JP2008166483A - Grid matching method and exposure system - Google Patents

Abstract

To improve the efficiency of management of grid error data.
A general machine (No. machine) which is an exposure apparatus other than the specific machine, based on a specific machine (No. A) which is a specific exposure apparatus among a plurality of exposure apparatuses which expose a substrate through a mask pattern. B) to manage error data related to grid errors, and based on the error data, calculate a relative error between any two exposure apparatuses to be grid-matched among the plurality of exposure apparatuses. To the corresponding exposure apparatus.
[Selection] Figure 4

Description

  The present invention relates to, for example, a grid matching method used in a photolithography process for manufacturing a semiconductor element, a liquid crystal display element, an imaging element, a thin film magnetic head, and the like, and information storing a program for causing a computer to execute the method. The present invention relates to a recording medium and an exposure system.

  Many devices such as semiconductor devices, liquid crystal display devices, image pickup devices (CCD: charge coupled devices), thin film magnetic heads, etc. are manufactured by overlaying multiple layers of patterns onto a substrate using multiple exposure devices and exposing and transferring them. Is done. Therefore, when the second and subsequent layers are exposed and transferred onto the substrate, each shot area on which the pattern has already been formed and the pattern image of the mask are aligned, that is, the alignment between the substrate and the reticle. (Alignment) must be performed accurately. For this reason, on the substrate on which the first layer pattern of the stage coordinate system is exposed, alignment marks called alignment marks are formed in a manner attached to each shot area (chip pattern area). . When the substrate on which the alignment mark is formed is carried into the exposure apparatus, the mark position (coordinate value) on the stage coordinate system is measured by the mark measurement apparatus provided in the exposure apparatus. Next, based on the measured position of the mark and the design position of the mark, alignment for positioning (positioning) one shot area on the substrate with respect to the reticle pattern is performed.

  As an alignment method, from the viewpoint of improving throughput, the regularity of shot arrangement on a substrate is statistically analyzed as disclosed in, for example, JP-A-61-44429 and JP-A-62-84516. Enhanced global alignment (EGA), which is precisely specified by the method, is the mainstream. EGA measures the position of the alignment mark for a plurality (for example, about 7 to 15) of sample shots selected in advance, and minimizes the error from the measured value and the design position of the alignment mark. In order to calculate the position coordinates (shot array) of all shot regions on the substrate by performing statistical calculation using the least square method or the like, the substrate stage is stepped according to the calculated shot array. Is. By this EGA, mainly linear errors (residual rotation error of the substrate, orthogonality error of the stage coordinate system (or shot array), linear expansion / contraction (scaling) of the substrate, offset of the substrate (center position) ( Translation) etc.) is removed.

  In addition, when overlay exposure is performed between multiple exposure units (units), there is a grid error in the stage between the exposure units (an error between the stage coordinate systems that define the wafer movement position in each exposure unit). Therefore, an overlay error occurs. In such a case, if the error generated in the shape and arrangement of the shot area on the wafer, which is a cause of the overlay error, is a linear component, it can be removed by the above-described EGA wafer alignment. If it is a nonlinear component, it is difficult to remove it. This is because the EGA method treats the shot area arrangement error on the wafer as linear. Grid compensation matching (GCM) is known as a technique for removing such non-linear errors in shot shapes and shot arrangements.

  In this GCM, during the exposure sequence (exposure processing for a process wafer), EGA measurement is performed again based on the EGA result to extract nonlinear components, and the extracted nonlinear components are averaged over a plurality of wafers. The value is held as a map correction value, and in the subsequent exposure sequence, this map correction value is used to correct the shot position (see JP-A-2001-345243). A reference wafer is used in advance separately from the exposure sequence. The nonlinear component (deviation amount for each shot) is measured and stored as a map correction file, and each shot position is corrected using the map correction file in the exposure sequence (Japanese Patent Laid-Open No. 2002). No. -353121) is known. Further, a difference (nonlinear error component) between the position of the shot array after the linear error component is removed by the EGA method and each design position (nonlinear error component) is evaluated based on a predetermined evaluation function, and the evaluation result A function that determines a function that expresses the nonlinear component based on the above and corrects the shot arrangement based on the function is also known (see Japanese Patent Application Laid-Open No. 2004-265957).

  In these conventional grid matching methods, a grid error is obtained as an error based on a design value (ideal lattice), a map correction file is created, stored and managed in each exposure apparatus, and each exposure apparatus The correction is performed so as to cancel the error, that is, close to the ideal lattice. However, when an error based on the grid error of the exposure apparatus of the original process is reflected in the exposed and transferred pattern in the original process (a process in which the exposure process is performed prior to the overlay), the current process ( Even if the grid error of the exposure apparatus of the current process is corrected with reference to the ideal grid, it is difficult to sufficiently remove the pattern overlay error. It is. Therefore, it is desirable to correct the grid error that occurs in the exposure apparatus in the current process with reference to the grid error that occurs in the exposure apparatus in the original process. In this case, it is necessary to relatively manage the grid error of the original process and the grid error of the current process. For any two combinations of all the exposure apparatuses constituting the exposure system, the grid error between both exposure apparatuses Managing the difference is cumbersome and numerous. As described above, conventionally, it is necessary to manage an enormous amount of grid error data, and there is a problem that the amount of data is extremely large, the capacity of the storage device is compressed, or the communication amount of error data is large.

The present invention has been made in view of such problems of the prior art, and an object of the present invention is to reduce the amount of grid error data to be managed and improve the efficiency of grid error management.
JP-A 61-44429 JP-A-62-84516 JP 2001-345243 A JP 2002-353121 A JP 2004-265957 A

  According to the present invention, on the basis of a specific machine that is a specific exposure apparatus among a plurality of exposure apparatuses that expose a substrate through a mask pattern, the grid according to each of general machines that are exposure apparatuses other than the specific machine A grid matching method comprising: managing error data relating to an error; and calculating a relative error between any two exposure apparatuses to be grid matched among the plurality of exposure apparatuses based on the error data. Provided. In the present invention, error data for each of other general numbered machines is managed based on a specific number of exposure apparatuses, and a relative error between any two exposure apparatuses to be grid-matched is calculated. Therefore, there is no need to manage error data for the specific machine, the amount of data to be managed can be reduced, and when error data is acquired from another apparatus (for example, an exposure apparatus or a measurement apparatus) In addition, the amount of data communication when error data is output to another apparatus (for example, an exposure apparatus) can be reduced. Other aspects and effects of the present invention will become apparent through the following embodiments.

  According to the present invention, there is an effect that the management of grid error data can be made highly efficient.

[Exposure system]
First, the overall configuration of an exposure system (lithography system) according to the present embodiment will be described with reference to FIG. In the present embodiment, the exposure apparatus may be referred to as a number machine. FIG. 1 schematically shows the overall configuration of an exposure system 100 according to an embodiment of the present invention. The exposure system 100 includes n exposure apparatuses EX1 to EXn, a host computer system 200, a terminal server 300, an overlay measuring device 400, a GCM server (Grid Compensation Matching Server) 500 as a management server, and the like. Each of the exposure apparatuses EX1 to EXn, the terminal server 300, the overlay measuring device 400, and the GCM server 500 are connected to a local area network (LAN) 600. The GCM server 500 includes a storage device 520 connected via a communication interface 510 such as SCSI (Small Computer System Interface). The host computer system 200 is connected to the LAN 600 via the terminal server 300. As a result, communication paths among the exposure apparatuses EX1 to EXn, the host computer system 200, the terminal server 300, the overlay measuring device 400, and the GCM server 500 are secured.

[Exposure equipment]
Each of the exposure apparatuses EX1 to EXn is a step-and-repeat type projection exposure apparatus (stepper) that performs exposure while the reticle stage and wafer stage are stationary, and exposure is performed while the reticle stage and wafer stage are moved synchronously. Any of a step-and-scan type projection exposure apparatus (scanning stepper) for performing the above and other types of exposure apparatuses may be used, or these may be mixed. For example, the exposure apparatus EX1 is configured as shown in FIG. The other exposure apparatuses EX2 to EXn are the same as those of the exposure apparatus EX1, and the description thereof is omitted. The exposure apparatus EX1 sequentially transfers the pattern image formed on the reticle R onto the shot area on the wafer W while the reticle stage RST and the wafer stage WST are moved synchronously with respect to the projection optical system PL. This is a scanning exposure apparatus.

In the drawing, the illumination optical system ILS includes a light source LS, an optical integrator OPI, an illumination system aperture stop IAS, and a reticle blind mechanism RBL having a fixed blind FRBL and a movable blind MRBL. In the figure, the lens system and the like are omitted. Here, an ArF excimer laser light source (wavelength 193 nm) is used as the light source LS, but an ultrahigh pressure mercury lamp that emits g-line (wavelength 436 nm) and i-line (wavelength 365 nm), or a KrF excimer laser (wavelength). 248 nm), F ₂ laser (wavelength 157 nm), and other light sources. The laser light emitted from the light source LS passes through a variable dimmer (not shown), an illuminance adjustment unit, a beam shaping optical system, etc. (all not shown), and then an optical integrator (fly eye lens, internal reflection type integrator, diffractive optics). The element is incident on the OPI. On the exit surface (exit-side focal plane) of the optical integrator OPI, that is, the optical Fourier transform plane (plane that is optically conjugate with the pupil plane of the illumination system and the pupil plane of the projection optical system PL) with respect to the pattern surface of the reticle R An illumination system aperture stop plate BAL for changing illumination conditions according to the pattern to be transferred is arranged. The illumination system aperture stop plate BAL is composed of a disc configured to be rotatable around a rotation axis, and is a circular aperture stop for normal illumination, an aperture stop for annular illumination, and a plurality of (for example, four poles) eccentricity. A plurality of aperture stops IAS including an aperture stop for modified illumination including a small aperture and a small circular aperture stop for a small coherence factor (σ value) are arranged along the circumferential direction. The rotation axis of the illumination system aperture stop plate BAL is rotated by a drive motor (not shown), and the illumination condition can be changed by switching the aperture stop disposed on the exit surface of the optical integrator OPI.

  The light emitted from the optical integrator OPI and passing through one of the aperture stops IAS of the aperture stop plate BAL is a fixed blind (fixed illumination field stop) arranged on or near the conjugate plane with the pattern surface (lower surface) of the reticle R. ) Incident into a reticle blind mechanism RBL composed of FRBL and movable blind (movable illumination field stop) MRBL, and its cross-sectional shape is shaped into a slit extending in a direction (Y direction) perpendicular to the scanning direction (X direction). . The movable blind MRBL can arbitrarily change and set the illumination area IA by the illumination light IL irradiated on the reticle R. In order to change the light amount distribution (the size and shape of the secondary light source) of the illumination light IL on the pupil plane of the illumination optical system ILS or the projection optical system PL, for example, it is arranged in exchange in the illumination optical system ILS. An optical unit including at least one of a plurality of diffractive optical elements, a prism (conical prism, polyhedral prism, etc.) movable along the optical axis of the illumination optical system ILS, and a zoom optical system, a light source LS, an optical integrator OPI, If the optical integrator is a fly-eye lens, the intensity distribution of the illumination light on the incident surface, if the optical integrator is an internal reflection integrator, the incident angle range of the illumination light with respect to the incident surface, etc. May be variable.

  The light that has passed through the movable blind MRBL is emitted as illumination light IL through a condenser lens system (not shown) and the like, and the illumination area (illumination) of the pattern surface (lower surface) of the reticle R disposed on the object plane of the projection optical system PL. Field of view) Illuminates IA. The reticle R is attracted and held on the reticle stage RST, and a moving mirror MRr to which a length measuring laser beam from the reticle interferometer system IFR is irradiated is fixed to one end of the reticle stage RST. Positioning of the reticle R is performed by a reticle driving device (not shown) that translates the reticle stage RST in the XY plane perpendicular to the optical axis AX and rotates it slightly in the XY plane. When transferring an image of the pattern of the reticle R onto the wafer W, the reticle driving device scans the reticle stage RST in a predetermined scanning direction (X direction) at a constant speed. Above the reticle stage RST, a pair of alignment systems RAL for photoelectrically detecting a plurality of reticle alignment marks formed around the reticle R are provided along the scanning direction. The detection result of the alignment system RAL is used for positioning the reticle R with a predetermined accuracy with respect to the optical axis AX of the projection optical system PL. Interferometer system IFR projects a laser beam onto movable mirror MRr, receives the reflected beam, and measures the positional change of reticle R.

  A pattern image formed in the illumination area IA of the reticle R under the illumination light IL is converted into a predetermined projection magnification α (α is, for example, 1/4 or 1/5) via the telecentric projection optical system PL. ) Is projected onto a slit-like exposure region (a region conjugate with the illumination region IA with respect to the projection optical system PL) on the wafer W as a substrate disposed on the imaging plane of the projection optical system PL. Projection optical system PL has a plurality of optical elements such as lenses. Some of these optical elements have adjustable positions (positions in the X, Y, and Z axis directions) and attitudes (angles relative to the optical axis AX), and the positions or attitudes of these optical elements can be adjusted. As a result, the optical characteristics such as the magnification, field curvature, and distortion of the projection optical system PL can be adjusted. In the present specification, optical characteristics such as aberrations that affect the shape of the projected image (image distortion) such as magnification, field curvature, and distortion are referred to as distortion. The adjustment of the imaging characteristic is performed by the imaging characteristic control device ICC under the control of the main control device CNT.

  A wafer table WTB is provided on wafer stage WST on which wafer W is placed and moved two-dimensionally along the XY plane. Wafer holder WH for vacuum-sucking wafer W is provided on wafer table WTB. Wafer table WTB slightly moves and tilts wafer holder WH in the Z direction (optical axis AX direction) based on a measurement value of an autofocus mechanism (AF mechanism) (not shown). The movement coordinate position of wafer stage WST in the XY plane and the minute rotation amount by yawing are measured by wafer interferometer system IFW. This interferometer system IFW irradiates a moving mirror MRw fixed to wafer table WTB of wafer stage WST with a laser beam for length measurement from a laser light source (not shown), and reflects the reflected light and predetermined reference light. The coordinate position and minute rotation amount (yaw amount) of wafer stage WST are measured by interference. An off-axis alignment system ALG for measuring position information of wafer marks (alignment marks) formed on the wafer W is provided on the side of the projection optical system PL. The alignment system ALG will be described later.

  On the wafer table WTB of the wafer stage WST, a reference plate (not shown) used for calibration of an AF sensor provided in the AF mechanism, measurement of a baseline amount, and the like is attached. On the surface of the reference plate, a reference mark (fiscal mark) that can be detected by the alignment system RAL and other marks are formed together with the mark of the reticle R. The AF sensor is a sensor that measures the amount of deviation of the surface of the wafer W from the image plane of the projection optical system PL. The baseline amount is an amount indicating the distance between the reference position (for example, the center of the pattern image) of the reticle pattern image projected onto the wafer W and the center of the visual field of the alignment system ALG. The main controller CNT is composed of, for example, a microcomputer, and comprehensively controls each part of the exposure apparatus EX1. In the present embodiment, the main controller CNT also controls a coater / developer (not shown) provided in the exposure apparatus EX1. In FIG. 1, the exposure apparatus EX1 is described as being connected to the LAN 600, but strictly speaking, the main controller CNT is connected to the LAN 600. The main control device CNT communicates with the host computer system 200 via the LAN 600 and the terminal server 300, and executes various control operations in accordance with commands from the host computer system 200.

[Alignment system]
As shown in FIG. 3, the alignment system ALG includes a measurement sensor 410 and a measurement control device 450. Measurement sensor 410 is a sensor that measures the position of a mark (for example, an alignment mark) formed on wafer W placed on wafer stage WST. Here, as an example, a sensor used in an FIA (Field Image Alignment) method will be described, but a sensor used in an LSA (Laser Step Alignment) method or an LIA (Laser Interferometric Alignment) method may be used. The LSA sensor is a sensor that irradiates a mark formed on a substrate with a laser beam and measures the position of the mark using the diffracted / scattered light. A laser beam having slightly different wavelengths is irradiated from two directions onto a diffraction grating-like mark formed on the surface, and the resulting two diffracted beams interfere with each other, and the position information of the mark is detected from the phase of the interference light. It is a sensor. The measurement sensor 410 may be provided with two or more of these three types of sensors so that the measurement sensor 410 can be used properly according to the respective characteristics and circumstances.

  In FIG. 3, illumination light IL <b> 10 is guided to the measurement sensor 410 from an illumination light source such as an external halogen lamp via an optical fiber 411. The illumination light IL10 is applied to the field division stop 413 through the condenser lens 412. Although not shown, the field division diaphragm 413 has a mark illumination diaphragm having a wide rectangular opening at the center thereof and a pair of narrow rectangular openings arranged so as to sandwich the mark illumination diaphragm. And a focus detection slit. The illumination light IL10 is split by the field division diaphragm 413 into a first light beam for mark illumination that illuminates the alignment mark region on the wafer W and a second light beam for focus position detection prior to alignment. The illumination light IL20 divided in this way is transmitted through the lens system 414, reflected by the half mirror 415 and the mirror 416, reflected by the prism mirror 418 through the objective lens 417, and formed on the wafer W. The mark area including M is irradiated. The reflected light of the surface of the wafer W when irradiated with the illumination light IL20 is reflected by the prism mirror 418, passes through the objective lens 417, is reflected by the mirror 416, and then passes through the half mirror 415. Thereafter, the light reaches the beam splitter 420 via the lens system 419, and the reflected light is branched in two directions. The first branched light that has passed through the beam splitter 420 forms an image of the mark M on the index plate 421. Then, the light from the image and the index mark on the index plate 421 enters the image sensor 422 made up of a two-dimensional CCD, and the image of the mark M and the index mark is formed on the light receiving surface of the image sensor 422.

  On the other hand, the second branched light reflected by the beam splitter 420 enters the light shielding plate 423. The light shielding plate 423 shields light incident on a predetermined rectangular area and transmits light incident on an area other than the rectangular area. Therefore, the light shielding plate 423 shields the branched light corresponding to the first light flux and transmits the branched light corresponding to the second light flux. The branched light transmitted through the light shielding plate 423 is incident on the line sensor 425 made of a one-dimensional CCD in a state where the telecentricity is destroyed by the pupil division mirror 424, and an image of the focus detection slit is formed on the light receiving surface of the line sensor 425. Imaged. Here, since telecentricity is ensured between the wafer W and the image sensor 422, when the wafer W is displaced in a direction parallel to the optical axes of the illumination light and the reflected light, the wafer W is connected to the light receiving surface of the image sensor 422. The image of the imaged mark M is defocused without changing the position on the light receiving surface of the image sensor 422. On the other hand, since the telecentricity of the reflected light incident on the line sensor 425 is broken as described above, if the wafer W is displaced in a direction parallel to the optical axes of the illumination light and the reflected light, the line sensor The image of the focus detection slit imaged on the light receiving surface 425 is displaced in the direction intersecting the optical axis of the branched light. If the amount of deviation of the image on the line sensor 425 with respect to the reference position is measured using such properties, the position (focal position) of the wafer W in the optical axis direction is detected.

  Note that the grid error measurement process by the measurement sensor 410 is performed on the reference wafer by the exposure apparatus EXi to which the grid error is to be measured. The measurement sensor 410 outputs imaged image data (gradation data for each pixel) from the image sensor 422 to the measurement control device 450. The measurement control device 450 integrates the captured image data in a direction orthogonal to the measurement method (here, the X direction) to obtain a one-dimensional signal, for example, a folded autocorrelation process, a template matching process using a predetermined template, or Mark center deriving processing such as edge position measurement processing (processing for obtaining the contour of a mark, processing for detecting the edge position of each mark element from the obtained contour, processing for obtaining the center of the mark from the detected edge position), etc. Position information (here, the X coordinate value) in the measurement direction of M is obtained. The Y coordinate value is similarly obtained, and the coordinate value (X, Y coordinate value) of the mark is stored and held together with the identification information of the mark in a storage device (not shown) provided in the measurement control device 450. The measurement control device 450 transmits the measurement result to the GCM server 500 via the LAN 600.

[Overlay measuring instrument]
The overlay measuring instrument 400 includes a measurement sensor similar to the measurement sensor 410 of the alignment sensor ALG described above and a measurement apparatus including a stage device configured similarly to the wafer stage WST described above, and a measurement similar to the measurement control apparatus 450 described above. It is configured with a control device. At least one measuring device is provided for this exposure system, and is provided in-line in the production line constituting the exposure system. However, a plurality of measuring devices may be provided in the exposure system. In this case, each measurement device may be provided separately from the exposure devices EX1 to EXn, or may be provided in-line in a coater / developer attached to each exposure device EX1 to EXn. . In addition, the measuring device is not necessarily provided inline in the production line, and may be provided off-line. As an example, the measurement sensor here is a sensor used in an FIA (Field Image Alignment) method, but is a sensor used in an LSA (Laser Step Alignment) method or an LIA (Laser Interferometric Alignment) method. May be. In addition, a measurement sensor can also provide two or more sensors among these three types of sensors so that they can be used properly according to their characteristics and circumstances.

[GCM server]
The GCM server 500 is a lithography system support apparatus configured by a medium-sized computer system (for example, a minicomputer or an engineering workstation) having excellent computing power. The GCM server 500 communicates with the host computer system 200 via the LAN 600 and the terminal server 300 in addition to communication with the exposure apparatuses EX1 to EXn via the LAN 600. Further, the GCM server 500 reads / writes data from / to the storage device 520 as necessary during communication with the exposure apparatuses EX1 to EXn. The GCM server 500 is provided with an input / output device (not shown) including a display as a man-machine interface and a pointing device such as a keyboard and a mouse. Further, the GCM server 500 is provided with a drive device (not shown) for an information recording medium such as a CD (compact disc), DVD (digital versatile disc), MO (magneto-optical disc) or FD (flexible disc). Connected with. A function as the GCM server 500 is realized by setting and reading an information recording medium storing a program and data for the GCM server in the drive device. The GCM server 500 expresses grid error data relating to the grid errors of the exposure apparatuses EX1 to EXn sent from the measurement control apparatus 450 in a database in the storage device 520 in a predetermined format, and stores and manages the data.

[Terminal server and host computer system]
The terminal server 300 is configured as a gateway processor for absorbing differences between the communication protocol in the LAN 600 and the communication protocol of the host computer system 200. The function of the terminal server 300 enables communication between the host computer system 200 and the exposure apparatuses EX1 to EXn, the overlay measuring device 400, the GCM server 500, and the like connected to the LAN 600. The host computer system 200 is a manufacturing execution system (MES) that includes a large computer. Here, the manufacturing management system means that all processes, equipment, conditions, and work data of each product flowing on the production line are managed and analyzed by a computer, thereby improving quality, improving yield and reducing work errors. It is a system that supports efficient production. The host computer system 200 may be other than the MES, and for example, a dedicated computer may be used.

[Selection of specific unit]
In the present embodiment, a specific number to be used as a reference in the exposure system is selected from the numbers EX1 to EXn constituting the exposure system. Specific units that should be the standard are hereinafter referred to as specific units, and the remaining units are referred to as general units. The specific unit may be any of the units, but each unit included in the exposure system is provided by various vendors (manufacturers) or has different specifications and performance. If there is a mixture of things, it is recommended to select one with a low stage grid error adjustment capability. However, since the specific number machine is a standard number machine, it is desirable to select a high-precision number machine with little change in grid error over time. In the following description, for convenience, the machine EX1 will be described as a specific machine, and EX2 to EXn will be described as general machines.

[Collecting grid error data]
Next, a process for collecting grid error data of each exposure apparatus EXi (i = 1 to n) will be described. The grid error is measured by measuring the reference pattern using a reference wafer (test wafer, common wafer) on which a predetermined reference pattern (reference mark) is formed. Here, as an example, a case where the grid error of the machine EX1 is measured will be described as an example. The reference wafer is generally manufactured by the following procedure. First, a thin film of silicon dioxide (or silicon nitride, polysilicon, etc.) is formed on almost the entire surface of a silicon substrate (wafer), and then a photosensitive agent (resist) is formed on the entire surface of the silicon dioxide film by a resist coating device (coater). Apply. Then, the resist-coated substrate is loaded onto a wafer holder of a reference exposure apparatus (a highly reliable exposure apparatus adjusted with high accuracy), and a reference reticle (not shown) is an enlarged pattern of a reference mark pattern. A special reticle on which is formed) is loaded on the reticle stage, and the pattern of the reticle for the reference wafer is exposed and transferred to each shot area on the silicon substrate in accordance with a predetermined shot arrangement.

  A reference mark pattern (wafer alignment mark (search for the alignment of the actual wafer) (search), which is preferably the same number of shot areas as the actual wafer loaded on the exposure apparatus to be used) on the silicon substrate. An image of alignment marks, fine alignment marks, etc.) is transferred and formed. Next, the silicon substrate is unloaded from the wafer holder and developed using a developing device (developer). Next, after the development processing is performed on the silicon substrate using an etching apparatus until the substrate surface is exposed, the resist remaining on the silicon substrate surface is removed using, for example, a plasma ashing apparatus. As a result, a reference wafer is formed in which a reference mark (wafer alignment mark) is formed in the silicon dioxide film on the silicon substrate as a recess corresponding to each of a plurality of shot regions arranged in the same manner as the actual wafer. As described above, the reference wafer is not limited to one in which a mark is formed on a silicon dioxide film by patterning, and a reference wafer in which a mark is formed as a recess in a silicon substrate may be used. Since the reference wafer is used for quality control of a plurality of exposure apparatuses in the exposure system, a plurality of exposure apparatuses can use various shot map data (size and arrangement data of each shot area on the wafer). If there is a possibility, it is desirable to create for each shot map data.

  Next, processing of the main controller CNT included in the exposure apparatus EX1 will be described with respect to an operation when creating a database regarding grid errors using the reference wafer manufactured as described above. First, a reference wafer is loaded onto the wafer holder WH of the exposure apparatus EX1 using a wafer loader (not shown). Next, search alignment of the reference wafer loaded on the wafer holder WH is performed. Specifically, for example, the alignment sensor ALG detects at least two search marks positioned in the peripheral portion substantially symmetrically with respect to the reference wafer center. These search marks are detected by sequentially positioning the wafer stage WST and setting the magnification of the alignment system ALG to a low magnification so that each search mark is positioned within the detection field of the alignment system ALG. . Based on the detection result of the alignment sensor ALG (relative positional relationship between the index center of the alignment sensor ALG and each search mark) and the measured value of the laser interferometer IFW when each search mark is detected, the stage coordinates of the two search marks Find the position coordinates on the system. Thereafter, the residual rotation error of the reference wafer is calculated from the position coordinates of the two search marks, and the wafer holder WH is slightly rotated so that the residual rotation error becomes substantially zero. Thereby, the search alignment of the reference wafer is completed.

  Next, the position coordinates on the stage coordinate system of all shot areas on the reference wafer are measured. Specifically, the position coordinates of the fine alignment mark (reference mark) on the wafer W on the stage coordinate system, that is, the position of the shot area, in the same manner as the measurement of the position coordinates of each search mark during the search alignment described above. Find the coordinates. The reference mark is detected by setting the magnification of the alignment system ALG to a high magnification. Next, statistical calculation (EGA calculation) using the least square method as disclosed in Japanese Patent Application Laid-Open No. 61-44429 and the like based on the measured position coordinates of the shot area and the position coordinates on each design. And calculate six parameters (corresponding to six parameters of rotation regarding the arrangement of each shot area on the reference wafer, scaling in the X and Y directions, orthogonality, and offset in the X and Y directions), and the calculation result And the position coordinates (array coordinates) of all shot areas are calculated based on the design position coordinates of each shot area, and the calculation result, that is, the position coordinates of all shot areas on the reference wafer is stored. Next, the linear component and the non-linear component of the positional deviation amount (error amount) are separated for all shot regions on the reference wafer. Specifically, the difference between the calculated position coordinates of each shot area and the respective design position coordinates is calculated as a linear component of the positional deviation amount, and the position coordinates of all shot areas actually measured are The residual obtained by subtracting the linear component from the difference from the design position coordinate is calculated as a nonlinear component of the positional deviation amount.

  Next, the calculated nonlinear component is sent to the GCM server 500 as grid error data. For the other machines EX2 to EXn, the grid error is similarly measured using the reference wafer and sent to the GCM server 500 in the same manner. In the GCM server 500, based on the grid error data sent from each of the units EX1 to EXn, the difference of the grid error data is calculated for each of the general units EX2 to EXn based on the specific unit EX1, and each of the general units EX2 The grid error data (difference data) is registered in the database in the storage device 520 together with the identification information of .about.EXn.

  Grid error data sent from each machine EX1 to EXn to the GCM server, grid error data (difference data) managed in the GCM server 500, and grid error data (difference) sent from the GCM server 500 to each general machine EX2 to EXn May be a delta map composed of X and Y coordinates of sample points and an error vector or other map format data, but in order to reduce the amount of information and communication, a three-dimensional model as shown below Substituting each value of these map data into the equation (Equation 1), using the least square method, obtain an approximate expression related to the grid error for the unit, and use the coefficient (k parameter) for each term of the approximate expression. You may make it manage or transmit / receive as the expressed grid error or the correction value or adjustment amount corresponding to this.

Δx (x, y) = k ₁ + k ₃ x + k ₅ y + k ₇ x ² + k ₉ xy + k ₁₁ y ²
+ K ₁₃ x ³ + k ₁₅ x ² y + k ₁₇ xy ² + k ₁₉ y ³
Δy (x, y) = k ₂ + k ₄ y + k ₆ x + k ₈ y ² + k ₁₀ yx + k ₁₂ x ^{2
_{+ K 14 y 3 + k 16}} y 2 x + k 18 yx 2 + k 20 x 3 ... ( Equation 1)
In the present embodiment, in the GCM server 500, the grid error for each of the general machines EX2 to EXn with the specific machine EX1 as a reference is referred to as grid error data with reference to the design value (hereinafter also referred to as absolute grid error data for convenience). ) Rather than the relative value (difference) with respect to the specific machine EX1. In this regard, conventionally, for any two combinations of all of the units included in the exposure system, without distinguishing between a specific unit and a general unit, the absolute grid error data from one is used as the absolute value of the other. The relative grid error data obtained by subtracting the dynamic grid error data was managed by the GCM server. In other words, in the prior art, the GCM server stores and manages all grid errors for any two combinations of these units, and precedes the current process unit (the unit that will perform the exposure process) to be superposed. The relative grid error data (correction value) corresponding to the machine of the original process to be performed (the machine immediately before being the superposition source) is sent to the machine of the current process. However, in the prior art, it is necessary to store and manage grid error data for any two combinations of all the units, and the amount of data is large, which is not efficient. In this embodiment, since the GCM server 500 manages the relative grid error data of the general machines EX2 to EXn with respect to the specific machine EX1, there is no need to store and manage grid error data for the specific machine EX1. , Storage capacity can be reduced.

  FIG. 4 is a diagram schematically showing the concept of relative management of grid errors of a general car based on a specific car. In the same figure, the exposure system is composed of six units of No. A to No. F, No. A is selected as a specific No., and the remaining general Nos. B to F are indicated by solid arrows in the figure. The manner of managing the relative grid error data for the specific machine A is shown. Between general units (Unit B and Unit C, Unit B and Unit D, Unit B and Unit E, Unit B and Unit F, Unit C and Unit D, Unit C and Unit E, Unit C and Unit F, Unit D and Unit E, No. D and No. F, No. E and No. F) relative difference in grid error, as indicated by the dotted arrows in the figure, from one to the other of the grid error data for a specific No. for each general No. Can be easily obtained by subtracting.

  In addition, after exposing and developing a pattern of a predetermined reticle (for example, the above-described reference reticle) on the wafer W by the specific machine EX1, the pattern is superimposed on the wafer W using the same reticle and exposed by the general machine EX2. The wafer W may be transferred and developed, the pattern overlay error may be measured using the overlay measuring device 400, and the overlay error may be added to the grid error data described above.

  In addition, in the above description, the specific unit in each unit constituting the exposure system is selected as one unit, but a plurality of units may be selected. In this case, the grid error difference of each general unit is managed for each specific unit. However, as shown in FIG. 5, in addition to managing the difference in grid error of each general unit (units D to F) with respect to one specific unit (unit A), for the one specific unit (unit A) The difference in grid error of other specific units (units B and C) is also managed, and the difference in grid error of each general unit (units D to F) with respect to the other specific unit (units B and C) is calculated. You may make it obtain | require indirectly by. In FIG. 5, Units A to C are specific units, Units D to F are general units, solid arrows indicate grid error data that is directly managed, and dotted arrows are obtained indirectly by calculation. The grid error data obtained is shown.

[Grid compensation]
Next, a grid error correction process when performing an exposure process on the wafer W will be described. First, a predetermined exposure command including a process program ID and exposure history data is sent from the host computer system 200 to the GCM server 500 and a machine on which exposure processing is to be performed. It is assumed that the wafer W exposure history data is stored in the internal storage device of the host computer system 200. In the exposure history data, the ID of the machine that performed the original process (here, the machine EX2 is taken as an example), information such as the process program ID when the original process was processed, and the machine that performs the current process (here, As an example, information such as ID of the machine EX3) and process program ID used in the current process are included. Hereinafter, the grid error data related to the machines EX2 and EX3 may be referred to as data DD2 and DD3, respectively. When the GCM server 500 receives the exposure command, based on the original process machine ID and the current process machine ID, both the original process and the current process machine ID are general machines. Data DD2 and data DD3 of current machine EX3 are acquired. Next, the data DD2 related to the machine EX2 of the original process is subtracted from the data DD3 related to the machine EX3 of the current process to obtain the difference (DD3-DD2) of the grid error data of the machine EX3 with respect to the machine EX2. In addition, when the machine of the original process is a specific machine (here, machine EX1), the calculation of the difference between such machines is unnecessary, and the grid error data for the general machine may be used as it is. Next, the GCM server 500 sends the data (DD3-DD2) to the corresponding number machine EX3.

  In the machine EX3, when an exposure command is received from the host computer system 200, first, the exposure apparatus EX3 is used using a reticle loader (not shown) according to the process program related to the process program ID included in the exposure command from the host computer system 200. The reticle R is loaded on the reticle stage RST, and the wafer W is loaded on the wafer holder WH of the exposure apparatus EX3 using a wafer loader (not shown). Next, search alignment of the wafer W loaded on the wafer holder WH is performed. Specifically, at least two search marks positioned at the peripheral portion of the wafer W in a substantially symmetrical manner with respect to the center of the wafer W are detected using the alignment system ALG. These search marks are detected by sequentially positioning the wafer stage WST and setting the magnification of the alignment system ALG to a low magnification so that each search mark is positioned within the detection field of the alignment system ALG. . Then, based on the detection result of alignment system ALG (relative positional relationship between the index center of alignment system ALG and each search mark) and the measured value of laser interferometer IFW when each search mark is detected, the stage coordinates of the two search marks Find the position coordinates on the system. Next, the residual rotation error of the wafer W is calculated from the position coordinates of the search mark, and the wafer holder WH is slightly rotated so that the residual rotation error becomes substantially zero. Thereby, the search alignment of the wafer W is completed.

  Next, the position coordinates on the stage coordinate system of all shot areas on the wafer W are measured. Specifically, the position coordinates of the fine alignment mark (wafer mark) on the wafer W on the stage coordinate system, that is, the position of the shot area, in the same manner as the measurement of the position coordinates of each search mark at the time of the search alignment described above. Find the coordinates. The wafer mark is detected by setting the magnification of the alignment system ALG to a high magnification. Next, statistical calculation (EGA calculation) using the least square method as disclosed in Japanese Patent Application Laid-Open No. 61-44429 and the like based on the measured position coordinates of the shot area and the position coordinates on each design. And calculate six parameters (corresponding to six parameters of rotation regarding the arrangement of each shot area on the reference wafer, scaling in the X and Y directions, orthogonality, and offset in the X and Y directions), and the calculation result The position coordinates (array coordinates) of all shot areas are calculated based on the design position coordinates of each shot area, and the calculation result, that is, the position coordinates of all shot areas on the wafer W is stored. Next, based on the grid error data sent from the GCM server 500, the calculated position coordinates of all shot regions are corrected, and the position coordinates corrected for the grid error are calculated. Next, the wafer W (wafer stage WST) to the acceleration start position (scanning start position) for exposure of each shot area based on the position information and the measured baseline of each shot area obtained in this way. ) Movement is performed, and an image of the pattern of the reticle R is exposed and transferred to the first shot area. Thereafter, similarly, the exposure transfer is repeatedly performed on the other shot areas sequentially in a step-and-scan manner. Thereby, the image of the reticle pattern is accurately superimposed on each shot area on the wafer W and exposed and transferred.

[Selection of exposure equipment to be the standard for overlay]
By the way, a semiconductor device is manufactured by superposing several to a dozen layers. Here, it is general that the numbering machine that is a reference for superposition, that is, the numbering machine that is a target for matching the grid error in the current process, is a numbering machine that has exposed the layer immediately before the current process. However, when the machine that exposed the previous layer is used as a reference for superposition, the previous layer is also exposed by matching the grid error with the previous layer (2 layers before when viewed from the current process). Errors can occur cumulatively. In addition, since the overlay standard is different for each layer, in the exposure process of each layer, the machine that exposed the previous layer is identified, the grid error at that time is managed, and the machine that exposed the previous layer It is necessary to calculate the difference between the grid errors, and information management becomes complicated.

  Therefore, in the present embodiment, a specific layer is arbitrarily selected from a plurality of layers in lot units (for example, in units of 25 wafers), grid matching is performed using the specific layer as a reference, and all the other layers are selected. The layer is exposed. That is, on the basis of a machine that performs exposure processing on a specific layer or that performs exposure processing (hereinafter may be referred to as a specific layer machine), other machines that perform exposure processing on all other layers (hereinafter referred to as general layer machines). The grid error of (sometimes) is matched with the specific layer machine. For this reason, the GCM server 500 uses the specific layer machine or the specific layer machine selected from the general machine as a reference based on the difference of the grid error of each general machine with respect to the specific machine. And a grid error correction value is output to each general layer machine. Specifically, as shown in FIG. 6, for example, as the specific layer machine, a machine that exposes the first layer (layer 1) is selected, and the remaining layers (layer 2 to layer n) are selected. Match to layer 1. However, the specific layer machine may be a machine that exposes layers other than the first layer, and is not limited to a machine related to a layer that has already been subjected to the exposure process, and the exposure process is performed after the current process. The number machine related to the layer may be used. As a result, the cumulative error described above is reduced, and information management is facilitated.

  In this embodiment, since the difference of the grid error of the general unit with respect to the specific unit is managed, by selecting the specific unit as the specific layer unit, the relative about each general unit managed Distortion data can be used as it is, that is, it can be used as it is without performing a difference operation, which is convenient.

[Others]
A semiconductor device as a micro device is a step of designing a function and performance of a device, a step of manufacturing a reticle based on the design step, a step of manufacturing a wafer from a silicon material, and the exposure system of the above-described embodiment. It is manufactured through a step of exposing and transferring a pattern to a wafer, a device assembly step (including a dicing process, a bonding process, and a package process), an inspection step, and the like.

The embodiment described above is described for facilitating understanding of the present invention, and is not described for limiting the present invention. Therefore, each element disclosed in the above embodiment is intended to include all design changes and equivalents belonging to the technical scope of the present invention.
For example, not only an exposure system used for manufacturing a semiconductor device, but also an exposure system used for manufacturing a liquid crystal display device, a plasma display, a thin film magnetic head, an imaging device (such as a CCD), a micromachine, and a DNA chip, and a reticle. Alternatively, the present invention can also be applied to an exposure system for manufacturing a mask. In other words, the present invention can be applied regardless of the exposure method and the application.

It is a block diagram which shows the whole structure of the exposure system of embodiment of this invention. It is a figure which shows schematic structure of the exposure apparatus with which the exposure system of embodiment of this invention is provided. It is a figure which shows schematic structure of the alignment system with which the exposure apparatus of embodiment of this invention is provided. It is a figure which shows notionally the management method of the grid error data of embodiment of this invention. It is a figure which shows notionally the management method of the grid error data at the time of selecting the multiple specific machine of embodiment of this invention. It is a figure which shows notionally the case where the distortion of each layer of embodiment of this invention is matched with a specific layer.

Explanation of symbols

  EX1 to EXn ... exposure apparatus (unit No.), 200 ... host computer system, 400 ... superposition measuring instrument, 500 ... management server (GCM server), W ... wafer, R ... reticle, CNT ... main controller, 410 ... measurement sensor 450 ... Measurement control device.

Claims (9)

With reference to a specific machine that is a specific exposure apparatus among a plurality of exposure apparatuses that expose the substrate through a mask pattern, error data relating to grid errors relating to each of the general machine that is an exposure apparatus other than the specific machine is obtained. A process to manage;
Calculating a relative error between any two exposure apparatuses of the plurality of exposure apparatuses to be grid-matched based on the error data.   The error data is error data obtained by measuring a reference pattern formed on a reference board with a measurement device provided in the specific machine in the specific machine, and the reference board in the measurement apparatus provided in the general machine in the general machine. The grid matching method according to claim 1, wherein the grid matching method is difference data from error data obtained by measuring the reference pattern.   When exposing and transferring a pattern of three or more layers on a substrate, the one exposure is based on one exposure apparatus arbitrarily selected from a plurality of exposure apparatuses used for the overlay exposure. 3. The grid matching method according to claim 1, wherein error data relating to each of exposure apparatuses other than the apparatus is calculated, and the error data is output to each corresponding exposure apparatus.   The grid matching method according to any one of claims 1 to 3, wherein a plurality of the specific machines are selected from the exposure apparatus.   The grid matching method according to claim 1, wherein the error data is data expressed in the form of a coefficient of a predetermined polynomial.   6. The grid matching method according to claim 1, wherein the error data is data including a nonlinear error component from which a linear error component is removed.   A computer-readable information recording medium storing a program for causing a computer to execute the grid matching method according to claim 1. A plurality of exposure apparatuses for exposing the substrate through a mask pattern;
A management server that manages error data related to grid errors related to each of the general number machines that are exposure apparatuses other than the specific number machine, with a specific number machine that is a specific number of exposure apparatuses among the plurality of exposure apparatuses as a reference. A featured exposure system. Each exposure apparatus has a measurement device that measures an error of a pattern formed on a substrate, and a control device that sends error data measured by the measurement device to the management server via a communication line,
The management server manages difference data between the error data sent from the control device included in the specific car and the error data sent from the control device included in the general car. Item 9. The exposure system according to Item 8.

Images