicated by the trajectories 64 and 14. Therefore, the detected images of the patterns 64 and 11 are bright and the detected images of the patterns 64 and 12 are dark. Thus, a potential contrast is obtained. If the brightness and potential of the detected image are calibrated in advance, the pattern potential can be measured from the detected image. The defect portion of the pattern can be evaluated from this potential distribution.

If there is a floating part in the die, charge is applied by the precharge unit to charge the floating part, and the potential difference from the part that is electrically conductive and grounded is calculated. Can be generated. The potential contrast data in this state can be acquired and analyzed to find a floating portion. It can be used as a defect detection method when there is a killer defect or the like. The potential contrast data may be converted into a potential contrast image and compared with a potential contrast image of another die pattern, or may be compared with a potential contrast image acquired from design data such as CAD.

  FIG. 65 shows a schematic configuration of a defect inspection apparatus provided with a precharge unit according to another embodiment of the present invention. Note that the same components as those in the embodiment of FIG. 61 are denoted by the same reference numerals, and detailed description thereof is omitted. In this embodiment, as shown in FIG. 65, the photoelectron emitting material is not coated on the glass surface of the UV lamps 6. Instead, the photoelectron emission plate 65. 1 is arranged above the wafer W in the sample chamber 61. 1, and the UV lamp 6. 5 is arranged at a position where the emitted ultraviolet light is irradiated to the photoelectron emission plate 65. Is done. A negative electrode of a power source 71, 7 is connected to the photoelectron emission plate 65.1, and a positive electrode of the power source is connected to the stage 61.2. The photoelectron emission plate 65. 1 may be made of a metal such as Au, or may be made as a plate coated with such a metal.

  The operation of the embodiment of FIG. 65 is the same as that of the embodiment of FIG. In the embodiment of FIG. 65 as well, since the photoelectrons can be irradiated onto the surface of the wafer W in a timely manner, the same effects as in the embodiment of FIG.

  FIG. 66 shows a schematic configuration of a defect inspection apparatus provided with a precharge unit according to still another embodiment of the present invention. In addition, the same code | symbol is attached | subjected about the component similar to embodiment of FIG.61 and FIG.65, and detailed description is abbreviate | omitted. In the embodiment of FIG. 66, as shown in the figure, a transparent window member 66. 1 is provided on the side wall of the sample chamber 61. 1, and ultraviolet rays emitted from the UV lamps 6. A UV lamp 61.5 is disposed outside the sample chamber 61.2 so that the photoelectron emission plate 651 disposed above the wafer W in the sample chamber 61.1 is irradiated. In the embodiment of FIG. 66, the UV lamps 6 and 5 are arranged outside the sample chambers 61 and 1 to be evacuated, so that it is not necessary to consider the vacuum resistance performance of the UV lamps 6 and 5, and FIGS. Compared with the embodiment, the choices of the UV lamp 61.5 can be expanded.

  66 is the same as that of the embodiment of FIGS. 61 and 65. Also in the embodiment of FIG. 66, photoelectrons can be irradiated onto the surface of the wafer W in a timely manner, so that the same effects as those of the embodiment of FIGS. 61 and 65 can be obtained.

  Although the above is each said embodiment, the defect inspection apparatus provided with the precharge unit by this invention is not limited only to the said example, It can change arbitrarily suitably within the scope of the summary of this invention. . For example, the semiconductor wafer W is taken as an example of the sample to be inspected, but the sample to be inspected according to the present invention is not limited to this, and any one that can detect a defect with an electron beam can be selected. For example, a mask on which an exposure pattern for a wafer is formed, a transmission type mask (stencil mask), or the like can be used as an inspection target. Needless to say, the present invention can be used not only for semiconductor processes but also for inspection or evaluation related to micromachines and liquid crystals.

Moreover, as the electron beam apparatus for defect inspection, the configuration of FIGS. 61 to 66 is shown, but the electron optical system and the like can be arbitrarily changed. For example, the electron beam irradiation means (59, 6, 59, 8) of the illustrated defect inspection apparatus is a type in which a primary electron beam is incident on the surface of the wafer W obliquely from above. A primary electron beam deflecting unit may be provided below 13 so that the primary electron beam is incident on the surface of the wafer W perpendicularly. As such a deflecting means, for example, there is a Wien filter that deflects a primary electron beam by a field E × B in which an electric field and a magnetic field are orthogonal to each other.

  Furthermore, as means for emitting photoelectrons, any means other than the combination of the UV lamps 6 and 5 and the photoelectron emission members 61 and 6 or the photoelectron emission plate 65 and 1 shown in FIGS. 61 to 66 can be adopted. Of course.

  The flow of the flowchart in FIG. 63 is not limited to this. For example, the sample determined to be defective in Steps 63 and 7 is not subjected to defect inspection in other areas. However, the processing flow may be changed so that defects are detected over the entire area. Good. Further, if the irradiation region of the primary electron beam is enlarged and the entire inspection region of the sample can be covered by one irradiation, Steps 63 and 10 and Steps 6 and 11 can be omitted.

  Further, in FIG. 63, when it is determined in step 63 or 7 that the wafer has a defect, the operator is immediately warned of the presence of the defect in step 63 or 8 and the post-processing is performed (step 63 or 9), but the defect information is recorded. In addition, after the end of the batch processing (after step 63 and 12 affirmative determination), the processing flow may be changed so as to report defect information of a wafer having defects.

  As described in detail above, according to the defect inspection apparatus and the defect inspection method according to the embodiment of FIGS. 61 to 66, the energy different from the primary electron beam, that is, the electron having lower energy than the primary electron beam is supplied to the sample. As a result, positive charge-up of the sample surface due to secondary electron emission is reduced, and as a result, image defects of the secondary electron beam due to charge-up can be eliminated, and the defect of the sample can be made with higher accuracy. It is possible to obtain an excellent effect that it is possible to inspect.

  Further, if the defect inspection apparatus of FIGS. 61 to 66 is used in the device manufacturing method, the defect inspection of the sample is performed using the defect inspection apparatus as described above, so that the yield of the product is improved and the defective product is improved. It is possible to obtain an excellent effect of preventing shipping.

  The above is a case where the electron energy for precharge is mainly low energy of 100 eV or less and softly irradiates the sample surface. After charging, image acquisition may be performed in a positive charging mode, a negative charging mode, or a reflected electron mode. In the negative charging mode, precharging may be performed with the same energy as the landing energy of the electron beam at the time of inspection.

  It is also effective to coat the sample surface with a conductive thin film in order to suppress charging. The film thickness at this time is 1-100 mm, preferably 1-10 mm, more preferably 1-3 mm. Furthermore, if the image is acquired after cleaning the sample surface by sputter etching or the like, a cleaner image can be obtained. The conductive thin film coat and sputter etching may be used independently or in combination with precharge. For example, pre-charge may be performed after sputter etching to acquire an image, or pre-charge may be performed after coating the conductive thin film after sputter etching.

2-5) Vacuum exhaust system The vacuum exhaust system consists of a vacuum pump, a vacuum valve, a vacuum gauge, vacuum piping, etc., and the electron optical system, detector section, sample chamber, and load lock chamber are evacuated according to a predetermined sequence. Do. In each part, the vacuum valve is controlled so as to achieve a required degree of vacuum. The vacuum level is constantly monitored, and when an abnormality occurs, emergency control of the isolation valve, etc. is performed by an interlock function to ensure the vacuum level. As the vacuum pump, a turbo molecular pump is used for main exhaust, and a roots type dry pump is used for roughing. The pressure of the inspection place (electron beam irradiation part) is 10 ⁻³ to 10
⁻⁵ Pa, preferably 10 ⁻⁴ to 10 ⁻⁶ Pa, one digit lower than that, is practical.

2-6) Control system The control system mainly consists of a main controller, control controller, and stage controller. The main controller is equipped with a man-machine interface, through which operator operations are performed (various instructions / commands, recipe input, inspection start instructions, automatic and manual inspection mode switching, manual inspection mode, etc. Input of all necessary commands at the time). In addition, communication with the host computer in the factory, control of the evacuation system, sample transfer of wafers, alignment control, command transmission to other control controllers and stage controllers, reception of information, etc. are also performed by the main controller. . In addition, acquisition of image signals from an optical microscope, stage vibration correction function that feeds back stage fluctuation signals to the electron optical system to correct image deterioration, and the Z direction of the sample observation position (axial direction of the secondary optical system) It has an automatic focus correction function that detects the displacement, feeds back to the electron optical system, and automatically corrects the focus. Transmission / reception of a feedback signal and the like to the electron optical system and transmission / reception of a signal from the stage are performed via a control controller and a stage controller, respectively.

  The control controller is mainly responsible for the control of the electron beam optical system (control of high-precision power sources for electron guns, lenses, aligners, Wien filters, etc.). Specifically, each operation mode, such as ensuring that the irradiation area is always irradiated with a constant electron current even when the magnification changes, and automatically setting the voltage to each lens system and aligner corresponding to each magnification. Control (interlocking control) such as automatic voltage setting for each lens system and aligner corresponding to is performed.

  The stage controller mainly controls the movement of the stage to enable precise movement in the X and Y directions in the order of μm (± 5 μm or less, preferably ± 1 μm or less, more preferably ± 0.5 μm or less). ing. In this stage, the rotational direction control (θ control) is also performed within an error accuracy of about ± 10 seconds, preferably within ± 1 second, more preferably within ± 0.3 seconds. The configuration of the control system will be specifically described below.

2-6-1) Configuration and function This apparatus captures and displays a specified position of a wafer with an electron microscope or an optical microscope, and functions to detect and classify a defect by imaging the specified position of a wafer with an electron microscope, And a function of imaging and displaying the position where the defect is detected with an electron microscope or an optical microscope. In addition, for the realization and maintenance of the above functions, electron optical system control, vacuum system control, wafer transfer control, component unit single operation, imaging function, automatic defect inspection processing, apparatus abnormality detection, apparatus activation / And a stop processing function.

The auxiliary functions are as follows.
(1) Electro-optical system control function (a) Lens voltage application control (a-1) Interlocking control (a-2) Voltage application by application function (a-3) Multipole lens interlocking voltage application (a-4) Wobble control (B) Electron beam output adjustment (b-1) Preheat (Gun)
(B-2) Heat up (Gun)
(B-3) Emission current control (BIAS control)
(2) Vacuum system control function (a) Individual chamber evacuation / atmosphere release (b) Specified chamber batch evacuation / atmosphere release (3) Wafer transfer control function Step operation of the following operation / Fully automatic operation (a) Wafer load ( b) Wafer unloading (4) Component unit operation function (5) Imaging function

Select the following two input systems for imaging:
(A) CCD camera ・ Optical microscope low magnification (pixel size: 2.75 μm / pix)
・ High magnification of optical microscope (pixel size: 0.25 μm / pix)
(B) TDI camera (b-1) TDI-still
(B-2) TDI-scan
EB × 80 (pixel size: 0.2 μm / pix)
EB × 160 (pixel size: 0.1 μm / pix)
EB × 320 (pixel size: 0.05 μm / pix)
EB × 480 (pixel size: 0.03 μm / pix).

  Furthermore, there is a user mode designation function as a function for restricting items that can be operated according to the skill and knowledge level of the operator to prevent accidents due to erroneous operations. This user mode is designated by a user ID and password that are input when a GUI (Graphical User Interface) is started.

The user mode includes a maintenance mode, recipe creation mode, and operator mode. Operation is performed in maintenance mode during start-up work and maintenance work after device installation, and operations and procedures required in recipe creation mode when creating recipes At the time of automatic defect inspection, inspection is performed using a recipe already created in the operator mode. The relationship between each user mode and the device operation mode is as shown in FIG. here,
Maintenance mode. . . . Single component operation, wafer transfer, vacuum system control, electron optical system control, observation (light microscope imaging, TDI imaging), defect inspection, review recipe creation mode. . . . . Wafer transfer, observation (optical microscope imaging, TDI imaging), defect inspection, review operator mode. . . . . Automatic defect inspection (automatic control of necessary functions such as wafer transfer) and review.

  In this apparatus, there are apparatus constants and recipes as variable parameters necessary for operation. An apparatus constant is defined as a parameter that absorbs an error inherent to the apparatus (such as an attachment error), and a recipe is defined as a parameter that defines various conditions for automatically performing defect inspection. The device constant is set at the time of start-up work and after maintenance work, and is basically not changed thereafter.

  Recipes are classified into transport recipes, alignment recipes, die map recipes, focus map recipes, and inspection recipes, and defect inspection is performed according to these recipes, so setting work is performed before the inspection process is performed, and multiple patterns are set. Saved.

As a procedure for creating the recipe, as shown in FIG. 68, the first step is to transfer the wafer onto the stage (wafer load). After installing the wafer cassette in the system, a wafer search is performed to detect the presence / absence of wafers in each slot in the cassette. The notch direction is designated, and the wafer is loaded by the procedure shown in FIGS. These conditions are stored in the transfer recipe. The arrangement direction of the dies on the wafer loaded on the stage does not necessarily coincide with the scanning direction of the TDI camera (FIG. 71). In order to make this coincide, an operation of rotating the wafer on the θ stage is necessary, and this operation is called alignment (FIG. 72). In the alignment recipe, the alignment execution condition after being loaded on the stage is stored.

  A die map (FIG. 73) indicating the arrangement of dies is created at the time of alignment, and the die size and the position of the origin die (which serves as a starting point indicating the position of the die) are stored in the die map recipe.

2-6-2) Alignment Procedure As an alignment (positioning) procedure, first, coarse positioning is performed at a low magnification of the optical microscope, and then detailed positioning is performed at the high magnification of the optical microscope and finally by an EB image.

A. Image taken with optical microscope at low magnification (1) <Specifying first, second, third search die and template>
(1-1) First Search Die Designation and Template Designation The user moves the stage so that the lower left corner of the die located below the wafer is located near the center of the camera. get. This die is a die serving as a positioning reference, and the coordinates of the lower left corner are the coordinates of the feature points. In the future, we will measure the exact position coordinates of any die on the substrate by performing pattern matching on this template image. As the template image, an image that has a unique pattern within the search area must be selected.

  In this embodiment, the lower left corner is the pattern matching template image acquisition position. However, the present invention is not limited to this, and an arbitrary position in the die may be selected as the feature point. However, in general, it is preferable to select any of the four corners because the corners are easier to specify the coordinates than the points on the inside of the die or on the sides. Similarly, in the present embodiment, the pattern matching template image is acquired for the die located below the wafer, but it is natural that any die may be selected so that alignment can be easily performed.

(1-2) Second search die designation The die next to the right of the first search die is the second search die, and the stage is moved by the user operation so that the lower left corner of the second search die is located near the center of the camera. After the position is determined, the exact coordinates of the pattern of the second search die that matches the template image specified by the first search die by automatically performing pattern matching using the template image acquired in (1-1) above. Get the value.

  In the present embodiment, the die on the right side of the first search die is described as an example of the second search die. However, the second search die of the present invention is not limited to this. . In short, it is only necessary to select a point that can accurately grasp the positional relationship of the die in the row direction from the reference point that grasps the position map of the accurate feature point by pattern matching. Therefore, for example, the die next to the left of the first search die can be used as the second search die.

(1-3) Third search die designation The die next to the second search die is the third search die, and the stage is moved by the user operation so that the lower left corner of the third search die is located near the center of the camera. After the position is determined, the exact coordinates of the pattern of the third search die that matches the template image specified by the first search die by automatically performing pattern matching using the template image acquired in (1-1) above Get the value.

  In the present embodiment, the upper die adjacent to the second search die is described as an example of the third search die. However, it goes without saying that the third search die of the present invention is not limited to this. In short, it is only necessary to be able to grasp the positional relationship including the distance of the coordinates of the specific point of the die in the column direction with reference to the die that grasped the exact coordinates of the feature points. Therefore, the die adjacent to the upper side of the first search die can be preferably applied as an alternative.

(2) <Light microscope low magnification Y direction pattern matching>
(2-1) From the relationship between the pattern match coordinates (X2, Y2) of the second search die and the pattern match coordinates (X3, Y3) of the third search die (dX, dY) Is calculated.

dX = X3-X2
dY = Y3-Y2
(2-2) Using the calculated movement amount (dX, dY), move the stage to the coordinates (XN, YN) where the pattern of the adjacent die on the first search die exists (expected).

XN = X1 + dX
YN = Y1 + dY
* (X1, Y1): Coordinates of the pattern of the first search die (2-3) The pattern that is currently being observed by taking the image at low magnification after moving the stage and performing pattern matching using the template image Are obtained, and 1 is set as the initial value of the number of detected dies (DN).

(2-4) The amount of movement (dX, dY) from the pattern coordinates (X1, Y1) of the first search die to the coordinates (XN, YN) of the pattern currently being imaged is calculated.
dX = XN-X1
dY = YN-Y1
(2-5) The stage is moved starting from the first search die by a movement amount (2 * dX, 2 * dY) twice the calculated movement amount (dX, dY).

  (2-6) After moving the stage, the image is picked up at low magnification, and pattern matching is executed using the template image, thereby updating the exact coordinate values (XN, YN) of the currently observed pattern, The number of detected signals is doubled. See FIG. 74 for this.

(2-7) The steps (2-4) to (2-6) are repeatedly executed toward the upper part of the wafer until the Y coordinate value specified in advance is exceeded.
In the present embodiment, an example has been described in which a double movement amount is repeated in order to increase accuracy, reduce the number of times of processing (number of repetitions), and shorten the processing time. However, there is a problem in accuracy. If the processing time is to be further reduced, it may be executed at a high magnification of an integral multiple such as 2 or more, such as 3 or 4 times. Conversely, if there is no problem, the movement may be repeated with a fixed movement amount in order to further improve the accuracy. In any of these cases, it goes without saying that this is also reflected in the detected number.

(3) <Light microscope low magnification θ rotation>
(3-1) The amount of movement from the pattern coordinates (X1, Y1) of the first search die to the exact coordinate values (XN, YN) of the last searched die pattern and the number of dies detected so far ( DN) and the rotation amount (θ) and the Y-direction die size (YD) are calculated (FIG. 75).
reference).

dX = XN-X1
dY = YN-Y1
θ = tan ⁻¹ (dX / dY)
YD = sqrt ((dX) ² + (dY) ² ) / DN
* Sqrt (A) = √A
(3-2) The θ stage is rotated by the calculated rotation amount (θ).

B. Imaging with optical microscope high magnification (1) The same procedure as (1) for optical microscope low magnification is executed using an optical microscope high magnification image.
(2) The same procedure as in (2) for light microscope low magnification is executed using a light microscope high magnification image.

(3) The same procedure as (3) for light microscopic low magnification is executed.
(4) <Tolerance value check after optical microscope magnification θ rotation>
(4-1) [First search die, light magnifying power template specification]
The coordinates (X′1, Y′1) of the first search die after rotation are calculated from the coordinates (X1, Y1) before rotation and the rotation amount (θ), and the stage is moved to the coordinates (X′1, Y′1). After moving and locating, get a template image for pattern matching.

X′1 = x ₁ * cos θ−y ₁ * sin θ
Y′1 = x ₁ * sin θ + y ₁ * cos θ
(4-2) Light-magnification high-magnification Y-direction pattern matching The pattern currently being observed by moving in the Y direction by dY from the coordinates (X′1, Y′1) of the first search die after rotation and executing pattern matching The exact coordinate values (XN, YN) are obtained.

(4-3) A movement amount (dX, dY) from the coordinates (X′1, Y′1) of the first search die after rotation to the coordinates (XN, YN) of the pattern currently being imaged is calculated.

dX = XN−X′1
dY = YN−Y′1
(4-4) The stage is moved from the first search die as a starting point by a movement amount (2 * dX, 2 * dY) twice the calculated movement amount (dX, dY).

  (4-5) After moving the stage, the image is captured at optical magnification, and the pattern matching is executed using the template image, thereby updating the exact coordinate values (XN, YN) of the currently observed pattern.

(4-6) Steps (4-3) to (4-5) are repeatedly executed toward the upper portion of the wafer until the Y coordinate value designated in advance is exceeded.
(4-7) Calculate the rotation amount of θ Move from the coordinate (X′1, Y′1) of the first search die after rotation to the exact coordinate value (XN, YN) of the die pattern searched last The rotation amount (θ) is calculated using the amount.

dX = XN-X1
dY = YN-Y1
θ = tan ⁻¹ (dX / dY)
(4-8) Light microscopic magnification θ allowable value check It is confirmed that the rotation amount (θ) calculated in (4-7) is within a predetermined value or less. If not, the steps (4-1) to (4-8) are executed again after the θ stage is rotated using the calculated rotation amount (θ). However, if it does not fall within the allowable range even if (4-1) to (4-8) are repeatedly executed a specified number of times, the process is interrupted as an error.

C. EB image alignment (1) <Y search first die, EB template designation>
The same procedure as that in (1) for optical magnification is performed using the EB image.

(2) <EB Y-direction pattern matching>
A procedure similar to that in (2) of the optical magnification is executed using the EB image.
(3) <EB θ rotation>
A procedure similar to that in (3) of the optical magnification is performed using the EB image.

(4) <Allowable value check after EB θ rotation>
A procedure similar to that in (4) of the optical magnification is performed using the EB image.
(5) If necessary, execute (1) to (4) using a high-magnification EB image. (6) The coordinates (X1, Y1) of the first search die and the coordinates (X2, Y2) of the second search die. Calculate approximate value of die size (XD) in X direction from Y2) dX = X2-X1
dY = Y2-Y1
XD = sqrt ((dX) ² + (dY) ² )
* Sqrt (A) = √A
D. Die map recipe creation (1) <X search first die, EB template designation>
Move the stage by user operation so that the lower left corner of the die located at the left edge of the wafer is located near the center of the TDI camera, and after determining the position, obtain a template image for pattern matching. As the template image, an image that has a unique pattern within the search area must be selected.

(2) <EB X direction pattern matching>
(2-1) Using the approximate X-direction die size (XD), move the stage to the coordinates (X1 + XD, Y1) where the pattern of the die next to the X search first die exists (and is expected).

  (2-2) After moving the stage, an EB image is captured with a TDI camera, and pattern matching is executed using a template image to obtain exact coordinate values (XN, YN) of the currently observed pattern, Further, 1 is set as the initial value of the number of detected die (DN).

(2-3) X Search The movement amount (dX, dY) from the pattern coordinates (X1, Y1) of the first die to the coordinates (XN, YN) of the pattern currently being imaged is calculated.
dX = XN-X1
dY = YN-Y1
(2-4) The stage is moved from the X search first die as much as the movement amount (2 * dX, 2 * dY) twice the calculated movement amount (dX, dY). (2-5) Stage movement After that, an EB image is taken with a TDI camera, and pattern matching is executed using a template image, thereby updating the exact coordinate values (XN, YN) of the currently observed pattern and reducing the number of detected dies to 2 Double.

(2-6) Steps (2-3) to (2-5) are repeatedly executed in the right direction of the wafer until the X coordinate value designated in advance is exceeded.
(3) <Calculate X-direction tilt>
The amount of movement from the pattern coordinates (X1, Y1) of the first die of the X search to the exact coordinate values (XN, YN) of the die pattern searched last and the number of dies detected so far (DN) are used. The stage direct error (Φ) and the X-direction die size (XD) are calculated.

dX = XN-X1
dY = YN-Y1
Φ = tan ⁻¹ (dY / dX)
XD = sqrt ((dX) ² + (dY) ² ) / DN
* Sqrt (A) = √A
(4) <Die map creation>
Thus, the X direction die size (XD) is obtained, and the die map (ideal die arrangement information) is created together with the Y direction die size (YD) obtained when the rotation amount (θ) is calculated in advance. To do. The die map shows the ideal die placement. On the other hand, dies on different substrates are affected by mechanical errors of the stage (components such as guides and assembly errors), interferometer errors (for example, due to assembly problems such as mirrors) and image distortion due to charge-up. However, it is not always possible to observe the available arrangement, but it is necessary to grasp the error between the actual die position and the ideal arrangement on the die map, and take this error into account. Make inspections while making corrections.

E. Focus recipe creation procedure Next, a focus recipe creation procedure will be described. The focus recipe stores information on an optimum focus position at a position on a mark on a plane of a sample such as a substrate or various conditions related to the focus position in a predetermined format such as a table. In the focus map recipe, a focus condition is set only at a designated position on the wafer, and a focus value between the designated positions is linearly complemented (see FIG. 76). The focus recipe creation procedure is as follows.

(1) Select a focus measurement target die from the die map. (2) Set a focus measurement point in the die. (3) Move the stage to each measurement point, and based on the image and contrast value, focus value (CL12 Adjust the voltage manually.

  The die map created by the alignment process is ideal position information calculated from the die coordinates at both ends of the wafer, and an error occurs between the die position on the die map and the actual die position due to various factors (FIG. 77) The procedure for creating parameters for absorbing this error is called fine alignment, and the fine alignment recipe stores error information between the die map (ideal die placement information) and the actual die position. Is done. The information set here is used at the time of defect inspection. In the fine alignment recipe, the error is measured only for the die designated on the die map, and the error between the designated dies is linearly complemented.

F. Fine alignment procedure (1) Specify the error measurement target die for fine alignment from the die map. (2) Select the reference die from the error measurement target die, and set the position of this die as the point where the error from the die map is zero. 3) Capture the lower left corner of the reference die with a TDI camera and obtain a template image for pattern matching. * Select a unique pattern as a template image in the search area. (4) Lower left (die map) of neighboring error measurement die Acquire the coordinates (X0, Y0) above and move the stage. After the movement, an image is taken with a TDI camera, and pattern matching is executed using the template image of (3), thereby obtaining exact coordinate values (X, Y).

(5) Save the error between the coordinate value (X, Y) acquired by pattern matching and the coordinate value (X0, Y0) on the die map. (6) Execute (4) to (5) for all the error measurement target dies. To do.

2-6-3) Defect Inspection As shown in FIG. 78, the defect inspection is performed by setting the conditions of the electron optical system (setting the imaging magnification, etc.) and moving the stage while irradiating the electron beam. (FIG. 79) is performed, and in accordance with the set inspection conditions (array inspection conditions, random inspection conditions, inspection area), defect inspection is performed in real time by the inspection dedicated processing unit (IPE).

In the inspection recipe, the conditions of the electron optical system, the inspection target die, the inspection area, the inspection method (random / array), and the like are set (A and B in FIG. 80).
In addition, in order to acquire a stable image for defect inspection, EO correction that suppresses blurring of a captured image due to positional deviation, speed unevenness, and the like, and an error between an ideal die map arrangement and an actual die position are absorbed. Die position correction and focus adjustment that complements the focus value of the entire area of the wafer using focus values measured in advance at finite measurement points are simultaneously performed in real time.

  In the defect inspection scanning operation, in addition to inspecting the entire area of the inspection target die (FIG. 81), as shown in FIG. 82, thinning inspection can be performed by adjusting the amount of step movement in the direction perpendicular to the scanning direction. (Reduced inspection time).

  After the inspection is completed, the number of defects, the position of the die containing the defect, the defect size, the position of the defect in each die, the defect type, the defect image, and the comparison image are displayed on the display as the inspection result. It is possible to confirm and reproduce past test results by saving to a file.

  By selecting and specifying various recipes during automatic defect inspection, the wafer is loaded according to the transfer recipe, the wafer is aligned on the stage according to the alignment recipe, focus conditions are set according to the focus map recipe, and inspection is performed according to the inspection recipe. And the wafer is unloaded according to the transfer recipe (A and B in FIG. 83).

2-6-4) Control System Configuration This apparatus is configured by a plurality of controllers as shown in FIG. The main controller controls the GUI unit / sequence operation of the device (EBI), receives operation commands from the factory host computer or GUI, and gives necessary instructions to the VME controller and IPE controller. The VME controller manages the operation of the equipment (EBI) components and gives instructions to the stage controller and the PLC controller according to instructions from the main controller. The IPE controller acquires defect inspection information from the IPE node computer, classifies the acquired defects, and displays an image according to an instruction from the main controller. The IPE node computer acquires an image output from the TDI camera and performs defect inspection.

  In response to an instruction from the VME controller, the PLC controller drives a device such as a valve, acquires sensor information, and performs abnormality monitoring such as a vacuum degree abnormality that requires constant monitoring. Upon receiving an instruction from the VME controller, the stage controller moves in the XY directions and rotates the wafer installed on the stage.

By configuring such a distributed control system, it is not necessary to change the software and hardware of the host controller by keeping the interface between the controllers the same when the terminal device configuration device is changed. Further, even when the sequence operation is added / modified, it is possible to flexibly cope with a configuration change by minimizing changes in the upper software and hardware.

2-6-5) User Interface Configuration FIG. 85 shows a device configuration of the user interface unit.
(1) Input unit A device that accepts input from a user, and includes a “keyboard”, “mouse”, and “joy pad”.

(2) Display unit A device that displays information to the user, and consists of two monitors.
Monitor 1: Display acquired image with CCD camera or TDI camera Monitor 2: GUI display Coordinate system In this device, the following three coordinate systems are defined.

(1) Stage coordinate system [X _S , Y _S ]
Reference coordinate system for position indication during stage position control With the lower left corner of the chamber as the origin, the X coordinate value increases in the right direction and the Y coordinate value increases in the upward direction.

There is only one coordinate system in the apparatus.
The position (coordinate value) indicated in the stage coordinate system is the center of the stage (wafer center).

That is, when the coordinate value [0, 0] is specified in the stage coordinate system, the stage center (wafer center) moves so as to overlap the origin of the stage coordinate system.
The unit is [μm], but the minimum resolution is λ / 1024 (≈0.618 [μm]).

* Λ: Wavelength of the laser used in the laser interferometer (λ ≒ 632.991 [μm])
(2). Wafer coordinate system [X _W , Y _W ]
Reference coordinates for designating the observation (imaging / display) position on the wafer The X coordinate value increases in the right direction and the Y coordinate value increases in the upward direction with the wafer center as the origin.

The position (coordinate value) indicated in the wafer coordinate system is the imaging center of the imaging device (CCD camera, TDI camera) selected at that time.
There is only one coordinate system in the apparatus.

The unit is [μm], but the minimum resolution is λ / 1024 (≈0.618 [μm]).
* Λ: Wavelength of the laser used in the laser interferometer (λ ≒ 632.991 [μm])
(3). Die coordinate system [X _D , Y _D ]
Reference coordinates for defining the observation (imaging / display) position in each die The X coordinate value increases in the right direction and the Y coordinate value increases in the upper direction with the lower left corner of each die as the origin. This coordinate system exists for each die. The unit is [μm], but the minimum resolution is λ / 1024 (≈≈0.618 [μm]).

* Λ: Wavelength of the laser used in the laser interferometer (λ ≒ 632.991 [μm])
The dies on the wafer are numbered (numbered), and the die that serves as a reference for numbering is called the origin die. By default, the die closest to the wafer coordinate system origin is set as the origin die, but the position of the origin die can be selected by the user's specification.

  The relationship between the coordinate values in each coordinate system and the observed (displayed) position is as shown in FIG. * The relationship between the coordinates specified by the user interface and the stage movement direction is as follows.

(1) Joystick & GUI arrow button The direction indicated by the joystick and GUI arrow button is regarded as the direction that the operator wants to see, and the stage is moved in the direction opposite to the indicated direction.
Direction: Right . . . Stage movement direction: Left (image moves to the left = field of view moves to the right)
Direction: Up ... Stage movement direction: Down (Image moves down = Field of view moves up)
(2) Directly inputting coordinates on the GUI The coordinates directly input on the GUI are regarded as a place the operator wants to see on the wafer coordinate system, and the stage is moved so that the corresponding wafer coordinates are displayed at the center of the captured image. .

2-7) Description of Other Functions and Configuration FIG. 87 shows an overall configuration diagram of the present embodiment. However, a part of the configuration is omitted for illustration. In the figure, the inspection apparatus has a primary column 87. 1, a secondary column 87 • 2 and a chamber 87 • 3. An electron gun 87.4 is provided inside the primary column 871, and a primary optical system 87.5 is disposed on the optical axis of the electron beam (primary beam) irradiated from the electron gun 87-4. The Also, stages 87 and 6 are installed inside the chambers 87 and 3, and the sample W is placed on the stages 87 and 6.

  On the other hand, in the secondary column 87. 2, on the optical axis of the secondary beam generated from the sample W, an objective lens 87 · 7, a numerical aperture 87 · 8, a Wien filter 87 · 9, a second lens 87. 10. Field apertures 87 and 11, third lenses 87 and 12, fourth lenses 87 and 13, and detectors 87 and 14 are arranged. The numerical apertures 87 and 12 correspond to aperture stops, and are thin plates made of metal (such as Mo) with circular holes. The aperture is arranged so as to be the primary beam focusing position and the focal position of the objective lens 87. Therefore, the objective lenses 87 and 7 and the numerical apertures 87 and 8 constitute a telecentric electron optical system.

  On the other hand, the outputs of the detectors 87 and 14 are input to the control units 87 and 15, and the outputs of the control units 87 and 15 are input to the CPUs 87 and 16. The control signals of the CPUs 87 and 16 are input to the primary column control units 87 and 17, the secondary column control units 87 and 18, and the stage drive mechanisms 87 and 19. The primary column control units 87 and 17 perform lens voltage control of the primary optical system 87 and 5, and the secondary column control units 87 and 18 include the objective lenses 87 and 7, the second lenses 87 and 10 to the fourth lenses 87 and 13 lens voltage control and electromagnetic field control applied to the Wien filters 87 and 9 are performed.

  The stage drive mechanisms 87 and 19 transmit stage position information to the CPUs 87 and 16. Further, the primary column 87. 1, the secondary column 87 2, and the chamber 87 3 are connected to an evacuation system (not shown), and are evacuated by a vacuum molecular pump of the evacuation system so that the inside is in a vacuum state. Is maintained.

(Primary beam) The primary beam from the electron gun 87.4 is incident on the Wien filter 87.9 while receiving the lens action by the primary optical system 87.5. Here, LaB ₆ that can extract a large current with a rectangular cathode is used as the tip of the electron gun. The primary optical system 72 uses a rotational axis asymmetric quadrupole or octupole electrostatic (or electromagnetic) lens. This can cause convergence and divergence in the X-axis and Y-axis, respectively, like a so-called cylindrical lens. This lens is composed of 2, 3, or 4 stages, and by optimizing each lens condition, the beam irradiation area on the sample surface can be arbitrarily rectangular or elliptical without losing irradiation electrons. Can be molded.

  Specifically, when an electrostatic quadrupole lens is used, four cylindrical rods are arranged around the optical axis. Opposing electrodes are equipotential, and a reverse voltage characteristic is given at a phase shifted by 90 degrees around the optical axis.

  In addition, you may use the lens of the shape which divided the circular plate normally used as an electrostatic deflector into 4 parts instead of a cylindrical shape as a quadrupole lens. In this case, the lens can be reduced in size. The trajectory of the primary beam that has passed through the primary optical system 72 is bent by the deflection action of the Wien filters 87 and 9. The Wien filters 87 and 9 make the electric field perpendicular to the electric field, and if the electric field is E, the magnetic field is B, and the velocity of the charged particle is v, only charged particles satisfying the Wien condition of E = vB go straight. Bending the trajectory of charged particles. For the primary beam, a force FB caused by a magnetic field and a force FE caused by an electric field are generated, and the beam trajectory is bent. On the other hand, since the force FB and the force FE work in opposite directions with respect to the secondary beam, they cancel each other, so the secondary beam goes straight.

  The lens voltage of the primary optical system 87/5 is set in advance so that the primary beam forms an image at the opening of the numerical aperture 87/8. The numerical apertures 87 and 8 prevent an extra electron beam scattered in the apparatus from reaching the sample surface and prevent the sample W from being charged up or contaminated. Further, since the numerical apertures 87 and 8 and the objective lenses 87 and 7 constitute a telecentric electron optical system, the primary beam transmitted through the objective lenses 87 and 7 becomes a parallel beam and is uniform and uniform on the sample W. Irradiate like That is, Koehler illumination referred to as an optical microscope is realized.

(Secondary beam) When the sample is irradiated with the primary beam, secondary electrons, reflected electrons, or backscattered electrons are generated as secondary particles from the beam irradiation surface of the sample.
The secondary particles pass through the lens while receiving the lens action by the objective lenses 87 and 7. By the way, the objective lenses 87 and 7 are composed of three electrodes. The bottom electrode is designed to form a positive electric field with the potential on the sample W side, draw electrons (especially secondary electrons with small directivity), and efficiently guide them into the lens. Yes. The lens action is performed by applying a voltage to the first and second electrodes of the objective lenses 87 and 7 to bring the third electrode to zero potential. On the other hand, the numerical apertures 87 and 8 are arranged at the focal position of the objective lenses 87 and 7, that is, the back focus position from the sample W. Therefore, the electron beam emitted from the center of the field of view (off-axis) also becomes a parallel beam and passes through the center position of the numerical apertures 87 and 8 without being distorted.

  The numerical apertures 87 and 8 serve to suppress lens aberrations of the second lenses 87 and 10 to the fourth lenses 87 and 13 with respect to the secondary beam. The secondary beam that has passed through the numerical apertures 87 and 8 passes straight as it is without being subjected to the deflection action of the Wien filters 87 and 9. Note that by changing the electromagnetic field applied to the Wien filters 87 and 9, only the electrons having a specific energy (for example, secondary electrons, reflected electrons, or backscattered electrons) from the secondary beam are detected by the detectors 87 and 14. Can lead to.

  When secondary particles are imaged only by the objective lenses 87 and 7, the lens action becomes strong and aberrations are likely to occur. Therefore, one image formation is performed together with the second lenses 87 and 10. The secondary particles obtain intermediate images on the field apertures 87 and 11 by the objective lenses 87 and 7 and the second lenses 87 and 10, respectively. In this case, the magnification magnification necessary for the secondary optical system is usually insufficient. Therefore, the third lens 87 12 and the fourth lens 87 13 are added as lenses for enlarging the intermediate image. To do. The secondary particles are enlarged and imaged by the third lenses 87 and 12 and the fourth lenses 87 and 13, respectively, and here, the secondary particles are imaged three times in total. The third lens 87 and 12 and the fourth lens 87 and 13 may be combined and imaged once (total of two times).

  The second lenses 87 and 10 to the fourth lenses 87 and 13 are all rotationally symmetrical lenses called unipotential lenses or Einzel lenses. Each lens has a configuration of three electrodes. Usually, the outer two electrodes are set to zero potential, and the lens action is performed with a voltage applied to the center electrode. In addition, field apertures 87 and 11 are arranged at the intermediate image forming point. Similarly to the field stop of the optical microscope, the field apertures 87 and 11 limit the field of view to the necessary range. However, in the case of an electron beam, the extraneous beams are converted into the third lens 87 and 12 and the fourth lens 87 13 to prevent charge-up and contamination of the detectors 87 and 14. The enlargement magnification is set by changing the lens condition (focal length) of the third lens 87 12 and the fourth lens 87 13.

  The secondary particles are enlarged and projected by the secondary optical system and imaged on the detection surfaces of the detectors 87 and 14. The detectors 87 and 14 include an MCP that amplifies electrons, a fluorescent plate that converts electrons into light, a relay between the vacuum system and the outside, a lens and other optical elements for transmitting an optical image, and an image sensor (CCD). Etc.). The secondary particles are imaged and amplified on the MCP detection surface, and the electrons are converted into optical signals by the fluorescent screen, and converted into photoelectric signals by the imaging device.

  The control units 87 and 15 read the image signals of the sample from the detectors 87 and 14 and transmit them to the CPUs 87 and 16. The CPUs 87 and 16 perform pattern defect inspection from the image signal by template matching or the like. The stages 87 and 6 can be moved in the XY directions by stage drive mechanisms 87 and 19. The CPUs 87 and 16 read the positions of the stages 87 and 6, output drive control signals to the stage drive mechanisms 87 and 19, drive the stages 87 and 6, and sequentially detect and inspect images.

  As described above, in the inspection apparatus according to the present embodiment, the numerical apertures 87 and 8 and the objective lenses 87 and 7 form a telecentric electron optical system. Can be irradiated uniformly. That is, Kohler illumination can be easily realized.

  Further, for the secondary particles, all the chief rays from the sample W are incident on the objective lenses 87 and 7 perpendicularly (parallel to the optical axis of the lens) and pass through the numerical apertures 87 and 8, so that There is no light, and the image brightness around the sample is not lowered. In addition, the so-called lateral chromatic aberration, in which the image forming position is different depending on the energy variation of the electrons (particularly, the secondary electron has a large lateral chromatic aberration due to the large energy variation). By arranging the numerical apertures 87 and 8 at the focal position, this lateral chromatic aberration can be suppressed.

Further, since the enlargement magnification is changed after passing through the numerical apertures 87 and 8, the field of view on the detection side can be changed even if the set magnifications of the lens conditions of the third lens 87 and 10 and the fourth lenses 87 and 13 are changed. A uniform image can be obtained on the entire surface. In the present embodiment, a uniform image without unevenness can be obtained. However, usually, when the enlargement magnification is increased, the brightness of the image is lowered. Therefore, in order to improve this, when changing the magnification condition by changing the lens conditions of the secondary optical system, the effective field of view on the sample surface determined accordingly, and the electron beam irradiated on the sample surface, The lens conditions of the primary optical system are set so that they have the same size.

  In other words, if the magnification is increased, the field of view is narrowed accordingly, but at the same time, by increasing the irradiation density of the electron beam, the signal density of the detected electrons is increased even when projected by the secondary optical system. It is always kept constant and the brightness of the image does not decrease.

  In the inspection apparatus according to the present embodiment, the Wien filter 87 or 9 that bends the trajectory of the primary beam and advances the secondary beam straight is used. However, the present invention is not limited thereto, and the trajectory of the primary beam goes straight. An inspection apparatus using a Wien filter that bends the beam trajectory may be used. Here, E × B is used, but only a magnetic field may be used. At this time, for example, the primary electron incident direction and the direction toward the signal electron detector may be the same and may have a Y-shaped configuration.

  In this embodiment, a rectangular beam is formed from a rectangular cathode and a quadrupole lens. However, the present invention is not limited to this. For example, a rectangular beam or an elliptical beam may be created from a circular beam. The rectangular beam may be taken out through. Further, a linear beam or a plurality of beams may be used, and these may be used by scanning.

2-7-1) Control Electrode Between the objective lenses 87 and 7 and the wafer W, there is an electrode (25 and 8 in FIG. 25-1) having a shape substantially axisymmetric with respect to the irradiation optical axis of the electron beam. Has been placed. Examples of the shape of this electrode are shown in FIGS. 88 and FIG. 89 are perspective views of the electrodes 88. 1 and 89. 1, FIG. 88 is a perspective view showing a case where the electrodes 88. 1 have an axially symmetric cylindrical shape, and FIG. -It is a perspective view which shows the case where 1 is an axisymmetric disk shape.

  In this embodiment, the electrode 881 is described as a cylindrical shape as shown in FIG. 88. However, if it is substantially axially symmetric with respect to the irradiation optical axis of the electron beam, a disk-shaped electrode as shown in FIG. It may be 89.1. Further, in order to generate an electric field for preventing discharge between the objective lenses 87 and 7 (25 and 7 in FIG. 25-1) and the wafer W, the voltage applied to the wafer W (this Since it is grounded in the embodiment, a predetermined voltage (negative potential) lower than 0 V) is applied by the power supplies 25 and 9. The potential distribution between the wafer W and the objective lenses 97 and 7 at this time will be described with reference to FIG.

  FIG. 90 is a graph showing a voltage distribution between the wafer W and the objective lenses 87 and 7. In the figure, the voltage distribution from the wafer W to the positions of the objective lenses 87 and 7 is shown with the position on the irradiation optical axis of the electron beam as the horizontal axis. In the conventional electron beam apparatus without the electrode 88. 1, the voltage distribution from the objective lens 87 · 7 to the wafer W is the maximum value of the voltage applied to the objective lens 87 · 7, and the wafer W is grounded. It is changing gently. (Narrow line in FIG. 90) On the other hand, in the electron beam apparatus of the present embodiment, the electrode 88 • 1 is disposed between the objective lens 87 • 7 and the wafer W, and the electrode 88 • 1 is connected to the wafer W. Since a predetermined voltage (negative potential) lower than the applied voltage is applied by the power supplies 25 and 9, the electric field of the wafer W is weakened (thick line in FIG. 90). Therefore, in the electron beam apparatus according to the present embodiment, the electric field is not concentrated in the vicinity of the vias 25 and 13 (FIG. 25-1) in the wafer W, so that a high electric field does not occur. Even if the vias 25 and 13 are irradiated with the electron beam and the secondary electrons are emitted, the emitted secondary electrons are not accelerated to the extent that the residual gas is ionized. It is possible to prevent the electric discharge occurring between W.

In addition, since the discharge between the objective lenses 87 and 7 and the vias 25 and 13 (FIG. 25-1) can be prevented, the pattern of the wafer W or the like is not damaged by discharge. In the above-described embodiment, discharge between the objective lenses 87 and 7 and the wafer W having the vias 25 and 13 can be prevented. However, since a negative potential is applied to the electrodes 88 and 1, a negative potential is applied. Depending on the size, the detection sensitivity of secondary electrons by the detectors 87 and 14 may decrease. Therefore, when the detection sensitivity is lowered, as described above, a series of operations of irradiating an electron beam and detecting secondary electrons are performed a plurality of times, and the obtained detection results are cumulatively added, averaged, etc. To obtain a predetermined detection sensitivity (signal S / N ratio). In this embodiment, as an example, the detection sensitivity is described as a signal-to-noise ratio (S / N ratio).

  Here, the secondary electron detection operation will be described with reference to FIG. This figure is a flowchart showing the secondary electron detection operation of the electron beam apparatus. First, secondary electrons from the sample to be inspected are detected by the detectors 87 and 14 (step 91.1). Next, it is determined whether the signal-to-noise ratio (S / N ratio) is equal to or greater than a predetermined value (steps 9 and 2). If the signal-to-noise ratio is greater than or equal to the predetermined value in step 91.2, detection of secondary electrons by detectors 87 and 14 is sufficient, and the secondary electron detection operation is completed.

  On the other hand, if the signal-to-noise ratio is less than the predetermined value in step 91.2, a series of operations of irradiating an electron beam and detecting secondary electrons is performed 4N times and an averaging process is performed (steps 91.3). ). Here, since the initial value of N is set to “1”, the detection operation of secondary electrons is performed four times for the first time in Steps 9 and 3.

  Next, “1” is added to N and counted up (steps 9 and 4), and it is determined again in step 9 and 2 whether the signal-to-noise ratio is equal to or greater than a predetermined value. Here, if the signal-to-noise ratio is less than the predetermined value, the process proceeds to step 91.3 again, and this time the secondary electron detection operation is performed eight times. Then, N is counted up, and Steps 9 · 2 to 9 · 4 are repeated until the signal-to-noise ratio becomes a predetermined value or more.

  In the present embodiment, the discharge prevention for the wafer W having the vias 25 and 13 has been described by applying a predetermined voltage (negative potential) lower than the applied voltage to the wafer W to the electrodes 88. The detection efficiency of secondary electrons may decrease. Therefore, when the sample to be inspected is a type of sample to be inspected that is unlikely to generate a discharge between the objective lens 87 and 7 such as a wafer without vias, the detection efficiency of the secondary electrons in the detectors 87 and 14 is high. In this way, the voltage applied to the electrode 88 • 1 can be controlled.

  Specifically, even when the sample to be tested is grounded, the voltage applied to the electrode 881 is set to a predetermined voltage higher than the voltage applied to the sample to be tested, for example, + 10V. . At this time, the distance between the electrode 88 • 1 and the sample to be inspected is set such that no discharge occurs between the electrode 88 • 1 and the sample to be inspected.

  In this case, the secondary electrons generated by the electron beam irradiation to the sample to be inspected are accelerated toward the detectors 87 and 14 by the electric field generated by the voltage applied to the electrodes 88 and 1. The electric field generated by the voltage applied to the objective lenses 87 and 7 is further accelerated toward the detectors 87 and 14 and is subjected to a convergence action, so that many secondary electrons are incident on the detectors 87 and 14. Detection efficiency can be increased.

  Furthermore, since the electrode 88 • 1 is axially symmetric, it also has a lens function for converging the electron beam applied to the sample to be inspected. Therefore, the primary electron beam can be narrowed more narrowly by the voltage applied to the electrode 88. In addition, since the primary electron beam can be narrowed down by the electrode 88. 1, a lower aberration objective lens system can be configured by combining with the objective lenses 87 and 7. To the extent that such lens action is possible, the electrodes 88.

  According to the electron beam apparatus of the above-described embodiment, the electron beam irradiation surface of the sample to be inspected is substantially axisymmetric with respect to the electron beam irradiation axis between the sample to be inspected and the objective lens. Since the electrode for controlling the electric field strength at is provided, the electric field between the sample to be inspected and the objective lens can be controlled.

  Between the specimen to be inspected and the objective lens, it has a shape that is substantially axisymmetric with respect to the irradiation axis of the electron beam, and includes an electrode that weakens the electric field intensity on the irradiation surface of the electron beam of the specimen to be inspected. It is possible to eliminate the discharge between the sample to be inspected and the objective lens. In addition, since the voltage applied to the objective lens is not changed or the like is not changed, secondary electrons can be efficiently passed through the objective lens, so that detection efficiency can be improved and a signal with a good S / N ratio can be obtained. it can.

  Depending on the type of the sample to be inspected, it is possible to control a voltage for weakening the electric field strength on the electron beam irradiation surface of the sample to be inspected. For example, when the sample to be inspected is a type of sample to be inspected that easily discharges between the objective lens, the voltage of the electrode is changed to further weaken the electric field intensity on the electron beam irradiation surface of the sample to be inspected. Thus, discharge can be prevented.

  Depending on the presence or absence of vias in the semiconductor wafer, the voltage applied to the electrodes can be changed, that is, the voltage for reducing the electric field intensity on the electron beam irradiation surface of the semiconductor wafer can be changed. For example, when the sample to be inspected is a type of sample to be inspected that easily discharges between the objective lens, the electric field intensity at the electron beam irradiation surface of the sample to be inspected is made weaker by changing the electric field by the electrode. Thus, it is possible to prevent discharge particularly in the via and the periphery of the via. In addition, since discharge between the via and the objective lens can be prevented, the pattern of the semiconductor wafer or the like is not damaged by discharge. In addition, since the potential applied to the electrode is lower than the charge applied to the sample to be inspected, the electric field strength on the electron beam irradiation surface of the sample to be inspected can be weakened, and discharge to the sample to be inspected can be prevented. Since the potential applied to the electrode is a negative potential and the sample to be inspected is grounded, the electric field intensity on the surface irradiated with the electron beam of the sample to be inspected can be weakened and discharge to the sample to be inspected can be prevented.

  So far, the use of the control electrode mainly for the purpose of preventing discharge has been described, but the control electrode can be used for energy selection of secondary electrons emitted from the wafer. That is, in order to obtain an image with a high resolution, when only secondary electrons having an energy of a certain level or more with the highest signal detection efficiency are detected, a predetermined negative voltage is applied to the control electrode. It can be applied and used as an energy barrier for secondary electrons. Since a negative potential is applied to the control electrode, a force that repels secondary electrons toward the sample works. The secondary electrons that cannot cross the potential barrier return to the sample, and only the secondary electrons that exceed the potential barrier are detected by the detector, and an image with a desired resolution can be obtained.

2-7-2) Potential Application Method In FIG. 92, the potential application mechanism 92.1 is based on the fact that the secondary electron information (secondary electron generation rate) emitted from the wafer depends on the potential of the wafer. The generation of secondary electrons is controlled by applying a potential of ± several V to the stage mounting table on which the wafer is placed. In addition, this potential application mechanism also serves to reduce the energy initially possessed by the irradiated electrons so that the irradiated electron energy is about 100 to 500 eV on the wafer.

As shown in FIG. 92, the potential application mechanism 92.1 includes a voltage application device 92.4 electrically connected to the mounting surface 92.3 of the stage device 92.2, a charge-up investigation and voltage determination system. (Hereinafter referred to as survey and decision system) 92.5. The investigation and determination system 92.5 is connected to the monitor 92-7 electrically connected to the image forming unit 92-6 of the detection system of the electro-optical device 13, 8 (FIG. 13), and to the monitor 92.7. Operators 92 and 8 and CPUs 92 and 9 connected to the operators 92 and 84 are provided. CPU 92.9 is
A signal is supplied to the voltage applying device 92.4.

The potential application mechanism is designed to search for a potential at which a wafer to be inspected is difficult to be charged and apply the potential.
As a method for inspecting an electrical defect of an inspection sample, it is also possible to use the fact that the voltage of the part is different between the part that is originally electrically insulated and the part that is energized.

  First, by applying a charge to the sample in advance, the voltage of the part that is inherently electrically isolated and the part that is inherently electrically insulated, but the part that is energized for some reason By generating a voltage difference with the voltage and then irradiating the beam of the present invention, data having a voltage difference is acquired, and the acquired data is analyzed to detect that the current state is energized.

2-7-3) Electron Beam Calibration Method In FIG. 93, the electron beam calibration mechanism 93.1 is installed at a plurality of locations on the side of the wafer mounting surface 93/3 on the rotary table 93-2. And a plurality of Faraday cups 9/4 and 93.5 for measuring beam current. The Faraday cup 93.4 is for a thin beam (about φ2 μm), and the Faraday cup 93.5 is for a thick beam (about φ30 μm). In the Faraday cup 93.4 for thin beams, the beam profile is measured by step-feeding the rotary table 93.2. The Faraday cup 93.5 for a thick beam measures the total current amount of the beam. The Faraday cups 93, 4, 93, 5 are arranged so that the upper surface is at the same level as the upper surface of the wafer W placed on the placement surface 93, 3. In this way, the primary electron beam emitted from the electron gun is constantly monitored. This is because the electron gun cannot always emit a constant electron beam, and the amount of emission changes during use.
2-7-4) Cleaning of electrode When the electron beam apparatus of the present invention is operated, the target substance is floated and attracted to the high-pressure region by proximity interaction (charge of particles near the surface). Organic materials are deposited on the various electrodes used for deflection. Insulators that gradually accumulate due to surface charging adversely affect the formation and deflection mechanism of the electron beam, so the deposited insulators must be removed periodically. Periodic removal of the insulator is performed by using an electrode in the vicinity of the region where the insulator is deposited, such as hydrogen, oxygen, fluorine, and a compound containing them such as HF, O ₂ , H ₂ O, and C _M F _{N in} vacuum. By generating plasma and maintaining the plasma potential in the space at a potential (several kV, for example, 20 V to 5 kV) at which sputtering occurs on the electrode surface, only organic substances are removed by oxidation, hydrogenation, and fluorination. In addition, pollutants on the surface of the electrode and the insulator can be removed by passing expectations that have a cleaning effect.

2-7-5) Alignment Control Method An alignment control device 94. 1 shown in FIG. 94 is a device for positioning the wafer W with respect to the electro-optical device 94 2 using a stage device. Approximate alignment by wide-field observation using 3 (measurement with a lower magnification than with the electron optical system), high magnification alignment using the electron optical system of the electron optical device 94-2, focus adjustment, inspection area setting, pattern alignment, etc. Control is to be performed. Inspecting the wafer at a low magnification using the optical system in this way is to automatically inspect the wafer pattern by observing the wafer pattern with a narrow field of view using an electron beam. This is because it is necessary to easily detect the alignment mark with an electron beam when performing.

The optical microscope 94. 3 is provided in the housing (may be provided so as to be movable in the housing), and a light source for operating the optical microscope is also provided in the housing (not shown). An electron optical system that performs high-magnification observation shares the electron optical system (primary optical system and secondary optical system) of the electron optical device 94-2. A schematic diagram of the configuration is as shown in FIG. In order to observe the observation point on the wafer at a low magnification, the observation point on the wafer is moved into the field of view of the optical microscope by moving the X stage of the stage apparatus in the X direction. The optical microscope 94.3 visually recognizes the wafer with a wide field of view, and the position to be observed on the wafer is displayed on the monitor 94.5 via the CCD 94.4, so that the observation position is roughly determined. In this case, the magnification of the optical microscope may be changed from a low magnification to a high magnification.

Next, the stage device is moved by a distance corresponding to the distance δx between the optical axis of the electro-optical device 94. 2 and the optical axis of the optical microscope 94. Move to the field of view of the optical device. In this case, the distance between the axis O ₃ -O _{3 of} the electron optical device and the optical axis O ₄ -O ₄ of the optical microscope 94 3 (in this embodiment, both are displaced only in the direction along the X axis). (Although it may be displaced in the Y-axis direction and the Y-axis direction.) Δx is known in advance, so that the observed point can be moved to the visual recognition position by moving the value δx. it can. After the movement of the observation point to the visual recognition position of the electron optical device is completed, the electron optical system performs SEM imaging of the observation point at a high magnification and stores the image, or the image is stored in the monitor 94 or 7 via the CCD 94 or 6. Display.

Thus, after displaying the observation point of the wafer on the monitor at a high magnification by the electron optical system, the positional deviation in the rotation direction of the wafer with respect to the rotation center of the rotary table of the stage apparatus is detected by a known method. A deviation δθ in the rotation direction of the wafer with respect to the optical axis O ₃ -O ₃ is detected, and a positional deviation in the X axis and Y axis directions of a predetermined pattern related to the electro-optical device is detected. Then, the wafer alignment is performed by controlling the operation of the stage devices 94 and 8 based on the detected value and the data of the inspection mark provided on the wafer obtained separately or the data on the shape of the pattern of the wafer. The alignment range is within ± 10 pixels in the XY coordinates. Preferably it is within ± 5 pixels, more preferably within ± 2 pixels.

2-7-6) EO correction Outline In order to image a beam from a wafer with TDI, the position of the wafer needs to be accurately positioned. However, since the wafer is actually on the XY stage, it is mechanically positioned. The accuracy is several hundred μ to several tens of nm, and the response speed is several seconds to several ms.

  On the other hand, the design rule is refined toward several tens of nanometers. Therefore, a wiring having a line width of several tens of nanometers and a via having a diameter of several tens of nanometers are inspected to detect shape defects and electrical defects and to have a diameter of several ten nanometers. Trash detection is required. Taking an image relying only on the mechanical positioning is a significant obstacle to obtaining an accurate image because the order of response time and positioning accuracy is far from the order of the design rule and imaging accuracy.

  The imaging sequence is executed by a combination of steps (x-axis) and constant speed scan (y-axis), and relatively dynamic control (y-axis) generally has a large control residual to prevent image blurring. From the meaning, more advanced control is required.

  In view of these items, it is a matter of course that the XY stage has high accuracy and excellent responsiveness. In addition, in order to realize beam control accuracy and speed for the imaging unit that cannot be achieved by the stage, EO correction is performed. It has the function of.

The basic method is that the position of the wafer on the stage is accurately determined within a time delay of several microseconds in the order of sub-nm by using a laser interferometer system and a bar mirror installed on the xy axis. It recognizes and drives a mechanical actuator by an automatic control loop, and is positioned with a time delay and a residual in the target position. The control residual as a result of positioning by this control is obtained by the difference between the target position generated inside the control device and the current position obtained by the laser interferometer system. On the other hand, after passing through a number of electrodes, the beam is guided to the imaging device via the correction deflection electrode. The correction deflection electrode has a sensitivity capable of deflection of about several hundred μm or less, preferably one hundred μm or less, more preferably several tens of μm or less in terms of the distance on the wafer, and a voltage is applied thereto. Thus, it is possible to deflect the beam to an arbitrary position two-dimensionally. After the control residual is calculated by the arithmetic unit, the control residual is converted into a voltage by the D / A converter, and the correction deflection electrode is applied in a direction to cancel the residual. With the above configuration, correction close to the resolution of the laser interferometer can be performed.

  As another method, the above-described means is used for the X axis (step direction) and the transfer clock of the TDI which is the image sensor is transferred for the Y axis (scan direction) in synchronization with the moving speed of the stage.

  FIG. 95 shows the concept of EO correction. An instruction 95.1 to the target position is output and applied to a control feedback loop 95.2 including a machine actuator. This part corresponds to the stage. The result of driving and position displacement is fed back by the position detector 95.3, and the position displacement of the drive system converges to the target position from the position indication, but because the gain of the control system is finite, A residual occurs. The current position is detected in the sub-nm order by a position output system 95.4 (in this case using a laser interferometer), the difference from the position pointing device 95/1 is detected by a residual detector 95/5, and high pressure and high speed The amplifier 95.6 is applied to the deflecting electrode 95.7, and a voltage is applied in a direction that cancels the residual. In the absence of this function, the fluctuation generated as 95.8 is 95.9. It has the function to reduce like.

  FIG. 96 presents a specific device configuration. The XY stage 96.1 realizes smooth servo characteristics by driving the X-axis and detecting the rough position and speed by the servo motor 96.2 for driving the X-axis and the encoder 96.3. In this example, a servo motor is used, but a similar configuration is possible in an actuator such as a linear motor or an ultrasonic motor. 96.6. Is a power amplifier for driving the motor. Precise position information on the X-axis realizes a position detection function with sub-nm resolution by combining mirrors 96 and 7, interferometers 96 and 8, receivers 96 and 9, laser light sources 96 and 10, and interferometer boards 96 and 11. doing.

The Y-axis has the same function as the orthogonal X-axis, and is composed of servo motors 96 and 12, amplifiers 96 and 13, mirrors 96 and 14, interference types 9 and 5, and receivers 96 and 16.
The XY stage controllers 96 and 17 control these devices in an integrated manner to enable two-dimensional operation of the stage, and have an accuracy of 1000 μm to 1 nm, preferably 100 μm to 2 nm, more preferably 1 μm. An accuracy of ˜2 nm, more preferably 0.1 μm to 2 nm is achieved, and a response speed of several thousand ms or less, preferably several tens ms or less, more preferably several ms or less is achieved. On the other hand, the X reference value and the Y reference value are output from the XY stage controllers 96 and 17 to the EO corrector 96 and 18, and the position information output in the 32-bit binary format from the interference type 96 and 11 is output at high speed. The EO corrector 96/18 receives the current position via the buffer board 96/19. After performing the calculation inside, the voltage is amplified by the high-voltage high-speed amplifiers 96, 20, 96, 21 and then applied to the deflection electrodes 96, 22, and the deflection is performed to correct the residual difference, thereby minimizing the positional deviation. The information electron beam is guided to TDI (imaging device) 96/23. As will be described later, 96 and 24 are portions for generating timing signals for determining the transfer speed of the TDI 96 and 23.

Next, the function for generating the target position in the scanning direction in this apparatus will be described. The EO correction is a function for obtaining a difference between the target position and the actual position and correcting the position by deflecting the electron beam so as to cancel the difference, but the correction range is limited to a range of approximately several tens of μm. This is determined by the electrode sensitivity, the dynamic range of the high-voltage high-speed amplifier, the noise level, the number of bits of the D / A converter, and the like. However, the actual position of the stage at the time of scanning causes a significant deviation from the target position as compared with that at the time of stopping due to the finite gain of the control loop. When traveling at 20 mm / s, the deviation from the target position is about 400 μm, and even if the difference is calculated and output as it is, the correction range is greatly exceeded and the system is saturated.

In order to prevent this phenomenon, this apparatus uses the following means to avoid this problem. FIG. 97 illustrates this concept.
97.1 is the target position of the stage, and since it is a constant speed movement during scanning, it increases linearly with time. On the other hand, the mechanical position 97.2 of the actual controlled result stage includes a mechanical vibration of several microns and has a steady deviation 97.3 of about 400 μm. As a means for removing this steady-state deviation, it is conceivable to smooth the position information during actual driving using a filter. In this case, a delay always occurs due to the time constant of the filter, and ripples can be ignored. If the time constant is given, the measurement start area is greatly limited, leading to a significant increase in the overall measurement time. Therefore, in this plan, in order to detect this steady-state deviation, at least the difference between the current position and the target position at the time of the previous scan is accumulated at least about 2 16 in this embodiment, and this is calculated as the number of samples. The average value 97.4 of the steady deviation between the target position and the current position is obtained by subtracting, and the target position 97.6 is obtained by subtracting the average value 97.4 from the target position 97.5 at the time of this scan. An arithmetic operation was performed, and a configuration capable of performing EO correction within the dynamic range as shown in 98/1 of FIG. 98 was realized. Note that the number of integrations is not limited to this value as long as the target accuracy can be obtained, and a smaller number of integration levels may be used.

  FIG. 99 shows a block diagram. The target value 99.1 is subtracted from the current position 99.2, and the previous integration operation is executed during scanning in the block 99.3. On the other hand, the average value of the steady deviation obtained in the same manner as the previous time is output from 99.4. The subtracter 99/5 subtracts 99.1 to 99.4 to obtain the combined target position 99.6, and subtracts this value from the current position 99.7 from the interferometer to correct the EO without delay in response or ripple. Data is realized.

  FIG. 100 shows the structure of 99.3 block difference average detection in FIG. Accumulation is performed at 100 · 1, 100 · 2, the word of the data selector 100 · 4 is selected by the value of the accumulation counter 100 · 3, the division phase is executed, and the average value of the steady deviation is output. ing.

  FIG. 101 describes an idea of a TDI transfer clock. The TDI is an image sensor that aims to improve sensitivity and reduce random noise by connecting photoelectric elements in multiple stages in the scanning direction and transmitting the charge of each image sensor to subsequent elements. As shown in FIG. It is important that the imaging target on the stage and the pixels on the TDI have a one-to-one correspondence. If this relationship breaks, image blurring occurs. The cases of being in a synchronous relationship are shown in 1-1, 1-2, 2-1, 2-2, and the cases of being out of sync are shown in 3-1, 3-2, 4-1, 4-2. Since the transfer of TDI is performed to the next stage in synchronization with an external pulse, this can be realized by generating a transfer pulse when the stage moves by one pixel.

However, since the position information output of the current mainstream laser interferometer is a format that outputs a 32-bit binary output in synchronization with its own internal clock of 10 MHz, it cannot be easily realized as it is. If the resolution is several tens of nm, the accuracy of the transfer pulse is important, and high-speed and high-precision digital processing is required. The method devised in this case is shown in FIG. In the figure, the position information of the interferometer and the synchronization signal of 10 MHz are introduced into this circuit from the buffer 102. 10 MHz clock 102.2 is synchronized by PLL 102.3 100
A MHz clock is generated and supplied to each circuit. A method is employed in which arithmetic processing is executed for every 10 states of the synchronization signals 102 and 4. The current position information is held in 102.5, and the previous value is held in 102.6. The difference between the two is calculated by 102.7, and the position difference for every 10 states is output from 102.8. This difference value is loaded as a parallel value into the parallel serial converters 102 and 9 and the difference is output as the number of serial pulses from 102 and 10 in synchronization with the clock of 100 MHz. The functions 102 and 11 have the same function, but are combined with 102 and 12 and 102 and 13 so that the operation can be performed without a break every 10 states. As a result, a serial pulse corresponding to the position difference is output from the sum circuits 102 and 10 to the counters 102 and 14 every 10 MHz. The resolution of the laser interferometer is 0.6 nm, and one pixel is 48 nm.
Then, if the comparators 102 and 15 are set to 80, 19 pulses are output at the timing when the counter becomes one pixel phase or the like. By using this signal as a transfer pulse from the outside of the TDI, even if the stage speed fluctuates, it is possible to operate in synchronization with it and to prevent blurring and blurring.

  FIG. 103 shows a timing chart. Reference numeral 1 denotes interferometer coordinate (position) information, and numerals indicate positions as examples. Reference numeral 2 denotes a synchronization signal of 100 MHZ created by the PLL. Bank A is the operation timing of the parallel serial converters 102 and 9, and bank B is that of 102 and 11 as well. After the latch timing 7 for storing the position information, the difference calculation timing 8 is executed, the value is loaded into the parallel serial converters 102 and 9, and the output of 4 is executed using the time of one cycle of the next 10M clock 3. Bank B executes the same operation at a timing delayed by one cycle of 10M clock 3 and realizes 6 pulses without difficulty.

2-7-7) Image Comparison Method FIG. 104 shows a schematic configuration of a defect inspection apparatus according to a modification of the present invention. This defect inspection apparatus is the above-described projection type inspection apparatus. The electron gun 104. 1 that emits a primary electron beam, the electrostatic lens 104. 2 that deflects and shapes the emitted primary electron beam, and the molded primary An E × B deflector 104 3 that deflects an electron beam so that it strikes the semiconductor wafer W substantially perpendicularly in a field perpendicular to an electric field E and a magnetic field B, and an objective lens 104 that forms an image of the deflected primary electron beam on the wafer W 4. A stage 104. 5 provided in a sample chamber (not shown) that can be evacuated to a vacuum and movable in a horizontal plane with the wafer W placed thereon, secondary electrons emitted from the wafer W by irradiation of the primary electron beam An electrostatic lens 104,6 of a projection system for projecting and projecting a line and / or a reflected electron beam at a predetermined magnification, and a detector 104,7 for detecting the formed image as a secondary electron image of the wafer And configured to include controls the entire apparatus, a control unit 104 · 8 for executing a process for detecting defects of the wafer W based on the secondary electron image detected by the detector 104 · 7. The secondary electron image includes contributions from not only secondary electrons but also scattered electrons and reflected electrons. Here, the secondary electron image is referred to as a secondary electron image.

  Further, between the objective lenses 104 and 4 and the wafer W, there are interposed deflection electrodes 104 and 9 for deflecting the incident angle of the primary electron beam to the wafer W by an electric field or the like. The deflection electrodes 104 and 9 are connected to deflection controllers 104 and 10 for controlling the electric field of the deflection electrodes. The deflection controllers 104 and 10 are connected to the control units 104 and 8, and control the deflection electrodes so that an electric field corresponding to a command from the control units 104 and 8 is generated by the deflection electrodes 104 and 9. The deflection controllers 104 and 10 can be configured as a voltage control device that controls the voltage applied to the deflection electrodes 104 and 9.

The detectors 104 and 7 can have any configuration as long as the secondary electron image formed by the electrostatic lenses 104 and 6 can be converted into a post-processable signal. For example, as shown in detail in FIG. 62, the detectors 104 and 7 are connected to the microchannel plates 62 and 1.
And the fluorescent screens 62 and 2, the relay optical systems 62 and 3, and the imaging sensors 62 and 4 made up of a large number of CCD elements. The microchannel plate 62. 1 has a large number of channels in the plate, and generates a larger number of electrons while the secondary electrons imaged by the electrostatic lenses 104 and 6 pass through the channels. . That is, secondary electrons are amplified. The fluorescent screen 62.2 converts the secondary electrons into light by emitting fluorescence with the amplified secondary electrons. The relay lenses 62 and 3 guide this fluorescence to the CCD image sensor 62 and 4, and the CCD image sensor 62 and 4 converts the intensity distribution of secondary electrons on the surface of the wafer W into an electrical signal for each element, that is, digital image data. Output to the control unit 104. Here, the microchannel plate 62.1 may be omitted, and in this case, blur due to spreading between the microchannel plate 62.1 and the phosphor screen can be reduced. For example, an image of 0.2 can be increased to 0.3 to 0.6 with MTF.

  As illustrated in FIG. 104, the control units 104 and 8 can be configured by a general-purpose personal computer or the like. The computer includes a control unit main body 104/11 for executing various controls and arithmetic processes according to a predetermined program, a CRT 104/12 for displaying a processing result of the main body 104/11, and a keyboard and a mouse for an operator to input commands. The control units 104 and 8 may be configured by hardware dedicated to the defect inspection apparatus, a workstation, or the like.

  The control unit main bodies 104 and 11 include various control boards such as a CPU, RAM, ROM, hard disk, and video board (not shown). On a memory such as a RAM or a hard disk, there are assigned secondary electron image storage areas 104 and 14 for storing electrical signals received from the detectors 104 and 7, that is, digital image data of a secondary electron image of the wafer W. Yes. Further, on the hard disk, there are reference image storage units 104 and 15 for storing reference image data of a wafer having no defect in advance. Further, on the hard disk, in addition to a control program for controlling the entire defect inspection apparatus, secondary electron image data is read from the storage areas 104 and 14, and defects on the wafer W are automatically detected based on the image data according to a predetermined algorithm. Defect detection programs 104 and 16 to be detected are stored. As will be described in detail later, the defect detection programs 104 and 16 match the reference image read from the reference image storage units 104 and 15 with the actually detected secondary electron beam image, and detect the defect portion. It has a function of automatically detecting and displaying a warning to the operator when it is determined that there is a defect. At this time, the secondary electron images 104 and 17 may be displayed on the display units of the CRTs 104 and 12.

  Next, the operation of the defect inspection apparatus according to the embodiment will be described with reference to the flowcharts of FIGS. First, as shown in the flow of the main routine in FIG. 105, the wafer W to be inspected is set on the stage 104/5 (step 105.1). As described above, this may be a mode in which all the wafers stored in the loader are automatically set on the stages 104 and 5 one by one.

  Next, images of a plurality of areas to be inspected that are displaced from each other while partially overlapping each other on the XY plane of the surface of the wafer W are acquired (step 105.2). As shown in FIG. 108, the plurality of regions to be inspected for image acquisition include, for example, reference numbers 108.2a, 108.2b,. . . , 108.2k,. . . It can be seen that the positions are shifted while partially overlapping around the wafer inspection pattern 108. 3. For example, as shown in FIG. 109, 16 images 109.1 (inspected images) of the inspected region are acquired. Here, in the image shown in FIG. 109, a rectangular cell corresponds to one pixel (or a block unit larger than the pixel), and a black cell corresponds to an image portion of a pattern on the wafer W. Details of steps 105.2 will be described later with reference to the flowchart of FIG.

  Next, the image data of the plurality of regions to be inspected acquired in step 105.2 are compared with the reference image data stored in the storage units 104 and 15 (steps 105 and 3 in FIG. It is determined whether or not there is a defect on the wafer inspection surface covered by the inspection area. In this step, so-called matching processing between image data is executed, and details thereof will be described later with reference to a flowchart of FIG.

  If it is determined from the comparison result in Steps 105 and 3 that there is a defect in the wafer inspection surface covered by the plurality of inspected areas (Yes in Step 105.4), the operator is warned of the presence of the defect (Step S5). 105.5). As a warning method, for example, a message notifying the presence of a defect may be displayed on the display unit of the CRT 104 or 12, or at the same time, enlarged images 104 and 17 of a pattern having a defect may be displayed. Such a defective wafer may be immediately taken out of the sample chamber and stored in a storage place different from the wafer having no defect (steps 105 and 6).

  If it is determined as a result of the comparison processing in Steps 105 and 5 that the wafer W is free of defects (No in Steps 105 and 4), whether there is still an area to be inspected for the wafer W that is currently inspected. It is determined whether or not (steps 105 and 7). When the area to be inspected remains (Yes in Steps 105 and 7), the stages 104 and 5 are driven, and the wafer W is moved so that the other area to be inspected now falls within the irradiation area of the primary electron beam. (Step 105.8). Thereafter, the process returns to Step 105.2 to repeat the same processing for the other inspection areas.

  When there is no area to be inspected (No at Step 105.7), or after the defective wafer extraction process (Step 105, 6), the wafer W currently being inspected is the final wafer. It is determined whether there is no uninspected wafer in a loader (not shown) (steps 105 and 9). If it is not the final wafer (No at Steps 105 and 9), the inspected wafer is stored in a predetermined storage location, and a new uninspected wafer is set on the stage 104 and 5 instead (Steps 105 and 10). Thereafter, the process returns to Step 105.2 to repeat the same processing for the wafer. If it is the final wafer (Yes at Steps 105 and 9), the inspected wafer is stored in a predetermined storage location, and all processes are completed. Each wafer or each wafer is assigned an identification number, and the wafer being inspected is recognized and monitored. For example, duplicate inspection of wafers is prevented.

Next, the processing flow of steps 105.2 will be described with reference to the flowchart of FIG. In the figure, first, an image number i is set to an initial value 1 (step 106.1). This image number is an identification number sequentially assigned to each of the plurality of inspection area images. Next, the image position (X _i , Y _i ) is determined for the set inspection area of image number i (step 106.2). This image position is defined as a specific position in the area for defining the inspection area, for example, a center position in the area. At this time, since i = 1, the image position (X ₁ , Y ₁ ) is obtained, which corresponds to the center position of the inspected area 108 2a shown in FIG. The image positions of all the image areas to be inspected are determined in advance, and are stored, for example, on the hard disk of the control units 104 and 8, and are read out in steps 106 and 2.

Next, the deflection controller is arranged so that the primary electron beam passing through the deflection electrodes 104 and 9 in FIG. 104 is irradiated to the inspected image region at the image position (X _i , Y _i ) determined in step 106. 104 and 10 apply a potential to the deflection electrodes 104 and 9 (steps 106 and 3 in FIG. 106).

Next, the primary electron beam is emitted from the electron gun 104. 1, and the set wafer W is passed through the electrostatic lens 104 • 2, the E × B deflector 104 • 3, the objective lens 104 • 4 and the deflection electrode 104 • 9. Irradiate the surface (step 106.4). At this time, the primary electron beam is deflected by the electric field generated by the deflection electrodes 104 and 9 and is irradiated over the entire image area to be inspected at the image position (X _i , Y _i ) on the wafer inspection surface 108. When the image number i = 1, the inspected area is 108 · 2a.

  Secondary electrons and / or reflected electrons (hereinafter referred to as “secondary electrons”) are emitted from the region to be inspected irradiated with the primary electron beam. Therefore, the generated secondary electron beam is imaged on the detectors 104 and 7 at a predetermined magnification by the electrostatic lens 104 and 6 of the magnifying projection system. The detectors 104 and 7 detect the imaged secondary electron beam, and convert it into an electrical signal, that is, digital image data for each detection element (steps 106 and 5). Then, the detected digital image data of the image number i is transferred to the secondary electronic image storage areas 104 and 14 (steps 106 and 6).

Next, the image number i is incremented by 1 (steps 106 and 7), and it is determined whether or not the incremented image number (i + 1) exceeds a predetermined value i _MAX (steps 10 and 8). This i _MAX is the number of inspected images to be acquired, and is “16” in the above-described example of FIG.

If the image number i does not exceed the fixed value i _MAX (No at Step 106.8), the process returns to Step 106.2 again, and the image position (X _{i + 1} , Y _{i + 1} ) for the incremented image number (i + 1). ) Again. This image position is a position moved from the image position (X _i , Y _i ) determined in the previous routine by a predetermined distance (ΔX _i , ΔY _i ) in the X direction and / or Y direction. In the example of FIG. 108, the inspected area is a position (X ₂ , Y ₂ ) moved only in the Y direction from (X ₁ , Y ₁ ), and becomes a rectangular area 108. Note that the values of (ΔX _i , ΔY _i ) (i = 1, 2,... I _MAX ) are actually empirical from the field of view of the detectors 104 and 7 on the wafer inspection surface 108. It can be determined as appropriate from the data on how much the distance is shifted and the number and area of the areas to be inspected.

Then, the processing of steps 106 - 2 to 106 - 7 sequentially repeatedly executed for i _MAX number of regions to be inspected. As shown in FIG. 108, these inspection regions are inspected image regions 108 and 2k at the image position (X _k , Y _k ) moved k times on the inspection surface 108 and 1 of the wafer. The position is shifted while partially overlapping. In this way, the 16 pieces of inspected image data illustrated in FIG. 109 are acquired in the image storage areas 104 and 14. The acquired images 109.1 (inspected images) of the plurality of regions to be inspected are partially or completely the images 109.2 of the pattern 108.3 on the wafer inspection surface 108.1, as illustrated in FIG. You can see that

If the incremented image number i exceeds i _MAX (YES at step 106.8), this subroutine is returned and the process proceeds to the main routine comparison process.
The image data transferred to the memory in Steps 106 and 6 is composed of secondary electron intensity values (so-called solid data) for each pixel detected by the detectors 104 and 7, but the subsequent comparison step (Step 105). Since the matching calculation is performed with the reference image in 3), it can be stored in the storage areas 104 and 14 in a state where various calculation processes are performed. Such arithmetic processing includes, for example, normalization processing for matching the size and / or density of the image data with the size and / or density of the reference image data, and an isolated pixel group having a predetermined number of pixels or less as noise. There is a process to remove. Furthermore, instead of simple solid data, the data may be compressed and converted into a feature matrix in which the features of the detection pattern are extracted within a range that does not reduce the detection accuracy of the high-definition pattern. As such a feature matrix, for example, a two-dimensional inspection area composed of M × N pixels is divided into m × n (m <M, n <N) blocks, and secondary electrons of the pixels included in each block. There is an m × n feature matrix or the like in which the sum of intensity values (or a normalized value obtained by dividing this sum by the total number of pixels in the entire region to be inspected) is used as each matrix component. In this case, the reference image data is also stored in the same expression. The image data referred to in the embodiment of the present invention includes not only simple data but also image data whose features are extracted by an arbitrary algorithm.

  Next, the flow of processing in steps 105.3 will be described with reference to the flowchart in FIG. First, the CPUs of the control units 104 and 8 read the reference image data from the reference image storage units 104 and 15 (FIG. 104) onto a working memory such as a RAM (step 107.1). This reference image is represented by reference numeral 109/3 in FIG. Then, the image number i is reset to 1 (steps 107 and 2), and the image data to be inspected with the image number i is read from the storage areas 104 and 14 onto the working memory (steps 107 and 3).

Next, the read reference image data and the image i data are matched to calculate a distance value D _i between them (steps 107 and 4). This distance value D _i represents the degree of similarity between the reference image and the inspected image i, and the greater the distance value, the greater the difference between the reference image and the inspected image. Any value can be used as long as the distance value D _i represents the similarity. For example, when the image data is composed of M × N pixels, the secondary electron intensity (or feature amount) of each pixel is regarded as each position vector component in the M × N dimensional space, and the reference image vector in the M × N dimensional space. And the Euclidean distance or correlation coefficient between the image i vectors. Of course, a distance other than the Euclidean distance, such as a so-called city area distance, can also be calculated. Furthermore, since the amount of calculation becomes enormous when the number of pixels is large, the distance value between the image data represented by the m × n feature vector may be calculated as described above.

Next, it is determined whether or not the calculated distance value D _i is smaller than a predetermined threshold value Th (step 107.5). This threshold value Th is experimentally obtained as a reference for determining a sufficient match between the reference image and the image to be inspected. If the distance value D _i is smaller than the predetermined threshold Th (Yes in Step 107.5), it is determined that the inspection surface 1034 of the wafer W is “no defect” (Step 107 · 6), and this subroutine is returned. . That is, if at least one of the images to be inspected substantially matches the reference image, it is determined that there is no defect. Thus, since it is not necessary to perform matching with all the images to be inspected, high-speed determination is possible. In the case of the example in FIG. 109, it can be seen that the inspected image in the third row and the third column is substantially coincident with the reference image with no positional deviation.

When the distance value D _i is equal to or larger than the predetermined threshold Th (No at Step 107/5), the image number i is incremented by 1 (Step 107/7), and the incremented image number (i + 1) exceeds a certain value i _MAX . It is determined whether or not (steps 107 and 8).

When the image number i does not exceed the predetermined value i _MAX (No at Step 107.8), the process returns to Step 107.3 again, image data is read for the incremented image number (i + 1), and the same processing is repeated.

When the image number i exceeds a certain value i _MAX (Yes at Step 107/8), it is determined that the inspection surface 1034 of the wafer W is “defect” (Step 107/9), and this subroutine is returned. . That is, if all of the inspected images do not substantially match the reference image, it is determined that “there is a defect”.

Although the above is each embodiment of a stage apparatus, this invention is not limited only to the said example, It can change arbitrarily suitably within the scope of the summary of this invention.
For example, the semiconductor wafer W is taken as an example of the sample to be inspected, but the sample to be inspected according to the present invention is not limited to this, and any one that can detect a defect with an electron beam can be selected. For example, a mask or the like on which a pattern for exposing a wafer is formed can be an inspection target.

Further, the present invention can be applied not only to an apparatus that performs defect detection using a charged particle beam other than electrons, but also to an arbitrary apparatus that can acquire an image capable of inspecting a defect of a sample.
Further, the deflection electrodes 104 and 9 can be placed not only between the objective lenses 104 and 4 and the wafer W but also at any position as long as the irradiation region of the primary electron beam can be changed. For example, between the E × B deflector 104 • 3 and the objective lens 104 • 4, between the electron gun 104 • 1 and the E × B deflector 104 • 3, and the like. Furthermore, the deflection direction may be controlled by controlling the field generated by the E × B deflector 104. That is, the function of the deflection electrodes 104 and 9 may be combined with the E × B deflector 104 and 3.

  In the above-described embodiment, when matching image data, either matching between pixels or matching between feature vectors is used. However, both may be combined. For example, first, high-speed matching is performed using feature vectors with a small amount of calculation, and as a result, high-similarity and high accuracy are achieved by a two-stage process of matching images with high-similarity with more detailed pixel data. Can be made compatible.

  In the embodiment of the present invention, the position shift of the image to be inspected is handled only by shifting the position of the irradiation region of the primary electron beam, but the process of searching for the optimum matching area on the image data before or during the matching process It is also possible to combine the present invention with (for example, detecting and matching regions having high correlation coefficients). According to this, a large positional shift of the image to be inspected can be dealt with by the positional shift of the irradiation region of the primary electron beam according to the present invention, and a relatively small positional shift can be absorbed by the subsequent digital image processing. The accuracy of detection can be improved.

  Furthermore, although the configuration of FIG. 104 is shown as the electron beam apparatus for defect inspection, the electron optical system and the like can be arbitrarily changed. For example, the electron beam irradiation means (104 · 1, 104 · 2, 104 · 3) of the defect inspection apparatus shown in FIG. 104 is a type in which a primary electron beam is incident on the surface of the wafer W from vertically above. However, the E × B deflectors 104 and 3 may be omitted, and the primary electron beam may be incident on the surface of the wafer W obliquely.

  Further, the flow of the flowchart of FIG. 105 is not limited to this. For example, for the sample determined to have a defect in step 105.4, the defect inspection in the other area is not performed, but the processing flow may be changed to detect the defect covering the entire area. Good. Further, if the irradiation region of the primary electron beam is enlarged and almost all the inspection region of the sample can be covered by one irradiation, Steps 105 and 7 and Steps 105 and 8 can be omitted.

  As described above in detail, according to the defect inspection apparatus of the present embodiment, images of a plurality of inspection regions displaced from each other while partially overlapping on the sample are obtained, and the inspection regions of these inspection regions are obtained. Since the defect of the sample is inspected by comparing the image with the reference image, it is possible to obtain an excellent effect of preventing the deterioration of the defect inspection accuracy due to the positional deviation between the image to be inspected and the reference image.

  Furthermore, according to the device manufacturing method of the present invention, since the defect inspection of the sample is performed using the defect inspection apparatus as described above, it is possible to improve the product yield and prevent the defective product from shipping. Is obtained.

2-7-8) Device Manufacturing Method Next, an embodiment of a semiconductor device manufacturing method according to the present invention will be described with reference to FIGS. FIG. 110 is a flowchart showing an embodiment of a semiconductor device manufacturing method according to the present invention. The manufacturing process of this embodiment includes the following main processes. (1) Wafer manufacturing process for manufacturing a wafer (or wafer preparation process for preparing a wafer) (Step 110. 1)
(2) Mask manufacturing process for manufacturing a mask to be used for exposure (or mask preparation process for preparing a mask) (Step 110, 2)
(3) Wafer processing process for performing necessary processing on the wafer (steps 110 and 3)
(4) Chip assembly process for cutting out chips formed on the wafer one by one and making them operable (steps 1110 and 4).
(5) A chip inspection process for inspecting the completed chip (step 110/5).

Each of the main processes described above further includes several sub-processes. Among these main processes, the wafer processing process (3) has a decisive influence on the performance of the semiconductor device. In this process, designed circuit patterns are sequentially stacked on a wafer to form a large number of chips that operate as memories and MPUs. This wafer processing process includes the following processes.
(A) A thin film forming process for forming a dielectric thin film to be an insulating layer, a wiring portion, or a metal thin film for forming an electrode portion (using CVD, sputtering, etc.)
(B) Oxidation process for oxidizing the thin film layer and the wafer substrate (C) Lithography process for forming a resist pattern using a mask (reticle) to selectively process the thin film layer and the wafer substrate (D) Resist pattern Etching process (eg using dry etching technology) to process thin film layers and substrates according to
(E) Ion / impurity implantation / diffusion process (F) Resist stripping process (G) Process for inspecting a processed wafer The wafer processing process is repeated as many times as necessary to manufacture a semiconductor device that operates as designed.

FIG. 111 is a flowchart showing a lithography process that is the core of the wafer processing process of FIG. 110. This lithography process includes the following steps.
(A) A resist coating process for coating a resist on the wafer on which the circuit pattern has been formed in the preceding process (steps 111 and 1)
(B) Step of exposing the resist (Step 111.2)
(C) Development process of developing the exposed resist to obtain a resist pattern (steps 11 and 3)
(D) Annealing process for stabilizing the developed resist pattern (steps 11 and 4)
The semiconductor device manufacturing process, wafer processing process, and lithography process are well known and need no further explanation.

  When the defect inspection method and the defect inspection apparatus according to the present invention are used in the inspection step (G), even a semiconductor device having a fine pattern can be inspected with high throughput, so that 100% inspection is possible, and the yield of products is improved. It is possible to prevent shipment of defective products.

2-7-9) Inspection The inspection procedure in the inspection step (G) will be described with reference to FIG. In general, a defect inspection apparatus using an electron beam is expensive and has a lower throughput than other process apparatuses. Therefore, important processes that are considered to require inspection at present (for example, etching,
After film formation or CMP (Chemical Mechanical Polishing) flattening process, etc., and in the wiring process, it is used for finer wiring process parts, that is, 1 to 2 processes of the wiring process and the gate wiring process of the previous process. ing. In particular, it is important to find shape defects and electrical defects such as wiring having a design rule of 100 nm or less, that is, a line width of 100 nm or less, and via holes having a diameter of 100 nm or less, and feeding back to the process.

  The wafer to be inspected is positioned on the ultra-precision XY stage through the atmospheric transfer system and the vacuum transfer system, and then fixed by an electrostatic chuck mechanism or the like, and thereafter, defect inspection or the like is performed according to the procedure of FIG. . First, as necessary, the position of each die is confirmed and the height of each location is detected and stored by an optical microscope. In addition to this, the optical microscope acquires an optical microscope image of a desired location such as a defect and is used for comparison with an electron beam image. Next, the conditions of the electron optical system are set, and the information set by the optical microscope is corrected using the electron beam image to improve the accuracy.

  Next, recipe information corresponding to the type of wafer (after which process, whether the wafer size is 200 mm or 300 mm, etc.) is input into the apparatus, and the inspection location designation, electron optical system setting, inspection condition setting, etc. After performing the above, defect inspection is normally performed in real time while acquiring an image. Cell-to-cell comparison, die comparison, and the like are inspected by a high-speed information processing system equipped with an algorithm, and the results are output to a CRT or stored in a memory as necessary.

  Defects include particle defects, shape abnormalities (pattern defects), and electrical (disconnections such as wiring or vias and poor conduction) defects, etc., which can be distinguished from each other, the size of defects, and killer defects (use of chips). It is also possible to automatically classify critical defects that are impossible) in real time. In particular, it is effective for classifying the defects such as wiring having a line width of 100 nm or less and vias having a diameter of 100 nm or less. Detection of an electrical defect is achieved by detecting a contrast abnormality. For example, a place with poor conduction is normally positively charged by electron beam irradiation (about 500 eV), and the contrast is lowered, so that it can be distinguished from a normal place. The electron beam irradiation means in this case is a low potential (energy) electron beam generation means (thermoelectron generation, UV / photoelectron) provided to make contrast due to potential difference stand out separately from the electron beam irradiation means for normal inspection. ). Before irradiating the inspection target region with the inspection electron beam, this low potential (energy is, for example, 100 eV or less) electron beam is generated and irradiated. In the case of a projection method that can be positively charged by irradiating an inspection electron beam, it is not necessary to provide a low-potential electron beam generating means depending on the specifications. Further, it is possible to detect a defect from a difference in contrast caused by applying a positive or negative potential to a sample such as a wafer with respect to a reference potential (which occurs because the flowability varies depending on the forward direction or reverse direction of the element).

  The contrast due to the potential difference may be converted into an image of a signal effective for displaying the potential contrast data and displayed. The potential now trust image can be analyzed to identify structures that are at voltages higher or lower than expected, ie, poor insulation or poor conduction or defects. For example, a potential contrast image is acquired from each different die on the wafer, and the defect is recognized by detecting the difference. In addition, by generating image data equivalent to the potential contrast image of the die to be inspected from design data such as CAD data, and detecting the difference between this image data and the potential contrast image acquired from the die to be inspected on the wafer. Recognize defects.

  It can also be used for line width measurement equipment and alignment accuracy measurement. Information on wafers to be inspected, for example, cassette numbers, wafer numbers (or lot numbers), etc., are all stored and managed in what position and state they are currently in. Therefore, there is no trouble that the inspection is mistakenly performed twice or more or the inspection is not performed.

2-8) Inspection method
2-8-1) Outline The basic flow of inspection is shown in FIG. First, after carrying the wafer including the alignment operation 113. 1, a recipe in which conditions relating to the inspection are set is created (113. 2). At least one type of recipe is required for a wafer to be inspected, but a plurality of recipes may exist for one wafer to be inspected in order to cope with a plurality of inspection conditions. Further, when there are a plurality of wafers to be inspected with the same pattern, a plurality of wafers may be inspected with one kind of recipe. 113 indicates that it is not necessary to create a recipe immediately before the inspection operation when inspecting with a recipe created in the past. Hereinafter, in FIG. 113, the inspection operations 113 and 4 inspect the wafer according to the conditions and sequence described in the recipe. Defect extraction is performed immediately every time a defect is discovered during an inspection operation,
a) Operation for performing defect classification (113/5) and adding extracted defect information and defect classification information to the result output file b) Operation for adding the extracted defect image to the image-only result output file or file c) Position of the extracted defect The operation for displaying the defect information on the operation screen is executed almost in parallel.

When inspection is completed for each wafer to be inspected,
a) Operation to close and save result output file b) Operation to send inspection result when communication from outside requests inspection result c) Operation to discharge wafer is executed almost in parallel.

When the setting for continuously inspecting the wafer is made, the next wafer to be inspected is transported and the series of operations are repeated.
Hereinafter, the details of the flow in FIG. 113 will be described.

(1) Recipe creation A recipe is a setting file for conditions related to inspection, and can be saved. The equipment is set using the recipe at the time of inspection or before inspection, but the conditions related to the inspection described in the recipe are:
a) Inspection target die b) Die internal inspection area c) Inspection algorithm d) Detection conditions (conditions necessary for defect extraction such as inspection sensitivity)
e) Observation conditions (magnification, lens voltage, stage speed, inspection order, and other conditions necessary for observation). A specific c) inspection algorithm will be described later.

  Among them, as shown in FIG. 114, the operator designates the die to be inspected on the die map screen displayed on the operation screen as shown in FIG. In the example of FIG. 114, the die 1 on the wafer end face and the die 2 that is clearly determined to be defective in the previous process are grayed out and deleted from the inspection target, and the remainder is used as the inspection target die. It also has a function of automatically specifying an inspection die based on the distance from the wafer end face and the quality information of the die detected in the previous process.

  In addition, the setting of the inspection area inside the die includes an image acquired by the operator using an optical microscope or an EB microscope with respect to the die internal inspection area setting screen displayed on the operation screen as shown in FIG. Specify with an input device such as a mouse. In the example of FIG. 115, a region 115.1 pointed by a solid line and a region 115.2 pointed by a broken line are set.

In the region 115. 1, almost the entire die is set as a setting region. The inspection algorithm is an adjacent die comparison method (die-die inspection), and details of detection conditions and observation conditions for this region are set separately. In the area 115. 2, the inspection algorithm is an array inspection (inspection), and details of detection conditions and observation conditions for this area are set separately. That is, a plurality of inspection areas can be set, and each inspection area can be set with its own inspection algorithm and inspection sensitivity. Also, the inspection areas can be overlapped, and different inspection algorithms can be simultaneously processed for the same area.

(2) Inspection operation The inspection is performed by subdividing the wafer to be inspected into a scan width as shown in FIG. The scanning width is substantially determined by the length of the line sensor, but is set so that the end portions of the line sensor slightly overlap. This is in order to determine the continuity between the lines when the detected defects are finally integrated, and to secure a margin for image alignment when performing the comparative inspection. The overlapping amount is about 16 dots for a 2048 dot line sensor.

  The scanning direction and sequence are schematically shown in FIG. That is, in order to shorten the inspection time, a bidirectional operation A, a unidirectional operation B due to machine limitations, and the like can be selected by the operator.

  It also has a function of automatically calculating and inspecting the operation of reducing the scanning amount based on the inspection target die setting of the recipe. FIG. 118A is an example of scanning when there is one inspection die 118. 1, and unnecessary scanning is not performed.

2-8-2) Inspection algorithm The inspection algorithm performed by this device is roughly classified as follows. Array inspection (cell inspection)
2. Random inspection (die inspection)
There are two types. As shown in FIG. 118-2, the die is divided into a cell portion 118.2 having a periodic structure mainly used for a memory and a random portion 118.3 having no periodic structure. Since the cell units 118 and 2 having a periodic structure have a plurality of comparison targets in the same die, they can be inspected by comparing cells in the same die. On the other hand, since the random parts 118 and 3 do not have a comparison target in the same die, it is necessary to compare the dies. Random inspection is further classified as follows according to the comparison target.

a) Adjacent die comparison method (Die-Die inspection)
b) Standard die comparison method (Die-Any Die inspection)
c) CAD data comparison method (Cad Data-Any Die inspection)
A method generally referred to as a golden template method represents the above b) and c). In the reference die comparison method, the reference die is a golden template, and in the CAD data comparison method, CAD data is a golden template.

The operation of each algorithm is described below.
2-8-2-1) Array inspection (cell inspection)
Array inspection is applied to inspection of periodic structures. An example is a DRAM cell.

  In the inspection, a reference image as a standard is compared with the image to be inspected, and the difference is extracted as a defect. The reference image and the image to be inspected may be a binary image or a multi-value image in order to improve detection accuracy.

The defect may be the difference itself between the reference image and the image to be inspected, but it is a secondary determination for preventing erroneous detection based on difference information such as the difference amount of the detected difference and the total area of the pixels having the difference. May be performed.

  In the array inspection, the reference image and the image to be inspected are compared in units of structure periods. That is, comparison may be made in units of one structure period while reading out images collectively acquired by a CCD or the like. If the reference image is n structure period units, n structure period units can be compared simultaneously.

One example of a reference image generation method is shown in FIG. 119. Here, since an example of comparison in units of one structural cycle will be described, one structural cycle unit generation is represented. It is also possible to set the number of periods to n in the same way.
As a premise, the inspection direction in FIG. Period 4 is the inspection period. Since the operator inputs the magnitude of the period while viewing the image, periods 1 to 6 in FIG. 119 can be easily recognized.

  A reference period image is generated by adding and averaging the periods 1 to 3 immediately before the inspection period in each pixel. Even if a defect exists in any one of 1 to 3, the influence is small because the average process is performed. A defect is extracted by comparing the formed reference periodic image and the inspection periodic image 4.

  Next, when the periodic image 5 to be inspected is inspected, the reference periodic images are generated by averaging the periods 2 to 4. In the same manner, a periodic image to be inspected is generated from images obtained before acquiring the periodic image to be inspected, and the inspection is continued.

2-8-2-2) Random inspection (die inspection)
The random inspection can be applied without being limited to the die structure. In the inspection, a reference image serving as a reference is compared with an image to be inspected, and the difference is extracted as a defect. The reference image and the image to be inspected may be a binary image or a multi-value image in order to improve detection accuracy. The defect may be the difference itself between the reference image and the image to be inspected, but based on the difference information such as the difference amount of the detected difference and the total area of the pixels having the difference, a secondary determination is made to prevent erroneous detection. You can go. Random inspection can be classified according to how to obtain a reference image. The operation is described below.

A. Adjacent die comparison method (Die-Die inspection)
The reference image is a die adjacent to the image to be inspected. A defect is judged by comparing with two dies adjacent to the image to be inspected. That is, in FIG. 120 and FIG. 121, the switches 121 and 4 and the switches 121 and 5 are set so that the memories 121 and 1 and the memories 121 and 2 of the image processing apparatus are connected to the paths 121 and 41 from the cameras 121 and 3. Has the following steps.

a) A step of storing the die image 1 in the memory 121 · 1 from the path 121 · 41 according to the scanning direction S.
b) A step of storing the die image 2 in the memory 121 · 2 from the route 121 · 41.

  c) While obtaining the die image 2 from the paths 121 and 42 simultaneously with the above b), the obtained die image 2 is compared with the image data stored in the memory 121 · 1 having the same relative position on the die and the difference is obtained. Step to seek.

d) A step of storing the difference of c).
e) A step of storing the die image 3 from the path 121 · 41 into the memory 121 · 1.
f) While obtaining the die image 3 from the paths 121 and 42 simultaneously with the above e), the obtained die image 3 is compared with the image data stored in the memory 121 and the relative position of the die is the same, and the difference is obtained. Step to seek.

g) A step of storing the difference of f) above.
h) A step of determining a defect of the die image 2 from the results stored in the above d) and g).
i) Repeat steps a) to h) in subsequent dies.

  Depending on the setting, before obtaining the difference in c) and f) above, the position alignment of the two images to be compared: correction is made so as to eliminate the position difference. Or density alignment: Correct so that there is no density difference. Alternatively, both processes may be performed.

B. Standard die comparison method (Die-Any Die inspection)
The reference die is designated by the operator. The reference die is a die image existing on the wafer or a die image stored before inspection. First, the reference die is scanned or transferred, and the image is stored in a memory to be a reference image. That is, in FIG. 121 and FIG. 122, it has the following steps.

a) An operator selects a reference die from a die of a wafer to be inspected or a die image stored before inspection.
b) When the reference die is present on the wafer to be inspected, the switches 121 and 4 are connected so that at least one of the memory 121 or 1 or the memory 121 or 2 of the image processing apparatus is connected to the path 121 or 41 from the camera 121 or 3; Setting the switches 121 and 5;

  c) When the standard die is a die image stored before the inspection, at least one of the memory 121 · 1 or the memory 121 · 2 of the image processing apparatus stores the reference image that is a die image from the memory 121 · 6. A step of setting the switches 121 and 4 and the switches 121 and 5 so as to be connected to the routes 121 and 7.

d) When the standard die is present on the wafer to be inspected, scanning the standard die and transferring a reference image, which is a standard die image, to the memory of the image processing apparatus.
e) A step of transferring a reference image, which is a standard die image, to the memory of the image processing apparatus without scanning, when the standard die is a die image stored before inspection.

  f) A step of obtaining a difference by comparing an image obtained by sequentially scanning an image to be inspected, an image in a memory to which a reference image as a standard die image is transferred, and image data having the same relative position on the die.

g) A step of determining a defect from the difference obtained in f) above.
h) Subsequently, as shown in FIG. 124, the same portion is inspected for the entire wafer with respect to the scanning position of the reference die and the die origin of the die to be inspected, and the scanning position of the reference die is changed until the entire die is inspected. While repeating d) to g) above.

  Depending on the setting, before obtaining the difference in f) above, the position alignment of the two images to be compared: correction is made so that there is no position difference. Or density alignment: correction is made so that the density difference is eliminated. Alternatively, both processes may be performed.

The reference die image stored in the memory of the image processing apparatus in the above d) or e) may be all the reference dies or may be inspected while being updated as a part of the reference die.
C. CAD data comparison method (CAD Data-Any Die inspection)
In the semiconductor manufacturing process shown in FIG. 123, a reference image is created from CAD data, which is an output of a semiconductor pattern design process by CAD, and used as a standard image. The reference image may be an entire die or a partial object including an inspection part.

The CAD data is usually vector data, and cannot be used as a reference image unless converted into raster data equivalent to image data obtained by a scanning operation. In this way, the following conversion is performed for the CAD data processing operation.

a) Vector data as CAD data is converted into raster data.
b) The above a) is performed in units of image scanning width obtained by scanning the inspection die during inspection.
c) The above b) converts image data whose relative position on the die is the same as the image to be obtained by scanning the inspection die.

d) The above c) is performed by overlapping the inspection scan and the conversion work.
The above a) to d) are examples in which conversion in units of image scanning width is performed for speeding up, but inspection is possible without fixing the conversion units to the image scanning width. In addition, at least one of the following functions is added as an additional function to the work of converting vector data into raster data.

a) Multi-value function of raster data.
b) A function of setting the multi-value gradation weight and offset in consideration of the sensitivity of the inspection apparatus with respect to a).

c) A function of performing image processing for processing pixels such as expansion and contraction after vector data is converted into raster data.
In FIG. 121, the inspection step by the CAD data comparison method is shown.

a) A step of converting CAD data into raster data by the computer 1 and generating a reference image by the additional function and storing it in the memory 121/6.
b) A step of setting the switches 121 and 4 and the switches 121 and 5 so that at least one of the memory 121 and the memory 121 and 2 of the image processing apparatus is connected to the path 121 and 7 from the memory 121 and 6.

c) A step of transferring the reference image in the memory 121.6 to the memory of the image processing apparatus.
d) A step of comparing the image obtained by sequentially scanning the image to be inspected, the image in the memory to which the reference image is transferred, and the image data having the same relative position on the die to obtain a difference.

e) A step of determining a defect from the difference obtained in d) above.
f) Subsequently, as shown in FIG. 124, using the scanning position of the standard die as a reference image, the same portion of the die to be inspected is inspected on the entire wafer, and the above-mentioned a ) To e) are repeated.

  Depending on the setting, before obtaining the difference in d) above, correction is made so that the position alignment of the two images to be compared: the position difference is eliminated. Or density alignment: correction is made so that the density difference is eliminated. Alternatively, both processes may be performed.

In c), the reference die image stored in the memory of the image processing apparatus may be all the reference dies or may be inspected while being updated as a part of the reference die.
2-8-2-2 ') Method of performing cell inspection and die inspection simultaneously So far, the algorithm of array inspection (cell inspection) and random inspection for inspecting the periodic structure has been explained. It is also possible to do it simultaneously. That is, the cell portion and the random portion are processed separately, and the cell portion compares the cells in the die, and at the same time, the random portion compares the adjacent die, reference die, or CAD data. In this way, the inspection time can be greatly shortened and the throughput is improved.

In this case, it is preferable that the cell portion inspection circuit is provided separately and independently. If inspection is not performed at the same time, it is possible to have one inspection circuit, set the software for cell inspection and random inspection to be switchable, and execute comparative inspection by switching software. is there. In other words, when pattern inspection is performed by applying a plurality of processing algorithms, these algorithms may be prepared separately and processed at the same time, or an algorithm corresponding to them may be provided to provide a single circuit. You may make it process by switching by. In any case, the present invention can be applied to cases where there are a plurality of types of cell portions, which are compared with each other and further leather is formed with dies or die and CAD data for random portions. is there.

2-8-2-3) Focus Mapping FIG. 125 shows the basic flow of the focus function. First, after carrying the wafer including the alignment operation, a recipe in which conditions and the like related to inspection are set is created. One of these recipes is a focus map recipe, and autofocus is performed during inspection and review operations according to the focus information set here. The focus map recipe creation procedure and autofocus operation procedure will be described below.

Focus map recipe creation procedure The focus map recipe has an independent input screen in the example, and the operator executes the following steps to create the recipe, but the input screen is provided for another purpose. It can also be added.

a) A step of inputting focus map coordinates such as a die position for inputting a focus value and a pattern in the die. Switch 126.1 in FIG.
b) A step of setting a die pattern necessary for automatically measuring the focus value. This step can be skipped if the focus value is not automatically measured.

c) A step of setting the best focus value of the focus map coordinates determined in the above a).
Among these, in step a), the operator can designate an arbitrary die, but it is possible to select all dies or select every n dies. In addition, the input screen can be selected by the operator as a schematic representation of the die arrangement on the wafer or an image using a real image.

  Of these, in step c), the operator manually sets the focus switch 126.2 in conjunction with the voltage value of the focus electrode (switch 126-3 in FIG. 126). Selection and setting are performed in a mode mode (switches 126 and 4 in FIG. 126) for automatically obtaining a focus value.

Focus Value Automatic Measurement Procedure The procedure for obtaining the focus value automatically in the above step c) is, for example, in FIG. 127. a) Obtain the image at the focus position Z = 1 and calculate its contrast.

b) The above a) is also performed with Z = 2, 3, and 4.
c) Regression from the contrast values obtained in a) and b) above to obtain a contrast function (FIG. 127)
d) Z which obtains the maximum value of the contrast function is obtained by calculation, and this is set as the best focus value.

For example, the die pattern necessary for automatically measuring the focus value shows good results when a line and space as shown in FIG. 128 is selected, but the contrast can be measured regardless of the shape if there is a monochrome pattern.

  The best focus value of one point can be obtained by performing steps a) to d). The data format at this time is (X, Y, Z) XY: coordinates for obtaining focus, Z: set of best focus values, and the number of focus map coordinates (X, Y, Z) determined by the focus map recipe is Will exist. This is called a focus map file in a part of the focus map recipe.

Auto Focus Operation Procedure A method for setting the focus to the best focus during the inspection operation and the review operation for acquiring an image from the focus map recipe is performed in the following steps.

  a) The position information is further subdivided based on the focus map file 1 created at the time of creating the focus map recipe, the best focus at this time is obtained by calculation, and the subdivided focus map file 2 is created.

b) The calculation of a) is performed with an interpolation function.
c) The interpolation function of b) is specified by the operator when creating a focus map recipe by linear interpolation, spline interpolation, or the like.

d) Monitor the XY position of the stage and change the voltage of the focus electrode to the focus value described in the focus map file 2 suitable for the current XY position.
More specifically, in FIG. 129, the black circle is the focus value of the focus map file 1, the white circle is the focus value of the focus map file 2,
1. Interpolation is performed between the focus values of the focus map file using the focus value of the focus map file.
2. The best focus is maintained by changing the focus position Z according to the scanning. At this time, the value is held until the next change position during the focus map file (white circle).

2-8-2-4) Measurement of litho margin Hereinafter, an embodiment relating to litho margin measurement will be described.
(1) Embodiment 10 (litho margin measurement 1)
Overview 1. Determine the range of exposure machine conditions and the best conditions. The target is focus.

  2. The application method of the inspection apparatus is not limited to the electron beam mapping method or the scanning method. That is, a method using light, an electron beam method, and a method using an arbitrary combination of these with a mapping method or a scanning method may be used.

3. Application of Reference Die Comparison Method (Die-Any Die Inspection) FIG. 130 is a flowchart showing the operation of the first embodiment. Hereinafter, description will be given based on this figure.
In steps 130 and 1, as shown in FIG. 131 as an example, the wafer was exposed two-dimensionally while changing the conditions using the focus condition and the exposure time condition as parameters. Also, one shot = 1 die image pattern.

Many stepper exposure machines have a function of automatically changing parameters and performing exposure, which is generally called TEST exposure. Therefore, this function may be used as it is.
In Steps 130 and 2, development, resist stripping, etching, CVD, CMP, plating, and the like can be considered. In particular, in observation with an electron beam, the resist is difficult to be charged and observed. The process from resist stripping to plating is performed. Preferably, resist observation is desirable.

  Details of steps 130 and 3 will be described with reference to FIG. This process uses the function of measuring the contrast of the image set by the operator of the inspection apparatus performing steps 130 and 4, and registers the minimum line & space part of the die pattern as an area for contrast measurement, and performs the following operations. did.

  First, an upper limit Db and a lower limit Da of the exposure time were obtained. Since the contrast value was extremely low at the exposure time of Db or more and the exposure time of Da or less, it is excluded from the inspection object. The gray portion in FIG.

  Next, an upper limit Fb and a lower limit Fa of the focus value were obtained. Since the contrast value is extremely low between the focus value of Fb or more and the focus value of Fa or less, it is excluded from the inspection object. The gray portion in FIG.

  Next, the die that is the intersection of the middle die row Ds of Da and Db and the middle die row Fs of Fa and Fb was selected as the best exposure condition shot. All processes for selecting these best exposure condition shots are performed automatically.

In Steps 130 and 4, the standard die in FIG. 132 is used as a reference image, and the white die is used as an image to be inspected by a standard die comparison method (Die-Any Die inspection).
Steps 130 and 5 determine exposure conditions using the inspection results 130 and 4. That is, if the exposure conditions are inappropriate, for example, the line and space of the die pattern do not resolve, or the edge of the pattern becomes obtuse, a difference from the reference image occurs, resulting in a pattern defect. Utilizes the detected effect. Of course, in addition to the exposure conditions, pattern defects and particles caused by exposure errors may be detected. In this case, reinspection is performed. However, it is not a problem because the frequency of occurrence is low.

The specific procedure of steps 130 and 5 is as follows:
1) Since priority is given to obtaining the focus margin, the exposure time is fixed to Ds in FIG. 132, and the relationship between the focus value and the number of defects is obtained (FIG. 133).
2) At this time, since the criterion for determining the focus value is a condition in which no defect occurs due to the exposure condition, as a conclusion, the focus value permitted as the exposure condition is F1 to F2.

  3) Specifically, what kind of numerical value / unit of exposure machine expression F1 and F2 can be determined by transferring the position of the die and its exposure condition through the communication path connected by Rs232c or Ethernet from the exposure machine. Can be calculated easily. This apparatus has a function of displaying the image by converting it into a value that can be directly input to the exposure machine, as well as determining whether the exposure condition is acceptable.

  4) If a dedicated communication path or a communication path such as the SEMI standard is used, the result of this inspection apparatus can be fed back to the exposure apparatus. The above procedure is further performed by changing the exposure condition (exposure time) to determine the focus and exposure margins.

(2) Embodiment 11 (litho margin measurement 2)
Outline Obtains the range of exposure machine conditions and the best conditions. The target is focus.

1. The application method of the inspection apparatus is not limited to the electron beam mapping method or the scanning method. An optical method, an electron beam method, and a method combining these with a mapping method or a scanning method can be used.

2. Application of CAD data comparison method (Cad Data-Any Die inspection).
FIG. 134 is a flow showing the operation of the second embodiment. Hereinafter, description will be given based on this figure.
In Steps 134 and 1, as shown in FIG. 135 as an example, the conditions were changed using the focus condition and the exposure time condition as parameters, and the wafer was exposed two-dimensionally. Also, one shot = 1 die image pattern.

Many stepper exposure machines have a function of automatically changing parameters and performing exposure, which is generally called TEST exposure. Therefore, this function may be used as it is.
In Steps 134 and 2, development, resist stripping, etching, CVD, CMP, plating, and the like can be considered. In particular, in observation with an electron beam, it is difficult to charge and observe the resist. The process from resist stripping to plating is performed. Preferably, it is better to observe the resist stage.

  In Steps 143-3, a reference image that is to be in the best possible state is generated from CAD data of the exposed shot pattern. At this time, raster data which is image data is multivalued. As shown in FIG. 136, for example, in patterns A, B, and C having different line widths, the pattern C is finer than the pattern B, but the white level of the pattern was empirically compared. When the pattern C white level is closer to black than the pattern B, and when the pattern black level is compared, the pattern C black level is closer to white than the pattern B. Therefore, the image data is multi-valued in consideration of the pattern shape, density, pattern position on the wafer, and the like, not the binary value that appears as black and white as the image.

  At the same time, considering the setting conditions of the observation system, the effect of charge-up, magnetic field, etc., when comparing the image actually obtained by observation and the image data generated from CAD data, it is not recognized as a pseudo defect. Image processing is performed on image data generated from CAD data.

In Steps 134 and 4, the image generated in 134 and 3 is used as a reference image, and the die on the wear is used as an image to be inspected, and die comparison is performed.
Steps 134 and 5 determine exposure conditions using the inspection results of 134 and 4. That is, if the exposure conditions are inappropriate, for example, the line and space of the die pattern does not resolve, or the edge of the pattern becomes obtuse, a difference from the reference image occurs, resulting in a pattern defect. Utilizes the detected effect. Of course, in addition to the exposure conditions, pattern defects and particles caused by exposure errors may be detected. In this case, reinspection is performed. However, it is not a problem because the frequency of occurrence is low.

The specific procedure of Steps 134.5 is as follows:
1) Since priority is given to obtaining the focus margin, the exposure time is set to a fixed value obtained empirically, and the relationship between the focus value and the number of defects in this case is obtained (FIG. 137).

  2) At this time, since the criterion for determining the focus value is a condition in which no defect occurs due to the exposure condition, as a conclusion, the focus value permitted as the exposure condition is F1 to F2.

  3) Specifically, what kind of numerical value / unit of exposure machine expression F1 and F2 can be determined by transferring the position of the die and its exposure condition through the communication path connected by Rs232c or Ethernet from the exposure machine. Can be calculated easily. This apparatus has a function of displaying the image by converting it into a value that can be directly input to the exposure machine, as well as determining whether the exposure condition is acceptable.

4) If a dedicated communication path or a communication path such as the SEMI standard is used, the result of this inspection apparatus can be fed back to the exposure apparatus.
Although the lithography margin measurement of the exposure conditions has been described above, a reticle or stencil mask that is an exposure mask may be inspected. In this case, the inspection for determining the exposure conditions can be simplified.

3 Other embodiments
3-1) Modification of Stage Device FIG. 138 shows a modification of the stage device in the detection device according to the present invention.

  On the upper surface of the Y-direction movable portion 138.2 of the stage 138.1, partition plates 138.4 extending substantially horizontally in the + Y direction and the -Y direction (left-right direction in FIG. 139) are attached, and the X-direction movable portion 138 is attached. A diaphragm 138/5 having a small conductance is always formed between the upper surface of 4 and the upper surface. In addition, a similar partition plate 138. 6 extends on the upper surface of the X direction movable portion 138. 4 in the ± X direction (left and right direction in FIG. 138 (A)), and the stage bases 138. The throttle portions 138 and 8 are always formed between the upper surface of each of them. The stage bases 138 and 7 are fixed on the bottom wall in the housings 138 and 9 by a known method.

  For this reason, the diaphragms 138 and 5 and 138 and 8 are always formed regardless of the position of the sample stage 138 and 10, so that the guide surfaces 138 and 11 are moved when the movable parts 138 and 138 and 4 are moved. Even if gas is released from 138.12, the movement of the released gas is hindered by the throttling portions 138.5 and 138.8. Therefore, the pressure increase in the spaces 138.13 near the sample irradiated with the charged beam is very small. I can hold it down.

  A differential exhaust groove as shown in FIG. 140 is formed around the hydrostatic bearings 138 and 14 on the side surface and the lower surface of the movable portion 138 and 2 of the stage and the lower surface of the movable portion 138 and 4. Since the evacuation is performed by the groove, when the throttle portions 138, 5 and 138 and 8 are formed, the discharge gas from the guide surface is mainly exhausted by these differential exhaust portions. For this reason, the pressures in the spaces 138 and 15 and 138 and 16 inside the stage are higher than the pressure in the chamber C. Therefore, if the space 138.15,138.16 is not only exhausted by the differential exhaust grooves 140.1 and 140.2, but a place to be evacuated is provided separately, the pressure of the space 138.15,138.16 is increased. The pressure increase in the vicinity of the sample 138 and 13 can be further reduced. For this purpose, vacuum exhaust passages 138.17 and 138.18 are provided. The exhaust passage passes through the stage bases 138 and 7 and the housings 138 and 9 and communicates with the outside of the housings 138 and 9. Further, the exhaust passages 138 and 18 are formed in the X direction movable portions 138 and 4, and open to the lower surfaces of the X direction movable portions 138 and 4.

  In addition, when the partition plates 138, 3, 138, 6 are installed, it is necessary to enlarge the chamber so that the chamber and the partition plate do not interfere with each other, but this point can be improved by making the partition plate a stretchable material or structure. Is possible. In this embodiment, the partition plate is made of rubber or has a bellows shape, and its end in the moving direction is the X direction movable portion 138.4 in the case of the partition plate 138.3, and the partition plate 138.6. In this case, it can be considered that the housings 138 and 9 are respectively fixed to the inner walls. Reference numerals 138 and 19 denote lens barrels.

FIG. 141 shows a second modification of the stage device. In this embodiment, cylindrical partitions 141 and 2 are configured so that a constriction portion is formed between the front end of the lens barrel, that is, the charged beam irradiation unit 141 and 1 and the upper surface of the sample W. In such a configuration, even if gas is released from the XY stage and the pressure in the chamber C increases, the interiors 141 and 3 of the partition are partitioned by the partitions 141 and 2 and are exhausted by the vacuum pipes 141 and 4. Therefore, a pressure difference is generated between the inside of the chamber C and the interiors 141 and 3 of the partition, and the pressure increase in the spaces 141 and 3 inside the partition can be suppressed to a low level. The gap between the partitions 141 and 2 and the sample surface varies depending on how much pressure is maintained in the chamber C and around the irradiation units 141 and 1, but is approximately several tens of μm to several mm. In addition, the inside of partition 141 * 2 and vacuum piping are connected by the well-known method.

  Further, in the charged beam irradiation apparatus, a high voltage of about several kV may be applied to the sample W, and if a conductive material is placed in the vicinity of the sample, there is a risk of causing discharge. In this case, if the material of the partitions 141 and 2 is made of an insulating material such as ceramics, no discharge occurs between the sample W and the partitions 141 and 2.

  The ring members 141 and 5 arranged around the sample W (wafer) are plate-shaped adjustment parts fixed to the sample tables 141 and 6, and a charged beam is irradiated to the end of the sample such as a wafer. Even so, the height is set to be the same as that of the wafer so that the minute gaps 141 and 7 are formed over the entire periphery of the front end of the partitions 141 and 2. As a result, no matter what position of the sample W is irradiated with the charged beam, a constant minute gap 952 is always formed at the tip of the partition 141, 2, and the pressure in the space 141, 3 around the tip of the lens barrel is stabilized. Can be kept in.

  FIG. 142 shows another modification. A partition 142. 1 incorporating a differential exhaust structure is provided around the charged beam irradiating units 141. 2 of the lens barrels 138 and 19. The partition 142. 1 has a cylindrical shape. A circumferential groove 142. 2 is formed inside the partition 142. 1, and an exhaust passage 142. 3 extends upward from the circumferential groove. The exhaust passage is connected to the vacuum pipes 142 and 5 via the internal spaces 142 and 4. A minute gap of about several tens of μm to several mm is formed between the lower end of the partition 142. 1 and the upper surface of the sample W.

  In such a configuration, even if the gas is released from the stage as the stage moves and the pressure in the chamber C rises and the gas tries to flow into the tip, that is, the charged beam irradiating unit 141. 2, the partition 142. Since the conductance is very small by narrowing the gap with the sample W, the gas is prevented from flowing in and the amount of inflow decreases. Furthermore, since the inflowing gas is exhausted from the circumferential groove 142/2 to the vacuum pipe 142/5, there is almost no gas flowing into the space 141/6 around the charged beam irradiation unit 141/2, and the charged beam irradiation. The pressure of the parts 141 and 2 can be maintained at a desired high vacuum.

  FIG. 143 shows still another modification. A partition 143. 1 is provided around the chamber C and the charged beam irradiation unit 141. 1 to separate the charged beam irradiation unit 141. 1 from the chamber C. The partition 143.1 is connected to the refrigerator 143.3 through a support member 142.3 made of a material having good thermal conductivity such as copper or aluminum, and is cooled to about −100 ° C. to −200 ° C. ing. The members 143 and 4 are for inhibiting heat conduction between the cooled partitions 143 and 1 and the lens barrels 138 and 19, and are made of a material having poor heat conductivity such as ceramics or a resin material. The members 143 and 5 are made of a non-insulator such as ceramics and are formed at the lower end of the partition 143 and 1 to prevent the sample W and the partition 143 and 1 from being discharged.

  With such a configuration, the gas molecules that are about to flow into the charged beam irradiation unit from the chamber C are blocked by the partition 143-1 and are frozen and collected on the surface of the partition 143.1 even if it flows. Therefore, the pressure of the charged beam irradiation unit 143, 6 can be kept low.

As the refrigerator, a refrigerator such as cooling with liquid nitrogen, a He refrigerator, a pulse tube refrigerator or the like can be used.
In FIG. 144, yet another modification is shown. Both movable parts of the stage are provided with partition plates 144, 1, 144, 2 in the same manner as shown in FIG. 138. Even if the sample stage 144, 3 is moved to an arbitrary position, these partitions are provided. As a result, the spaces 144 and 4 in the stage and the interior of the chamber C are partitioned through the apertures 144 and 5 and 144 and 6, respectively. Further, partitions 144 and 7 similar to those shown in FIG. 141 are formed around the charged beam irradiation unit 141. 1, and the space in the chamber C and the charged beam irradiation unit 141. 8 is partitioned. For this reason, when the stage is moved, even if the gas adsorbed on the stage is released into the spaces 144 and 4 and the pressure in this portion is increased, the pressure increase in the chamber C is kept low, and the pressure in the spaces 144 and 9 is increased. Is further reduced. Thereby, the pressure of the charged beam irradiation spaces 144 and 9 can be kept low. Further, by using a partition 142. 1 with a built-in differential exhaust mechanism as shown in the partition 144. 7 or a partition cooled by a refrigerator as shown in FIG. 142, the spaces 144. It becomes possible to maintain it stably at a low pressure.

According to these embodiments, the following effects can be obtained.
(1) The stage device can exhibit highly accurate positioning performance in a vacuum, and the pressure at the charged beam irradiation position is unlikely to increase. That is, it is possible to perform processing with a charged beam on the sample with high accuracy.
(2) The gas released from the static pressure bearing support can hardly pass through the partition to the charged beam irradiation region side. As a result, the degree of vacuum at the charged beam irradiation position can be further stabilized.
(3) It becomes difficult for the emitted gas to pass to the charged beam irradiation region side, and the degree of vacuum in the charged beam irradiation region can be easily kept stable.
(4) The inside of the vacuum chamber is divided into three chambers: a charged beam irradiation chamber, a static pressure bearing chamber, and an intermediate chamber thereof through a small conductance. Then, the evacuation system is configured so that the pressure in each chamber becomes the charged beam irradiation chamber, the intermediate chamber, and the static pressure bearing chamber in order from the lowest. The pressure fluctuation to the intermediate chamber is further suppressed by the partition, and the pressure fluctuation to the charged beam irradiation chamber is further reduced by the other partition, and the pressure fluctuation can be reduced to a level that is not substantially problematic. .
(5) It is possible to suppress a pressure increase when the stage moves.
(6) It is possible to further suppress the pressure increase when the stage moves.
(7) Since it is possible to realize an inspection apparatus with high accuracy of stage positioning and stable vacuum in the charged beam irradiation area, an inspection apparatus with high inspection performance and no risk of contaminating the sample is provided. can do.
(8) Since an exposure apparatus with high stage positioning performance and a stable vacuum degree in the charged beam irradiation region can be realized, an exposure apparatus with high exposure accuracy and no risk of contaminating the sample is provided. be able to.
(9) A fine semiconductor circuit can be formed by manufacturing a semiconductor with an apparatus having high accuracy in positioning of the stage and a stable degree of vacuum in the charged beam irradiation region.

It is obvious that the stage apparatus of FIGS. 138 to 144 can be applied to the stages 13 and 6 of FIG.
With reference to FIGS. 145 to 147, another embodiment of the XY stage according to the present invention will be described. It should be noted that reference numerals indicating the common components in the conventional example and the embodiment in FIG. 148 are the same. In this specification, “vacuum” is a vacuum called in the technical field, and does not necessarily indicate an absolute vacuum.

In FIG. 145, another embodiment of the XY stage is shown. The tip of the lens barrel 145. 1 that irradiates the charged beam toward the sample, that is, the charged beam irradiation unit 145. 2 is attached to the housing 145. 3 that defines the vacuum chamber C. A sample W placed on a movable table in the X direction (left and right direction in FIG. 145) of the XY stages 145 and 4 is arranged immediately below the lens barrel 145 and 1. This sample W can be accurately irradiated with a charged beam at an arbitrary position on the sample surface by a highly accurate XY stage 145.

  The bases 145 and 5 of the XY stages 145 and 4 are fixed to the bottom wall of the housings 145 and 3, and the Y table 145 and 6 that moves in the Y direction (direction perpendicular to the paper surface in FIG. 145) is placed on the bases 145 and 5. It is listed. On both side surfaces (left and right side surfaces in FIG. 145) of the Y tables 145, 6 are formed on the sides facing the Y table of the pair of Y direction guides 145, 7 and 145, 8 mounted on the bases 145,. A projecting portion is formed in the recessed groove. The concave groove extends in the Y direction over substantially the entire length of the Y direction guide. Static pressure bearings 145, 9, 145, 10, 145, 11, 145, 12 having a known structure are provided on the upper, lower, and side surfaces of the protrusions that protrude into the concave grooves, respectively, By blowing out the high-pressure gas, the Y tables 145 and 6 are supported in a non-contact manner with respect to the Y-direction guides 145 and 7 and 145 and 8, and can smoothly reciprocate in the Y-direction. Further, linear motors 145 and 13 having a known structure are arranged between the bases 145 and 5 and the Y tables 145 and 6, and the Y-direction drive is performed by the linear motors. High pressure gas is supplied to the Y table by flexible pipes 145, 14 for supplying high pressure gas, and the static pressure bearings 145, 10 to 145, 9, and 145 are passed through gas passages (not shown) formed in the Y table. High pressure gas is supplied to 12 to 145 The high-pressure gas supplied to the hydrostatic bearing is jetted into a gap of several microns to several tens of microns formed between the opposing guide surfaces of the Y-direction guide, and the Y table is guided in the X and Z directions with respect to the guide surface. It plays the role of accurately positioning in the direction (vertical direction in FIG. 145).

  On the Y table, X tables 145 and 14 are placed so as to be movable in the X direction (left and right direction in FIG. 145). A pair of X direction guides 145 and 15 (145 and 16) (only 145 and 15 are shown) having the same structure as the Y direction guides 145 and 145 and 8 for the Y table are provided on the Y table 145 and 6.・ It is provided with 14 in between. A groove is also formed on the side of the X direction guide facing the X table, and a protrusion projecting into the groove is formed on the side of the X table (side facing the X direction guide). The concave groove extends substantially over the entire length of the X-direction guide. The static pressure bearings 145, 9, 145, 10, 145, 17, 145, 11, 145, 12, 145, 18 are provided on the upper, lower and side surfaces of the X-direction table 145, 14 protruding into the groove. The same hydrostatic bearings (not shown) are provided in the same arrangement. Between the Y table 145 and 6 and the X table 145 and 14, linear motors 145 and 19 having a known structure are arranged, and the X table is driven in the X direction by the linear motor. High pressure gas is supplied to the X tables 145 and 14 through the flexible pipes 145 and 20, and the high pressure gas is supplied to the static pressure bearings. The high pressure gas is ejected from the static pressure bearing to the guide surface of the X direction guide, so that the X tables 145 and 14 are supported with high precision and non-contact with the Y direction guide.

  The vacuum chamber C is exhausted by vacuum pipes 145, 21, 145, 22, 145, 23 connected to a vacuum pump or the like having a known structure. The inlet sides (inside the vacuum chamber) of the pipes 145, 22, 145, 23 pass through the bases 145, 5 and open on the top surface near the position where the high-pressure gas is discharged from the XY stages 145, 4. The pressure in the vacuum chamber is prevented as much as possible from rising due to the high-pressure gas ejected from the static pressure bearing.

A differential pumping mechanism 145, 24 is provided around the tip of the lens barrel 145, 1, that is, around the charged beam irradiation unit 145, 2. The pressure is made sufficiently low. That is, the annular members 145 and 26 of the differential exhaust mechanisms 145 and 24 attached around the charged beam irradiation unit 145 2
It is positioned with respect to the housing 145.3 so that a minute gap (several microns to several hundreds of microns) 145.27 is formed between the sample W side surface) and the sample. 145 and 28 are formed. The annular grooves 145 and 28 are connected to a vacuum pump (not shown) by exhaust pipes 145 and 29. Therefore, the minute gaps 145 and 27 are exhausted through the annular grooves 145 and 28 and the exhaust ports 145 and 29, and gas molecules try to enter the spaces 145 and 25 surrounded by the annular members 145 and 26 from the vacuum chamber C. Will be exhausted. Thereby, the pressure in the charged beam irradiation spaces 145 and 25 can be kept low, and the charged beam can be irradiated without any problem. The annular groove may have a double structure or a triple structure depending on the pressure in the chamber and the pressure in the charged beam irradiation spaces 145 and 25.

  Generally, dry nitrogen is used as the high-pressure gas supplied to the hydrostatic bearing. However, if possible, it is preferable to use a higher purity inert gas. This is because if impurities such as moisture and oil are contained in the gas, these impurity molecules adhere to the inner surface of the housing and the surface of the stage components that define the vacuum chamber, and the degree of vacuum is deteriorated. This is because the degree of vacuum of the charged beam irradiation space is deteriorated by adhering to the surface. In the above description, the sample W is not normally placed directly on the X table, but has functions such as holding the sample in a removable manner and performing a slight position change with respect to the XY stage 145. However, the presence / absence of the sample stage and its structure are not related to the gist of the present embodiment, and are omitted for the sake of simplicity.

  In the charged beam apparatus described above, since the stage mechanism of the static pressure bearing used in the atmosphere can be used almost as it is, the high-precision XY stage equivalent to the high-precision stage for the atmosphere used in the exposure apparatus or the like is almost This can be realized for an XY stage for a charged beam apparatus at the same cost and size. The structure and arrangement of the static pressure guide and the actuator (linear motor) described above are only one embodiment, and any static pressure guide or actuator that can be used in the atmosphere can be applied.

  Next, FIG. 146 shows numerical examples of the sizes of the annular members 145 and 26 of the differential exhaust mechanism and the annular grooves formed thereon. In this example, the annular groove has a double structure of 146.1 and 146.2, which are separated in the radial direction.

The flow rate of the high-pressure gas supplied to the static pressure bearing is usually about 20 L / min (atmospheric pressure conversion). Assuming that the vacuum chamber C is evacuated with a dry pump having an evacuation rate of 20000 L / min through a vacuum pipe having an inner diameter of 50 mm and a length of 2 m, the pressure in the vacuum chamber is about 160 Pa (about 1.2 Torr). . At this time, if the dimensions of the annular member 146. 3 and the annular groove of the differential exhaust mechanism are set as shown in FIG. 146, the pressure in the charged beam irradiation space 141 • 1 is 10 ⁻⁴ Pa (10 ^−6. Torr).

In FIG. 147, another embodiment of the XY stage is shown. Dry vacuum pumps 147 and 4 are connected to vacuum chamber C defined by housing 147 and 1 through vacuum pipes 147 and 2 and 147 and 3, respectively. In addition, the annular grooves 147 and 6 of the differential exhaust mechanism 147 and 5 are connected to turbo molecular pumps 147 and 9 which are ultrahigh vacuum pumps via vacuum pipes 147 and 8 connected to the exhaust ports 147 and 7. . Further, the interiors of the lens barrels 147 and 10 are connected to the turbo molecular pumps 147 and 13 via the vacuum pipes 147 and 12 connected to the exhaust ports 147 and 11, respectively. These turbo molecular pumps 147, 9, 147, 13 are connected to the dry vacuum pumps 147, 4 by vacuum pipes 147, 14, 147, 15. In the figure, the roughing pump of the turbo molecular pump and the vacuum exhaust pump of the vacuum chamber are combined with one dry vacuum pump. However, the flow rate of the high-pressure gas supplied to the static pressure bearing of the XY stage, the volume of the vacuum chamber and the internal Depending on the surface area and the inner diameter and length of the vacuum piping, they may be evacuated by a separate dry vacuum pump.

  A high-purity inert gas (N 2 gas, Ar gas, etc.) is supplied to the static pressure bearing of the XY stage through the flexible pipes 147, 16, 147, 16. These gas molecules ejected from the hydrostatic bearing diffuse into the vacuum chamber and are exhausted by the dry vacuum pumps 147 and 4 through the exhaust ports 147, 18, 147, 19, 147, and 20. Further, these gas molecules that have entered the differential pumping mechanism and the charged beam irradiation space are sucked from the annular groove 147,6 or the tip of the lens barrel 147,10, and are passed through the exhaust ports 147,7 and 147,11 to become a turbo. The gas is exhausted by the molecular pumps 147 and 9 and 147 and 13, and is exhausted by the dry vacuum pump 147 and 4 after being discharged from the turbo molecular pump. Thus, the high purity inert gas supplied to the static pressure bearing is collected by the dry vacuum pump and discharged.

  On the other hand, the exhaust ports of the dry vacuum pumps 147 and 4 are connected to the compressors 147 and 22 through the pipes 147 and 21, and the exhaust ports of the compressors 147 and 22 are connected to the pipes 147, 23, 147, 24, 147, 25. And the flexible pipes 147, 16, 147, and 17 through regulators 147, 26, 147, and 27. For this reason, the high-purity inert gas discharged from the dry vacuum pumps 147 and 4 is pressurized again by the compressors 147 and 22, adjusted to an appropriate pressure by the regulators 147, 26, 147, and 27, and then again XY Supplied to the static pressure bearing of the table.

  Since the gas supplied to the hydrostatic bearing needs to be as highly pure as possible and contain as little water and oil as possible, turbo molecular pumps, dry pumps, and compressors have gas flow paths. It is required to have a structure in which moisture and oil are not mixed. In addition, cold traps, filters, etc. (147, 28) are provided in the middle of the compressor discharge side pipes 147, 23 to trap impurities such as moisture and oil mixed in the circulating gas and supply them to the static pressure bearings. It is also effective to prevent this from happening.

  By doing so, the high purity inert gas can be circulated and reused, so that the high purity inert gas can be saved and the inert gas is not spilled into the room where the apparatus is installed. It is possible to eliminate the risk of accidents such as suffocation.

  Note that high-purity inert gas supply systems 147 and 29 are connected to the circulation piping system, and when the gas circulation is started, the vacuum chamber C and the vacuum piping 147, 8, 147, 12, 147, 14, 147 The role of filling high-purity inert gas in all circulation systems including 15, 147, 2, 147, 3 and pressure side piping 147, 21, 147, 23, 147, 24, 147, 25, 147, 30; It plays a role of supplying the shortage when the flow rate of the circulating gas decreases for some reason. Further, by providing the dry vacuum pump 147.4 with a function of compressing to the atmospheric pressure or higher, the dry vacuum pump 147.4 and the compressor 147.22 can be combined with one pump.

  Furthermore, a pump such as an ion pump or a getter pump can be used instead of the turbo molecular pump for the ultra-high vacuum pump used for exhausting the lens barrel. However, when these reservoir pumps are used, a circulation piping system cannot be constructed in this portion. Of course, other types of dry pumps such as a diaphragm type dry pump can be used instead of the dry vacuum pump.

FIG. 149 schematically shows the optical system and detector of the charged beam apparatus according to the present embodiment. Although the optical system is provided in the lens barrel, the optical system and the detector are merely examples, and any optical system and detector can be used as necessary. The optical system 149.1 of the charged beam apparatus includes a primary optical system 149.3 that irradiates the sample W placed on the stage 149.2 and a secondary electron emitted from the sample. And a second optical system 149.4. The primary optical system 149.3 includes an electron gun 149.5 that emits a charged beam, a lens system 149.6 that includes a two-stage electrostatic lens that focuses the charged beam emitted from the electron gun 149.5, and a deflection. 149.7, a Wien filter that deflects the charged beam so that its optical axis is perpendicular to the target surface, that is, an E.times.B separator 149.8, and a lens system 149.9 comprising two stages of electrostatic lenses. As shown in FIG. 149, they are arranged with the electron gun 149. 5 at the top, and the optical axis of the charged beam is inclined with respect to a line perpendicular to the surface (sample surface) of the sample W. Has been placed. The E × B deflectors 149 and 8 include electrodes 149 and 10 and magnets 149 and 11.

  The secondary optical system 149. 4 is an optical system into which secondary electrons emitted from the sample W are input, and is a two-stage electrostatic lens disposed above the E × B type deflectors 149 and 8 of the primary optical system. The lens system 149 * 12 which consists of these is provided. The detectors 149 and 13 detect the secondary electrons sent through the secondary optical system 149 and 4. Since the structures and functions of the constituent elements of the optical system 149.1 and the detectors 149.13 are the same as those of the prior art, detailed description thereof will be omitted.

  The charged beam emitted from the electron gun 149/5 is shaped by the square aperture of the electron gun, reduced by the two-stage lens system 149/6, the optical axis is adjusted by the polarizer 149-7, and the E × B deflection is performed. An image is formed in a square having a side of 1.925 mm on the deflection center plane of the device 149. The E × B deflectors 149 and 8 have a structure in which the electric field and the magnetic field are orthogonal to each other in a plane perpendicular to the normal line of the sample, and the relationship between the electric field, the magnetic field, and the electron energy satisfies a certain condition. Sometimes the electrons are caused to travel straight, and at other times they are deflected in a predetermined direction due to the mutual relationship between the electric field, magnetic field and electric field energy. In FIG. 149, it is set so that the charged beam from the electron gun enters the sample W perpendicularly and the secondary electrons emitted from the sample advance straight in the direction of the detectors 149 and 13. The shaped beam deflected by the E × B polarizer is reduced to 1/5 by the lens systems 149 and 9 and projected onto the sample W. The secondary electrons having the pattern image information emitted from the sample W are enlarged by the lens systems 149, 9, 149, 4 and the secondary electron images are formed by the detectors 149, 13. This four-stage magnifying lens is a distortion-free lens because the lens systems 149 and 9 form symmetrical doublet lenses and the lens systems 149 and 12 also form symmetrical doublet lenses.

According to the present embodiment, the following effects can be obtained.
(1) Using a stage having a structure similar to that of a static pressure bearing type stage generally used in the atmosphere (a stage supporting a static pressure bearing that does not have a differential exhaust mechanism) Processing with a charged beam can be performed stably.
(2) It is possible to minimize the influence of the charged beam irradiation area on the degree of vacuum, and it is possible to stabilize the processing of the charged beam on the sample.
(3) It is possible to provide an inspection apparatus with high accuracy in stage positioning performance and a stable vacuum degree in the charged beam irradiation area at low cost.
(4) It is possible to provide an exposure apparatus with high accuracy in positioning the stage and a stable degree of vacuum in the charged beam irradiation area at low cost.
(5) A fine semiconductor circuit can be formed by manufacturing a semiconductor with an apparatus having high accuracy in positioning of the stage and a stable degree of vacuum in the charged beam irradiation region.

3-2) Other Embodiments of Electron Beam Apparatus Furthermore, as another method in consideration of solving the problem of the mapping projection method, a plurality of primary electron beams are used, and the plurality of electron beams are two-dimensionally (XY). (Raster scan) while irradiating the observation area on the sample surface, and the secondary electron optical system employs a mapping projection method.

  This method has the advantages of the above-mentioned mapping projection method, and is a problem of this mapping method. (1) It is easy to charge up on the sample surface to collectively irradiate electron beams, (2) This method The fact that there is a limit to the electron beam current obtained in (1), which hinders the improvement of the inspection speed, can be solved by scanning a plurality of electron beams. That is, since the electron beam irradiation point moves, the charge easily escapes and the charge-up is reduced. Further, the current value can be easily increased by increasing the number of the plurality of electron beams. In the embodiment, when four electron beams are used, one electron beam current is 500 nA (electron beam diameter 10 μm), and a total of 2 μA is obtained. The number of electron beams can be easily increased to about 16, and in this case, 8 μA can be obtained in principle. The scanning of a plurality of electron beams only needs to be performed so that the irradiation amount of the plurality of electron beams is uniform in the irradiation region. Therefore, scanning of other shapes such as a Lissajous figure is not limited to the raster scanning as described above. It may be in shape. Therefore, the scanning direction of the stage need not be perpendicular to the scanning direction of the plurality of electron beams.

3-2-1) Electron gun (electron beam source)
A thermal electron beam source is used as the electron beam source used in this embodiment. The electron emission (emitter) material is LaB6. Other materials can be used as long as the material has a high melting point (low vapor pressure at high temperature) and a small work function. Two methods are used to obtain a plurality of electron beams. One is a method of obtaining a plurality of electron beams by drawing a single electron beam from one emitter (one protrusion) and passing a thin plate (open plate) with a plurality of holes. In this method, a plurality of protrusions are formed on a single emitter, and a plurality of electron beams are drawn directly therefrom. In either case, the electron beam utilizes the property of being easily emitted from the tip of the protrusion. Other electron beam sources such as a thermal field emission type electron beam or a Schottky type can also be used. Furthermore, the electron gun may emit a rectangular or linear beam, and in order to create its shape, it may be opened, and the shape of the electron generator (chip or filament, etc.) of the electron source is rectangular or linear. Anyway.

  The thermoelectron beam source is a system that emits electrons by heating the electron emission material. The thermoelectrolytic emission electron beam source emits electrons by applying a high electric field to the electron emission material. This is a system in which electron emission is stabilized by heating the emission part.

  FIG. 150A is a schematic diagram of an electron beam apparatus according to another embodiment. On the other hand, B in FIG. 150 is a schematic plan view T showing a mode in which a sample is scanned with a plurality of primary electron beams. The electron gun 150. 1 operable under the space charge limiting condition forms a multi-beam as indicated by reference numeral 150. 2 in FIG. The multi-beam 150.2 is composed of primary electron beams 150.3 which are eight circular beams arranged on the circumference.

  A plurality of primary electron beams 150. 3 generated by the electron gun 150. 1 are focused using lenses 150. 5 and 150. 6, and an E × B separator 150. 9 is incident at right angles to the sample W. A plurality of primary electromagnetic rays 150. Focused on the sample W by a primary optical system including these elements 150 · 4, 150 · 5, 150 · 6, 150 · 9, a lens 150 · 10 and an objective lens 150 · 11. 3 is used for scanning on the sample W by a two-stage deflector (not shown, included in the primary optical system) provided downstream of the lenses 150 and 6.

  The scanning of the sample W is performed in the x-axis direction with the main surface of the objective lens 150 or 11 as the deflection center. As shown in FIG. 150B, the primary electron beams 150 and 3 of the multi-beams 150 and 2 are arranged apart from each other on the circumference, and are on the y-axis orthogonal to the x direction that is the scanning direction. When projected, the distance between primary electron beams 150 and 3 adjacent to each other (measured at the center of each primary electron beam) is designed to be equal. At this time, the primary electron beams 150 and 3 adjacent to each other may be separated from each other, may be in contact with each other, or may partially overlap each other.

  The overlapping pitch may be set to an arbitrary value of 100 μm or less, preferably 50 μm or less, more preferably 10 μm or less. By making the pitch less than the beam shape, the beams can be brought into contact with each other to form a linear shape. Alternatively, a rectangular or linear beam formed from the beginning may be used.

  As shown in FIG. 150B, the respective primary electron beams 150.3 constituting the multi-beams 150.2 are arranged apart from each other, so that the current density limit value of each primary electron beam 150.3 is set. That is, the limit current density value that does not cause charging of the sample W can be maintained equivalent to the case where a single circular beam is used, thereby preventing the S / N ratio from being lowered. Further, since the primary electron beams 150 and 3 are separated from each other, the space charge effect is small.

  On the other hand, the multi-beams 150.2 can scan the sample W at a uniform density over the entire surface of the field of view 150.12 in one scan. Thereby, an image can be formed with high throughput, and the inspection time can be shortened. In FIG. 150B, if reference numeral 150.2 indicates a multi-beam at the start point of scanning, reference numeral 150.13 indicates a multi-beam at the end point of scanning.

  The sample W is placed on a sample table (not shown). This stage is continuously moved along a direction y orthogonal to the scanning direction x during scanning in the x direction (for example, scanning with a width of 200 μm). Thereby, raster scanning is performed. A driving device (not shown) for moving the stage on which the sample is placed is provided.

  Secondary electrons generated from the sample W during scanning and emitted in various directions are accelerated in the direction of the optical axis by the objective lenses 150 and 11, and as a result, secondary electrons emitted from various points in various directions respectively. Are converged finely, and the distance between the images is increased by the lenses 150, 10, 150, 11, 150, 14, 150, 15. Secondary electron beams 150, 16 formed through a secondary optical system including these lenses 150, 10, 150, 11, 150, 14, 150, 15 are projected onto the light receiving surfaces of detectors 150, 17, An enlarged image of the field of view is formed.

  Detectors 150 and 17 included in the optical optical system multiply the secondary electron beam by an MCP (microchannel plate), convert it to an optical signal by a scintillator, and convert it to an electrical signal by a CCD detector. A two-dimensional image of the sample W can be formed by an electrical signal from the CCD. Each primary electron beam 150. 3 has a dimension of at least two CCD pixels.

  By operating the electron gun 150.1 under the space charge limiting condition, the shot noise of the primary electron beam 150.3 can be reduced by about one digit compared with the case of operating under the temperature limiting condition. Therefore, the shot noise of the secondary electron signal can be reduced by an order of magnitude, so that a signal with a good S / N ratio can be obtained.

  According to the electron beam apparatus of the present embodiment, the S / N ratio is lowered by maintaining the current density limit value of the primary electron beam that does not cause charging of the sample to be equal to that when a single circular beam is used. The inspection time can be shortened by forming an image with high throughput while preventing the above-mentioned problem.

  In addition, the device manufacturing method according to the present embodiment can improve the yield by evaluating the wafer after the completion of each wafer process using the electron beam apparatus.

  FIG. 151 is a diagram showing details of the electron beam apparatus according to the embodiment of FIG. 150A. Four electron beams 151 and 2 (151 and 3 to 151 and 6) emitted from the electron gun 151 and 1 are shaped by the aperture stops 151 and 7, and the Wien filter is formed by the two-stage lenses 151 and 8 and 151 and 9. The image is formed in an ellipse of 10 μm × 12 μm on the deflection center planes 151 and 10, and is raster-scanned by the deflectors 151 and 11 in the direction perpendicular to the paper surface of the figure, and a rectangular area of 1 mm × 0.25 mm as the entire four electron beams The image is formed so as to cover uniformly. A plurality of electron beams deflected by E × B 151 · 10 are crossed over by an NA aperture, reduced to 1/5 by the lenses 151 · 11, covering the sample W by 200 μ × 50 μm, and perpendicular to the sample surface. Is irradiated and projected (called Koehler illumination). The four secondary electron beams 151 and 12 having the information of the pattern image (sample image F) emitted from the sample are enlarged by the lenses 15 and 11, 151 and 13, and 151 and 14, and are entirely displayed on the MCP 151 and 15. As a rectangular image (enlarged projection image F ′) synthesized by the four electron beams 151 and 12. The enlarged projected images F ′ by the secondary electron beams 151 and 12 are sensitized 10,000 times by the MCPs 151 and 15, converted into light by the fluorescent part, and synchronized with the continuous moving speed of the sample by the TDI-CCDs 151 and 16. It was acquired as a continuous image by the image display units 151 and 17 and output to a CRT or the like.

  The electron beam irradiation unit needs to irradiate the sample surface with the electron beam in a rectangular or elliptical shape with as little uniformity as possible and with less irradiation unevenness. Irradiation is necessary. Conventional non-uniformity of electron beam irradiation is about ± 10%, and the non-uniformity of image contrast is large. Further, since the electron beam irradiation current is as small as about 500 nA in the irradiation region, there is a problem that high throughput cannot be obtained. In addition, compared with the scanning electron microscope (SEM) method, this method has a problem in that an image formation failure due to charge-up is likely to occur because a large image observation area is collectively irradiated with an electron beam.

  A primary electron beam irradiation method of this embodiment mode is shown in FIG. The primary electron beam 152. 1 is composed of four electron beams 152. 2 to 152. 5, and each beam has an elliptical shape of 2 μm × 2.4 μm, and a rectangular of 200 μm × 12.5 μm per each. The areas are raster-scanned and added together so that they do not overlap to irradiate a rectangular area of 200 μm × 50 μm as a whole. The beam 151 · 2 arrives at 151 · 2 ′ in a finite time, and then returns to a position immediately below 15 · 1 shifted by the beam spot diameter (10 μm) with almost no time loss, and again in the same finite time as 151 · Move parallel to 2-151.2 'and directly below 151.2' (151-3 'direction) and repeat this for 1/4 of the rectangular irradiation area indicated by the dotted line (200 .mu.m.times.12.5 .mu.m). After scanning, the first point 152.1 is returned to and repeated at high speed.

The other electron beams 152 · 3 to 152 · 5 are repeatedly scanned at the same speed as the electron beam 152 · 2 and uniformly irradiate the rectangular irradiation region (200 µ x 50 µm) as a whole at high speed.
As long as the irradiation can be performed uniformly, the raster scan may not be performed. For example, scanning may be performed so as to draw a Lissajous shape. Therefore, the moving direction of the stage need not be the direction A shown in the figure. That is, it is not necessary to be perpendicular to the scanning direction (the high-speed scanning direction in the horizontal direction in the figure).

  In this embodiment, the unevenness of electron beam irradiation can be irradiated at about ± 3%. The irradiation current was 250 nA per electron beam, and 1.0 μA was obtained with four electron beams as a whole on the sample surface (twice as compared with the conventional method). By increasing the number of electron beams, current can be increased and high throughput can be obtained. Further, since the irradiation point is smaller than the conventional one (about 1/80 in area) and moved, the charge-up can be suppressed to 1/20 or less of the conventional one.

Although not shown in the figure, in addition to the lens, this apparatus includes a limited field stop, a deflector (aligner) having four or more poles for adjusting the axis of the electron beam, astigmatism. A unit necessary for illumination and imaging of an electron beam such as a corrector (stigmator) and a plurality of quadrupole lenses (quadrupole lenses) for shaping the beam shape are provided.

3-2-2) Electrode Structure FIG. 153 shows an electron beam apparatus having an electrode structure for preventing dielectric breakdown in an electron optical system using an electrostatic lens that irradiates a sample with an electron beam.

  To use a high-sensitivity, high-resolution electron beam device that uses an electron beam to inspect the surface state of a fine sample that has not been able to obtain sufficient sensitivity and resolution only by optical inspection. Is being studied.

  Such an electron beam apparatus is an inspection object by emitting an electron beam from an electron beam source and accelerating or converging the emitted electron beam by an electrostatic optical system such as an electrostatic lens. Incident on the sample. Next, by detecting a secondary electron beam emitted from the sample by the incidence of the electron beam, a signal corresponding to the detected secondary electron beam is generated, and for example, data of the sample is formed by this signal. The surface state of the sample is inspected based on the formed data.

  In an electron optical system using an electrostatic lens such as an electrostatic lens used in such an electron beam apparatus, an electrode that generates an electric field for accelerating or converging the electron beam has multiple stages in the optical axis direction of the electron beam. It is arranged. A predetermined voltage is applied to each of these electrodes, and the electron beam is accelerated or converged to a predetermined point on the optical axis by an electric field generated by the potential difference between the electrodes.

  In the conventional electron beam apparatus, a part of the electron beam emitted from the electron beam source may collide with the electrode regardless of the electric field in the electron optical system using the electrostatic lens. In this case, when the electron beam collides with the electrode, a secondary electron beam is emitted from the electrode itself. The amount of secondary electron beam emitted from the electrode varies depending on the material of the electrode or the material coating the electrode. When the secondary electron beam emitted from this electrode increases, this secondary electron beam is accelerated by the electric field of the electrode, ionizes the residual gas in the device, and this ion collides with the electrode. A secondary electron beam is emitted. Therefore, when a large amount of secondary electron beams are emitted, discharge easily occurs between the electrodes, and the probability of causing dielectric breakdown between the electrodes increases.

  For example, when the probability of dielectric breakdown was compared between the case where the electrode was coated with aluminum and the case where the electrode was coated with gold, the probability of dielectric breakdown between the electrodes was slightly higher in the case where aluminum was negative. Aluminum has a work function of 4.2 [eV], and gold has a work function of 4.9 [eV]. Here, the work function is the minimum energy (unit: eV) necessary for taking out one electron beam in a metal into a vacuum.

  Further, when the electrode is coated with gold and the sample of the electron beam apparatus is a semiconductor wafer, the gold is sputtered by the collision of the electron beam with the coated gold, and the surface of the semiconductor wafer Gold may adhere to the surface. When gold adheres to the semiconductor surface, gold is diffused into the silicon crystal in a later thermal process, and the performance of the transistor is deteriorated. Therefore, in this case, the electron beam apparatus is not suitable for inspection of a semiconductor wafer.

  On the other hand, for example, in an electrostatic lens of an electron optical system using an electrostatic lens, an electrostatic lens having a short focal length can be obtained by shortening the distance between the electrodes. When the focal length is short, the aberration coefficient of the electrostatic lens becomes small and the aberration becomes low, so that the electrostatic lens has a high resolution and the resolution of the evaluation apparatus is improved.

  In addition, an electrostatic lens having a short focal length can be obtained by increasing the potential difference applied between the electrodes of the electrostatic lens. Therefore, as in the case where the distance between the electrodes is shortened, the electrostatic lens has low aberration and high resolution, and the resolution of the electron beam apparatus is improved. Therefore, if the distance between the electrodes is shortened and the potential difference between the electrodes is increased, the electrostatic lens can synergistically achieve high resolution with low aberration. However, when the distance between the electrodes is shortened and the potential difference between the electrodes is increased, there is a problem that electric discharge is likely to occur between the electrodes and the probability of causing dielectric breakdown between the electrodes increases.

  Conventionally, the insulation between electrodes has been maintained by inserting an insulating material between the electrodes and supporting the electrode with the insulating material. Further, the insulation performance of the surface of the insulating material has been improved by increasing the shortest creepage distance (insulating surface length) of the insulating material between the electrodes. For example, the shortest creepage distance between the electrodes has been increased by forming the surface of the insulating material into a pleated shape in the interelectrode direction.

  However, in general, the processing of the surface of the insulating material is more difficult than metal processing, and the processing cost becomes expensive. Further, if the surface of the insulating material has a pleated shape or the like, the surface area of the insulating material becomes large. Therefore, when the inside of the electron beam apparatus is vacuum, the amount of gas released from the insulating material may increase. Therefore, the degree of vacuum is deteriorated, and the breakdown voltage between the electrodes is often lowered.

  The embodiment of FIG. 153 has been proposed to solve such a problem. Hereinafter, an electron beam apparatus according to this embodiment that can prevent dielectric breakdown between electrodes of the electrostatic optical system will be described. When applied to a mapping projection type evaluation apparatus having an electrostatic optical system, the configuration and operation of the mapping projection type evaluation apparatus and a device manufacturing method using the apparatus will be described.

  In FIG. 153, in the projection type evaluation apparatus 153-1, the electron beam that irradiates the sample has a predetermined emission surface, and the secondary electron beam emitted from the sample by the electron beam irradiation also has a predetermined emission surface. have. An electron beam having a two-dimensional area, for example, a rectangular radiation surface, is emitted from the electron beam source 155.3, and is magnified by a predetermined magnification by the electrostatic lens system 155.3. The expanded electron beam is incident on the E × B type deflector 153,4 from obliquely above, and the semiconductor wafer 153.5, which is a sample, is generated by a field where the electric field and magnetic field of the E × B deflector 153,4 are orthogonal to each other. (The solid line in FIG. 153).

  The electron beam deflected toward the semiconductor wafers 153 and 5 by the E × B type deflectors 153 and 4 is decelerated by the electric field generated by the voltage applied to the electrodes in the electrostatic objective systems 153 and 6, An image is formed on the semiconductor wafer 153/5 by the electric objective lens system 153/6.

  Next, the secondary electron beam generated by the electron beam irradiation to the semiconductor wafers 153 and 5 is accelerated in the direction of the detectors 153 and 7 by the electric field of the electrostatic objective lens system 153 and 6 (dotted line in FIG. 153). , And enters the E × B deflector 153. The E × B deflectors 153 and 4 direct the accelerated secondary electron beam toward the electrostatic intermediate lens system 153 and 8, and then detect the secondary electron beam by the electrostatic intermediate lens system 153 and 8. The secondary electron beam is detected by being incident on the detectors 153 and 7. The secondary electron beams detected by the detectors 153 and 7 are converted into data and transmitted to the display devices 153 and 9, and an image of the secondary electron beam is displayed on the display devices 153 and 9. Inspect the pattern.

Next, regarding the configuration of the electrostatic lens system 153 and 3, the electrostatic objective lens system 153 and 6, the electrostatic intermediate lens system 153 and 8, and the E × B type deflector 153 and 4 in the mapping projection type evaluation apparatus 153 and 1. This will be described in detail. Electrostatic lens systems 153 and 3 through which electron beams pass, and electrostatic objective lens system 1
The electrostatic intermediate lens systems 153 and 8 through which the secondary electron beam passes and a plurality of electrodes for generating a predetermined electric field are included. Moreover, platinum is coated on the surface of all these electrodes. Furthermore, the surfaces of the electrodes 153 and 10 of the E × B deflectors 153 and 4 are also coated with platinum.

  Here, with reference to FIG. 154, the dielectric breakdown occurrence probability for each metal coating the electrode will be described. The dielectric breakdown occurrence rate is represented by a relative magnitude relationship for each metal. Further, in the projection type evaluation apparatus, the other inspection conditions except for the type of metal coating the electrode were the same.

  First, comparing the probability of dielectric breakdown occurring when the metal coated with the electrode is aluminum and gold, the probability of dielectric breakdown of the electrode was slightly lower in the case of gold. Therefore, gold is more effective in preventing dielectric breakdown. Furthermore, comparing the probability of dielectric breakdown occurring when the metal coated with the electrode is gold and platinum, the probability of occurrence of dielectric breakdown of the electrode was even lower in the case of platinum.

  Here, the work function of each metal is 4.2 [eV] for aluminum, 4.9 [eV] for gold, and 5.3 [eV] for platinum. The work function of a metal is the minimum energy (unit: eV) required to take out one electron beam in the metal into a vacuum. That is, the larger the work function value, the more difficult it is to extract the electron beam.

  Therefore, in the mapping projection type evaluation apparatus 153-1, when an electron beam emitted from the electron beam source 153-1 collides with the electrode, a metal having a large work function value (a metal having a large work function value is used as a main material). If the electrode is coated on the electrode, the number of secondary electron beams emitted from the electrode is reduced, so that the probability of dielectric breakdown of the electrode is also reduced. Therefore, a metal having a large work function is good to some extent. Specifically, if the work function of the metal coated on the electrode is 5 [eV], the probability of occurrence of dielectric breakdown of the electrode can be kept low.

  Further, as in this embodiment, when the sample to be inspected is the semiconductor wafer 153.5, and the metal coated on the electrode is gold, the electron beam collides with gold, Gold sometimes adhered on the pattern of the semiconductor wafers 153.5. Therefore, in this embodiment, when the metal to be coated on the electrode is platinum, platinum does not adhere on the pattern of the semiconductor wafer 153.5, and even if platinum adheres, the device performance It will not deteriorate. Furthermore, the probability of occurrence of dielectric breakdown of the electrode can be lowered, which is more preferable.

  Next, an example of the shape and configuration of the electrode will be described with reference to FIGS. In FIG. 155, electrodes 155. 1 are electrodes of electrostatic lenses included in the electrostatic lens systems 153 and 3, the electrostatic objective lens systems 153 and 6, and the electrostatic intermediate lens systems 153 and 8.

  The electrode 155. 1 has a disk shape in which a passage hole through which an electron beam or a secondary electron beam can pass is substantially in the center. In the projection type evaluation apparatus 153 · 1 of this embodiment, the electrode 155. A predetermined voltage is applied to 155. 1 by a power supply device (not shown).

  FIG. 156 is a partial sectional view of the surface portion of the electrode 155. The surfaces of the electrodes 153 and 10 of the E × B deflectors 153 and 4 may have the same configuration as the surface of the electrodes 155 and 1. The material of the electrode 155.1 is composed of silicon copper (silicon bronze) 156.1, and titanium 156.2 is sputtered to a thickness of 50 nm on the silicon copper 156.1 processed into a required size and shape. Coating is performed, and further, platinum 156.3 is sputter-coated to a thickness of 200 nm on titanium 156.2 to form electrode 155.1.

  Here, with reference to FIG. 157 and FIG. 158, an electrode configuration for preventing dielectric breakdown between electrodes when the potential difference between the electrodes is large will be described in detail in this embodiment. The electrodes 157, 1, 157, and 2 in FIG. 157 are electrodes included in, for example, the electrostatic objective lens system 153 and 6, and the electrodes are coated with platinum as described above. In addition, a predetermined voltage is applied to the electrodes 157. In this embodiment, a high voltage, for example, a voltage of 15 kV is applied to the electrodes 157 and 2 on the semiconductor wafers 153 and 5 side, and a voltage of 5 kV is applied to the electrodes 157 and 1.

  The passage hole 157. 3 through which the electron beam or the secondary electron beam passes is located at the center of the electrodes 157, 1, 157, and 2). Is formed. By this electric field, the electron beam is decelerated and converged, and is irradiated onto the semiconductor wafers 153 and 5. At this time, since the potential difference between the electrodes is large, the electrostatic objective lens system 153, 6 can be an electrostatic objective lens having a short focal length. Therefore, the electrostatic objective lens system 153, 6 has low aberration and high resolution.

  Insulating spacers 157, 4 are inserted between the electrodes 157, 1, 157, 2, and the insulating spacers 157, 4 support the electrodes 157, 1, 157, 2 substantially vertically. The shortest creepage distance between the electrodes of the insulating spacers 157 and 4 is substantially the same as the distance between the electrodes in the supported electrode portion. That is, the surfaces of the insulating spacers 157 and 4 between the electrodes are not creased or the like in the direction between the electrodes, but are almost straight.

  The electrodes 157 and 2 have a first electrode surface 157 and 5 that is the shortest distance between the electrodes, and a second electrode surface 157 and 6 that has a longer interelectrode distance than the first electrode surface 157 and 5; Between the first electrode surfaces 157 and 5 and the second electrode surfaces 157 and 6, there are steps 157 and 7 (FIG. 158) in the direction between these two electrodes. The insulating spacers 157 and 4 support the electrodes 157 and 2 by the second electrode surfaces 157 and 6.

  Since the electrodes 157 and 2 have such a shape, the insulating spacers 157 and 4 are not processed into a pleated shape in the inter-electrode direction while keeping the shortest distance between the electrodes at a predetermined distance. It becomes possible to make the shortest creepage distance of 4 longer than the shortest distance between the electrodes. In addition, since a large electric field is not applied to the surfaces of the insulating spacers 157 and 4, a structure in which creeping discharge hardly occurs can be obtained.

  Therefore, the electrostatic objective lens system 135, 6 can be an electrostatic objective lens having a short focal length, and can have high resolution with low aberration, and the insulation performance between the electrodes of the insulating spacers 157, 4 does not deteriorate. Insulation breakdown between electrodes can be prevented. Further, since the metal electrodes 157 and 2 are processed so as to be provided with the steps 157 and 7, the processing cost is lower than that of processing the insulating spacers 157 and 4. In addition, the surface of the insulating spacers 157 and 4 in the inter-electrode direction has almost no uneven portions, and the amount of gas released from the insulating spacers 157 and 4 does not increase. Furthermore, since the corners of the opening ends 157 and 8 of the passing holes 157 and 3 of the electrodes 157 and 1 and the opening ends 157 and 9 of the passing holes 157 and 3 of the electrodes 157 and 2 are given curvature, Electric field is not concentrated on both corners, and dielectric breakdown between the electrodes can be further prevented. Furthermore, since the curvature is given to the corner portion between the electrodes of the steps 157 and 7 of the electrodes 157 and 2, the electric field does not concentrate on the corner portion, and the dielectric breakdown between the electrodes can be further prevented.

In this embodiment, the steps 157 and 7 are provided on the electrodes 157 and 2. However, the electrodes 157 and 1 may be processed so as to have a step in the direction of the electrodes 157 and 2, or the electrodes 157 and 2 may be processed. Instead of this, only the electrode 157. 1 may be processed so as to have a step in the direction of the electrode 157. In the electrostatic objective lens systems 153 and 6, the electrodes in which the insulating spacers 157 and 4 are inserted have been described. However, in other electrostatic lens systems, when there is an electrode having a large potential difference, the electrostatic lens system includes By applying, dielectric breakdown between electrodes can be prevented.

  The embodiment described with reference to FIGS. 153 to 158 is used for the inspection process in the device manufacturing method already described, thereby evaluating a semiconductor wafer without causing dielectric breakdown between electrodes of the electrostatic lens system. Is possible.

3-3) Embodiment relating to vibration damping device In this embodiment, an electron beam that performs at least one of processing, manufacturing, observation, and inspection of a substance by irradiating the target position of the substance with an electron beam More particularly, the present invention relates to an electron beam apparatus in which unnecessary mechanical vibration generated in a mechanical structure for positioning an electron beam is reduced, a method for damping the same, and processing, manufacturing, observation, and semiconductor device using the apparatus. The present invention relates to a semiconductor manufacturing process including a step of executing at least one of inspections.

  In general, methods for observing the fine structure of a substance using an electron beam include an inspection device for inspecting a defect of a pattern formed on a wafer or the like, a scanning electron microscope (SEM), etc., but the observation resolution is μm. Since it is ˜several tens of nanometers, it is necessary to observe the vibration from outside sufficiently. Also, in an apparatus that performs exposure using an electron beam, in order to deflect the electron beam and accurately irradiate the target position with a beam, a vibration isolator for sufficiently removing vibration from the outside is used, In addition, it is necessary to increase the rigidity in order to minimize the fluctuation caused by the mechanical resonance caused by the structure of the lens barrel portion. In order to increase the rigidity of the structure, there are physical dimensional constraints imposed by the electron optical system, so it is difficult to improve the rigidity by reducing the size, and therefore the rigidity can be improved by increasing the thickness and size of the lens barrel. Often done. However, the improvement in rigidity by this method has many disadvantages, including constraints on the degree of freedom in design, including the weight of the device, shape restrictions, the increase in size of the vibration isolation table, and the economic aspect. It was.

  In view of the above facts, the present embodiment is suitable so that unnecessary vibration due to resonance of the mechanical structure for positioning the beam can be maintained with high accuracy without necessarily improving the rigidity of the mechanical structure. By making it possible to attenuate, an electron beam apparatus that realizes relaxation of design constraints, reduction in size and weight of the apparatus, and improvement in economic efficiency, and efficient production using the apparatus in a semiconductor device manufacturing process, Provide a semiconductor manufacturing process that enables inspection, processing, observation, etc.

  FIG. 159 shows a configuration when the present embodiment is applied to an electron beam inspection apparatus that inspects a defect of a semiconductor wafer using an electron beam. The electron beam inspection apparatus 159. 1 shown in the figure is of a so-called mapping projection type, and has a mechanical structure of an A block and a B block protruding obliquely upward from the A block. A primary electron beam irradiation means for irradiating the primary electron beam is arranged in the B block, and a mapping projection optical system for mapping and projecting the secondary electron beam and an intensity of the secondary electron beam are detected in the A block. Imaging means. The A block is connected to the bottom fixed base 159.

  The primary electron beam irradiation means arranged in the B block includes an electron beam source 159, 3 composed of a cathode and an anode for emitting and accelerating the primary electron beam, and a rectangular opening 159, 4 for shaping the primary electron beam into a rectangle. And quadrupole lenses 159 and 5 for reducing the primary electron beam to form an image. Below the A block, an E × B deflector 159, 7 that deflects the reduced primary electron beam so that it strikes the semiconductor wafer 159, 6 in a direction perpendicular to the electric field E and the magnetic field B, and an aperture aperture (NA). ) 159.8, and objective lenses 159.9 for imaging the primary electron beam that has passed through the aperture aperture on the wafer 159.6.

Here, the primary electron beam reduced by the quadrupole lens 159.
For example, an image of 500 μm × 250 μm, for example, is formed on the deflection main surfaces 9 and 7, and simultaneously, a crossover image of the electron beam source 159 and 3 is formed on the aperture apertures 159 and 8 so that the Kohler proof condition is satisfied. For example, an image of 100 μm × 50 μm is formed on the wafer 159 and 6 by the objective lenses 159 and 9, and the area is illuminated.

  The wafers 159 and 6 are disposed in a sample chamber (not shown) that can be evacuated to a vacuum, and are disposed on stages 159 and 10 that are movable in an XY horizontal plane. Here, the relationship between the A block and the B block and the XYZ orthogonal coordinate system is shown in FIG. The wafer surface is on the XY horizontal plane, and the Z axis is substantially parallel to the optical axis of the mapping projection optical system. The stages 159 and 10 move in the XY horizontal plane with the wafers 159 and 6 placed thereon, whereby the inspection surfaces of the wafers 159 and 6 are sequentially scanned by the primary electron beam. The stages 159 and 10 are placed on the fixed bases 159 and 2.

  The mapping projection optical system disposed on the upper part of the A block includes intermediate electrostatic lenses 159 and 11 and projection electrostatic lenses 159 and 12 and diaphragms 159 and 13 disposed between these lenses. The secondary electron beam, the reflected electron beam, and the scattered electron beam emitted from the wafer 159, 6 by the irradiation of the primary electron beam are enlarged and projected at a predetermined magnification (for example, 200 to 300 times) by the mapping projection optical system, An image is formed on the lower surface of microchannel plates 159 and 14 to be described later.

  The imaging means arranged at the top of the A block includes microchannel plates 159 and 14, fluorescent screens 159 and 15, relay lenses 159 and 16, and imaging units 159 and 17. The microchannel plates 159 and 14 are provided with a number of channels in the plate, and a larger number of secondary electron beams imaged by the electrostatic lenses 159 and 11 and 159 and 12 pass through the channels. The electron beam is generated. That is, the secondary electron beam is amplified. The fluorescent screens 159 and 15 emit fluorescence with intensity corresponding to the intensity of the secondary electron beam when irradiated with the amplified secondary electron beam. That is, the intensity of the secondary electron beam is converted into the intensity of light. Relay lenses 159 and 16 are arranged so as to guide the fluorescence to the imaging units 159 and 17. The imaging units 159 and 17 are configured by a number of CCD imaging elements for converting the light guided by the relay lenses 159 and 16 into electric signals. In order to improve the S / N ratio of the detection signal, it is preferable to use a so-called TDI detector. Although irradiation with the primary electron beam generates not only a secondary electron beam but also a scattered electron beam and a reflected electron beam, they are collectively referred to as a secondary electron beam here.

  By the way, the lens barrel 160. 1 composed of the mechanical structure of the A block and the B block connected to the A block usually has one or more natural vibration modes. The resonance frequency and resonance direction of each natural vibration mode are determined by the shape, mass distribution, size, internal machine layout, and the like. For example, as shown in FIG. 160 (b), the lens barrel 160.1 has at least the mode 1 of the natural vibration 160.2. In this mode 1, the lens barrel 160.1 swings at a frequency of 150 Hz, for example, substantially along the Y direction. An example of the transfer function of the lens barrel in this case is shown in FIG. In FIG. 161, the horizontal axis represents frequency and the vertical axis represents the logarithm of vibration amplitude A. This transfer function has a gain of a resonance magnification of 30 dB (about 30 times) at a resonance frequency of 150 Hz. Therefore, even when a minute vibration is applied from the outside, if the vibration includes a frequency component in the vicinity of 150 Hz, the frequency component is amplified about 30 times in this example to vibrate the lens barrel. As a result, harmful events such as blurring of the map are generated.

  In the prior art, in order to prevent this, the entire lens barrel is placed on a vibration isolation table to dampen external vibrations, and / or the thickness and structure of the lens barrel are reviewed to lower the resonance magnification. A major measure was taken.

  In the present embodiment, in order to avoid this, as shown in FIG. 160 (c), an actuator 160/4 for applying pressure vibration 160.3 to the lens barrel so as to cancel the vibration 160.2 is provided at the base of the A block. Install. The actuators 160 and 4 are electrically connected to vibration damping circuits 159 and 18.

  FIG. 162 shows a schematic configuration of the actuators 160 and 4 and the vibration damping circuits 159 and 18. As shown in the figure, the actuators 160 and 4 are composed of piezoelectric elements 162 and 4 in which a dielectric material 162 and 1 having a piezoelectric effect are sandwiched between electrodes 162 and 2 and 162 and 3, and the piezoelectric elements are connected to electrodes 162 and 3. And support bases 162 and 5 fixed to the fixed bases 159 and 2 for supporting from the side. The piezoelectric elements 162 and 4 are sandwiched between the A block of the lens barrel 160 and the support bases 162 and 5, the electrodes 162 and 2 are on the outer wall of the A block, and the electrodes 162 and 3 are the support bases 162.・ It is adhered to 5. As a result, the piezoelectric elements 162 and 4 are subjected to a positive pressure when the lens barrel 160. 1 is directed and a negative pressure when the lens barrel 160. The piezoelectric elements 162 and 4 are installed at effective positions to suppress vibrations 160 and 2 of the lens barrel 160 and 1. For example, it is preferable that the vibrations 160. 2 are arranged so that the directions of the vibrations 160. 2 are orthogonal to the electrodes 162.

The vibration attenuating circuits 159 and 18 include variable inductances 162 and 6 and resistors 162 and 7 connected in series between the electrodes 162 and 2 and 162 and 3 of the piezoelectric elements 162 and 4. Since the variable inductances 162 and 6 have an inductance L, the resistors 162 and 7 have a resistance value R _D , and the piezoelectric elements 162 and 4 have an electric capacitance C, the piezoelectric elements 162 and 4 and the vibration damping circuits 159 and 18 connected in series are , Which is equivalent to the series resonant circuit indicated by reference numerals 162 and 8. The resonant frequency f ₀ ′ of this series resonant circuit is
fo ′ = 1 / {2π (LC) ^1/2 }
It is represented by In the present embodiment, each parameter is set so that the resonance frequency f ₀ ′ of the series resonance circuit substantially matches the resonance frequency f ₀ of the lens barrel 160. That is, for a given capacitance C of the piezoelectric element 162,4,
fo = 1 / {2π (LC) ^1/2 }
The inductance L of the variable inductances 162 and 6 is adjusted so that is established. Actually, the capacitance C of the piezoelectric elements 162 and 4 is small in forming a resonance circuit in accordance with the mechanical resonance frequency, and therefore, in many cases, a very large inductance L is required. A resonant circuit can be realized by forming an equivalently large inductance using an operational amplifier or the like.

Further, the value R _D of the resistors 162 and 7 is selected so that the Q value of the resonance frequency component of the series resonance circuit substantially matches the Q value of the resonance component having a peak in the transfer function shown in FIG. The series resonant circuits 162 and 8 thus produced have electrical frequency characteristics indicated by reference numerals 161 and 1 in FIG.

  The electron beam inspection apparatus 159. 1 shown in FIG. 159 is controlled and managed by the control units 159 and 19. As illustrated in FIG. 159, the control units 159 and 19 can be configured by a general-purpose personal computer or the like. This computer includes a control unit main body 159/20 for executing various controls and arithmetic processes according to a predetermined program, a CRT 159/21 for displaying the processing results of the main body 159/20, and a keyboard and mouse for an operator to input commands. The control units 159 and 19 may be configured from hardware dedicated to the electron beam inspection apparatus, a workstation, or the like.

The control unit main bodies 159 and 20 include various control boards such as a CPU, RAM, ROM, hard disk, and video board (not shown). On a memory such as a RAM or a hard disk, secondary electron beam image storage areas 159 and 23 for storing electric signals received from the imaging units 159 and 17, that is, digital image data of secondary electron beam images of the wafer 159 and 6. Is assigned. Further, on the hard disk, there are reference image storage units 159 and 24 for storing reference image data of a wafer having no defect in advance. In addition to the control program for controlling the entire electron beam inspection apparatus, defect detection programs 159 and 25 are stored on the hard disk. The defect detection programs 159 and 25 control the movement of the stages 159 and 10 in the XY plane and perform various arithmetic processes such as addition on the digital image data received from the imaging units 159 and 17 during this period. The secondary electron beam image is reconstructed on the storage areas 159 and 23 from the obtained data. Further, the defect detection program 159/25 reads the secondary electron beam image data configured on the storage areas 159/23, and automatically detects defects on the wafer 159/6 based on the image data according to a predetermined algorithm. To do.

  Next, the operation of this embodiment will be described. A primary electron beam is emitted from the electron beam source 159. 3, and is set through the rectangular apertures 159. 4, the quadrupole lens 159. 5, the E × B deflector 159. 7 and the objective lens 159. Irradiate on the surface. As described above, a region to be inspected having a size of 100 μm × 50 μm, for example, is illuminated on the wafer 159. 6, and a secondary electron beam is emitted. This secondary electron beam is enlarged and projected onto the lower surfaces of the multi-channel plates 159 and 14 by the intermediate electrostatic lenses 159 and 11 and the projecting electrostatic lenses 159 and 12, picked up by the image pickup units 159 and 17, and on the wafers 159 and 6 A secondary electron beam image of the projected area is obtained. By driving the stages 159 and 10 and sequentially moving the wafers 159 and 6 in the XY horizontal plane for each predetermined width and executing the above procedure, an image of the entire inspection surface can be obtained.

If an external force including a vibration component of the resonance frequency f ₀ (150 Hz) is applied to the lens barrel 160 • 1 while the enlarged secondary electron beam image is being captured, the lens barrel 160 • 1 is transferred with its transfer function. The vibration component is amplified at a fixed resonance magnification (30 dB) to cause natural vibration. This vibration 160. 2 applies positive and negative pressure to the piezoelectric elements 162. The piezoelectric elements 162 and 4 once convert the vibration energy of the lens barrel 160 and 1 into electric energy and output it. Since the inductances 162 and 6 (L) and the resistors 162 and 7 (R _D ) are connected in series to the electrodes 162 and 2 and 162 and 3 of the piezoelectric elements 162 and 4, a resonance circuit is formed. At f ₀ , the capacitive impedance of the piezoelectric elements 162 and 4 and the inductive impedance L of the inductances 162 and 6 cancel each other, and the impedance of the resonance circuit is effectively only the resistance R _D. Therefore, at the time of resonance, almost all of the electric energy output from the piezoelectric elements 162 and 4 is consumed by the resistors 162 and 7 (R _D ).

  Thus, the piezoelectric elements 162 and 4 generate forces so as to cancel the external force applied from the lens barrel 160 and 1 to the piezoelectric elements 162 and 4, cancel the vibrations 160 and 2 generated by mechanical resonance, and the resonance magnification. Can be lowered. Since the secondary electron beam is magnified, the fluctuation of the map due to vibration becomes even larger. In this embodiment, however, the blur of the map due to such fluctuation can be prevented.

  As shown in FIG. 163, the resonance component of the transfer function 161.1 (corresponding to FIG. 161) of the lens barrel 160.1 as a mechanical structure is a series resonance circuit 162.8 having an electrical frequency characteristic 163-1. The lens barrel 160. 1 has an overall transfer function 163 · 2 having a low resonance magnification as a whole.

As described above, when a good secondary electron beam image having no mapping blur is obtained, the electron beam inspection apparatus 159. 1 according to this embodiment performs processing for inspecting the defects of the wafer 159. 6 from the image. Do. As this defect inspection processing, a so-called pattern matching method or the like can be used. In this method, the reference image read from the reference image storage units 159 and 24 is matched with the actually detected secondary electron beam image, and a distance value representing the similarity between the two is calculated. When this distance value is smaller than a predetermined threshold value, it is determined that the degree of similarity is high and “no defect” is determined. On the other hand, when the distance value is equal to or greater than a predetermined threshold value, it is determined that the degree of similarity is low and “defect” is determined. When it is determined that there is a defect, a warning may be displayed to the operator. At this time, you may make it display the secondary electron beam image 159 * 26 on the display part of CRT159 * 21. The pattern matching method may be used for each partial region of the secondary electron beam image.

  Other than the pattern matching method, for example, there is a defect inspection method shown in FIGS. FIG. 164 (a) shows an image 164 • 1 of the die detected first and an image 164 • 2 of the other die detected second. If the third detected image of another die is determined to be the same as or similar to the first image 164. 1, it is determined that the portion 164. 3 of the second die image 164. 2 has a defect. , Defective part can be detected.

  FIG. 164 (b) shows an example of measuring the line width of the pattern formed on the wafer. When the actual pattern 164/4 on the wafer is scanned in the direction 164/5, the actual intensity signal of the secondary electron beam is 164/6, and this signal is a threshold level 164 determined in advance by calibration. The width 166.4 of the portion continuously exceeding 7 can be measured as the line width of the pattern 164/4. When the measured line width is not within the predetermined range, it can be determined that the pattern has a defect.

  FIG. 164 (c) shows an example in which the potential contrast of the pattern formed on the wafer is measured. In the configuration shown in FIG. 159, axially symmetric electrodes 164, 9 are provided above the wafers 159, 6 and, for example, a potential of −10 V is applied to the wafer potential of 0 V. The equipotential surface of −2 V at this time has a shape shown by 14 · 10. Here, it is assumed that the patterns 164, 11 and 164, 12 formed on the wafer have potentials of −4V and 0V, respectively. In this case, since the secondary electron beam emitted from the patterns 164 and 11 has an upward velocity corresponding to the kinetic energy of 2 eV at the −2 V equipotential surface 164 and 10, the potential barrier 164. As shown by the trajectories 164 and 13, they escape from the electrodes 164 and 9 and are detected by the detector. On the other hand, the secondary electron beam emitted from the patterns 164 and 12 cannot be detected through the potential barrier of −2V and is driven back to the wafer surface as indicated by the trajectories 164 and 14. Therefore, the detected images of the patterns 164 and 11 are bright, and the detected images of the patterns 164 and 12 are dark. Thus, a potential contrast is obtained. If the brightness and potential of the detected image are calibrated in advance, the pattern potential can be measured from the detected image. The defect portion of the pattern can be evaluated from this potential distribution.

As described above, by performing each measurement as described above on a good secondary electron beam image having no mapping blur obtained by the present embodiment, a more accurate defect inspection can be realized.
When the electron beam inspection apparatus described so far as the present embodiment is used in the wafer inspection process in the device manufacturing method, it is possible to prevent the detection image from being deteriorated due to the vibration of the mechanical structure. Inspection is possible, and shipment of defective products can be prevented.

  In addition, this Embodiment is not limited only to what was demonstrated above, In the range of the summary of this invention, it can change arbitrarily suitably. For example, the mechanical resonance frequency and mode are not necessarily one, and a plurality of them are generally generated. In this case, the necessary number of actuators 160 and 4 can be installed at each important point of the lens barrel. It becomes possible. For example, when the mechanical structure block A shown in FIG. 160 (b) has vibration in the X direction as well as vibration in the Y direction 160.2, a separate actuator may be installed so as to cancel the vibration in the X direction. it can. Further, when there are independent natural vibrations in the B block and the D block, actuators may also be installed in these blocks.

  The vibration attenuating circuits 159 and 18 do not have to be equivalent to the series resonance circuits 162 and 8. When the mechanical natural vibration has a plurality of resonance frequencies in the same vibration direction, the electric frequency characteristics of the circuit have a plurality of characteristics. It can be countered with a resonance frequency.

  The actuator can be installed not only in the lens barrel but also in parts necessary for accurately positioning the beam position, for example, the XY stages 159 and 10 or optical parts of various optical instruments.

  Although the semiconductor wafer 159, 6 has been taken as an example of the sample to be inspected in the electron beam inspection apparatus of the present embodiment, the sample to be inspected is not limited to this, and any one that can detect defects with an electron beam is available. Selectable. For example, a mask or the like on which an exposure pattern for a wafer is formed can be the inspection target.

  Furthermore, the present embodiment can be applied to all electron beam application apparatuses that irradiate a target position of a material with a beam. In this case, the application range can be expanded not only to the inspection of the substance but also to an apparatus that performs at least one of processing, manufacturing, and observation. Naturally, the concept of a substance here is not only a wafer and the above-described mask, but also any object that can be inspected, processed, manufactured, and observed by a beam. Similarly, the device manufacturing method can be applied not only to the inspection during the manufacturing process of the semiconductor device, but also to the process itself of manufacturing the semiconductor device with a beam.

  In addition, although the structure shown in FIG. 159 was shown as an electron beam inspection apparatus of this Embodiment, the electron optical system etc. can be changed arbitrarily suitably. For example, the electron beam irradiating means of the electron beam inspection apparatus 159.1 is a type in which the primary electron beam is incident on the surface of the wafer 159.6 from vertically above, but the E.times.B deflector 159.7 is omitted. The primary electron beam may be incident obliquely on the surface of the wafer 159.

3-4) Embodiment Regarding Holding of Wafer In this embodiment, an electrostatic chuck for electrostatically holding and holding a wafer in an electron beam apparatus, a combination of a wafer and an electrostatic chuck, particularly a deceleration electric field objective lens The present invention relates to a combination of an electrostatic chuck and a wafer that can be used in the used electron beam apparatus, and a device manufacturing method using the electron beam apparatus including the combination of an electrostatic chuck and a wafer.

  In a known electrostatic chuck that electrostatically holds and fixes a wafer, an electrode layer disposed on a substrate is formed by a plurality of mutually insulated electrodes, and voltage is sequentially applied from one electrode to the other. A power supply device is provided. An electron beam apparatus using a decelerating electric field objective lens is known.

  When a wafer in the middle of a process is evaluated by an electron beam apparatus using a decelerating electric field objective lens, it is necessary to apply a negative high voltage to the wafer. In this case, if a negative high voltage is suddenly applied, the device in the middle of the process may be destroyed. Therefore, it is necessary to gradually apply the voltage.

On the other hand, most wafers have an insulating film such as SiO ₂ or a nitride film attached to the side and back surfaces of the wafer, so that there is a problem that no voltage is applied when trying to apply a zero or low potential to the wafer. It was. Furthermore, a wafer whose center is convexly distorted toward the electrostatic chuck can be attracted and fixed relatively easily, but a wafer whose center is distorted into a concave on the chuck side is chucked only at the periphery in a single-pole electrostatic chuck. There is a problem that the central portion is held without being chucked.

In order to solve the above problems, this embodiment can be used with a decelerating electric field objective lens, and can be used to chuck a wafer whose side and back surfaces are covered with an insulating film and whose center is distorted concavely toward the chuck side. In addition to providing an electric chuck and a combination of a wafer and an electrostatic chuck, a device manufacturing method for evaluating a wafer in the middle of a process using such an electrostatic chuck or a combination of a wafer and an electrostatic chuck is provided. To do.

  FIG. 165 is a plan view of the electrostatic chuck 1410 according to the present embodiment, in which the wafer is removed and the electrode plate 165. 1 is viewed. FIG. 166 is a schematic sectional view in the vertical direction along the line MM of the electrostatic chuck of FIG. 165, and shows a state where a wafer is placed and no voltage is applied. As shown in FIG. 166, the electrostatic chuck 165.2 has a laminated structure including a substrate 166.1, an electrode plate 166.2, and an insulating layer 166.3. The electrode plates 166, 2 include first electrodes 165, 2 and second electrodes 165, 3. The first electrodes 165 and 2 and the second electrodes 165 and 3 are separated so that voltages can be applied separately, and are formed as thin films so that they can move at high speed without generating eddy currents in a magnetic field.

The first electrode 165. 2 is composed of a central part and a part of the peripheral part of the circular electrode plate 166. 2 in a plan view, and the second electrode 165. 3 is composed of the remaining horseshoe-shaped peripheral part of the electrode plate. An insulating layer 166.3 is disposed above the electrode plate 166.2. The insulating layers 166 and 3 are formed of a sapphire substrate having a thickness of 1 mm. Since sapphire is a single crystal of alumina and does not have any small holes like alumina ceramics, it has a high dielectric breakdown voltage. For example, a 1 mm thick sapphire substrate can sufficiently withstand a potential difference of 10 ⁴ V or more.

  The voltage is applied to the wafers 166 and 4 through the contacts 166 and 5 having a knife-edge metal portion. As shown in FIG. 166, the two contacts 166. 5 are brought into contact with the side surfaces of the wafers 166. The reason why the two contacts 166 and 5 are used is that when only one contact is used, there is a possibility that the continuity cannot be obtained and that the force to push the wafers 166 and 4 to one side is not generated. Although the insulation layer (not shown) is destroyed to make it conductive, there is a risk of scattering of particles when discharging, so the contacts 166, 5 are connected to the power source 166, 7 via the resistors 166, 6 However, a large discharge was not generated. If the resistors 166 and 6 are too large, a conduction hole is not formed. If the resistors 166 and 6 are too small, a large discharge occurs and particles are scattered. Therefore, an allowable resistance value is determined for each insulating layer (not shown). This is because the thickness of the insulating layer varies depending on the history of the wafer, and it is necessary to determine an allowable resistance value for each wafer.

FIG. 167 (a) shows a time chart of voltage application. As shown by line A, 4 kV is applied to the first electrode at time t = 0. At time t = t ₀ when both the central portion and the peripheral portion of the wafer are chucked, 4 kV is applied to the second electrode as indicated by line B. At time t = t ₁ , the voltage C of the wafer is gradually increased (decreased) and controlled to reach −4 kV at time t = t ₂ . The voltage of the first electrode and the second electrode is gradually decreased from time t = t ₁ to time t = t ₂ , and is set to 0 V at time t = t ₂ .

At the time t = t _{3 when} the evaluation of the wafer held by chucking is held, the wafer voltage C is set to 0 V, and the wafer is taken out.
When the electrostatic chuck holds and holds the wafer with a potential difference of 2 kV even if there is no potential difference of 4 kV, a voltage A ′ of 2 kV is applied to the first electrode and the second electrode, respectively, as shown by a one-dot chain line in FIG. , B ′ are applied. When −4 kV is applied to the wafer, −2 kV is applied to the first electrode and the second electrode, respectively. In this manner, voltage application can be prevented from applying more voltage to the insulating layer 2104 than necessary, so that the insulating layer can be prevented from being broken.

FIG. 168 is a block diagram showing an electron beam apparatus including the electrostatic chuck described above. The electron beam emitted from the electron beam source 168. 1 is subjected to removal of an unnecessary beam by the opening of the anode 168 • 2 that determines the aperture aperture (NA), and is reduced by the condenser lens 168 • 7 and the objective lens 168 • 13. An image is formed on the wafer 166 · 4 to which −4 kV is applied, and the wafer 166 · 4 is scanned by the deflectors 168 · 8 and 168 · 12. The secondary electron beams emitted from the wafers 166 and 4 are collected by the objective lenses 168 and 13, and bent to the right by about 35 ° by the E × B separators 168 and 12, and the secondary electron beam detectors 168 and 10 are bent. And an SEM image on the wafer is obtained. In the electron beam apparatus of FIG. 168, reference numerals 168, 3, 168, and 5 are axis alignment instruments, 168 and 4 are astigmatism correction instruments, 168 and 6 are aperture plates, 168 and 11 are shields, and 168 and 14 are electrodes. . The electrostatic chuck described with reference to FIGS. 166 and 167 is disposed below the wafers 166 and 4.

  By using this embodiment for an inspection process in a device manufacturing method, even a semiconductor device having a fine pattern can be inspected with high throughput, 100% inspection can be performed, product yield can be improved, and shipment of defective products can be prevented. is there.

  The method of increasing or decreasing the voltage applied to the electrostatic chuck is not limited to that shown in FIG. For example, a voltage that changes exponentially as shown in FIG. In short, any voltage may be used as long as it reaches the predetermined voltage in time.

  As described above, the first to twelfth embodiments of the present invention have been described in detail. In any of the embodiments, the term “predetermined voltage” means a voltage at which measurement such as inspection is performed. It shall be.

  Moreover, although various embodiments described so far use an electron beam as a charged particle beam, the present invention is not limited to this, but a charged particle beam other than an electron beam, a neutron beam having no charge, a laser beam, or the like. It is also possible to use uncharged particle beams such as electromagnetic waves.

When the charged particle beam device according to the present invention is operated, the target substance floats and is attracted to the high-pressure region by proximity interaction (charge of particles near the surface), and thus is used for the formation and deflection of charged particle beams. Organic materials are deposited on various electrodes. Organic substances that gradually accumulate due to surface charging adversely affect the formation of charged particle beams and the deflection mechanism, so these deposited organic substances must be removed periodically. Therefore, in order to periodically remove the deposited organic material, hydrogen, oxygen, or fluorine and alternatives including HF, H ₂ O, and the like are used in vacuum using an electrode near the region where the organic material is deposited. By creating plasma such as C _M F _N and maintaining the plasma potential in the space at a potential (several kV, for example, 20 V to 5 kV) at which sputtering occurs on the electrode surface, only organic substances are oxidized, hydrogenated, or fluorinated. It is preferable to remove.

3-5) Embodiment of E × B Separator FIG. 169 shows an E × B separator 169. 1 according to the present embodiment. The E × B separator 169. 1 includes an electrostatic deflector and an electromagnetic deflector. In FIG. 169, the E × B separator 169. 1 is on the xy plane orthogonal to the optical axis (axis perpendicular to the drawing: z-axis). It is shown as a cross-sectional view. The x-axis direction and the y-axis direction are also orthogonal.

  The electrostatic deflector includes a pair of electrodes (electrostatic deflection electrodes) 169, 2 provided in a vacuum vessel, and generates an electric field E in the x-axis direction. These electrostatic deflection electrodes 169, 2 are attached to the vacuum walls 169, 4 of the vacuum vessel via insulating spacers 169, 3, and the distance D between these electrodes is the y-axis of the electrostatic deflection electrodes 169, 2 The direction length is set to be smaller than 2L. With such a setting, the range in which the electric field strength formed around the z-axis is uniform can be made relatively large. Ideally, if D <L, the electric field strength is in a uniform range. Can be made larger.

  That is, since the electric field strength is not uniform in the range of D / 2 from the edge of the electrode, the region where the electric field strength is almost uniform is 2LD of the central portion excluding the non-uniform edge region. It becomes an area. For this reason, in order for a region having a uniform electric field strength to exist, it is necessary to satisfy 2L> D. Furthermore, by setting L> D, a region having a uniform electric field strength becomes larger.

  An electromagnetic deflector for generating a magnetic field M in the y-axis direction is provided outside the vacuum walls 169 and 4. The electromagnetic deflector includes electromagnetic coils 169 and 5 and electromagnetic coils 169 and 6, which generate magnetic fields in the x-axis direction and the y-axis direction, respectively. The coil 169, 6 alone can generate the magnetic field M in the y-axis direction, but a coil that generates a magnetic field in the x-axis direction is provided in order to improve the orthogonality between the electric field E and the magnetic field M. That is, the orthogonality between the electric field and the magnetic field can be improved by canceling the + x-axis direction generated by the coil 169.6 by the -x-axis direction magnetic field component generated by the coil 169.6.

  These magnetic field generating coils 169 and 5 and 168 and 6 are provided outside the vacuum vessel, so that each of them is divided into two parts, attached from both sides of the vacuum walls 169 and 4, and screwed at the portions 169 and 7. It is only necessary to tighten and integrate them.

  The outermost layers 169 and 8 of the E × B separator are configured as permalloy or ferrite yokes. Similarly to the coils 169, 5, 169, 6, the outermost layers 169, 8 may be divided into two parts and attached to the outer periphery of the coils 169, 6 from both sides, and may be integrated by screwing or the like at the portions 169, 7.

FIG. 170 shows a cross section orthogonal to the optical axis (z axis) of the E × B separator 170. 1 according to the present embodiment. 170 is different from the E × B separator of the embodiment shown in FIG. 169 in that six poles of electrostatic deflection electrodes 170. 1 are provided. These electrostatic deflection electrodes 170. 1 have an angle θ _i (i = 0, 1, 1) between the line connecting the center of each electrode and the optical axis (z-axis) and the direction of the electric field (x-axis direction). when a 2,3,4,5), a voltage k · cosθ _i (k proportional to cos [theta] _i is a constant) is supplied. However, θ _i is an arbitrary angle.

  In the embodiment shown in FIG. 170 as well, only the electric field E in the x-axis direction can be generated, so coils 169.5 and 169.6 that generate magnetic fields in the x- and y-axis directions are provided to correct the orthogonality. According to the present embodiment, it is possible to further increase the region where the electric field strength is uniform as compared with the embodiment shown in FIG.

In the E × B separator of the embodiment shown in FIGS. 169 and 170, the coil for generating the magnetic field is formed in a saddle type, but a toroidal type coil may be used.
In the E × B separator 169. 1 of FIG. 169, the parallel plates formed as the pair of electrodes of the electrostatic deflector that generates the electric field are longer in the direction perpendicular to the optical axis than the distance between the electrodes. Since the mold electrode is used, a region where a parallel electric field with a uniform intensity is generated around the optical axis is widened.

  In the E × B separator of FIGS. 169 and 170, a saddle type coil is used for the electromagnetic deflector, and the angle at which the coil is viewed from the optical axis is set to 2π / 3 on one side, so that a 3θ component is generated. Without this, a region where a parallel magnetic field with a uniform intensity is generated around the optical axis is widened. Furthermore, since the magnetic field is generated by the electromagnetic coil, a deflection current can be superimposed on the coil, thereby providing a scanning function.

Since the E × B separator of FIGS. 169 and 170 is configured as a combination of an electrostatic deflector and an electromagnetic deflector, the aberration of the electrostatic deflector and the lens system is calculated, and separately from this, the electromagnetic deflection is performed. The aberrations of the optical system can be obtained by calculating the aberrations of the optical system and the lens system and summing these aberrations.

3-6) Embodiment of Production Line FIG. 171 shows an example of a production line using the apparatus of the present invention. Read out information such as the lot number of the wafer to be inspected by the inspection apparatus 171, 1 and the history of the manufacturing apparatus through the manufacture from the memory provided in the SMIF or FOUP 171-2, or read the lot number as SMIF, FOUP or It can be recognized by reading the ID number of the wafer cassette. During the transfer of the wafer, the amount of moisture is controlled to prevent the metal wiring from being oxidized.

  The defect inspection apparatus 171. 1 can be connected to the network system of the production line. Through this network system 171. 3, the production line control computer 171. Information such as the lot number of the wafer to be inspected and the inspection result can be sent to the apparatuses 171 and 5 and another inspection apparatus. The manufacturing apparatus includes lithography-related apparatuses such as exposure apparatuses, coaters, curing apparatuses, developers, etc., film forming apparatuses such as etching apparatuses, sputtering apparatuses, and CVD apparatuses, CMP apparatuses, various measuring apparatuses, other inspection apparatuses, and review apparatuses. Etc. are included.

3-7) Embodiments Using Other Electrons In the present invention, an electron whose information on the surface of a substrate is obtained by irradiating a sample such as a substrate on which a wiring pattern having a book with a line of 100 nm or less is irradiated with an electron beam. It is an essential object to obtain a substrate surface image from the detected electrons and to inspect the sample surface. In particular, when irradiating a sample with an electron beam, an electron beam having an area including a certain imaging region is irradiated, and electrons emitted from the imaging region on the substrate are detected on a detector using a CCD or a CCD-TDI. An image of the imaging region is obtained by forming an image, and the obtained image is further compared with the SEM method by performing inspection by appropriately combining cell inspection and die comparison inspection according to the die pattern. Have proposed an inspection method and apparatus that realizes a very fast throughput. That is, according to the inspection method and inspection apparatus using the electron beam of the present invention, the optical inspection apparatus cannot sufficiently inspect the pattern defect of the wiring having a line width of 100 nm or less due to the low resolution, while the SEM type inspection. The system eliminates both problems of being unable to meet the demand for high throughput because it takes too much time for inspection, and enables inspection of wiring patterns having a line width of 100 nm or less with sufficient resolution and high throughput. ing.

  In the inspection of a sample, from the viewpoint of resolution, it is desirable to obtain an image of the substrate surface by colliding an electron beam with the substrate and detecting electrons emitted from the substrate. Therefore, in the embodiments of the present invention, the description has been given mainly with an example focusing on secondary electrons, reflected electrons, and backscattered electrons emitted from the substrate. However, the detected electrons can be anything as long as they can obtain information on the surface of the substrate. For example, by forming a reverse electric field near the substrate, the electrons are reflected near the substrate without directly colliding with the substrate. Mirror electrons (also referred to as reflected electrons in a broad sense) or transmitted electrons that pass through the substrate may be used. In particular, when mirror electrons are used, the electrons do not directly collide with the sample, so that there is an advantage that the influence of charge-up is extremely small.

  When mirror electrons are used, a negative potential lower than the acceleration voltage is applied to the sample to form a reverse electric field near the sample. This negative potential is preferably set to a value such that most of the electron beam is returned near the surface of the substrate. Specifically, it may be set to a potential lower by 0.5 to 1.0 V or more than the acceleration voltage of the electron gun. For example, in the case of the present invention, when the acceleration voltage is −4 kV, the voltage applied to the sample is preferably set to −4.00 kV to −4.050 kV. More desirably, it is preferably −4.005 kV to −4.020 kV, and more preferably −4.005 kV to −4.010 kV.

  When using transmission electrons, when the acceleration voltage is set to −4 kV, the voltage applied to the sample is 0 to −4 kV, preferably 0 to −3.9 kV, more preferably 0 to −3.5 kV. It is preferable to set to.

In addition, although not an electron beam, an X-ray may be used. The secondary system and die comparison of the present invention are sufficiently applicable.
Regardless of whether mirror electrons or transmission electrons are used, the electron gun, the primary optical system, the deflector for separating the primary electron beam and the detection electron beam, the detector using CCD or CCD-TDI, and image processing A device, an arithmetic unit for die comparison, and the like are used. An electron beam having a certain region such as an ellipse is used. Of course, a finely focused electron beam used for the SEM type may be used. Of course, the electron beam may be singular or plural. The deflector for separating the primary electron beam and the detection electron beam may be a Wien filter that forms both an electric field and a magnetic field, or a deflector having only a magnetic field. The detector uses a CCD or a CCD-TDI that forms an image of the imaging region on the detector and can perform a quick inspection. When an SEM type electron gun is used, a corresponding semiconductor detector or the like It is natural to use. When acquiring an image of the substrate surface and performing a die inspection, a cell inspection applied to a periodic pattern and a die-to-die comparison inspection applied to a random pattern are appropriately performed according to the die pattern. Use. Of course, all may be processed by comparison inspection between dies. In the case of comparative inspection between dies, dies on the same substrate may be compared, dies on different substrates may be compared, or die and CAD data may be compared. . Any suitable one may be used arbitrarily. Further, the substrate is aligned before the inspection. Measure the positional deviation of the substrate and correct the rotational angle deviation. At that time, a focus map may be created, and the inspection may be performed while correcting the position of the substrate on the plane and the shift of the focus in consideration of them during the inspection.

  In addition, when using the apparatus of the present invention in the manufacturing process, information on a wafer to be inspected is obtained from a computer connected to a network system and controlling the production line, inspection results are transmitted, and the production line is transmitted. It is desirable to reflect this in the production conditions of each device.

3-8) Embodiment Using Secondary Electrons and Reflected Electrons In this embodiment, the inspection target can be irradiated with a surface beam, and the secondary electrons and the reflected electrons are switched and used according to the inspection target. The present invention relates to a high-resolution and high-throughput projection type electron beam apparatus. A method of forming an image of the field of view by irradiating the field of view spread in at least a one-dimensional direction instead of one point on the sample in this way is called “mapping projection method”. This mapping projection type electron beam apparatus is a high-resolution and high-throughput apparatus that can avoid the space charge effect, has a high signal-to-noise ratio, and improves the image processing speed by parallel processing.

  Hereinafter, the case where the projection type electron beam apparatus of this embodiment is embodied as a defect inspection apparatus will be described in detail with reference to FIGS. 172 to 181. In these drawings, the same reference numerals or reference numerals indicate the same or corresponding components.

172 (A) and 172 (B), the electron gun EG of the defect inspection apparatus EBI has a thermoelectron emission type LaB ₆ cathode 1 that can be operated with a large current, and is directed from the electron gun EG in the first direction. The emitted primary electrons pass through the primary optical system including several stages of quadrupole lenses 2 and the beam shape is adjusted, and then pass through the Wien filter 172. The traveling direction of the primary electrons is changed by the Wien filter 172.1 to the second direction so as to be input to the wafer W to be inspected. The primary electrons leaving the Wien filter 172. 1 and traveling in the second direction are limited in beam diameter by the NA aperture plate 172. 2 and pass through the objective lens 172. The objective lens 172.3 is a high precision electrostatic lens.

As described above, in the primary optical system, an electron gun EG having a high brightness made of LaB ₆ is used, so that it has a low energy, a large current and a large area as compared with a conventional scanning type defect inspection apparatus. A primary beam can be obtained.

Since the wafer W is irradiated by the surface beam formed in a rectangular shape having a cross section of, for example, 200 μm × 50 μm by the primary optical system, a small area of a predetermined area on the wafer W can be irradiated. In order to scan the wafer W with this surface beam, the wafer W is placed on a high precision XY stage (not shown) corresponding to, for example, 300 mm, and the XY stage is moved two-dimensionally with the surface beam fixed. Let Further, since it is not necessary to narrow down the primary electrons to the beam spot, the surface beam has a low current density and the damage to the wafer W is small. For example, in the conventional beam scanning type defect inspection apparatus, the current density of the beam spot is 10 ³ A / cm ² , but in the illustrated defect inspection apparatus EBI, the surface beam current density is 0.1 A / cm ² to. It is only 0.01 A / cm ² . On the other hand, the dose is 1 × 10 ⁻⁵ C / cm ² in the conventional beam scanning method, whereas it is 1 × 10 ⁻⁴ C / cm ^{2 to} 3 × 10 ⁻⁵ C / cm ² in the present method. Yes, this method is more sensitive.

  Secondary electrons and reflected electrons emerge from the region of the wafer W irradiated with the surface beam-like primary electrons. The reflected electrons will be described later. First, detection of secondary electrons will be described. Secondary electrons emitted from the wafer W are enlarged by the objective lenses 172 and 3 so as to travel in the second opposite direction. After passing through the aperture plate 172. 2 and the Wien filter 172. 1, it is magnified again by the intermediate lens 172. 4 and further magnified by the projection lens 172. 5 and enters the secondary electron detection system D. In the secondary optical system that guides secondary electrons, the objective lenses 172 and 3, the intermediate lenses 172 and 4, and the projection lenses 172 and 5 are all high-precision electrostatic lenses, and the magnification of the secondary optical system is variable. Configured to be. Since the primary electrons are incident on the wafer W almost perpendicularly and the secondary electrons are taken out almost perpendicularly, the surface of the wafer W is not shaded by unevenness.

  The secondary electron detection system D that receives the secondary electrons from the projection lenses 172 and 5 includes a microchannel plate 172 and 6 that propagates the incident secondary electrons, and electrons emitted from the microchannel plate 172.6 as light. Fluorescent screens 192 and 7 for conversion and sensor units 172 and 8 for converting light emitted from the fluorescent screens 172 and 6 into electric signals are provided. The sensor units 172 and 8 have high-sensitivity line sensors 172 and 9 formed of a large number of two-dimensionally arranged solid-state imaging devices, and the fluorescence emitted from the fluorescent screens 172 and 7 is the line sensors 172 and 9. Is converted into an electrical signal and sent to the image processing units 172 and 10 for parallel, multistage and high-speed processing.

  While moving the wafer W and sequentially irradiating and scanning the individual areas on the wafer W with the surface beam, the image processing units 172 and 10 sequentially store data on the XY coordinates and the image of the area including the defect. The inspection result file including the coordinates and images of all the areas to be inspected including defects is generated for one wafer. In this way, inspection results can be managed collectively. When this inspection result file is read out, the defect distribution of the wafer and the defect detail list are displayed on the display of the image processing units 172 and 10.

  Actually, among the various components of the defect inspection apparatus EBI, the sensor units 172 and 8 are arranged in the atmosphere, but the other components are arranged in a lens barrel kept in a vacuum. In the embodiment, a light guide is provided on an appropriate wall surface of the lens barrel, and light emitted from the fluorescent screens 172 and 7 is taken out into the atmosphere via the light guide and relayed to the line sensors 172 and 9.

FIG. 173 shows a specific configuration example of the secondary electron detection system D in the defect inspection apparatus EBI of FIG. 172. A secondary electron image or a reflected electron image 173. 1 is formed on the incident surface of the microchannel plate 172 • 6 by the projection lens 172 • 5. The microchannel plate 172, 6 has, for example, a resolution of 16 μm, a gain of 10 ³ to 10 ⁴ , and an effective pixel of 2100 × 520. The electron is multiplied corresponding to the formed electronic image 173.1, and the fluorescent screen 172. 7 is irradiated. As a result, fluorescent light is emitted from the portion of the fluorescent screen 172 · 7 irradiated with electrons, and the emitted fluorescent light is emitted into the atmosphere through the light guide 173 · 2 having a low distortion (eg, distortion of 0.4%). The The emitted fluorescence is incident on the line sensors 172 and 9 through the optical relay lenses 173 and 3. For example, the optical relay lenses 173 and 3 have a magnification of 1/2, a transmittance of 2.3%, and a distortion of 0.4%, and the line sensors 172 and 9 have 2048 × 512 pixels. The optical relay lenses 173 and 3 form optical images 173 and 4 corresponding to the electronic images 173 and 1 on the incident surfaces of the line sensors 172 and 9. An FOP (fiber optic plate) can be used instead of the light guide 173.2 and the relay lens 173.3. In this case, the magnification is 1.times.

  The defect inspection apparatus EBI shown in FIG. 172 adjusts the acceleration voltage of the electron gun EG and the wafer voltage applied to the wafer W and uses the electron detection system D, so that in the case of secondary electrons, the positive charge mode and negative It can operate in either charging mode. Furthermore, by adjusting the acceleration voltage of the electron gun EG, the wafer voltage applied to the wafer W, and the objective lens conditions, the defect inspection apparatus EBI detects high-energy backscattered electrons emitted from the wafer W by irradiation of primary electrons. Can be operated in the backscattered electronic imaging mode. The backscattered electrons have the same energy as that when the primary electrons are incident on the sample such as the wafer W, and the energy is higher than that of the secondary electrons. There is a feature. As the electron detection system, an electron impact type detector such as an electron impact type CCD or an electron impact type TDI that outputs an electric signal corresponding to the intensity of secondary electrons or reflected electrons can be used. In this case, the electron impact type detector is installed and used at the imaging position without using the microchannel plates 172 and 6, the fluorescent screens 172 and 7, and the relay lenses 173 and 3 (or EOP). With this configuration, the defect inspection apparatus EBI can operate in a mode suitable for an inspection target. For example, in order to detect a metal wiring defect, a GC wiring defect, or a resist pattern defect, a negative charging mode or a backscattered electron imaging mode may be used, and a via conduction failure or a via bottom residue after etching may be detected. For the detection, a reflected electron imaging mode may be used.

FIG. 174 (A) is a diagram for explaining the requirements for operating the defect inspection apparatus EBI of FIG. 1 in the above three modes. The acceleration voltage of the electron gun EG is V _A , the wafer voltage applied to the wafer _W is V _W , the irradiation energy of primary electrons when irradiating the wafer W is E _IN , and the signal of secondary electrons incident on the electron detection system D Let E _OUT be the energy. The electron gun EG is configured to change the acceleration voltage V _A, and a variable wafer voltage V _W is applied to the wafer W from an appropriate power source (not shown). Therefore, when the acceleration voltage V _A and the wafer voltage V _W are adjusted and the electron detection system D is used, the defect inspection apparatus EBI has a range in which the secondary electron yield is larger than 1, as shown in FIG. Then, in the positive charging mode, in the range smaller than 1, it can operate in the negative charging mode. Further, by adjusting the acceleration voltage V _A , the wafer voltage V _W and the objective lens condition, the defect inspection apparatus EBI can operate in the reflected electron imaging mode using the energy difference between the secondary electrons and the reflected electrons. . Note that, in (B) in FIG. 174, the value of electron irradiation energy E _IN at the boundary between the positive charge area and the negative charge area is actually different depending on the sample.

Examples of values of V _A , V _W , E _IN and E _OUT for operating the defect inspection apparatus EBI in the backscattered electron imaging mode, the negative charging mode and the positive charging mode are as follows:
In the backscattered electron imaging mode, V _A = −4.0 kV
V _W = −2.5 kV
E _IN = 1.5 keV
E _OUT = 4 keV
In the negative charging mode, V _A = −7.0 kV
V _W = −4.0 kV
E _IN = 3.0 keV
E _OUT = 4 keV + α (α = energy width of secondary electrons)
In positive charging mode, _VA = -4.5kV
V _W = −4.0 kV
E _IN = 0.5 keV
E _OUT = 4 keV + α (α = energy width of secondary electrons)
It becomes.

  Actually, the detected amounts of secondary electrons and reflected electrons vary depending on the surface composition, pattern shape, and surface potential of the region to be inspected on the wafer W. That is, the secondary electron yield and the amount of reflected electrons differ depending on the surface composition of the object to be inspected on the wafer W, and the secondary electron yield and the amount of reflected electrons are large at the sharp points and corners of the pattern compared to the plane. Further, when the surface potential of the object to be inspected on the wafer W is high, the amount of secondary electron emission decreases. Thus, the electron signal intensity obtained from the secondary electrons and reflected electrons detected by the detection system D varies depending on the material, pattern shape, and surface potential.

  FIG. 175 shows a sectional shape of each electrode of the electrostatic lens used in the electron optical system of the defect inspection apparatus EBI shown in FIG. 172. As shown in FIG. 175, the distance from the wafer W to the microchannel plates 172 and 6 is, for example, 800 mm, and the objective lenses 172 and 3, the intermediate lenses 172 and 4, and the projection lenses 172 and 5 have a plurality of special shapes. An electrostatic lens having an electrode. Assuming that −4 kV is applied to the wafer W, +20 kV is applied to the electrode closest to the wafer W of the objective lens 172.3, and −1476 V is applied to the remaining electrodes. At the same time, −2450 V is applied to the intermediate lenses 172 and 4, and −4120 V is applied to the projection lenses 172 and 5. As a result, the magnification obtained by the secondary optical system is 2.4 times by the objective lens 172.5, 2.8 times by the intermediate lens 172/4, and 37 times by the projection lens 172/5, and 260 times in total. Become. In FIG. 175, reference numerals 175, 1, 175, 2 are field apertures for limiting the beam diameter, and reference numeral 175, 3 is a polarizer.

FIG. 176A is a diagram schematically showing a configuration of a multi-beam / multi-pixel type defect inspection apparatus EBI which is another embodiment of the projection type electron beam apparatus. The electron gun EGm in this defect inspection apparatus EBI is a multi-beam type electron gun having a cathode made of LaB ₆ and capable of emitting a plurality of primary electron beams 176. A plurality of primary electron beams 176. 1 emitted from the electron gun EGm are adjusted in beam diameter by an aperture plate 176. 2 in which a small hole is formed at a position corresponding to each primary electron beam. The position of each beam is adjusted by the axially symmetric lenses 176, 3 and 176.4 and travels in the first direction, passes through the Wien filter 172.1 and changes the traveling direction from the first direction to the second direction. It proceeds so as to enter the wafer W. Thereafter, each primary electron beam 176. 1 passes through the NA aperture plate 172. 2 and the objective lens 172. 3 to irradiate a predetermined region of the wafer W.

Secondary electrons and backscattered electrons 176, 5 emitted from the wafer W by irradiation with a plurality of primary electron beams 176, 1 are opposite to the second direction, as already described with reference to FIG. And enters the detection system D through the objective lens 172.3, NA aperture plate 172.2, Wien filter 172.1, intermediate lens 172.4, projection lens 172.5, and sensor unit 172. 8 is converted into an electrical signal.

  A plurality of deflectors 176, 6 for deflecting a plurality of primary electron beams 176, 1 are arranged between the axially symmetric lenses 176, 4 and the Wien filter 172, 1 on the downstream side when viewed from the electron gun EGm. . Therefore, in order to scan a certain region R on the wafer W by a plurality of primary electron beams 176. 1, the deflector 176 is moved while moving the wafer W in the Y-axis direction as shown in FIG. 6 simultaneously deflects a plurality of primary electron beams 176. 1 in the X-axis direction perpendicular to the Y-axis. As a result, the region R is raster-scanned by a plurality of primary electron beams 176.

  FIG. 177 (A) shows a schematic configuration of a multi-beam / monopixel type defect inspection apparatus EBI which is still another embodiment of the mapping projection type electron beam apparatus. In the figure, the electron gun EGm can emit a plurality of primary electron beams 176. 1, and the emitted plurality of primary electron beams 176. 1 are the same as described with reference to FIG. 176 (A). The wafer W is guided by the aperture plates 176, 2, the axially symmetric lenses 176, 3, 176, 4, the deflectors 176, 6, the Wien filter 172, 1, and the objective lens 172, 3 so as to travel in the first direction. Irradiate.

  Secondary electrons or reflected electrons 176. 5 emitted from the wafer W by being irradiated with a plurality of primary electron beams 176. 1 pass through the objective lens 172. 3, and then the Wien filter 172. After the traveling direction is changed, the light passes through the intermediate lenses 172 and 4 and the projection lenses 172 and 5 and enters the multi-detection system D ′. The multi-detection system D ′ shown in the figure is a secondary electron detection system, and includes a multi-opening plate 177. 1 in which the same number of holes as the n small holes formed in the opening electrode 176. -N pieces provided corresponding to each hole of the multi-aperture plate 177.1 so as to capture the secondary electrons that have passed through the n holes of 1 and convert them into electrical signals representing the intensity of the secondary electrons. A detector 177. 2, n amplifiers 177. 3 for amplifying the electrical signals output from the detectors 177. 2, and the electrical signals amplified by the respective amplifiers 177. And image processing units 172 and 10 ′ that store, display, and compare image signals of the scanned region R on the wafer W.

  In the defect inspection apparatus EBI shown in FIG. 177A, scanning of the region R by a plurality of primary electron beams 176.1 is performed as shown in FIG. 177B. That is, as shown in FIG. 177 (B), the region R is divided in the Y-axis direction by the number of primary electron beams 176 · 1, and small regions r1, r2, r3, r4 are assumed, and each primary electron beam is assumed. 176 · 1 is assigned to each of these small areas r1 to r4. Therefore, while moving the wafer W in the Y-axis direction, the deflectors 176 and 6 simultaneously deflect the primary electron beams 176 and 1 in the X-axis direction, respectively, and assign the small numbers assigned to the primary electron beams 176 and 1 respectively. The regions r1 to r4 are scanned. As a result, the region R is scanned by a plurality of primary electron beams 176.

Note that the primary optical system of the multi-beam is not limited to that shown in FIG. 176, and may be a multi-beam at the time of irradiation on the sample. For example, a single electron gun may be used.
In the defect inspection apparatus EBI described so far, it is preferable to use a mechanism that can place the wafer W on a stage and accurately position the stage in a vacuum chamber. In order to position the stage with high accuracy, for example, a structure in which the stage is supported in a non-contact manner by a static pressure bearing is employed. In this case, it is desirable to maintain a vacuum degree of the vacuum chamber by forming a differential exhaust mechanism for exhausting the high pressure gas in the range of the static pressure bearing so that the high pressure gas supplied from the static pressure bearing is not exhausted to the vacuum chamber. .

  FIG. 178 is a diagram showing an example of the configuration of a mechanism for accurately positioning the stage on which the wafer W is placed in the vacuum chamber, and the inert gas circulation piping system. In FIG. 178, the tip of the lens barrel 178 • 1 that irradiates the primary electrons toward the wafer W, that is, the primary electron irradiation portion 178 • 2 is attached to the housing 178 • 3 that defines the vacuum chamber C. A wafer W placed on a movable table in the X direction (left and right direction in FIG. 178) of the high-precision XY stage 178.4 is disposed immediately below the lens barrel 178.1. By moving the XY stages 178 and 4 in the X direction and the Y direction (the direction perpendicular to the paper surface in FIG. 178), it is possible to accurately irradiate the primary electrons to an arbitrary position on the surface of the wafer W.

  The bases 178 and 5 of the XY stages 178 and 4 are fixed to the bottom wall of the housing 178 and the Y table 178 and 6 that moves in the Y direction is placed on the bases 178 and 5. Projections are formed on both side surfaces (left and right side surfaces in FIG. 178) of the Y table 178.6, and these projections are formed on a pair of Y-direction guides 178.7a and 178.7b provided on the bases 178.5. It fits with the concave groove. Each concave groove extends in the Y direction over substantially the entire length of the Y direction guides 178, 7a, 178, 7b. A static pressure bearing (not shown) having a known structure is provided on each of the upper, lower and side surfaces of the protrusion protruding into the groove. By blowing high-pressure and high-purity inert gas (N2 gas, Ar gas, etc.) through these hydrostatic bearings, the Y tables 178 and 6 are not in contact with the Y-direction guides 178 and 7a and 178 and 7b. And can smoothly reciprocate in the Y direction. Between the bases 178 and 5 and the Y tables 178 and 6, linear motors 178 and 8 having a known structure are arranged to drive the Y tables 178 and 6 in the Y direction.

  An X table 178, 9 is mounted on the upper side of the Y table 178, 6 so as to be movable in the X direction. A pair of X direction guides 3178, 10a, 178, 10b having the same structure as the Y direction guides 178, 7a, 178, 7b for the Y tables 178, 6 so as to sandwich the X tables 178, 9 (178 in FIG. 178) -Only 10a is shown). A concave groove is also formed on the side of the X direction guide facing the X table 178, 9 and a protrusion projecting into the concave groove is formed on the side of the X table 178, 9 facing the X direction guide. Has been. These concave grooves extend over almost the entire length of the X-direction guide. Static pressure bearings (not shown) similar to the static pressure bearings for non-contact support of the Y tables 178 and 6 are provided on the upper, lower and side surfaces of the protrusions of the X-direction tables 178 and 9 protruding into the concave grooves. Is provided. By supplying a high-pressure and high-purity inert gas to these static pressure bearings and ejecting them from the static pressure bearings to the guide surfaces of the X-direction guides 178, 10a, 178, 10b, the X tables 178, 9 become X The direction guides 178 · 10a, 178 · 10b are supported in a non-contact manner with high accuracy. The Y tables 178 and 6 are provided with linear motors 178 and 11 having a known structure for driving the X tables 178 and 9 in the X direction.

  Since the stage mechanism with a static pressure bearing used in the atmosphere can be used almost as it is as the XY stage 178/4, an XY stage having the same accuracy as the high-precision stage for the atmosphere used in an exposure apparatus or the like is used. Therefore, it can be realized as an XY stage for a defect inspection apparatus with substantially the same cost and size. The wafer W is not placed directly on the X table 178/9, but is a sample stage having a function of holding the wafer W in a removable manner and performing a slight position change with respect to the XY stage 178/4. Usually placed on top.

The inert gas is supplied to the static pressure bearings through gas passages (not shown) formed in the flexible pipes 178, 12, 178, 13 and the XY stages 178, 4. The high-pressure inert gas supplied to the hydrostatic bearing is from several microns formed between the Y-direction guides 178, 7a, 1878, 7b and the opposing guide surfaces of the X-direction guides 178, 10a, 178, 10b. The Y table 178. 6 and the X table 178. The gas molecules of the inert gas ejected from the static pressure bearing diffuse into the vacuum chamber C, and are supplied to the dry vacuum pump 178 • through the exhaust ports 178 • 14, 178 • 15a, 178 • 15b and the vacuum pipes 178 • 16, 178 • 17. 18 is exhausted. The suction ports of the exhaust ports 178, 15a, 178, 15b pass through the bases 178, 5 and are provided on the upper surfaces thereof. As a result, the suction port opens near the position where the high-pressure gas is discharged from the XY stage 178.4, thereby preventing the pressure in the vacuum chamber C from rising due to the high-pressure gas ejected from the static pressure bearing. .

  The exhaust ports of the dry vacuum pumps 178 and 18 are connected to the compressors 178 and 20 via the pipes 178 and 19, and the exhaust ports of the compressors 178 and 20 are connected to the pipes 178, 21, 178, 22, 178, 23 and the regulator. It connects to flexible piping 178 * 12,178 * 13 via 178 * 24,178 * 25. For this reason, the inert gas discharged from the dry vacuum pumps 178 and 18 is pressurized again by the compressors 178 and 20, adjusted to an appropriate pressure by the regulators 178, 24, 178 and 25, and then again in the XY table. Supplied to the hydrostatic bearing. By doing so, the inert gas of high purity can be circulated and reused, so that the inert gas can be saved, and since the inert gas is not released from the defect inspection apparatus EBI, accidents such as suffocation due to the inert gas Can be prevented. In addition, removal means 178 and 26 such as cold traps and filters are provided in the middle of the discharge side pipes 178 and 21 of the compressors 178 and 20, and trap impurities such as moisture and oil mixed in the circulating gas. It is preferable not to supply to the hydrostatic bearing.

  Differential exhaust mechanisms 178 and 27 are provided around the tip of the lens barrel 178 and 1, that is, around the primary electron irradiation unit 178 and 2. This is because the pressure in the primary electron irradiation spaces 178 and 28 is sufficiently low even if the pressure in the vacuum chamber C is high. The annular members 178 and 29 of the differential evacuation mechanisms 178 and 27 attached around the primary electron irradiation units 178 and 2 are several microns to several hundreds between the lower surface (the surface facing the wafer W) and the wafer W. It is positioned with respect to the housing 178.3 so that a micron minute gap is formed.

  Annular grooves 178 and 30 are formed on the lower surfaces of the annular members 178 and 29, and the annular grooves 178 and 30 are connected to the exhaust ports 178 and 31. The exhaust ports 178 and 31 are connected to turbo molecular pumps 178 and 33, which are ultrahigh vacuum pumps, through vacuum pipes 178 and 32, respectively. Further, exhaust ports 178 and 34 are provided at appropriate positions of the lens barrels 178 and 1, and the exhaust ports 178 and 34 are connected to the turbo molecular pumps 178 and 36 via the vacuum pipes 178 and 35, respectively. These turbo molecular pumps 178/33, 178/36 are connected to dry vacuum pumps 178/18 by vacuum pipes 178/37, 178/38. Therefore, the gas molecules of the inert gas that have entered the differential pumping mechanisms 178 and 27 and the charged beam irradiation spaces 178 and 26 are turbo molecular pumps via the annular grooves 178 and 30, the exhaust ports 178 and 31, and the vacuum pipes 178 and 32. Since the gas is exhausted by 178/33, the gas molecules that have entered the spaces 178/28 surrounded by the annular members 178/29 from the vacuum chamber C are exhausted. Thereby, the pressure in primary electron irradiation space 178 * 28 can be kept low, and a primary electron can be irradiated without a problem. Gas molecules sucked from the tip of the lens barrel 178 • 1 are exhausted by the turbo molecular pumps 178 • 36 through the exhaust ports 178 • 34 and the vacuum pipes 178 • 35. The gas molecules discharged from the turbo molecular pumps 178/33, 178/36 are collected by the dry vacuum pumps 178/18 and supplied to the compressors 178/20.

The annular grooves 178 and 30 may have a double or triple structure depending on the pressure in the vacuum chamber C or the pressure in the primary electron irradiation spaces 178 and 28. Further, in the inspection apparatus shown in FIG. 178, the roughing pump of the turbo molecular pump and the vacuum exhaust pump of the vacuum chamber are combined with one dry vacuum pump, but it is supplied to the static pressure bearing of the XY stage. Depending on the flow rate of the high-pressure gas, the volume and inner surface area of the vacuum chamber, the inner diameter and length of the vacuum pipe, etc., it is possible to exhaust with a dry vacuum pump of another system.

  Generally, dry nitrogen is used as the high-pressure gas supplied to the static pressure bearings of the XY stage 178/4. However, it is preferable to use a higher purity inert gas if possible. This is because when impurities such as moisture and oil are contained in the gas, these impurity molecules adhere to the inner surfaces of the housings 178 and 3 defining the vacuum chamber C and the surfaces of the stage components, thereby deteriorating the degree of vacuum. This is because the degree of vacuum of the primary electron irradiation spaces 178 and 28 deteriorates due to adhesion to the surface of the wafer W. In addition, since it is necessary to prevent moisture and oil from being contained as much as possible, the turbo molecular pumps 178, 33, 178, and 36, the dry vacuum pumps 178 and 18, and the compressors 178 and 20 include moisture and oil in the gas flow path. It is required to have a structure that does not mix.

  As shown in FIG. 178, high-purity inert gas supply systems 178 and 19 are connected to the inert gas circulation piping system, and the vacuum chamber C and the vacuum piping 178. 16, 178, 15, 178, 32, 178, 35, 178, 37 and pressurized piping 178, 19, 178, 21, 178, 22, 178, 23, 178, 39 It plays the role of filling the active gas and supplying the shortage when the flow rate of the circulating gas decreases for some reason. In addition, by providing the dry vacuum pumps 178 and 18 with a function of compressing to atmospheric pressure or higher, the dry vacuum pumps 178 and 18 can also function as the compressors 178 and 20. Furthermore, it is also possible to use a pump such as an ion pump or a getter pump instead of the turbo molecular pumps 178 and 36 as an ultra-high vacuum pump used for exhausting the lens barrel 178. However, when these storage pumps are used, it becomes impossible to construct a circulation piping system. Instead of the dry vacuum pumps 178 and 18, it is also possible to use other types of dry pumps such as a diaphragm type dry pump.

FIG. 179 shows numerical examples of the sizes of the annular members 178 and 29 of the differential exhaust mechanisms 178 and 27 and the annular grooves 178 and 30 formed thereon. Here, an annular groove having a double structure separated in the radial direction is used. The flow rate of the high-pressure gas supplied to the static pressure bearing is usually about 20 L / min (atmospheric pressure conversion). Assuming that the vacuum chamber C is evacuated with a dry pump having an evacuation rate of 20000 L / min through a vacuum pipe having an inner diameter of 50 mm and a length of 2 m, the pressure in the vacuum chamber is about 160 Pa (about 1.2 Torr). . At this time, if the dimensions of the differential exhaust mechanisms 178 and 27, the annular members 178 and 29, the annular grooves 178 and 30 are set as shown in FIG. 179, the pressure in the primary electron irradiation space 56 is set to 10 ⁻⁴ Pa (10 ^-6 Torr).

  FIG. 180 schematically shows the overall configuration of an inspection system on which the defect inspection apparatus EBI described so far with reference to FIGS. 172 to 179 is mounted. As shown in the figure, the components of the path from the primary optical system of the defect inspection apparatus EBI to the detection system D through the wafer W and the secondary optical system are accommodated in a lens barrel 178. The cylinder 178. 1 is installed on the upper surface of the vibration isolation table 180. 1 supported by the active vibration isolation unit so as to prevent external vibration from being transmitted. The interior of the lens barrel 178 • 1 is kept in a vacuum by the evacuation system 180 • 2. A required voltage is supplied from the control power supply 180/3 to the components of the primary optical system and the secondary optical system inside the lens barrel 178/1 via the high-voltage cable 180/4.

An alignment mechanism 180/5 having an optical microscope and an autofocus means is provided at an appropriate position of the lens barrel 178.1, and each element constituting the primary optical system and the secondary optical system is placed on a predetermined optical axis. The primary electrons emitted from the electron gun are properly arranged and adjusted so as to automatically focus on the wafer W.

  On the upper surface of the vibration isolation table 180.1, an XY stage 178.4 having a chuck (not shown) for mounting and fixing the wafer W is installed, and the position of the XY stage 178.4 during the scanning period. Are detected by a laser interferometer at predetermined intervals. Further, a loader 180. 6 for accumulating a plurality of wafers W to be inspected and a wafer W in the loader 180. The transfer robots 180 and 7 are placed on the XY stages 178 and 4 in order to take out the wafer W from the lens barrel 178 and 1 after completion of the inspection.

  The operation of the entire system is controlled by a main controller 180/8 in which a required program is installed. The main controllers 180 and 8 include displays 180 and 9, and are connected to the detection system D via cables 180 and 10. As a result, the main controller 180, 8 receives the digital image signal from the detection system D via the cable 180, 10 and processes it by the image processing unit 172, 10, and the contents of the inspection result file obtained by scanning the wafer W The defect distribution and the like of the wafer W can be displayed on the displays 180 and 9. The main controller 180/8 displays the operation state of the system on the display 180/9 in order to control the operation of the entire system.

Although the stage on which the wafer W is placed has been described as being movable in the XY plane, in addition to this, the stage can be rotated about any axis perpendicular to the XY plane or passing through the XY plane. It may be. The inspection target is not limited to a wafer, but includes a sample that can be inspected by an electron beam such as a mask. Furthermore, a distributed defect inspection network can be constructed by mutually connecting the mapping projection electron beam apparatus according to this embodiment, the conventional beam scanning type defect review apparatus, the server, and the main controller via a LAN. it can.
As understood from the above description, this embodiment is
(1) Since the sample is irradiated with a surface beam, the throughput can be improved. For example, the defect inspection time per wafer can be shortened to about 1/7 as compared with a conventional beam scanning type inspection apparatus. it can,
(2) Since it is not necessary to focus the primary electrons to the beam spot, the space charge effect can be avoided and the sample is irradiated at a low current density, so that the sample is less damaged.
(3) Since the sample is irradiated with a surface beam, it can be inspected to a size smaller than one pixel.
(4) The electron gun acceleration voltage and the voltage applied to the sample are selected, and the objective lens is adjusted to operate in any one of the positive charging mode, the negative charging mode, and the reflection electron imaging mode. Since it can be performed, appropriate inspection can be performed according to the inspection site in the sample,
(5) By using the electrostatic lens, the primary optical system and / or the secondary optical system can be made small and highly accurate.
There are exceptional effects such as.

Claims (14)

Die, each having a pattern by irradiating electrons to the substrate disposed in a matrix, a method of detecting inspecting said substrate by a detector of secondary electrons generated from the substrate,
(A) a step for placing over said substrate stage,
(B) selecting a reference die serving as a positioning reference from the dies on the substrate placed on the stage, and obtaining a pattern matching template image including the coordinates of the feature points of the reference die When,
(C) any die definitive a row or column including said reference die, a step of pattern matching is executed, we obtain the coordinates of the feature points of the arbitrary die using the template image,
(D) based on the feature point coordinates between the coordinate and said reference die of said feature points of said any die acquired the in step (c), the substrate and the direction of the row or column including said reference die Calculating a deviation angle formed by one of the movement directions ;
(E) rotating the stage to correct the misalignment angle so that the direction of the row or column containing the reference die matches the one movement direction of the substrate ;
(F) after the step (e), rotating the detector to match the one movement direction of the substrate with the scanning direction of the detector;
(G) test method, characterized in that the electronic after said step (f) and a step of irradiating toward the substrate. The inspection method according to claim 1,
Before the step (a), further comprising the step of placing the means for irradiating the substrate with electrons, the stage, and the detector in a vacuum vessel,
The inspection method , wherein the step (f) of rotating the detector is performed by a rotating mechanism disposed outside the vacuum vessel in order to rotate the detector . The inspection method according to claim 2,
The vacuum vessel includes a first lens barrel, and a second lens barrel that is rotatably supported with respect to the first lens barrel and to which the detector is attached,
The inspection method , wherein the rotating mechanism is configured to rotate the second lens barrel . The inspection method according to any one of claims 1 to 3 ,
In steps (a) to (g), the substrate is irradiated with a primary electron beam having a two-dimensional cross section, and the secondary electron beam having information on the surface of the substrate irradiated with the primary electron beam is applied to the substrate. An inspection method that is executed in a mapping projection electron beam apparatus that detects a desired image by detecting with a detector . The inspection method according to claim 4 ,
The mapping projection type electron beam apparatus comprises E × B for separating the primary electron beam and the secondary electron beam,
The inspection method , wherein the E direction of ExB and the scanning direction of the detector are matched . The inspection method according to claim 5,
After the step (e) and before the step (f), the step of moving the NA to an optically optimal position in conjunction with a change in magnification of the mapping projection electron beam apparatus
An inspection method further comprising : The inspection method according to any one of claims 2 to 6, comprising:
The steps (a) to (g) are performed using a low magnification optical microscope, and then the steps (a) to (g) are performed again using a high magnification optical microscope. Inspection method characterized by The inspection method according to claim 1 ,
The step (b) of acquiring the coordinates of the feature points of the arbitrary die is as follows:
Based on the positional relationship between the coordinates of the feature points of the reference die and the coordinates of the exact feature points of the arbitrary die that have already been obtained by performing pattern matching with the template image, the feature points of the next arbitrary die Estimating the coordinates of
Performing pattern matching using the template image near the coordinates of the estimated feature points;
Obtaining the exact feature point coordinates of the next arbitrary die; and
The inspection method characterized by including the step which repeats . The inspection method according to claim 8,
The repeating step estimates the coordinate of the feature point of the next arbitrary die based on the positional relationship between the coordinate of the feature point of the reference die and the coordinate of the feature point of the die obtained in the previous step, An inspection method comprising repeating the step of obtaining the coordinates of the exact feature point of the next arbitrary die by pattern matching. The inspection method according to claim 1,
Obtaining a size in a direction orthogonal to the row or column from which the deviation angle was obtained;
Creating a die map based on the determined size;
An inspection method further comprising: The inspection method according to claim 10,
The step of obtaining a size in a direction orthogonal to the row or column from which the deviation angle is obtained includes:
Selecting a reference die to be a positioning reference, and obtaining a pattern matching template image including the coordinates of the feature points of the reference die;
Pattern matching is performed on any die in a row or column in a direction perpendicular to the row or column in which the deviation angle is obtained, including the reference die, using the template image, and the feature points of the arbitrary die Obtaining the coordinates;
The distance between the coordinate of the feature point of the reference die and the coordinate of the feature point of the arbitrary die and the number of dies included in the distance are obtained, and based on this, the row or column from which the deviation angle is obtained; Determining the size of the die in the orthogonal direction; and
The inspection method characterized by including. The inspection method according to any one of claims 1 to 11,
Obtaining a difference between a target position of the stage and an actual position of the stage, and applying a correction voltage for canceling the difference to the electrode for deflecting the electrons
An inspection method further comprising: The inspection method according to any one of claims 1 to 12,
The inspection method, wherein the secondary electrons are at least one of electrons generated from the substrate, reflected electrons, and backscattered electrons. The inspection method according to any one of claims 1 to 12,
The inspection method, wherein the secondary electrons are mirror electrons reflected near the surface of the substrate.

Images